Chemical, chromatic, and sensory attributes of 6 red wines produced with prefermentative cold soak

Food Chemistry 174 (2015) 110–118

Contents lists available at ScienceDirect

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

Chemical, chromatic, and sensory attributes of 6 red wines produced with prefermentative cold soak L. Federico Casassa ⇑, Esteban A. Bolcato, Santiago E. Sari Centro de Estudios de Enología, Estación Experimental Agropecuaria Mendoza, Instituto Nacional de Tecnología Agropecuaria (INTA), San Martín 3853, 5507 Luján de Cuyo, Mendoza, Argentina

a r t i c l e

i n f o

Article history: Received 30 July 2014 Received in revised form 27 October 2014 Accepted 28 October 2014 Available online 3 November 2014 Keywords: Prefermentative cold soak Wine colour Anthocyanins Tannins Sensory analysis

a b s t r a c t Six red grape cultivars, Barbera D’Asti, Cabernet Sauvignon, Malbec, Merlot, Pinot Noir and Syrah, were produced with or without prefermentative cold soak (CS). Cold soak had no effect on the basic chemical composition of the wines. At pressing, CS wines were more saturated and with a higher red component than control wines. After 1 year of bottle aging, CS wines retained 22% more anthocyanins than control wines, but tannins and total phenolics remained unaffected. Both saturation and the red component of colour were slightly higher in CS wines. From a sensory standpoint, CS only enhanced colour intensity in Barbera D’Asti and Cabernet Sauvignon wines, whereas it diminished colour intensity in Pinot Noir. Cold soak had no effect on perceived aroma, bitterness, astringency, and body of the wines. Principal Component Analysis suggested that the outcome of CS is contingent upon the specific cultivar to which the CS technique is applied. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction In the past 20 years, the winemaking technique known as prefermentative cold soak (CS) has gained widespread use for the production of red wines and is nowadays applied in most winegrowing regions for many different grape cultivars. Prefermentative CS consists of the contact of fermentation solids (skins, seeds and occasionally stems) with the must in a non-alcoholic and low-temperature environment prior to the onset of alcoholic fermentation (Casassa & Sari, in press; Delteil, 2004). The absence of ethanol is ensured by keeping the must at low temperatures, typically in the range of 5–10 °C, for a variable period of time, ranging from 3 to 5 h up to 10 days (Álvarez, Aleixandre, García, & Lizama, 2006; Gil-Muñoz et al., 2009; Gordillo, Lo´pez-Infante, Rami´rezPe´rez, Gonza´lez-Miret, & Heredia, 2010; Marais, 2003; OrtegaHeras, Pérez-Magariño, & González-Sanjosé, 2012; Reynolds, Cliff, Girard, & Kopp, 2001). As both anthocyanins and tannins are water-soluble, CS should theoretically favour the extraction of both phenolic classes (Delteil, 2004; Hernández-Jiménez, Kennedy, Bautista-Ortín, & Gómez-Plaza, 2012), assuming that the increased solubility outweighs the decreased cellular permeability observed al lower temperatures (Sacchi, Bisson, & Adams, 2005).

⇑ Corresponding author. Tel.: +54 261 4963020x175. E-mail address: [email protected] (L.F. Casassa). http://dx.doi.org/10.1016/j.foodchem.2014.10.146 0308-8146/Ó 2014 Elsevier Ltd. All rights reserved.

The extent of extraction of anthocyanin and tannins as a result of the application of CS has been found to be cultivar-dependent, with conflicting results reported in the literature. For example, some reports indicate an increase in phenolics (Busse-Valverde et al., 2010; González-Neves et al., 2013), a decrease (Budic-Leto, Tomislav, & Vrhovsek, 2003; González-Neves, Gil, Favre, & Ferrer, 2012) or no effect (Ortega-Heras et al., 2012; Pérez-Lamela, García-Falcón, Simal-Gándara, & Orriols-Fernández, 2007) upon application of CS to red wines. Similarly, glycosylated boundaroma compounds, also known as aroma precursors, are watersoluble and one of the claiming benefits of CS is to enhance the extraction of glycosylated bound aroma compounds thereby improving the aromatic potential of the wines during aging (Delteil, 2004). However, a specific look at the concentration of these compounds upon application of CS by two independent studies in Cabernet Sauvignon and Malbec wines have shown inconclusive results regarding this putative effect of CS on glycosylated aroma compounds (Casassa, 2007; McMahon, Zoecklein, & Jasinski, 1999). Prefermentative CS has generated an intense debate over its sensory effects on red wines as well. Unfortunately, most of this ongoing debate is based on anecdotal reports from winemakers, with wines produced under variable and often unreported conditions (length and temperature of CS, yeast inoculation, use of CO2 or external refrigeration). Indeed, these informal accounts are sometimes at odds with what formal research has reported. For

111

4.42 ± 0.04 a 4.42 ± 0.19 a 4.87 ± 0 .12 b 6.47 ± 0.11 c de e d c 3.77 ± 0.01 3.78 ± 0.02 3.72 ± 0.01 3.50 ± 0.02 b a a a 26.10 ± 0.09 24.10 ± 0.04 23.80 ± 0.15 24.00 ± 0.06 3.44 ± 0.15 c 4.41 ± .015 d 1.62 ± 0.09 a 2.92 ± 0.08 b a b d b 8.11 ± 0.13 10.02 ± 0.28 14.77 ± 0.03 9.33 ± 0.02 a

7.5 7.4 12.8 9.7 13-III-2008 7-III-2008 18-III-2008 15-III-2008

Different letters within a column indicate significant differences for Fisher’s LSD test and p < 0.05.

c d e a 1.95 ± 0.02 2.05 ± 0.03 2.67 ± 0.03 1.51 ± 0.02 d c b c 1.73 ± 0.03 1.56 ± 0.02 1.31 ± 0.02 1.61 ± 0.02 d c b c

18.5 11.8

1.93 ± 0.04 1.69 ± 0.02 1.43 ± 0.02 1.77 ± 0.03

9.90 ± 0.21 d 6.60 ± 0.15 c 2.96 ± 0.02 a 3.32 ± 0.04 b 25.80 ± 0.21 b 24.15 ± 0.13 a 4.96 ± 0.32 e 3.16 ± 0.11 bc 10.01 ± 0.09 b 11.22 ± 0.45 c 1.85 ± 0.03 e 1.15 ± 0.03 a 2.01 ± 0.04 da 1.22 ± 0.06 a

