Volatile composition and aroma profile of Uruguayan Tannat wines

Food Research International 69 (2015) 244–255

Contents lists available at ScienceDirect

Food Research International journal homepage: www.elsevier.com/locate/foodres

Volatile composition and aroma profile of Uruguayan Tannat wines Laura Fariña a,b,⁎, Valeria Villar a, Gastón Ares c, Francisco Carrau a, Eduardo Dellacassa d, Eduardo Boido a a

Sección Enología, Facultad de Química, Universidad de la República, Uruguay Departamento de Biología Molecular, Instituto de Investigaciones Biológicas Clemente Estable, Uruguay c Departamento de Ciencia y Tecnología de Alimentos, Facultad de Química, Universidad de la República, Uruguay d Cátedra de Farmacognosia y Productos Naturales, Facultad de Química, Universidad de la República, Uruguay b

a r t i c l e

i n f o

Article history: Received 7 October 2014 Accepted 20 December 2014 Available online 26 December 2014 Keywords: Projective mapping Napping GC–MS Odor Sensory characterization

a b s t r a c t Tannat is a red variety of Vitis vinifera that has become the major variety for the production of premium red wines in Uruguay. Due to its small cultivation around the world, research on the viticulture and enology of this variety is still necessary to improve wine quality. In this context, the aim of the present work was to characterize the aroma profile of Uruguayan Tannat wines using chemical and sensory methodologies. The volatile composition of ten Uruguayan Tannat wines, sold in the international market, was studied by gas-chromatography (GC–MS). Sixty two volatile compounds were identified using GC–MS, being alcohols and esters the most abundant compounds. Only few volatile compounds were found at concentrations higher than their odor threshold in all samples. Sensory characterization of wine aroma was characterized by a panel of wine professionals using projective mapping. Red fruits, fruity, dry fruits, and woody were the main descriptors used for describing similarities and differences in the aroma profile of the wines. Projective mapping sorted samples into four main groups. Partial least square regression (PLSR) enabled to explain many of the most important sensory descriptors (woody, earthy, phenolic, sulfur, chemical and microbiological) through volatile composition. © 2014 Elsevier Ltd. All rights reserved.

1. Introduction Aroma is one of the most important quality factors of wine and is one of the key determinants of consumer acceptance (Lockshin & Corsi, 2012; Rapp, 1998; Saénz-Navajas, Ballester, Pêcher, Peyron, & Valentin, 2013). Wine aroma is a complex sensory characteristic that is determined by more than 1300 volatile compounds, including alcohols, esters, acids, aldehydes, isoprenoids, lactones and ketones, with a wide concentration range (Villamor & Ross, 2013). Differences in the aromatic profile of wines are determined by changes in the type, proportion and concentration of these volatile compounds (Atanasova et al., 2005). Aroma characterization of wine is usually performed by gas chromatography–mass spectroscopy analyses, which enable the identification and quantification of volatile and non-volatile components (Francis & Newton, 2005). The type and concentration of these volatile compounds are responsible for the characteristic aroma of wine. In particular, concentration usually explains variation in aroma between certain types of wine which contain the same volatile compounds (Boido et al., 2003).

⁎ Corresponding author at: Departamento de Biología Molecular, Instituto de Investigaciones Biológicas Clemente Estable, Uruguay. Tel.: +598 29248194; fax: +598 29241906. E-mail address: [email protected] (L. Fariña).

http://dx.doi.org/10.1016/j.foodres.2014.12.029 0963-9969/© 2014 Elsevier Ltd. All rights reserved.

The contribution of volatile compounds to wine aroma depends on both their concentration and their perception threshold (Ferreira, Lopez, & Cacho, 2000). However, aroma perception of wine depends on the simultaneous perception of a large number of compounds. In this complex mixtures perceptual interactions between volatile compounds exist (Laffort & Dravnieks, 1982; Villamor & Ross, 2013), which lead to changes in qualitative and quantitative aromatic differences (Atanasova et al., 2005; Thomas-Danguin & Chastrette, 2002). For these reasons, in order to adequately evaluate the aroma profile of wine and understand which compounds are responsible for the characteristics notes are necessary to correlate volatile composition and sensory data (Francis & Newton, 2005; Green, Parr, Breitmeyer, Valentin, & Sherlock, 2011; Noble & Ebeler, 2002; Vilanova, Escudero, Graña, & Cacho, 2013). Descriptive analysis with highly trained panels has been the most widely used methodology for characterizing the aromatic profile of wine (De La Presa-Owens & Noble, 1995; Heymann & Noble, 1987; Noble, Williams, & Langron, 1984). In this methodology assessors are trained in the identification and quantification of specific notes, and to provide a qualitative and quantitative description of wine aroma (ASTM, 1992). Descriptive analysis allows obtaining detailed, robust, consistent and reproducible results, which are stable in time (Lawless & Heymann, 2010). However, creating and maintaining well-trained, calibrated sensory panels can be economically challenging and time consuming, particularly when dealing with a complex product such as wine (Varela & Ares, 2012). Moreover, due to extensive training highly

L. Fariña et al. / Food Research International 69 (2015) 244–255

trained assessors can perceive wine aroma differently from consumers, who have a unified and holistic impression of the product. In this context, several novel methodologies for sensory characterization have been developed in the last decade (Varela & Ares, 2012). These methodologies can be performed with trained, semi-trained or even naïve assessors, providing sensory maps similar to those obtained using classic descriptive analysis (Ares & Varela, 2014). Holistic methodologies are one of the most popular types of novel methodologies for sensory characterization (Varela & Ares, 2012). They rely on the evaluation of global similarities and differences among samples, encouraging the generation of a synthetic representation of the products, which is inhibited when assessors are asked to focus their attention on specific characteristics (Ares & Varela, 2014; Prescott, 1999). Projective mapping, also known as Napping®, is a holistic methodology for sensory characterization proposed by Risvik, McEwan, Colwill, Rogers, and Lyon (1994). In this methodology consumers are asked to provide a two dimensional projection of a group of samples, according to their own criteria (Varela & Ares, 2012). This methodology has been previously used for sensory characterization of red wine (Hopfer & Heymann, 2013; Perrin & Pagès, 2009; Torri et al., 2013). One of the main advantages of projective mapping for the evaluation of wine aroma is that it enables the evaluation of global differences among samples and the spontaneous identification of the main notes responsible for those differences. Considering that Uruguay is one of the few places in the world where Tannat is commonly grown, Uruguayan wine-making industry has established a strategy to produce high-quality Tannat wines using state-of-the-art viticultural technology (Carrau, 1997). However, due to its small cultivation around the world, research on the viticulture and enology of this variety is still necessary to better characterize its wine quality potential. Tannat is one of the varieties with the highest contents of anthocyanins and other polyphenolic compounds (Alcalde-Eon, Boido, Carrau, Dellacassa, & Rivas-Gonzalo, 2006; Boido et al., 2011) and has moderate intensity aromas which are usually described as raspberry, plum, quince, and small-berry-like (Varela & Gámbaro, 2006). In this context, the aim of the present work was to characterize the aroma profile of Uruguayan Tannat wines using physicochemical and sensory methodologies.

2. Materials and methods 2.1. Samples Ten commercial samples of Uruguayan 100% varietal Tannat wine, sold in the international market, were selected for the study. Samples were obtained directly from the wineries. Samples were selected to represent high quality Uruguayan Tannat wines, belonging to different price segments. Wines were bottled in 750 mL bottles and were conserved under 15 °C until their analysis. A description of the wines is shown in Table 1.