1.76 ± 0.02 b 1.99 ± 0.02 cd

Titratable acidity (g/L tartaric acid) pH Brix Laccase activity (U/ mL) Solid-to-juice ratio (%) Seeds/berry Berry volume (cm3)

17-IV-2008 17-III-2008

Grapes were transported upon harvest to the INTA Wine Research Center experimental winery. Grapes were crushed and destemmed (Metal Liniers model MTL 12, Mendoza, Argentina), and pumped into 100-L stainless steel tanks. Sulphur dioxide (SO2) was dosed during crushing at a rate of 80 mg/L for all the experiments. The experimental design consisted of two maceration treatments for each of the 6 cultivars, replicated three times (n = 3). Initial must temperatures upon crushing were registered

Barbera D’Asti Cabernet Sauvignon Malbec Merlot Pinot Noir Syrah

2.2. Winemaking

Berry weight (g)

Own-rooted Barbera D’Asti, Cabernet Sauvignon, Malbec, Merlot, Pinot Noir and Syrah grapes (Vitis vinifera L.) were obtained from a commercial vineyard property of INTA located in Luján de Cuyo, Mendoza, Argentina (33° 000 S, 68° 510 W). For Barbera D’Asti, Malbec and Merlot, the trellis system was vertical shoot positioning, whereas Cabernet Sauvignon, Pinot Noir and Syrah were trellised in overhead canopy positioning (‘‘parral’’); yields ranged from 7.4 to 18.5 th/ha (Table 1). For each cultivar, a total of 700 kg were manually harvested to 18-kg plastic boxes at selected dates depending on the cultivar (Table 1). Visual inspection of the grapes revealed variable, albeit fairly low degrees of Botrytis cinerea damage, with Barbera D’Asti and Pinot Noir having about 5% of affected clusters, and with Cabernet Sauvignon and Syrah showing about 2–3% of affected clusters. For the grape basic analysis, four independent samples of 30 berries each were taken at harvest for each cultivar and analysed independently for berry weight and volume, seeds/berry, Brix (Atago, Tokyo, Japan), pH (Orion model 701-A, Thermo Scientific, Waltham, MA, USA), titratable acidity, and laccase activity (Dubourdieu, Grassin, Deruche, & RibereauGayon, 1984). The solid-to-juice ratio was computed as the percentage ratio between the weight of solids (skins and seeds) and the total liquid weight (i.e. pulp).

Yields (tn/ ha)

2.1. Grapes

Harvest date

2. Materials and methods

Cultivar

example, the low temperatures at which the must is typically kept during CS (5–10 °C) are thought to favour the metabolism of non-Saccharomyces yeast over that of Saccharomyces cerevisiae, resulting in positive modifications of the flavor profile of the wines (Charpentier & Feuillat, 1998). While the viability of non-Saccharomyces yeast during CS have been established (Casassa & Sari, in press; Zott, Miot-Sertier, Claisse, Lonvaud-Funel, & MasneufPomarede, 2008), their impact on the sensory profile of the wines is often negative. Published reports have found that CS increased the ethyl-acetate content (González-Neves et al., 2013) and the acetaldehyde character (Casassa & Sari, in press) of the finished wines, both compounds with negative sensory connotations. Likewise, the volatile composition of control and CS wines as determined by SPME GC–MS was statistically undistinguishable for 31 out of 33 volatiles in Cabernet Sauvignon wines (Gardner, Zoecklein, & Mallikarjunan, 2011), which casts doubt on a practical sensory impact of this technique. The aim of the present study was to assess the chemical, chromatic and sensory effects of CS applied to 6 different grape cultivars grown in Mendoza, Argentina. Towards this end, experimental lots of Barbera D’Asti, Cabernet Sauvignon, Malbec, Merlot, Pinot Noir and Syrah were produced in triplicate by applying a control maceration of 14 days and also a CS treatment in the wines of each cultivar. The wines were analysed for basic and phenolic chemistry and chromatic parameters both at press and after 1 year of bottle aging. Complementarily, a formal sensory analysis of the wines was conducted.

Table 1 Harvest date, yields, physical and chemical composition of Barbera D’Asti, Cabernet Sauvignon, Malbec, Merlot, Pinot Noir and Syrah grapes used for the winemaking treatments. For berry data, values represent the mean (±SEM) of four independent sample replicates taken at harvest (n = 30 berries).

L.F. Casassa et al. / Food Chemistry 174 (2015) 110–118

112

L.F. Casassa et al. / Food Chemistry 174 (2015) 110–118

across all control tanks and found to be 21 ± 2.4 °C (n = 18). Control wines were produced with a maceration length of 14 days at 24 ± 2.5 °C; cap management consisted of two daily full-volume pump-overs followed by two daily punch downs (morning and afternoon, 1 min each). Cold soak with solid CO2 (CS) consisted of 4 days of CS at 9 ± 2.5 °C achieved by periodic additions of CO2. During crushing, 13 kg of CO2 pellets (Praxair SA, Mendoza, Argentina) per 100 kg of destemmed grapes were added into the CS tanks, which resulted on a temperature drop of 10.5 °C on average. Afterwards, CO2 pellets were added regularly at a rate of 5 kg/tank/day followed by two daily punch-downs (1 min each). A total of 33 kg of CO2 per each tank processed with CS were used. The 4-day CS period was followed by a 10-day maceration period under the same conditions as those described for control wines. All tanks were inoculated 5 h after crush with a commercial yeast (EC-1118; Lallemand Inc., Copenhagen, Denmark) at a rate of 0.3 g/L, following a hydration protocol previously detailed (Casassa & Sari, in press). The titratable acidity (TA) of all musts was adjusted to 7 g/L with food-grade tartaric acid, except for Barbera D’Asti in which the TA in the grapes was >7 g/L. After maceration length was completed, malolactic fermentation (MLF) was induced by inoculation of a commercial Oenococcus Oeni culture (VP-41, Lallemand Inc.). After MLF (malic acid <0.20 g/L), the wines were racked, adjusted to 30 mg/L of free SO2, and stored at 1 °C for 45 days to allow tartaric stabilisation. After this period, the wines were racked off the lees and bitartrate crystals and brought to room temperature for 48 h. Prior to bottling, free SO2 was adjusted to ensure 0.5 mg/L of molecular SO2. The bottles were stored horizontally in the Wine Research Center cellar and maintained at 12 ± 1 °C until needed. 2.3. Wine basic analysis Alcohol content, titratable acidity (TA), volatile acidity (VA), pH, and glycerol content were measured with a FOSS Wine-Scan (FT-120) rapid-scanning infrared Fourier-transform spectrometer with WineScan software Version 2.2.1 (FOSS, Hillerod, Denmark). During fermentation, density at 20 °C was measured directly with a densimeter (Fite, Buenos Aires, Argentina). Reducing sugars were determined following the official reference method (INV, 2013). Malic acid content was determined enzymatically (Vintessential Laboratories, Victoria, Australia). Free and total SO2 levels were determined using the aspiration method (Iland, Bruer, Edwards, Caloghiris, & Wilkes, 2012).

had extensive experience on wine sensory analysis and were part of the staff of the INTA Wine Research Center and the Department of Enology of a local university, were convened. Wines were analysed after 3 months of bottle aging. A total of six evaluation sessions were held throughout the experiment, one for each wine cultivar, with two additional introductory sessions devoted to terminology development and attribute definition, following published specifications (IRAM, 1997). Panelists defined by consensus two colour attributes (colour intensity and violet hue), four aroma attributes (overall aroma, fresh fruit, jammy, vegetal), one taste attribute (bitterness) and two mouthfeel attributes (astringency and body) for each of which a definition and a standard, if applicable, were provided (Supplemental Table 1). The intensity of each attribute was assessed using a non-structured 120 mm line scale containing two reference points located at 12 mm of each end of the line, and results were decoded manually. Wines were presented in triplicate in ISO wine glasses (ISO, 1977) covered with plastic lids to trap volatiles, following a completerandomised design (Jackson, 2009). To minimise sensory carryover, panelists were asked to rinse their mouth with mineral water and eat a cracker between samples following a sip and spit protocol (Colonna, Adams, & Noble, 2004). 2.6. Data analysis Basic fruit physical and chemical compositions were analysed by a one-way analysis of variance (ANOVA). The basic, phenolic and chromatic composition of the wines were analysed both at pressing and after 1 year of bottle aging by a fixed-effect twoway ANOVA with interactions, including as main effects the grape cultivar and the maceration technique (control and CS), as well as the cultivar  maceration technique interaction. In addition, the full data set was reevaluated by one-way ANOVA on a cultivar and maceration technique basis (Supplemental Tables 3–5). In all cases, Fisher’s LSD test was used as a post-hoc comparison of means with a 5% level for rejection of the null hypothesis. Data analysis was performed using XLSTAT (ver. 2011; Addinsoft, Paris, France). The data generated by the sensory analysis were analysed on a cultivar basis by a Student’s T-test for independent samples (5% level for rejection of the null hypothesis). The whole data set was also evaluated by Principal Component Analysis (PCA), using the correlation matrix without rotation, employing the Infostat statistical package (version 2012, Cordoba, Argentina).

2.4. Spectrophotometric analysis

3. Results and discussion

Spectrophotometric measurements to evaluate phenolics and wine colour were performed at pressing and after 1 year of bottle aging. Total phenols (expressed as gallic acid equivalents (GAE), mg/L), anthocyanins (mg/L malvidin-3-glucoside) and tannins (mg/L (+)-catechin) were measured as previously described (Casassa & Sari, in press; Fanzone et al., 2012). Characterisation of wine colour was undertaken by means of the Cie-Lab system using MSCV™ software (Grupo de Color de La Rioja, Logroño, Spain). Prior to each measure, the samples were centrifuged 30 min to 1600 g (Gelectronic G-49, Buenos Aires, Argentina) and then filtered through a 0.45-lm membrane (Sartorius, Goettingen, Germany). Spectrophotometric measures were performed in a Perkin-Elmer Lambda 3B spectrophotometer (Norwalk, CT, USA).

3.1. Basic physical and chemical composition of the grapes

2.5. Sensory analysis Two panels, one of eleven individuals (seven males and four females), and the other of fourteen individuals (seven males and seven females), with ages ranging from 26 to 66 years, all of which

Table 1 shows the harvest date and yields for each of the 6 cultivars as well as the basic physical and chemical composition of the grapes used for winemaking. Consistent with expectations, berry size and volume were higher for Barbera D’Asti and Malbec, whereas Cabernet Sauvignon had the smallest berries. The average number of seeds/berry was determined due to its influence on the wine’s tannin content (Harbertson, Kennedy, & Adams, 2002). Pinot Noir had the highest number of seeds/berry and, accordingly, the highest solid-to-juice ratio. Laccase activity was detected in all the samples but it only reached potentially problematic levels (5 U/mL) in Barbera D’Asti grapes, which were last to be harvested, and probably endured the highest pressure of Botrytis cinerea infection. Concerning the basic chemistry of the grapes, Barbera D’Asti had the lowest pH and highest TA, whereas Merlot had the highest pH and lowest TA. Notwithstanding, at the moment of application of the winemaking treatments the grapes were at commercial maturity.

L.F. Casassa et al. / Food Chemistry 174 (2015) 110–118

3.2. Sugar consumption and temperature profiles during cold soak and alcoholic fermentation Fig. 1 shows the temperature profile and the evolution of sugar consumption (measured as density at 20 °C) during cold soak and alcoholic fermentation for control and CS treatments of the 6 wine cultivars. Temperatures during CS ranged from 4 to 15 °C, with an average minimum of 6.8 °C. Temperature differentials during CS were more prominent in Barbera D’Asti (10 °C) and Malbec

113

(6 °C), relative to the remaining wines (3.2 °C on average) but these differentials are nonetheless a common feature of CS conducted with solid CO2 (Casassa & Sari, in press; Parenti, Spugnoli, Calamai, Ferrari, & Gori, 2004). Temperature differentials are a common feature of CS because CO2 pellets tend to aggregate into blocks, which result in areas of frozen and unfrozen must in the fermentation tank (Parenti et al., 2004). In general, there was no sugar consumption during CS, with only Cabernet Sauvignon and Syrah showing marginal variations

Fig. 1. Temperature profile and evolution of sugar consumption during cold soak and alcoholic fermentation for the control and cold soak treatments of the 6 cultivars. (a) Barbera D’Asti; (b) Cabernet Sauvignon; (c) Malbec; (d) Merlot; (e) Pinot Noir; (f) Syrah. The cold soak period is shown in grey colour. If not shown, error bars are obscured by the treatment symbol.