245

2.2. Volatile composition analysis 2.2.1. Chemical and reagents Pure standards were purchased from Sigma-Aldrich Corp. (Milwaukee, WI) and Extrasynthese (Genay Cèdex, France). Solvents were of spectrophotometric grade from Merck (USA). ISOLUTE ENV+ was purchased from Biotage AB (Uppsala, Sweden). All other chemicals were of analytical grade. 2.2.2. Sample preparation Volatiles were determined after adsorption and separate elution from an isolute ENV+ cartridge packed with 1 g of highly crosslinked styrene-divinyl benzene (SDVB) polymer (40–140 mm, cod. no. 9150100-C), as previously reported by Boido et al. (2003). The cartridges were sequentially equilibrated with methanol (15 mL) and distilled water (20 mL). A sample of 50 mL of wine, diluted with 50 mL of distilled water and containing 0.1 mL of internal standard (1-heptanol at 230 mg/L in a 50% hydroalcoholic solution), was applied with an appropriate syringe (4–5 mL/min) and the residue was washed with 15 mL of distilled water. The aroma compounds were eluted with 30 mL of dichloromethane. The solution was dried with Na2SO4 and concentrated to 1.5 mL on a Vigreux column, stored at 10 °C, and, immediately prior to GC–MS analysis, further concentrated to 150 μL under a gentle nitrogen stream. Sample preparation was performed in duplicate 2.2.3. Gas chromatography–mass spectrometry (GC–MS) analyses GC/MS analyses were conducted using Shimadzu QP 2010 Ultra mass spectrometry using a DB-WAX 30 (Agilent Technologies J&W, Santa Clara, CA, USA) bonded fused silica capillary column, coated with poly(ethylene glycol) (30 m × 0.25 mm i.d., 0.25 μm film thickness). The GC-oven was programmed from a starting temperature of 40 °C, which was retained 8 min, to 180 °C at 3 °C/min, then ramped to 220 °C at 5 °C/min, 220 °C (20 min); injector temperature, 250 °C; injection was performed in Split mode (1:50); volume injected, 1.0 μL; carrier gas, helium, 76 kPa (42.4 cm/s); interface temperature, 250 °C; energy, 70 eV; acquisition mass range, 35–500 amu. HRGC-FID and HRGC-MS instrumental procedures using an internal standard (1-heptanol) were applied for quantification, as described by Boido et al. (2003). The components of the wine aroma were identified by comparison of their linear retention indices (LRI), determined in relation to a homologous series of n-alkanes, with those from pure standards or using published data. Comparison of fragmentation patterns in the mass spectra with those stored on databases (Adams, 2007; McLafferty & Stauffer, 1991; NIST08, version 2.0, National Institute of Standards and Technology, Gaithersburg, MD, USA) was also performed. In cases where pure reference compounds were not used, the identification was indicated as tentative and the quantification was performed using the characteristic fragments (Loscos, Hernandez-Orte, Cacho, & Ferreira, 2007). 2.3. Sensory characterization

Table 1 Description of the Uruguayan Tannat wine samples considered in the study. Sample

Export price range (US$)

Vintage

Aged in oak barrel

Alcoholic degree

M1 M2 M3 M4 M5 M6 M7 M8 M9 M10

6–8 3–5 6–8 9–11 9–11 6–8 9–11 9–11 3–5 3–5

2008 2010 2010 2011 2007 2010 2011 2011 2009 2012

Yes No Yes Yes Yes Yes Yes Yes Yes No

13.0 13.8 14.0 15.0 13.5 13.5 13.5 14.7 13.5 12.5

Sensory characterization was performed by 30 wine professionals, including sommeliers, winemakers, and oenologists. Participants were recruited from the Uruguayan Sommelier Society and had a minimum of 2 year experience in the wine industry. The tests took place in standard sensory booths (ISO, 2007), under white lighting, controlled temperature (22–24 °C) and airflow conditions. Samples (30 mL) were presented at room temperature (20 °C) in clear 190 mL standard glasses (ISO, 1977), covered with a plastic cover and marked with three digit codes. Wines were presented following a William's Latin square design to minimize order and carry-over effects. Assessors were asked to smell the samples and to place them on an A3 white sheet (42 cm × 30 cm), according to their similarities

246

L. Fariña et al. / Food Research International 69 (2015) 244–255

and dissimilarities. They were asked to complete the task using their own criteria and they were told that there were no right or wrong answers. Additionally, it was explained to them that two samples close together on the sheet were very similar and that if they perceived two samples very different they had to place them very far from each other. After positioning the samples consumers were asked to provide a description of the aroma of each sample.

2.4. Data analysis 2.4.1. Volatile composition Principal component analysis (PCA) was performed on the matrix containing the average concentration of the identified volatile compounds for each sample.

3. Results and discussion 3.1. Volatile composition Sixty two volatile compounds were identified and quantified in the 10 commercial Tannat wines, including acids, alcohols, C6 compounds, esters, phenols and terpenes. Table 2 shows the mean, maximum and minimum concentration of the identified volatile compounds. Compounds are ordered according to their chemical structure. To assess the possible contribution of different components to wine aroma, the detection threshold and aroma descriptor for each compound reported in the literature are included.

2.4.2. Sensory characterization Projective mapping data were analyzed using multiple factor analysis (MFA). The X and Y coordinates of each sample on the sheet of each assessor were determined considering the left bottom corner of the sheet as the origin of coordinates. MFA was performed considering the coordinates of each consumer as a separate group of variables (Pagès, 2005). Confidences ellipses were constructed using truncated total bootstrapping considering the first four dimensions of the MFA (Cadoret & Husson, 2013). The words elicited by the assessors in the description phase were qualitatively analyzed. Words were grouped into categories, taking into account the aroma wheel proposed by Noble et al. (1987). Frequency of mention of each category was determined by counting the number of assessors who elicited words within each category. Categories mentioned by at least 5% of the consumers were retained for further analysis. The frequency table of the descriptors was considered as a group of supplementary variables in the MFA of projective mapping data (Pagès, 2005)

3.1.1. Alcohols Higher alcohols are secondary products of yeast metabolism and have been associated with pungent, sweet and fruity odors (Li, Tao, Kang, & Yin, 2006; López de Lema, Bellicontro, Mencarelli, Moreno & Peinado, 2012). As shown in Table 2, alcohols were the largest group of volatile compounds in the evaluated Tannat wine, in agreement with results reported by other authors for this and other red wine varieties (Boido et al., 2003; Dominguez & Agosin, 2010; García-Carpintero, Sánchez-Palomo, & González-Viñas, 2011; Gil, Caballeros, Arroyo, & Prodanov, 2006; Jiang, Xi, Luo, & Zhang, 2013). However, the average total concentration of alcohols in the evaluated Tannat wines (183 mg/L) was lower to the total concentration of alcohols reported for other red wine varieties (García-Carpintero et al., 2011; Jiang et al., 2013). As shown in Table 2, odor threshold was only reached for a small proportion of the identified alcohols, including 2 and 3-methyl1-butanol (for all the evaluated samples) and beta phenyl ethyl alcohol (only for some of the samples). This result is in agreement with Escudero, Campo, Fariña, Cacho, and Ferreira (2007) who reported that odor threshold in two Uruguayan Tannat wines was only reached for a small proportion of the identified alcohols.

2.4.3. Comparison of the methodologies The RV coefficient (Robert & Escoufier, 1976) was used to compare sample configurations from the PCA of volatile composition data and the MFA of projective mapping. This coefficient was calculated considering sample coordinates in the first four dimensions of PCA and MFA. The RV coefficient is a measure of the similarity between two factorial configurations, which takes the value of 0 if the configurations are uncorrelated, and the value of 1 if the configurations are homothetic. It depends on the relative position of the points in the configuration and therefore is independent of rotation and translation (Robert & Escoufier, 1976). The significance of the RV coefficient was tested using a permutation test, as suggested by Josse, Husson, and Pagès (2007). If the RV coefficient between two sample configurations is significant, it can be concluded that they are correlated and therefore information about the similarities and differences among samples is similar. Partial least square regression (PLSR) was used to study the relationship between volatile composition and perceived similarities in the aroma of the evaluated samples (Ergon, 2014). Volatile composition data were the explanatory variables (X-matrix) and sample coordinates in the first four dimensions of the MFA of projective mapping data were the response data (Y-matrix). The R2 coefficient, which indicates the proportion of variance in each variable contained in the Y-matrix explained by the model, was considered as indicator of goodness of fit. Standardized coefficients for each volatile compound were determined for each of the variables included in the Y-matrix. Confidence intervals at 95% were calculated using a jack-knife procedure. All data analyses were performed using R language (R Core Team, 2013).