114

L.F. Casassa et al. / Food Chemistry 174 (2015) 110–118

in the musts’ densities. Similarly, in a previous report, the application of CS with CO2 also avoided sugar consumption and thus fermentative metabolism during CS (Casassa & Sari, in press). All wines were fermented to dryness (<4 g/L residual sugars, Supplemental Table 2), during the 14-days of total maceration length for all treatments. 3.3. Basic chemical composition of the finished wines In general, the application of CS had no effect on the basic chemical composition of the wines: no significant differences between control and CS wines were found for ethanol content, pH, titratable acidity, malic acid, volatile acidity, residual sugars, and glycerol of the final wines. (Supplemental Tables 2 and 3). This reaffirms the concept that CS has a marginal effect on wines’ basic chemistry, as previously shown for Malbec wines produced with a similar protocol to that followed here (Casassa & Sari, in press). In the present work, we expand this finding to other wines of commercial and international recognition such a Cabernet Sauvignon, Merlot, Pinot Noir and Syrah. 3.4. Phenolic and chromatic composition of the finished wines The phenolic and chromatic composition of the wines was assessed at pressing and after 1 year of bottle aging to understand the immediate and long-term effect of CS on these parameters. At pressing (Table 2), a significant effect imparted by the cultivar was observed for all phenolic and chromatic parameters. Malbec and Syrah wines had the highest anthocyanin content, and Barbera D’Asti the lowest. For total phenols, the following trend was observed: Syrah > Cabernet Sauvignon, Malbec, Pinot Noir > Merlot > Barbera D’Asti. Accordingly, tannins were highest in Syrah and Pinot Noir wines and lowest in Barbera D’Asti wines. In general, the chromatic composition of the wines at pressing was in agreement with the observed anthocyanin content for each wine cultivar. For example, Pinot Noir and Barbera D’Asti had the lowest anthocyanin content, and, accordingly, the lowest saturation (C⁄) and red component (a⁄) and the highest lightness (L⁄). The application of CS only affected the chromatic composition but did not affect the anthocyanin and tannin content of the final wines. The lack thereof or even a negative effect of CS on the phenolic composition of the wines had been previously demonstrated in very dissimilar red wine cultivars such as Babic (Budic-Leto et al., 2003), Mencía (Pérez-Lamela et al., 2007), Sangiovese (Gambacorta et al., 2011), Syrah (González-Neves et al., 2013),

Tannat (González-Neves et al., 2012) and Tempranillo (Puertas, Guerrero, Jurado, Jimenez, & Cantos-Villar, 2008). Moreover, there are conflicting results on the effect of phenolic maturity on the outcome of CS whereby some authors have hypothesised an enhancement of phenolic extraction with unripe or low-phenolic grapes, whereas others indicate the exact opposite. Proponents of increased phenolic extraction by CS in unripe fruit argue that deficient phenolic extractability in unripe fruit may be overcome by the disrupting effect exerted by solid CO2 on cell walls (Llaudy, Zamora, Canals, Canals, & Cabanillas, 2005). Likewise, the same applies to cultivars or winegrowing conditions conducive to inherently low levels of phenolic compounds in their grapes and wines (Álvarez et al., 2006; Puertas, Jiménez, Cantos-Villar, & Piñeiro, 2013). Proponents of increased phenolic extraction by CS in ripe fruit argue that the extractive effect of solid CO2 is moderate at best, and is only realised when there is already an increased disruption of berry skin cell walls in ripe or overripe fruit (González-Neves et al., 2012; Ortega-Heras et al., 2012). In the present study, all 6 cultivars had sugar levels typical of ripe fruit, which are otherwise commonly achieved under regular winemaking conditions in Mendoza (Table 1). These results indicate a clear lack of a synergistic effect between fruit ripeness and the supposed CO2-mediated phenolic extraction under the experimental conditions herein described. Moreover, if CS is only effective at increasing phenolic extraction in low-phenolic cultivars but not in highphenolic fruit, the lack of positive results obtained for Cabernet Sauvignon, Malbec, Merlot and Syrah, which under Mendoza’s growing conditions typically yields wines high in phenolics (Fanzone et al., 2012) are to be expected. Herein, however, CS failed to enhance tannin and anthocyanin extraction in a low-phenolic grape such as Pinot Noir (Supplemental Table 4). This discrepancy had led us to hypothesise that under our experimental conditions, low temperatures during the 4 days of CS, which account for a portion of total maceration length (14 days), could have led to lower phenolic extraction relative to what would have been obtained at temperatures normally applied under regular maceration conditions (25–30 °C). Lower phenolic extraction is expected at lower maceration temperatures because both the permeability of the hypodermal cells releasing anthocyanins and the solubility of other phenolics in the wine matrix are lower at lower fermentation temperatures (Sacchi et al., 2005). In fact, an averaged (n = 6) integration of temperatures (by trapezoidal approximation) for control wines totaled 344 ± 20 °C, whereas for CS wines it totaled 295 ± 16 °C during the 14 days maceration lasted. Moreover, a Student’s T-test for independent samples revealed that the integrated

Table 2 Two-way ANOVA of the phenolic and chromatic composition at pressing of Barbera D’Asti, Cabernet Sauvignon, Malbec, Merlot, Pinot Noir and Syrah wines produced with a control and a cold soak treatment. Values represent the mean (±SEM) of three tank replicates. ANOVA parameter

Cultivar (cv.) Barbera D’Asti Cabernet Sauvignon Malbec Merlot Pinot Noir Syrah p-Value

Phenolic composition Anthocyanins (mg/L)

Total phenols (mg/L GAE)

Tannins (mg/L)

L⁄

133 ± 13 aa 238 ± 14 c 381 ± 52 d 213 ± 15 bc 163 ± 7 ab 366 ± 9 d <0.0001b

1523 ± 33 2449 ± 13 2540 ± 22 2073 ± 13 2505 ± 61 2834 ± 12 <0.0001

554 ± 45 a 955 ± 40 b 934 ± 52 b 1051.9 ± 37 b 1479 ± 72 c 1535 ± 62 c <0.0001

56.8 ± 2.1 37.7 ± 1.1 38.8 ± 0.9 50.4 ± 1.2 68.3 ± 1.2 38.9 ± 0.7 <0.0001

2302 ± 70 a 2211 ± 44 a 0.4039

1054 ± 94 a 1116 ± 77 a 0.1148

0.0014

0.0612

Maceration technique (MT) Control 231 ± 23 a Cold soak 266 ± 28 a p-Value 0.0666 cv.  MT interaction p-Value 0.2771 a b

Chromatic composition (Cie-Lab units)

a c c b c d

H⁄

a⁄

b d d c a d

1.7 ± 0.2 a 11.1 ± 1.2 d 3.7 ± 0.67 b 6.8 ± 0.2 c 4.3 ± 0.6 b 4.12 ± 0.6 b <0.0001

51.6 ± 2.1 62.6 ± 0.4 60.6 ± 0.4 56.4 ± 0.8 37.7 ± 1.4 61.3 ± 1.0 <0.0001

49.6 ± 2.7 b 47.4 ± 2.9 a 0.0004

54.3 ± 2.0 a 56.5 ± 2.4 b 0.0003

4.4 ± 0.6 a 6.17 ± 0.9 b <0.0001

54.1 ± 1.9 a 56.1 ± 2.3 b 0.0008

4.1 ± 0.8 a 6.2 ± 1.1 b 0.0003

<0.0001

<0.0001

0.0031

<0.0001

0.0211

Different letters within a column indicate significant differences for Fisher’s LSD test and p < 0.05. Significant p-values (p < 0.05) are shown in bold.

C⁄ c a a b d a

51.7 ± 2.1 63.9 ± 0.7 60.7 ± 0.4 56.8 ± 0.8 37.8 ± 1.4 61.5 ± 1.1 <0.0001

b⁄ b d d c a d

1.5 ± 0.3 a 12.4 ± 1.5 d 3.7 ± 0.7 ab 6.7 ± 0.3 c 2.8 ± 0.3 ab 3.8 ± 1.2 b <0.0001

115

L.F. Casassa et al. / Food Chemistry 174 (2015) 110–118

temperatures in control and CS wines were statistically different (p = 0.0003). The added negative effect of enzymatic oxidations during the CS period is also a possible factor explaining these results (Casassa & Sari, in press). As shown by the two-way ANOVA (Table 2), wines made with CS were slightly, though significatively, more saturated, and possessed a higher red colour component than control wines. Thus, the net effect of CS was to positively affect the chromatic properties of the wines at pressing, but interestingly, it had no effect on the anthocyanin and tannin content. This suggests that other phenolic-derived compounds, with effect on wine colour, may be responsible for some of these chromatic features favouring CS wines. Polymeric pigments are good candidates to explain these changes. Previously published research on Monastrel wines produced with CS had uncovered an enhanced formation of polymeric

anthocyanins as a result of CS (Álvarez et al., 2006). The polymeric pigment content, however, was not determined analytically in the present study. Table 3 shows a two-way ANOVA with the phenolic and chromatic composition of the wines after 1 year of bottle aging. Within all wine cultivars, an almost 3-fold drop (averaging 63%) in the anthocyanin content was observed from pressing to 1 year of bottle aging. Concomitantly, saturation and the red component decreased (23% and 18%, respectively) and lightness and hue increased (19% and 165%, respectively) during the same period. Tannins in the 1-year old wines decreased slightly (12%), but, overall, differences observed at pressing in the tannin content of the wines were maintained after bottle aging. Considering the maceration technique factor, CS increased anthocyanins by 22% relative to control wines, but had no effect on total phenols and

Table 3 Two-way ANOVA of the phenolic and chromatic composition after 1 year of bottle aging of Barbera D’Asti, Cabernet Sauvignon, Malbec, Merlot, Pinot Noir and Syrah wines produced with a control and a cold soak treatment. Values represent the mean (±SEM) of three tank replicates. ANOVA parameter

Cultivar (cv.) Barbera D’Asti Cabernet Sauvignon Malbec Merlot Pinot Noir Syrah p-Value

Phenolic composition Anthocyanins (mg/L)

Total phenols (mg/L GAE)

Tannins (mg/L)

L⁄

44 ± 4 aa 111 ± 19 bc 131 ± 22 c 102 ± 10 b 45 ± 2 a 118 ± 7 bc <0.0001b

1346 ± 16 2272 ± 12 2276 ± 44 1748 ± 18 2427 ± 40 2523 ± 12 <0.0001

360 ± 15 a 1046 ± 31 b 992 ± 53 b 1047 ± 39 b 1234 ± 84 c 1030 ± 39 b <0.0001

63.6 ± 1.9 41.2 ± 1.5 48.6 ± 1.1 63.6 ± 0.7 77.4 ± 1.1 46.9 ± 0.9 <0.0001

2090 ± 40 a 2103 ± 40 a 0.7262

962 ± 80 a 941 ± 63 a 0.5052

0.0021

0.0031

Maceration technique (MT) Control 83 ± 9 a Cold soak 101 ± 14 b p-Value 0.0132 cv.  MT interaction p-Value <0.0001 a b

Chromatic composition (Cie-Lab units)

a c c b d d

C⁄

H⁄

a⁄

b e d c a d

9.7 ± 0.7 c 14.7 ± 0.3 d 2.3 ± 0.6 a 28.9 ± 0.9 f 23.3 ± 1.4 e 6.1 ± 0.5 b <0.0001

37.2 ± 2.1 51.4 ± 1.1 74.4 ± 2.1 35.2 ± 0.8 21.4 ± 1.1 49.9 ± 0.6 <0.0001

57.4 ± 2.9 a 56.4 ± 3.2 a 0.1461

41.7 ± 2.4 a 42.9 ± 2.7 b 0.045

13.7 ± 2.1 a 14.6 ± 2.5 a 0.093

43.7 ± 3.9 a 46.1 ± 4.2 b 0.0089

9.2 ± 1.2 a 9.7 ± 1.5 a 0.1889

<0.0001

<0.0001

0.0003

0.0027

0.0393

c a b c d b

37.7 ± 2.1 53.2 ± 1.1 49.7 ± 1.1 40.2 ± 0.8 23.1 ± 1.1 50.1 ± 0.6 <0.0001

b⁄ b c d a b c

6.2 ± 0.1 b 13.6 ± 0.6 d 3.1 ± 0.8 a 19.4 ± 0.6 e 8.9 ± 0.1 c 5.4 ± 0.6 b <0.0001

Different letters within a column indicate significant differences for Fisher’s LSD test and p < 0.05. Significant p-values (p < 0.05) are shown in bold.