3.1.2. Esters Esters were the second most abundant group of compounds in Tannat wines (Table 2), in agreement with published data for different red wine varieties (Nykänen, 1986; Selli et al., 2004). Aliphatic ethyl esters and acetates are produced enzymatically by yeasts during alcoholic fermentation and their concentration has been reported to strongly depend on yeast strain, fermentation temperature, aeration and sugar content (Perestrelo, Fernandes, Albuquerque, Marques, & Camara, 2006). Moreover the production of ethyl lactate is a characteristic change of malolactic fermentation, while the concentration of aliphatic ethyl esters and acetates usually decrease during aging (Boido et al., 2009). The average total concentration of esters was 118 mg/L, which is similar to that reported for Bobal, Cabernet Sauvignon and Merlot wines (García-Carpintero et al., 2011; Jiang et al., 2013). As shown in Table 2, ethyl succinate and ethyl lactate showed the highest concentrations in the evaluated Tannat wines (Table 2). Despite the importance of the concentration of this group of compounds, Tannat wine has not been characterized by fruity descriptors (Varela & Gámbaro, 2006). As shown in Table 2, esters rarely exceeded their odor threshold. Ethyl hexanoate was the only ester that showed a concentration that exceeded its odor threshold for all samples, while odor threshold of phenyl ethyl acetate was exceeded in some of the evaluated samples. Although the concentrations of alcohols and their esters in wine are mainly determined by the yeast strain and fermentation conditions (Molina et al., 2009), it is important to highlight that amino acids are implicated in the biosynthesis path of these compounds (Hernandez-Orte, Cacho, & Ferreira, 2002). For this reason, the profile and concentration of alcohols and esters in wine are also affected by grape variety.

Table 2 Kovats' index (KI), and average, minimum and maximum concentrations of volatile compounds (μg/L) identified in 10 Uruguayan Tannat wines, commercialized in the international market. A description of how the compounds were identified, odor descriptor and odor thresholds reported in the literature are also included. KIa

Compound name

Odor threshold (μg/L)

Odor descriptor

Identificationb

Concentration (μg/L) Average

Terpenes T1 T2 T3 T4 T5 T6

6000 6000 50 400 3000 3000

Leafy, sweet, floral, creamy, earthy 1 Leafy, sweet, floral, creamy, earthy1 Rose2 Floral, pine3 Leafy, sweet, floral, creamy, earthy2 Leafy, sweet, floral, creamy, earthy2

Norisoprenoids N1 1505 N2 1626 N3 1730 N4 2646 N5 2669 N6 2699 N7 3142 N8 3262

Vitispirane Riesling acetal 1,6,6-Trimethyl-1,2-dihydronaphthalene (TDN) 3-Oxo-alpha-ionol 3-Hidroxy-7,8-dihydro-beta-ionol 3-Oxo-7,8-dihydro-alpha-ionol Dehydrovomifoliol Vomifoliol

N/A N/A 2 N/A N/A N/A N/A N/A

Woody, spicy4 Floral, raspberry4 Petrol5 Honey, apricots6 N/A N/A N/A N/A

TI (93, 121, 136)23 TI (123, 125, 158)23 TI (115, 142, 157)23 TI (108, 137, 165)24 TI (109, 121, 136)24 TI (108, 135, 150)24 TI (95, 124, 166)25 TI (124, 135, 150)24

Phenols P1 P2 P3 P4 P5 P6

1854 2021 2187 2187 2260 2769

Guaiacol 4-Ethylguaiacol 4-Ethylphenol 4-Vinylguaiacol 2,6-Dimetoxy-phenol 4-(4-Hydroxy-3-methoxyphenyl)-2-butanone (zingerone)

9.5 110 605 40 570 N/A

P7 P8 P9

2780 2823 2901

Ethyl-beta-4-hydroxy-3-methoxy-phenyl-propionate 4-Hydroxy-3-methoxyphenylethyl alcohol (homovanillyl alcohol) 4-(4-Hydroxy-3-methoxy-phenyl)butan-2-ol (zingerol)

N/A N/A N/A

Smoky, hospital7 Bretty flavors8 Bretty flavors8 Clove, curry9 Nutty, smokey10 Sweet, fruity, cooked pears11 N/A N/A N/A

C6 compounds C1 1363 C2 1371 C3 1389

1-Hexanol Trans-3-hexen-1-ol Cis-3-hexen-1-ol

2500 1000 400

Grass just cut12 Green13 Green, kiwi14

Alcohols AL1 AL2 AL3 AL4 AL5 AL6 AL7 AL8 AL9

1021 1048 1094 1152 1221 1716 1873 1906 3032

2-Methyl-2-butanol 1-Propanol 2-Methyl-1-propanol 1-Butanol 2 and 3-methyl 1-butanol 3-(Methylthio)-1-propanol Benzyl alcohol 2-Phenylethanol Tyrosol

306,000 40,000 150,000 40,000 1000 200,000 10,000 N/A

Esters E1 E2 E3 E4 E5

911 1018 1157 1239 1348

Isoamyl acetate Isobutyl acetate Ethyl butyrate Ethyl hexanoate Ethyl lactate

30 N/A N/A 14 60,000

A A A A TI (59, 68, 93) 21,22 TI (59, 68, 93)21,22

Maximum

1 14 11 25 6 14

0 0 2 7 2 8

1 82 18 138 12 36

2 21 2 39 8 7 9 37

0 3 0 25 6 3 6 22

9 46 9 50 11 13 13 72

A A A A A A

550 11 160 100 168 11

325 0 0 14 52 3

783 107 872 293 392 25

TI (89, 137, 150)26 TI (122, 137, 168)26 TI (123,137,196)26

15 168 9

9 88 4

21 342 12

A A A

1329 36 49

884 16 22

1797 54 69

Fusel alcohol, ripe fruit12 Like wine, nail polish12 Like wine, medicine12 Like wine, nail polish13 Sweet, potato14 Floral, rose, phenolic, balsamic12 Rose, talc, honey12 N/A

A A A A A A A A

512 235 9725 279 140,960 1700 708 22,298 6545

20 61 4472 156 122,099 737 86 3010 4490

707 501 13,595 382 166,425 2399 5322 40,099 8494

Banana13 N/A N/A Green apple15 Strawberry, raspberry16

A A A A A

35 26 47 208 29,071

8 14 20 161 11,317

148 70 85 275 49,550 247

(continued on next page)

L. Fariña et al. / Food Research International 69 (2015) 244–255

Trans-linalool oxide (furanoid) Cis-linalool oxide (furanoid) Linalool Alpha-terpineol Trans-linalool oxide (pyranoid) Cis-linalool oxide (pyranoid)

1458 1468 1555 1696 1739 1766

Minimum

248

Table 2 (continued) KIa

Compound name

Odor threshold (μg/L)

Odor descriptor

Identificationb

E6

1431

Ethyl octanoate

500

A

234

186

362

E7 E8 E9 E10 E11 E12 E13

1507 1640 1678 1813 1822 2041 2202

Ethyl 3-hydroxybutyrate Ethyl decanoate Diethyl succinate 2-Phenylethyl acetate Ethyl 4-hydroxybutyrate Diethyl malate Diethyl 2 hydroxy glutarate