Fig. 2. Sensory analysis of control and cold soak wines of (a) Barbera D’Asti; (b) Cabernet Sauvignon; (c) Malbec; (d) Merlot; (e) Pinot Noir; (f) Syrah after 3 months of bottle aging. ‘‘n’’ indicates the number of panelist involved in each analysis. (*) indicates significant differences for a Student’s T-test and p < 0.05.

116

L.F. Casassa et al. / Food Chemistry 174 (2015) 110–118

tannins after 1 year of bottle aging. In line with a slightly higher anthocyanin content remaining in CS wines, both saturation and the red component of wine colour were also higher in CS wines after 1 year of bottle aging. Therefore, the positive effect of CS on the chromatic composition of the wines was long lasting. As the formation of polymeric pigments is known to provide stable colour with spectral and sensory features different from that of the intact anthocyanins and tannins (Casassa & Harbertson, 2014), future studies along these lines that compare CS to control treatments are worth considering. Both at pressing and after 1 year of bottle aging, the two corresponding two-way ANOVAs (Tables 2 and 3), indicated that, with the only exception of the anthocyanin and tannin content at pressing, the cultivar  maceration technique interaction was always significative. The significance of these interactive effects suggests that the outcome of CS, as it relates to the phenolic and chromatic composition of the wines, is contingent upon the specific cultivar to which CS is applied. For example, a breakdown of the results by a one-way ANOVA showed that at pressing, CS increased wine

colour saturation (C⁄) in Barbera D’Asti, Cabernet Sauvignon, Merlot and Syrah, but the opposite was observed for Pinot Noir (Supplemental Table 4). After 1 year of bottle aging, lower colour saturation was also observed in Malbec wines processed with CS. Overall, this cultivar-specific effect of CS has been previously observed by González-Neves et al. (2012), who suggested that other factors such as berry size and the composition of the skin cell walls can modify the outcome of CS. Indeed, the specific composition of the skin cell walls has been found to affect both anthocyanin (Ortega-Regules, Ros-García, Bautista-Ortín, López-Roca, & Gómez-Plaza, 2008), and tannin extraction (reviewed in: Hanlin, Hrmova, Harbertson, & Downey, 2010). Specifically for anthocyanins, high extractability is predicted in cell walls with low concentrations of galactose, cellulose, rhamnose, xylose, and a low degree of pectin methylation (Ortega-Regules, Romero-Cascales, RosGarcía, López-Roca, & Gómez-Plaza, 2006). The results presented herein suggest again a cultivar-specific effect of CS, positive from the perspective of wine colour for Barbera D’Asti, Cabernet Sauvignon, Merlot and Syrah, and negative for Malbec and Pinot Noir.

Fig. 3. Principal Component Analysis carried out including the chemical and sensory data of the control and cold soak wine treatments of 6 cultivars. Only principal components (PC) with eigenvalues > 1 were retained for the analysis. (a) PC1 vs. PC2 and (b) PC1 vs. PC3.

L.F. Casassa et al. / Food Chemistry 174 (2015) 110–118

3.5. Sensory analysis The full set of wines, including the 6 wine cultivars and the two winemaking treatments, were analysed by a trained sensory panel after 3 months of bottle aging. The only positive sensory effect of CS was to enhance perceived colour intensity in Barbera D’Asti and Cabernet Sauvignon wines, but this effect was absent for the remaining wines (Fig. 2). Moreover, for Pinot Noir, the opposite was found, i.e. CS was detrimental for perceived colour. This last result is in line with the observed effects of CS on anthocyanins, colour saturation and the red component of Pinot Noir wines (Tables 2 and 3). Generally, the application of CS had no effect on perceived aroma, taste (bitterness), and mouthfeel sensations (astringency and body) for any of the 6 wine cultivars. Previous reports on the effect of CS on the sensory profile of the wines are often conflicting. This is to be expected, as CS has been typically applied to different cultivars, fruit maturity, and winemaking conditions (length, temperature, yeast inoculation, use of CO2 vs. external refrigeration). However, there seems to be consensus among selected studies that the overall sensory impact of CS is moderate at best, particularly when formal sensory techniques are applied to evaluate the wines. For example, a consumer panel (n = 54), evaluated control and CS wines (Cabernet Sauvignon) using a triangle test and failed to detect any differences between treatments (Gardner et al., 2011). In Pinot Noir wines produced with and without CS, a quantitative descriptive sensory analysis (n = 12) found no effect of CS on most aroma descriptors, with wines produced with CS at 10 °C being particularly bitter (Goldsworthy, 1993). In Malbec wines produced with CS with and without CO2, a sensory panel (n = 10), evaluated the wines from pressing up to 24 months of bottle aging (Casassa & Sari, in press). Relative to control wines without CS, this study found that CS without CO2 resulted in wines with lower colour and a noticeable acetaldehyde character, whereas CS with CO2 produced wines with less fruity character than control wines. All these three cultivars, Cabernet Sauvignon, Malbec and Pinot Noir, were also studied here, with results consistent with those detailed above. 3.6. Principal Component Analysis (PCA) Fig. 3 shows a PCA carried out including both the relevant chemical and sensory data. Three components with eigenvalues > 1 were retained for the analysis, accounting together for 76% of the variability of the data set. The first principal component (Fig. 3a), which explained 34.3% of the variability, separated the wines as a function of the wine cultivar, with Malbec, Syrah and Cabernet Sauvignon clustered together in the negative dimension of PC1, whereas Barbera D’Asti, Merlot, and Pinot Noir were located in the positive dimension of PC1. On the other hand, PC2 and PC3, which explained 23.6% and 16.5% of the variability, respectively, separated the wines as a function of the maceration technique. However, the effect of CS was not the same for all the wines, as previously observed for the phenolic and chromatic composition (Tables 2 and 3). For example, Fig. 3b shows that CS increased the anthocyanin content and colour intensity in Malbec, Merlot and Cabernet Sauvignon wines, but it caused the opposite effect on Pinot Noir and Syrah. In light of the higher contribution of PC1 to the total variance relative to PC2 and PC3, it can be concluded that the application of CS was a lesser factor in defining the chemical and sensory composition of the wines. Conversely, the wine cultivar was the driving factor on defining those features. In similar published studies in which more than one cultivar were produced with CS, the wine cultivar have also been found to be the determining factor in defining the chemical and sensory composition of the wines, over other factors such as the grape maturity, winemaking technique, and aging status of the wines (Busse-Valverde et al., 2010; Gil-Muñoz et al., 2009;