N/A 200 100,000 250 N/A 760,000 20,000

A A A A A A

250 55 4819 118 3419 607 940

43 34 1819 48 1138 124 589

441 82 9964 316 5840 1746 1639

E14

2386

Ethyl succinate

1,000,000

Sweet, banana, pineapple14 N/A Sweet, hazelnut oil14 Overripe melon, lavender13 Fruity, honeyed, floral14 Cotton candy17 Green12 Grape, green apple, marshmallow18 Toffee, coffee13

A

78,225

Acids AC1 AC2 AC3 AC4 AC5 AC6

1565 1625 1667 1843 2059 2278

2-Methylpropanoic acid Butyric acid 2- and 3-methylbutanoic acid Hexanoic acid Octanoic acid Decanoic acid

230 173 250 420 500 1000

Acid, fatty19 Cheese, rancid19 Blue, cheese19 Fatty, cheese19 Fatty19 Rancid, fat19

A A A A A A

899 506 1058 1064 1191 257

503 235 794 595 406 106

1451 1175 1337 1390 2062 550

Furfural 2-Furanmethanol Z-whiskey lactone Gamma-butyrolactone 2-Hydroxy-3,3-dimethyl-gamma-butyrolactone (pantolactone) 4-(Carboethoxy)-gamma-butyrolactone 3-Hydroxy butanone

770 2000 67 1000 2200 nd 30,000

Fusel alcohol, toasted bread12 Varnish19 Coconut10 Toasted burned12 Toasted bread, smoked20 Smokey, toasted17 Sour yogurt, sour milk12

A A A A A A A

747 578 169 1296 158 892 56

24 31 – 885 80 337 5

6471 1702 369 2086 234 1578 274

Concentration (μg/L) Average

51,298

TI indicates tentative identification assigned by comparing mass spectra and linear retention index with those obtained from the literature (mass fragments indicated between brackets). N/A, not available. a Linear retention index based on a series of n-hydrocarbons reported according to their elution order on Carbowax 20M. b Indicates identification performed by comparing mass spectra and retention time with those of authentic standards supplied by Aldrich (Milwaukee, WI), Fluka (Buchs, Switzerland) and Lancaster (Strasbourg, France). 1 Ribérau-Gayon, Boidron and Terrier (1975). 2 Escudero et al. (2007). 3 Riberau-Gayon, Glories, Maujean and Dubourdieu (2000). 4 Schneider, Razungles, Augier and Baumes (2001). 5 Sacks, Gates, Ferry, Lavin, Kurtz and Acree (2012). 6 González-Pombo, Fariña, Carrau, Batista-Viera and Brena (2014). 7 Maga (1973). 8 Fariña, Boido, Carrau and Dellacassa (2007). 9 Guth (1997). 10 Lopez, Aznar, Cacho and Ferreira (2002). 11 Duarte, Dias, Oliveira, Vilanova, Teixeira, Silva and Schwan (2010). 12 López de Lerma, Bellicontro, Mancarelli, Moreno and Peinado (2012). 13 Moreno (2005). 14 Gómez-Minguez, Cacho, Ferreira, Vicario and Heredia (2007). 15 Lorenzo, Pardo, Zalacain, Alonso and Salinas (2008). 16 Lloret, Boido, Lorenzo, Medina, Carrau, Dellacassa and Versini (2002). 17 Lee and Noble (2003). 18 Etievant (1991). 19 Dominguez and Agosin (2010). 20 Muñoz, Peinado, Medina and Moreno (2007). 21 Adams (2007). 22 McLafferty and Stauffer (1991). 23 Marais et al. (1992). 24 Winterhalter (1990).

Maximum

114,177

L. Fariña et al. / Food Research International 69 (2015) 244–255

Other compounds O1 1453 O2 1663 O3 2042 O4 1610 O5 2020 O6 2162 O7 1266

Minimum

L. Fariña et al. / Food Research International 69 (2015) 244–255

3.1.3. Acids Acids were the third most abundant group of volatile compounds in Tannat wines. The average total concentration of these compounds was 5 mg/L, similar to that reported for other grape varieties (Escudero et al., 2007; García-Carpintero et al., 2011). These compounds are mainly produced during fermentation and their concentration has been reported to depend on the initial composition of the must and fermentation conditions (Schreirer, 1979). As shown in Table 2, a total of 6 acids were identified. Octanoic acid, hexanoic acid and 2- and 3-methylbutanoic acid showed the highest concentration and exceeded their odor threshold. Although these acids have been associated with negative odors at concentrations higher than 20 mg/L, at low concentrations they have a positive contribution to quality by increasing the complexity of the wine aroma (Shinohara, 1985). 3.1.4. C6 compounds Three higher C-6 alcohols were identified in all the evaluated Tannat wine samples: 1-hexanol, trans-3-hexen-1-ol and cis-3-hexen-1-ol (Table 2). These compounds are associated with negative green and herbaceous odors (López de Lema et al., 2012). As shown in Table 2, these compounds did not reach odor threshold and therefore their contribution to wine aroma is expected to be insignificant. The biosynthesis pathway of C6 compounds is catalyzed by four enzymes, being lipoxygenase and hydroperoxide lyase the most relevant (Hatanaka, 1993). The relationship between cis and trans 3-hexen-1-ol has been reported to be a characteristic of the grape variety (Versini, Orriols, & Dalla Serra, 1994). In the case of all the evaluated Tannat wine samples the concentration of cis 3-hexen-1-ol was higher than the concentration of the trans compound, in agreement with results reported by Boido, Carrau, Lloret, Medina, and Dellacassa (2002). 3.1.5. Norisoprenoids and volatile phenolic compounds Norisoprenoids are volatile compounds that are formed through the degradation of carotenoid molecules from the grape (Marais, Van Wyk, & Rapp, 1992), which have been reported to have a large impact on the varietal character of wines (Guth, 1997). As shown in Table 2, 8 norisoprenoid compounds were identified in very low concentrations. Vomifoliol and 3-oxo-alpha-ionol showed the highest concentration in all the evaluated samples. Norisoprenoids are generally characterized by low odor thresholds and have been reported to act as enhancers of fruity, dried raisin or red plum notes (depending on the concentration) (Escudero et al., 2007). Volatile phenolic compounds are relevant contributors to wine aroma. Some of these compounds are shikimates derivates associated with varietal aroma (Guth, 1997). These compounds have been associated with pleasant, floral, tea and eucalyptus aroma (Genovese, Gambuti, Piombino, & Moio, 2007). The shikimates derivatives identified in the present work have been previously found in Tannat grapes (Boido, Fariña, Carrau, Dellacassa, & Cozzolino, 2013) and have been identified among aglycones of other red grapes like Shiraz and Muscat of Alexandria (Loscos et al., 2007; Schneider, Razungles, Augier & Baumes, 2001; Wirth, Guo, Baumes, & Günata, 2001). However, although zingerol has been previously reported by Wirth et al. (2001) as glycoconjugate in grapes of Shiraz variety this is the first time it has been identified in wine. Meanwhile, ethyl-beta-4-hydroxy-3methoxy-phenyl-propionate has been found in Tannat grapes after enzymatic hydrolysis (Boido et al., 2013) and in Chardonnay wines (Liberatore, Pati, Del Nobile, & La Notte, 2010). Phenolic compounds, such as 4-ethylguaiacol and 4-ethylphenol, are generated by yeast Brettanomyces and provide negative defective odors to wine (Chatonnet, Dubourdieu, & Boidron, 1995). These compounds were identified in one of the samples in concentrations higher than their odor threshold, which can be attributed to the presence of Brettanomyces during aging.

249

3.1.6. Terpenes Terpenes are secondary plant metabolites, which are associated with floral and citric aroma (Guth, 1997). In the present work, 6 terpenes were identified, being alpha-terpineol the most abundant. The concentration of these compounds was very low for all the evaluated samples. As shown in Table 2 the concentration of terpenes was always lower than 200 μg/L. Considering the odor threshold of these compounds, their contribution to the aroma profile of the evaluated Tannat wines is expected to be negligible. Similar results have been reported for other red wine varieties (García-Carpintero et al., 2011; Jiang et al., 2013). 3.2. Principal component analysis (PCA) of volatile compounds The first four principal components of the PCA performed on the concentration of the 62 identified volatile compounds accounted for 69.2% of the total variance (22.8%, 18.3%, 15.8% and 12.3% respectively). As shown in Fig. 1(a), samples were sorted into three main groups in the first and second principal components. The first group, composed of samples M2, M9 and M10, was located at negative values of PC1. These samples were mainly associated with volatile compounds generated during alcoholic fermentation, such as acids (AC3: 2- and 3-methylbutanoic acid and AC6: decanoic acid), alcohols (AL4: 1-butanol), and esters (E10: 2 phenyl ethyl acetate and E11: ethyl 4-hydroxybutyrate) (Fig. 1(b)). As shown in Table 1, samples M2, M9 and M10 had the lowest export prices, while M2 and M10 were not aged in oak barrel. The second group included samples M4 and M8, which were located at positive and high values of PC2. This principal component was positively correlated with varietal volatile compounds like phenols (P1: guaiacol, P6: zingerone, P7: ethyl-beta-4-hydroxy-3-methoxyphenyl-propionate, P8:homovanillyl alcohol), norisoprenoids (N4: 3oxo-alpha-ionol, N5: 3-hydroxy-7,8-dihydro-beta-ionol,N8: vomifoliol) and terpenes (T4: alpha-terpineol), as well as the fermentative compounds 2-phenylethanol (AL8) and tyrosol (AL9). Samples M4 and M8 belonged to the category with the highest export price range and were aged in oak barrels, which explains their high concentration of z-whiskey lactone (O3) (Guth, 1997). The third group of wines comprised five samples (M1, M3, M5, M6, and M7) which were located at positive values of PC1 and negative values of PC2. These samples were associated with esters generated during alcoholic fermentation: isoamyl acetate (E1), isobutyl acetate (E2), ethyl butyrate (E3), ethyl lactate (E5), ethyl octanoate (E6), ethyl 3-hydroxybutyrate (E7), ethyl decanoate (D8), and diethyl malate (E12) (Fig. 1(b)). PC3 and PC4 mainly sorted samples M1, M5, and M7 apart from the rest of the samples. Sample M1 was located at native values of PC3 and positive values of PC4, being characterized by its high concentration of 4-ethylguaiacol (P2) and 4-ethylphenol (P3), which have been associated with the presence of Brettanomyces in wine, as well as isoamyl acetate (E1), isobutyl acetate (E2), Riesling acetal (N2) and tyrosol (AL9). Sample M7 was characterized by high concentrations of 3-oxo-7,8dihydro-alpha-ionol (N6), ethyloctanoate (E6), 1-hexanol (C1), trans-linalool oxide (furanoid) (T1), 1-propanol (AL2), and 2 and 3-methyl 1-butanol (AL5); whereas ethyl hexanoate (E4), 2,6dimetoxy-phenol (P5), diethyl malate (E12), cis-linalool oxide (furanoid) (T2), and 2-methyl-2-butanol (AL1) were associated with M5. Meanwhile, sample M6, which was located at positive values of PC3 was characterized by ethyl 3-hydroxybutyrate (E7), trans-linalool oxide (furanoid) (T1), 3-hydroxy-7,8-dihydro-beta-ionol (N5), and 2-hydroxy-3,3-dimethyl-gamma-butyrolactone (O5). 3.3. Sensory characterization The main descriptors used for describing differences in the aroma profile of the wine samples were red fruit, fruity, dry fruit, and woody. These results are in agreement with those reported by Varela and Gámbaro (2006).

250

L. Fariña et al. / Food Research International 69 (2015) 244–255

(a)

8

8

6

M8

M4

6

4

4

0

M6

-2

M1

M5

M9

M2

PC4 (12.31 %)

PC2 (18.32 %)

M7

M1

2

M7

M2

2 M8

M9

0

M6

M3

M10

-2

-4

M4 M3

M10

-4

-6

M5

-8 -10

-8

-6

-4

-2

0

2

4

6

8

10

-6 -10

12

-8

-6

-4

-2

PC1 (22.76 %)

P6

0.75

P8

P7 T4

AL8 0.5

PC2 (18.32 %)

0.25

AL9

6

8

10

E14

P5 P9 N1

N7

AL4

E10

O1

C3T2

O7

AC1

-0.75

-0.5

-0.25

N2 E8

P3 E9

O6 E13

0 0.25 PC1 (22.76 %)

P3

AC6 E11

0

O1 N3

E10

AC3 O7 AL7 T1

AL3

T4AC1 E5

E13

P8

0.75

1

-1

AL2

O6 P6

E7 P7 N5 E14

E9

E6

-1

-0.75

-0.5

-0.25

0 0.25 PC3 (15.77 %)

P9AC2 AL5 O5

O3 E12 E4 P5

-0.75

0.5

P1P4

O4C3 AL4 AL8 AC4N7 N1 T3 E8

AC4

AC5 O5

C1 T2

N4

O2

AL6

N2 AC5

E5

N8

-0.5

E4

AL3 AL6

P2 AL9

-0.25

E6

E3

E2 E1

0.25

AC2

AL1

AL2 -0.75

P2

E7

AL5 O4

-0.5

T5 T6

AL7

AC6 P4

T1 E3N3 E2 E1

E12

N6

C2

C1 AC3

N6 0.5

O2

E11

0

-1 -1

4

C2

0.75

N4 O3 N8 P1

N5

T3

-0.25

2

1

1

PC4 (12.31 %)

(b)

0 PC3 (15.77 %)

T5

AL1

T6

0.5

0.75

1

Fig. 1. Principal component analysis performed on the concentration of the 62 volatile compounds identified in commercial Uruguayan Tannat wines: (a) representation of the samples and (b) representation of the compounds in the first four principal components.

The first four dimensions of the MFA explained 66.6% of the variance of the experimental data. Fig. 2 shows the representation of the samples and the projection of terms used to describe them, in the first four dimensions of the MFA. As shown in Fig. 2(b), the first dimension of the

MFA was positively correlated with defects (chemical, caustic, sulfur and microbiological) and negatively correlated to earthy and phenolic notes and to terms related to aging in oak barrels (woody). Meanwhile, the second dimension of the MFA was negatively related to red fruit

L. Fariña et al. / Food Research International 69 (2015) 244–255

251

(a)

(b)

Fig. 2. Sensory characterization of the aroma profile of Uruguayan Tannat wines using projective mapping: (a) sample representation and (b) projection of the descriptive terms in the first four coordinates of the multiple factor analysis.

notes and positively correlated to cooked vegetable odors. Categories were not clearly associated with the third and fourth dimensions of the MFA, except for floral and cooked vegetable notes. According to hierarchical cluster analysis, samples can be sorted into 5 main groups considering their position in the first four dimensions of the MFA. Samples M1, M2 and M10 were located apart from the rest of the samples in the first dimension of the MFA (Fig. 2(a)), being mainly described with terms related to defects (microbiological, sulfur, chemical) (Fig. 2(b)). However, as shown in Fig. 2(a) sample M1 was located in a distinct position in the third dimension due to its floral and microbiological notes. A second group of samples, composed of wines M4 and M5, was located at negative values of the first and second dimensions, and positive values of the fourth dimension, being characterized by their earthy, phenolic, woody, red fruit and yeasty notes. These two samples were aged in oak barrels and were included in the highest export price range. Samples M3, M6 and M9 were mainly characterized by their red fruit and cooked vegetable notes. Finally, samples M7 and M8 were sorted apart from the rest of the samples in the first two dimensions of the MFA (Fig. 2(a)). These samples were characterized by their intense aroma and their burnt and cooked vegetable notes and corresponded to the highest price range.