117

González-Neves et al., 2013; Moreno-Pérez, Vila-López, Fernández-Fernández, Martínez-Cutillas, & Gil-Muñoz, 2013). Similarly, the overall results of this study suggest that the grape cultivar is the most important driver of the chemical and sensory makeup of the wines, both at pressing and after 1 year of bottle aging. 4. Conclusions This study evaluated the chemical and sensory effects of CS conducted with solid CO2 in 6 different red wine cultivars from Mendoza, Argentina. Temperature differentials up to 10 °C within a single tank were registered during CS, which were ascribed to the aggregating effect of solid CO2 when applied to the must. Non-homogeneous distribution of temperature within a tank implies that only a portion of the must is treated, and while the aggregating effect of CO2 appears to be an intrinsic property derived from its use in winemaking, this effect can be minimised by regular punch-downs during CS. Cold soak had no effect on the ethanol content, pH, titratable acidity, volatile acidity, residual sugars, glycerol, anthocyanins and tannins of the wines. Conversely, CS had a positive effect on the colour saturation and red component of the wines. From a sensory standpoint, the only positive effect of CS was to enhance colour intensity in Barbera D’Asti and Cabernet Sauvignon wines, but had no effect on aroma attributes, bitterness, astringency, and body. Both a two-way ANOVA and a Principal Component Analysis suggested that the outcome of CS is contingent upon the specific cultivar to which CS is applied. In a current global context of both low-input viticulture and enology, the use of CO2 to conduct CS in red grapes appears to be of little merit, let aside cost, safety, storage and handling issues associated with its use in the winery. The results herein presented suggest that CS conducted with solid CO2 in the form of pellets and applied to ripe grapes of 6 different red cultivars has only a moderate (albeit positive) effect on the chromatic properties of Barbera D’Asti and Cabernet Sauvignon wines, but had no effect on the chemical and sensory composition of Malbec, Merlot, Pinot Noir and Syrah wines. Therefore, the application of CS causes slight to minor changes in the chemical and sensory make-up of the wines, which most likely will not outweigh the cost and logistics of its application. In light of the minor sensory changes induced by CS, we would like to hypothesise that these sensory changes may scape consumer’s notice. Acknowledgements Members of the sensory panels are acknowledged for their commitment with this study. Dr. Richard Larsen (Viticulture and Enology Program, Washington State University, Prosser, Washington, USA), and Dr. Greg Gasic are thanked for careful review of the manuscript and helpful suggestions. 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.2014. 10.146. References Álvarez, I., Aleixandre, J. L., García, M. J., & Lizama, V. (2006). Impact of prefermentative maceration on the phenolic and volatile compounds in Monastrell red wines. Analytica Chimica Acta, 563(1–2), 109–115. Budic-Leto, I., Tomislav, L., & Vrhovsek, U. (2003). Influence of different maceration techniques and ageing on proanthocyanidins and anthocyanins of red wine cv. Babic (Vitis vinifera, L.). Food Technology and Biotechnology, 41, 203–299. Busse-Valverde, N., Go´mez-Plaza, E., Lo´pez-Roca, J. M., Gil-Mun~oz, R., Ferna´ndezFerna´ndez, J. I., & Bautista-Orti´n, A. B. (2010). Effect of different enological

118

L.F. Casassa et al. / Food Chemistry 174 (2015) 110–118

practices on skin and seed proanthocyanidins in three varietal wines. Journal of Agricultural and Food Chemistry, 58(21), 11333–11339. Casassa, L. F. (2007). Efecto de dos variantes de maceración prefermentativa en frío (MPF) sobre la composición y características organolépticas de vinos Malbec. MS thesis. Facultad de Ciencias Agrarias – Universidad Nacional de Cuyo, Mendoza, Argentina. Casassa, L. F., & Sari, S. E. (in press). Sensory and chemical effects of two alternatives of prefermentative cold soak in Malbec wines during winemaking and bottle ageing. International Journal of Food Science & Technology Casassa, L. F., & Harbertson, J. F. (2014). Extraction, evolution, and sensory impact of phenolic compounds during red wine maceration. Annual Review of Food Science and Technology, 5(1), 83–109. Charpentier, C., & Feuillat, M. (1998). Métabolisme des levures cryotolérantes: ápplication à la maceration préfermentaire du Pinot noir en Bourgogne. Revue Francaise d’Oenologie, 170(170), 34–40. Colonna, A. E., Adams, D. O., & Noble, A. C. (2004). Comparison of procedures for reducing astringency carry-over effects in evaluation of red wines. Australian Journal of Grape and Wine Research, 10(1), 26–31. Delteil, D. (2004). La Macération Préfermentaire à Froid (MPF) des raisins méditerranéens et rhodaniens. Revue des Œnologues, 112(112), 29–32. Dubourdieu, D., Grassin, C., Deruche, C., & Ribereau-Gayon, P. (1984). Mise au point d´une mesure rapide de l´activité laccase dans les moûts et dans les vins par la méthode a la syringaldazine. Application a l´appréciation de l´état sanitaire des vendanges. Connaissance de la Vigne & du Vin, 18, 237–252. Fanzone, M., Zamora, F., Jofré, V., Assof, M., Gómez-Cordovés, C., & Peña-Neira, Á. (2012). Phenolic characterisation of red wines from different grape varieties cultivated in Mendoza province (Argentina). Journal of the Science of Food and Agriculture, 92(3), 704–718. Gambacorta, G., Antonacci, D., Pati, S., la Gatta, M., Faccia, M., Coletta, A., & La Notte, E. (2011). Influence of winemaking technologies on phenolic composition of Italian red wines. European Food Research and Technology, 233(6), 1057–1066. Gardner, D. M., Zoecklein, B. W., & Mallikarjunan, K. (2011). Electronic nose analysis of Cabernet Sauvignon (Vitis vinifera L.) grape and wine volatile differences during cold soak and postfermentation. American Journal of Enology and Viticulture, 62(1), 81–90. Gil-Muñoz, R., Moreno-Pérez, A., Vila-López, R., Fernández-Fernández, J., MartínezCutillas, A., & Gómez-Plaza, E. (2009). Influence of low temperature prefermentative techniques on chromatic and phenolic characteristics of Syrah and Cabernet Sauvignon wines. European Food Research and Technology, 228(5), 777–788. Goldsworthy, S. A. (1993). Pre-fermentation maceration of Pinot Noir wine. MS thesis, Lincoln University. González-Neves, G., Gil, G., Favre, G., Baldi, C., Hernández, N., & Traverso, S. (2013). Influence of winemaking procedure and grape variety on the colour and composition of young red wines. South African Journal of Enology and Viticulture, 34(1), 138–146. González-Neves, G., Gil, G., Favre, G., & Ferrer, M. (2012). Influence of grape composition and winemaking on the anthocyanin composition of red wines of Tannat. International Journal of Food Science & Technology, 47(5), 900–909. Gordillo, B., Lo´pez-Infante, M. I., Rami´rez-Pe´rez, P., Gonza´lez-Miret, M. L., & Heredia, F. J. (2010). Influence of prefermentative cold maceration on the color and anthocyanic copigmentation of organic Tempranillo wines elaborated in a warm climate. Journal of Agricultural and Food Chemistry, 58(11), 6797–6803. Hanlin, R. L., Hrmova, M., Harbertson, J. F., & Downey, M. O. (2010). Review: Condensed tannin and grape cell wall interactions and their impact on tannin extractability into wine. Australian Journal of Grape and Wine Research, 16(1), 173–188. Harbertson, J. F., Kennedy, J. A., & Adams, D. O. (2002). Tannin in skins and seeds of Cabernet Sauvignon, Syrah, and Pinot Noir berries during ripening. American Journal of Enology and Viticulture, 53(1), 54–59. Hernández-Jiménez, A., Kennedy, J. A., Bautista-Ortín, A. B., & Gómez-Plaza, E. (2012). Effect of ethanol on grape seed proanthocyanidin extraction. American Journal of Enology and Viticulture, 63(1), 57–61.