3.4. Comparison of the methodologies The RV coefficient between the samples' coordinates in the first four dimensions of the PCA of volatile compounds' concentration and the MFA of projective mapping data was 0.595. Although the RV coefficient was significant (p = 0.233), its value was low, suggesting that the PCA of physicochemical data was not able to fully explain global similarities and differences in the aroma profile of the wines. This can be explained considering differences in the perception threshold of the volatile compounds, as well as changes in odor perception due to interactions

Table 3 Percentage of variance explained on the first four dimension of the MFA of projective mapping data by a four component partial least squares model constructed considering projective mapping data as Y-matrix and the concentration of the identified volatile compounds as X-matrix. Dimension

Percentage of explained variance (%)

1 2 3 4

94.3 92.1 90.5 81.3

252

L. Fariña et al. / Food Research International 69 (2015) 244–255

(a)

(b)

(c)

(d)

L. Fariña et al. / Food Research International 69 (2015) 244–255

among volatile compounds (Atanasova et al., 2005; Laffort & Dravnieks, 1982). Also, it should be considered that in projective mapping the importance of the different odor notes to global similarities and differences among samples is determined by the assessors, whereas PCA gives the same importance to all the volatile compounds for the construction of sample configurations. Partial least squares (PLS) regression identified the volatile compounds that had the largest contribution to the perceived similarities and differences among samples. This statistical technique enabled the identification of compounds that influenced, positively or negatively, the position of samples in the first four dimensions of the MFA of projective mapping data. The percentage of variance explained by a four component PLS model is shown in Table 3 for the first four dimension of the MFA of projective mapping data. As shown in Fig. 3(a), the volatile compounds that positively contributed to the position of samples in the first dimension of the MFA were two alcohols (AL3: 2-methyl-1-propanol and AL6: 3-(methylthio)-1propanol), one ester (E10: 2-phenylethyl acetate) and one acid (AC1: 2-methylpropanoic). Some of these compounds have been related to defective odors, such as acid, fatty, medicinal, and potato (Table 1), in agreement with the fact that the first dimension was positively correlated to sulfur, chemical, and microbiological odors (Fig. 2(b)). Meanwhile, isoprenoids (vitispirane and -oxo-alpha-ionol), and volatile phenols (P6: zingerone, P7: ethyl-beta-4-hydroxy-3-methoxy-phenyl-propionate, and P8: 4-hydroxy-3-methoxyphenylethyl alcohol) were the compounds that had significant negative coefficients for the position of samples in the first dimension. The first dimension was negatively correlated to woody, earthy and phenolic descriptors (Fig. 2(b)), in agreement with the fact that these compounds have been associated with phenolic, smoky and woody odors (c.f. Table 1), The position of samples in the second dimension of the MFA of projective mapping data was positively determined by vomifoliol (N8), 4-hydroxy-3-methoxyphenylethyl alcohol (P8), ethyl butyrate (E3), alpha-terpineol (T4), and 2-furanmethanol (O2). Besides, the concentration of ethyl succinate (E14) was negatively correlated to this dimension (Fig. 3(b)). The individual descriptors of these compounds did not explain the odors that were correlated with the second dimension of the MFA. As shown in Fig. 2(b), this dimension was positively correlated with vegetable odors and negatively correlated to red fruit. The third dimension of the MFA of projective mapping data was negatively correlated to floral odors, which can be explained by the negative coefficients of the compounds trans-linalool oxide (T1), ethyl octanoate (E6), ethyl decanoate (E8), diethyl succinate (E9), diethyl-2-hydroxy glutarate (E13), and 4-(carboethoxy)-gamma-butyrolactone (O6). Also, this dimension was negatively correlated to the descriptors chemical and microbiological, which can be explained considering the negative coefficient of the volatile compound 4-ethylphenol (P3), associated with Brettanomyces. Volatile composition was not able to explain the position of the wines in the fourth dimension of the MFA. As shown in Fig. 3(d), none of the identified compounds showed significant coefficients for this dimension.

4. Conclusions Volatile composition and aroma profile of Uruguayan commercial Tannat wines were characterized. Fermentation compounds, and specifically alcohols and esters, were the most abundant volatile compounds in all samples. The main sensory descriptors used for describing similarities and differences in the aroma profile of the wines were red fruit, fruity, dry fruit and woody.

253

The application of multivariate statistical techniques, like PCA and hierarchical cluster analysis, for the analysis of volatile compounds of Tannat wines enabled the identification of different groups of samples with similar physicochemical characteristics. Furthermore, the application of a rapid methodology for sensory characterization enabled the evaluation of overall differences among wine samples without the need of having an extensively trained panel. This technique enabled grouping samples according to their similarities and differences and identified the main odor descriptors in commercial Tannat wines. Projective mapping is based on the evaluation of global differences among samples and therefore enables the identification of the main odors responsible for perceived differences among samples. This type of evaluation opens new possibilities for the characterization of wine aroma which better reflects consumers' experience with wine. Sample configurations obtained using multivariate techniques on physicochemical and sensory data were not highly correlated, mainly due to differences in odor threshold of volatile compounds, the existence of interactions among components that may have additive or suppression effects, as well as different weights given to perceived odors in projective mapping. This result stresses the importance of relying on sensory methodologies for sensory characterization. Finally, the application of PLS regression allowed the identification of volatile compounds that significantly affected the main aromatic notes that discriminated the evaluated wine samples. References Adams, R.P. (2007). Identification of essential oil components by gas chromatography/mass spectrometry (4th ed.). Carol Stream: Allured Publ. Corp. Alcalde-Eon, C., Boido, E., Carrau, F., Dellacassa, E., & Rivas-Gonzalo, J.C. (2006). Pigment profiles in monovarietal wines produced in Uruguay. American Journal of Enology and Viticulture, 57, 449–459. Ares, G., & Varela, P. (2014). Comparison of novel methodologies for sensory characterization. In P. Varela, & G. Ares (Eds.), Novel techniques in sensory characterization and consumer profiling (pp. 365–389). Boca Raton: CRC Press. ASTM (1992). Quantitative descriptive analysis (QDA). Philadelphia: ASTM Digital Library. http://dx.doi.org/10.1520/MNL10523M. Atanasova, B., Thomas-Danguin, T., Chabanet, C., Langlois, D., Nicklaus, S., & Etiévant, P. (2005). Perceptual interactions in odour mixtures: odour quality in binary mixtures of woody and fruity wine odorants. Chemical Senses, 30, 209–217. Boido, E., Carrau, F., Lloret, A., Medina, K., & Dellacassa, E. (2002). Effect of b-glycosidase activity of Oenococcus oeni on the glycosylated flavour precursors of Tannat wine during malolactic fermentation. Journal of Agricultural and Food Chemistry, 50, 2344–2349. Boido, E., Fariña, L., Carrau, F., Dellacassa, E., & Cozzolino, D. (2013). Characterization of glycosylated aroma compounds in Tannat grapes and feasibility of the near infrared spectroscopy application for their prediction. Food Analytical Methods, 6, 100–111. Boido, E., García-Marino, M., Dellacassa, E., Carrau, F., Rivas-Gonzalo, J.C., & EscribanoBailón, M.T. (2011). Characterisation and evolution of grape polyphenol profiles of Vitis vinifera L. cv. Tannat during ripening and vinification. Australian Journal of Grape and Wine Research, 17, 383–393. Boido, E., Lloret, A., Medina, K., Fariña, L., Carrau, F., Versini, G., et al. (2003). Aroma composition of Vitis vinífera cv. Tannat: the typical red wine from Uruguay. Journal of Agricultural and Food Chemistry, 51, 5408–5413. Boido, E., Medina, K., Versini, G., Fariña, L., Carrau, F., & Dellacassa, E. (2009). The effect of bacterial strain and aging on the secondary volatile metabolites produced during malolactic fermentation of Tannat red wine. Journal of Agricultural and Food Chemistry, 57, 6271–6278. Cadoret, M., & Husson, F. (2013). Construction and evaluation of confidence ellipses applied at sensory data. Food Quality and Preference, 28, 106–115. Carrau, F. (1997). The emergence of a new Uruguayan wine industry. Journal of Wine Research, 8, 179–185. Chatonnet, P., Dubourdieu, D., & Boidron, J.N. (1995). The influence of Brettanomyces/ Dekkera sp. yeasts and lactic acid bacteria on the ethylphenol content of red wines. American Journal of Enology and Viticulture, 46, 463–468. R Core Team (2013). R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. De La Presa-Owens, C., & Noble, A.C. (1995). Descriptive analysis of three white wine varieties from Penedès. American Journal of Enology and Viticulture, 46, 5–9. Dominguez, A., & Agosin, E. (2010). Gas chromatography coupled with mass spectrometry detection for the volatile profiling of Vitis vinifera cv. Carménère wines. Journal of Chilean Chemical Society, 55, 385–391. Duarte, W., Dias, D., Oliveira, J. M., Vilanova, M., Teixeira, J., Silva, J., & Schwan, R. (2010). Raspberry (Rubus idaeus L.) wine: Yeast selection, sensory evaluation and

Fig. 3. Standardized PLS-regression coefficients for explaining samples' coordinates in the first four dimensions of the multiple factor analysis performed on projective mapping data as function of the concentration of the identified volatile compounds: (a) first dimension, (b) second dimension, (c) third dimension, (d) fourth dimension. Vertical bars.

254

L. Fariña et al. / Food Research International 69 (2015) 244–255

instrumental analysis of volatile and other compounds. Food Research International, 43, 2303–2314. Ergon, R. (2014). Principal component regression (PCR) and partial least squares regression (PLSR). In D. Granato, & G. Ares (Eds.), Mathematical and statistical methods in food science and technology (pp. 143–174). Chichester, UK: John Wiley & Sons, Ltd. Escudero, A., Campo, E., Fariña, L., Cacho, J., & Ferreira, V. (2007). Analytical characterization of the aroma of five premium red wines. Insights into the role of odor families and the concept of fruitness of wines. Journal of Agricultural and Food Chemistry, 55, 4501–4510. Etievant, P.X. (1991). Volatile compound in food and beverages. New York: Marcel Dekker. Fariña, L., Boido, E., Carrau, F., & Dellacassa, E. (2007). Determination of volatile phenols in red wines by dispersive liquid–liquid microextraction and gas chromatography–mass spectrometry detection. Journal of Chromatography A, 1157, 46–50. Ferreira, V., Lopez, R., & Cacho, J.F. (2000). Quantitative determination of the odorants of young red wines from different grape varieties. Journal of Sciences of Food and Agriculture, 80, 1659–1667. Francis, I.L., & Newton, J.L. (2005). Determining wine aroma from compositional data. Australian Journal of Grape and Wine Research, 11, 114–126. García-Carpintero, E., Sánchez-Palomo, E., & González-Viñas, M.A. (2011). Aroma characterization of red wines from cv. Bobal grape variety grown in La Mancha region. Food Research International, 44, 61–70. Genovese, A., Gambuti, A., Piombino, P., & Moio, L. (2007). Sensory properties and aroma compounds of sweet Fiano wine. Food Chemistry, 103, 1228–1236. Gil, M., Caballeros, M., Arroyo, T., & Prodanov, M. (2006). Characterization of the volatile fraction of young wines from the denomination of origin “Vinos de Madrid” (Spain). Analytica Chimica Acta, 563, 145–153. Gómez-Minguez, M.J., Cacho, J.F., Ferreira, V., Vicario, I.M., & Heredia, F.J. (2007). Volatile composition of Zalema white wines. Food Chemistry, 100, 1464–1473. González-Pombo, P., Fariña, L., Carrau, F., Batista-Viera, F., & Brena, B. (2014). Aroma enhancement in wines using co-immobilized Aspergillus niger glycosidases. Food Chemistry, 143, 185–191. Green, J.A., Parr, W.V., Breitmeyer, J., Valentin, D., & Sherlock, R. (2011). Sensory and chemical characterisation of Sauvignon blanc wine: Influence of source of origin. Food Research International, 44, 2788–2797. Guth, H. (1997). Identification of character impact odorants of different white wine varieties. Journal of Agricultural and Food Chemistry, 45, 3027–3032. Hatanaka, A. (1993). The biogeneration of green odour by green leaves. Phytochemistry, 34, 120–1218. Hernandez-Orte, P., Cacho, J., & Ferreira, V. (2002). Relationship between varietal amino acid profile of grapes and wine aromatic composition. Experiments with model solutions and chemometric study. Journal of Agricultural and Food Chemistry, 2002, 2891–2899. Heymann, H., & Noble, A.C. (1987). Descriptive analysis of commercial Cabernet Sauvignon wines from California. American Journal of Enology and Viticulture, 38, 41–44. Hopfer, H., & Heymann, H. (2013). A summary of projective mapping observations — The effect of replicates and shape, and individual performance measurements. Food Quality and Preference, 28, 164–181. ISO (1977). Sensory analysis – apparatus – wine-tasting, ISO 3591:1977. Geneva: International Organization for Standardization. ISO (2007). Sensory analysis — General guidance for the design of test rooms, ISO 8589:2007. Geneve: International Organization for Standardization. Jiang, B., Xi, Z., Luo, M., & Zhang, Z. (2013). Comparison on aroma compounds in Cabernet Sauvignon and Merlot wines from four wine grape-growing regions in China. Food Research International, 51, 482–489. Josse, J., Husson, F., & Pagès, J. (2007). Testing the significance of the RV coefficient. IASC 07, August 30th–September 1st, 2007, Aveiro, Portugal. Laffort, P., & Dravnieks, A. (1982). Several models of suprathreshold quantitative olfactory interaction in humans applied to binary, ternary and quaternary mixtures. Chemical Senses, 7, 153–174. Lawless, H.T., & Heymann, H. (2010). Sensory evaluation of food. Principles and practices (2nd ed.). New York: Springer. Lee, S., & Noble, A. (2003). Characterization of Odor-Active Compounds in Californian Chardonnay Wines Using GC-Olfactometry and GC-Mass Spectrometry. Journal of Agricultural Food Chemistry, 51, 8036–8044. Li, H., Tao, Y.S., Kang, W.H., & Yin, C.L. (2006). Wine aroma analytical investigation progress on GC (review). Food Science and Biotechnology, 25, 99–104. Liberatore, M., Pati, S., Del Nobile, M., & La Notte, E. (2010). Aroma quality improvement of Chardonnay white wine by fermentation and ageing in barrique on lees. Food Research International, 43, 996–1002. Lloret, A., Boido, E., Lorenzo, D., Medina, K., Carrau, F., Dellacassa, E., & Versini, G. (2002). Aroma variation in Tannat wines: effect of malolactic fermentation on ethyl lactate level and its enantiomeric distribution. Italian Journal of Food Science, 14, 175–180. Lockshin, L., & Corsi, A.M. (2012). Consumer behaviour for wine 2.0: A review since 2003 and future directions. Wine Economics and Policy, 1, 2–23. Lopez, R., Aznar, M., Cacho, J., & Ferreira, V. (2002). Quantitative determination of minor and trace volatile compounds in wine by solid-phase extraction and gas chromatography with mass spectrometric detection. Journal of Chromatography A, 966, 167–177. López de Lerma, N., Bellicontro, A., Mencarelli, F., Moreno, J., & Peinado, R.A. (2012). Use of electronic nose, validated by GC–MS, to establish the optimum off-vine dehydration time of wine grapes. Food Chemistry, 130, 447–452. Lorenzo, C., Pardo, F., Zalacain, A., Alonso, G.L., & Salinas, M.R. (2008). Differentiation of cowinemaking wine by their aroma composition. European Food Research and Technology, 227, 777–787.