Iland, P., Bruer, N., Edwards, G., Caloghiris, S., & Wilkes, E. (2012). Chemical analysis of grapes and wine techniques and concepts (2nd ed.). Adelaide, Australia: Patrick Iland Wine Promotions Pty Ltd. INV (Instituto Nacional de Vitivinicultura) (2013). Argentinean Government. Resolución mayo de 1938 - Dirección Nacional de Química. URL: Accessed 09.12.13. IRAM (1997). Norma Argentina IRAM 20012: análisis sensorial: métodos para determinar el perfil del flavor. ISO (1977). ISO 3591. Sensory analysis-apparatus: Wine tasting glass. Geneva, Switzerland: International Organization for Standardization. Jackson, R. S. (2009). Wine tasting: A professional handbook (2nd ed.). San Diego, CA: Academic Press. Llaudy, M., Zamora, F., Canals, R., Canals, J., & Cabanillas, P. (2005). La maceración prefermentativa en frío: efectos sobre la extracción de color y los compuestos fenólicos e influencia del nivel de maduración de la uva. ACE Revista de Enología. Accessed 24.10.14. Marais, J. (2003). Effect of different wine-making techniques on the composition and quality of Pinotage wine. I. Low-temperature skin contact prior to fermentation. South African Journal of Enology and Viticulture, 24(2), 70–75. McMahon, H. M., Zoecklein, B. W., & Jasinski, Y. W. (1999). The effects of prefermentation maceration temperature and percent alcohol (v/v) at press on the concentration of Cabernet Sauvignon grape glycosides and glycoside fractions. American Journal of Enology and Viticulture, 50(4), 385–390. Moreno-Pérez, A., Vila-López, R., Fernández-Fernández, J. I., Martínez-Cutillas, A., & Gil-Muñoz, R. (2013). Influence of cold pre-fermentation treatments on the major volatile compounds of three wine varieties. Food Chemistry, 139(1–4), 770–776. Ortega-Heras, M., Pérez-Magariño, S., & González-Sanjosé, M. L. (2012). Comparative study of the use of maceration enzymes and cold prefermentative maceration on phenolic and anthocyanic composition and colour of a Mencía red wine. LWT – Food Science and Technology, 48(1), 1–8. Ortega-Regules, A., Romero-Cascales, I., Ros-García, J. M., López-Roca, J. M., & Gómez-Plaza, E. (2006). A first approach towards the relationship between grape skin cell-wall composition and anthocyanin extractability. Analytica Chimica Acta, 563(1–2), 26–32. Ortega-Regules, A., Ros-García, J. M., Bautista-Ortín, A., López-Roca, J. M., & GómezPlaza, E. (2008). Differences in morphology and composition of skin and pulp cell walls from grapes (Vitis vinifera L.): Technological implications. European Food Research and Technology, 227(1), 223–231. Parenti, A., Spugnoli, P., Calamai, L., Ferrari, S., & Gori, C. (2004). Effects of cold maceration on red wine quality from Tuscan Sangiovese grape. European Food Research and Technology, 218(4), 360–366. Pérez-Lamela, C., García-Falcón, M. S., Simal-Gándara, J., & Orriols-Fernández, I. (2007). Influence of grape variety, vine system and enological treatments on the colour stability of young red wines. Food Chemistry, 101(2), 601–606. Puertas, B., Guerrero, R. F., Jurado, M. S., Jimenez, M. J., & Cantos-Villar, E. (2008). Evaluation of alternative winemaking processes for red wine color enhancement. Food Science and Technology International, 14(5), 21–27. Puertas, B., Jiménez, M. J., Cantos-Villar, E., & Piñeiro, Z. (2013). Effect of the dry ice maceration and oak cask fermentation on colour parameters and sensorial evaluation of Tempranillo wines. International Journal of Food Science & Technology, 48(4), 835–842. Reynolds, A., Cliff, M., Girard, B., & Kopp, T. G. (2001). Influence of fermentation temperature on composition and sensory properties of Semillon and Shiraz wines. American Journal of Enology and Viticulture, 52(3), 235–240. Sacchi, K. L., Bisson, L. F., & Adams, D. O. (2005). A review of the effect of winemaking techniques on phenolic extraction in red wines. American Journal of Enology and Viticulture, 56(3), 197–206. Zott, K., Miot-Sertier, C., Claisse, O., Lonvaud-Funel, A., & Masneuf-Pomarede, I. (2008). Dynamics and diversity of non-Saccharomyces yeasts during the early stages in winemaking. International Journal of Food Microbiology, 125(2), 197–203.