Loscos, N., Hernandez-Orte, P., Cacho, J., & Ferreira, V.J. (2007). Release and formation of varietal aroma compounds during alcoholic fermentation from nonfloral grape odorless flavor precursors fractions. Agricultural Food Chemistry, 55, 6674–6684. Maga, J.A. (1973). Taste thresholds values for phenolic acids which can influence flavor properties of certain flours, grains and oilseeds. Cereal Science Today, 18, 326–330. Marais, J., Van Wyk, C.J., & Rapp, A. (1992). Effect of sunlight and shade on norisoprenoid levels in maturing Weisser Riesling and Chenin blanc grapes and Weisser Riesling wines. South African Journal of Enology and Viticulture, 13, 23–32. McLafferty, F.W., & Stauffer, D.B. (1991). The Wiley/NBS registry of mass spectral data (5th ed.). New York: Wiley. Molina, A., Guadalupe, V., Varela, C., Swiegers, J., Pretorius, I., & Agosin, E. (2009). Differential synthesis of fermentative aroma compounds of two related commercial wine yeast strains. Food Chemistry, 117, 189–195. Moreno, J. (2005). Influencia del tipo de envejecimiento sobre el perfil aromático de vinos generosos andaluces (Ph.D. thesis). University of Córdoba. Muñoz, D., Peinado, R.A., Medina, M., & Moreno, J. (2007). Biological aging of sherry wines under periodic and controlled microaerations with Saccharomyces cerevisiae var. capensis: Effect on odorant series. Food Chemistry, 100, 1188–1195. NIST Chemistry WebBook (2014). http://webbook.nist.gov/chemistry (Accessed March 2014) Noble, A.C., Arnold, R.A., Buechsenstein, J., Leach, E.J., Schmidt, J.O., & Stern, P.M. (1987). Modification of a standardized system of wine aroma terminology. American Journal of Enology and Viticulture, 38, 143–146. Noble, A.C., & Ebeler, S.E. (2002). Use of multivariate statistics in understanding wine flavor. Food Review International, 18, 1–21. Noble, A.C., Williams, A.A., & Langron, S.P. (1984). Descriptive analysis and quality ratings of 1976 wines from four Bourdeaux communes. Journal of the Science of Food and Agriculture, 35, 88–98. Nykänen, L. (1986). Formation and occurrence of flavor compounds in wine and distilled alcoholic beverages. American Journal of Enology and Viticulture, 37, 84–96. Pagès, J. (2005). Collection and analysis of perceived product inter-distances using multiple factor analysis: Application to the study of 10 white wines from the Loire Valley. Food Quality and Preference, 16, 642–649. Perestrelo, R., Fernandes, A., Albuquerque, F.F., Marques, J.C., & Camara, J.S. (2006). Analytical characterization of the aroma of Tinta Negra Mole red wine: Identification of the main odorants compounds. Analytica Chimica Acta, 563, 154–164. Perrin, L., & Pagès, J. (2009). Construction of a product space from the ultra-flash profiling method: Application to 10 red wines from the Loire Valley. Journal of Sensory Studies, 24, 373–395. Prescott, J. (1999). Flavour as a psychological construct: Implications for perceiving and measuring the sensory qualities of foods. Food Quality and Preference, 10, 349–356. Rapp, A. (1998). Volatile flavour of wine: Correlation between instrumental analysis and sensory perception. Nahrung/Food, 6, 351–363. Ribérau-Gayon, P., Boidron, J.N., & Terrier, A. (1975). Aroma of muscat grape varieties. Journal of Agricultural Food Chemistry, 23, 1042–1047. Ribereau-Gayon, P., Glories, Y., Maujean, A., & Dubourdieu, D. (2000). Varietal aroma. In Y. Ribereau-Gayon, A. Glories, Maujean, & D. Dubourdie (Eds.), The chemistry of wine: Stabilization and treatments. Handbook of Enology, vol. 2. (pp. 187–206). John Wiley & Sons Ltd. Risvik, E., McEwan, J.A., Colwill, J.S., Rogers, R., & Lyon, D.H. (1994). Projective mapping: A tool for sensory analysis and consumer research. Food Quality and Preference, 5, 263–269. Robert, P., & Escoufier, Y. (1976). A unifying tool for linear multivariate statistical methods: The RV coefficient. Applied Statistics, 25, 257–265. Saénz-Navajas, M.P., Ballester, J., Pêcher, C., Peyron, D., & Valentin, D. (2013). Sensory drivers of intrinsic quality of red wines: Effect of culture and level of expertise. Food Research International, 54, 1506–1518. Sacks, G., Gates, M., Ferry, F., Lavin, E., Kurtz, A., & Acree, T. (2012). Sensory threshold of 1,1,6-Trimethyl-1,2-dihydronaphtalene (TDN) and Concentractions in Young Riesling and Non-Riesling Wines. Journal of Agricultural Food Chemistry, 60, 2998–3004. Schneider, R., Razungles, A., Augier, C., & Baumes, R. (2001). Monoterpenic and norisoprenoidic glycoconjugates of Vitis vinifera L. cv. Melon B. as precursors of odorants in Muscadet wines. Journal of Chromatography A, 936, 145–157. Schreirer, P. (1979). Flavor composition of wines: A review. Critical Reviews in Food Science and Nutrition, 12, 59–111. Selli, S., Cabaroglu, T., Canbas, A., Erten, H., Nurgel, C., & Lepoutre, J.P. (2004). Volatile composition of red wine from cv. Kalecik Karasi grown in central Anatolia. Food Chemistry, 85, 207–213. Serra, S., Barakat, A., & Fuganti, C. (2007). Chemoenzymatic resolution of cis- and trans3,6-dihydroxy-a-ionone. Synthesis of the enantiomeric forms of dehydrovomifoliol and 8,9-dehydrotheaspirone. Tetrahedron: Asymmetry, 18, 2573–2580. Shinohara, T. (1985). Gas chromatographic analysis of volatile fatty acids in wines. Agricultural and Biological Chemistry, 49, 2211–2212. Thomas-Danguin, T., & Chastrette, M. (2002). Odour intensity of binary mixtures of odorous compounds. Comptes Rendus Biologies, 325, 767–772. Torri, L., Dinnella, C., Recchia, A., Naes, T., Tuorila, H., & Monteleone, H. (2013). Projective mapping for interpreting wine aroma differences as perceived by naïve and experienced assessors. Food Quality and Preference, 29, 6–15. Varela, P., & Ares, G. (2012). Sensory profiling, the blurred line between sensory and consumer science. A review of novel methods for product characterization. Food Research International, 48, 893–908. Varela, P., & Gámbaro, A. (2006). Sensory descriptive analysis of Uruguayan Tannat wine: Correlation to quality assessment. Journal of Sensory Studies, 21, 203–217.

L. Fariña et al. / Food Research International 69 (2015) 244–255 Versini, G., Orriols, I., & Dalla Serra, A. (1994). Aroma components of Galician Albariño, Loureira and Godello wines. Vitis, 33, 165–170. Vilanova, M., Escudero, A., Graña, M., & Cacho, J. (2013). Volatile composition and sensory properties of North West Spain white wines. Food Research International, 54, 562–568. Villamor, R., & Ross, C. (2013). Wine matrix compounds affect perception of wine aromas. Annual Review of Food Science and Technology, 4, 1–20.

255

Winterhalter, P. (1990). Bound terpenoids in te Juice of the Purple Passion Fruit (Passiflora edulis Sims). Journal of Agricultural Food Chemistry, 38, 452–455. Wirth, J., Guo, W., Baumes, R., & Günata, Z. (2001). Volatile compounds released by enzymatic hydrolysis of glycoconjugates of leaves and grape berries from Vitis vinifera muscat of Alexandria and Shiraz cultivars. Journal of Agricultural and Food Chemistry, 49, 2917–2923.