Monitoring the process to obtain red wine enriched in resveratrol and piceatannol without quality loss

Food Chemistry 122 (2010) 195–202

Contents lists available at ScienceDirect

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

Monitoring the process to obtain red wine enriched in resveratrol and piceatannol without quality loss Raúl F. Guerrero a, Belén Puertas a, Maria J. Jiménez a, Juan Cacho b, Emma Cantos-Villar a,* a b

IFAPA Rancho de la Merced. Ctra. Trebujena, Km. 3.2. Apdo. 589, CP 11471 Jerez de la Frontera, Cadiz, Spain Department of Analytical Chemistry, University of Zaragoza, 50009 Zaragoza, Spain

a r t i c l e

i n f o

Article history: Received 29 August 2009 Received in revised form 12 January 2010 Accepted 24 February 2010

Keywords: GC–olfactometry Oenological parameter Piceatannol Postharvest UVC Quality red wines trans-Resveratrol Viniferins

a b s t r a c t Stilbene-enriched wine is considered an interesting new food product with added value as a consequence of the numerous health-promoting properties ascribed to it, mainly from its trans-resveratrol content. Postharvest grapes have been treated with ultraviolet-C light to produce stilbene-enriched grapes that were later used in a conventional winemaking process to obtain a red wine enriched in stilbenes. By measuring oenological parameters and stilbene concentration it has been possible to monitor both the quality parameters and stilbenes throughout the process. The maximum concentration in trans-resveratrol and piceatannol was obtained after pressing, but there was significant loss from grape to wine. A significant increase in both piceatannol and trans-resveratrol concentration (up to 26 times and 3.2 times higher than in control, respectively) was achieved in bottled wine. Regarding the oenological parameters, the wines obtained possessed good quality, apart from a herbaceous aroma, which could not be identified by GC–olfactometry. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Numerous epidemiological studies have shown that long-term, moderate consumption of wine is linked to a lower level of cardiovascular illnesses. In 1992 a study conducted by Renaud and De Lorgeril revealed that the incidence of heart infarction in France is about 40% lower than in the rest of Europe; this was termed the ‘‘French paradox”, which appeared to be related to regular consumption of red wine (Renaud & De Lorgeril, 1992). Numerous other beneficial qualities, with positive effects on health, have been attributed to wine, including anti-oxidant, anti-carcinogenic and anti-spasmodic properties, enhancement or activation of bile secretion, antibacterial and antihistaminic agents (Pignatelli et al., 2006). The findings that red wine possesses more health-promoting activity than beer or spirits have led to research attention being focused on phenolic compounds; within this group, stilbenes (in particular, trans-resveratrol) seem to show high bioactivity. In the last few years, interest in resveratrol has increased greatly, due to its numerous health-promoting properties (according to the ISI Web of Knowledge, there have been 4768 bibliographic entries in the last 10 years). Resveratrol has been described as a compound capable of preventing or reducing a wide

* Corresponding author. Tel.: +34 671560353; fax: +34 956034610. E-mail address: [email protected] (E. Cantos-Villar). 0308-8146/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodchem.2010.02.057

range of diseases, such as cancer (Jang et al., 1997), cardiovascular diseases, and ischaemic damage (Bertelli, 2007); it can also increase the resistance to stress and prolong the lifespan of diverse organisms, from yeast to vertebrates (Guerrero, García-Parrilla, Puertas, & Cantos-Villar, 2009). Studies in mice have shown that obese animals, whose diet was supplemented with resveratrol, not only lived longer, but were more active and produced fewer cases of the negative effects of a high-calorie diet (Baur et al., 2006). Regarding bioavailability, numerous studies in animals and humans have shown that resveratrol is metabolised with relative difficulty. However, a relatively low dose of resveratrol obtained regularly from red wine or other dietary sources could be therapeutic in some cases (Bertelli, Bertelli, Gozzini, & Giovannini, 1998). Apart from resveratrol, the presence in wine of other stilbenes, such as piceid, astringin, piceatannol and viniferins, has been described. Piceatannol is rarely found in wines and its presence in wines is very interesting since it is bioactive and has a long plasma half-life. It exhibits a pronounced anti-oxidant activity and exerts immunosuppressive, anti-leukaemic, and anti-tumorigenic activities in various cell lines and animal models (Murias et al., 2005). It has also been demonstrated that viniferins present anti-inflammatory and anti-proliferative activities (Kitanaka et al., 1990). All these stilbenes show high bioactivity but they are present in concentrations even lower than that of resveratrol, and less research has been done on them.

196

R.F. Guerrero et al. / Food Chemistry 122 (2010) 195–202

Dietary sources of resveratrol are rather limited, grapes and their derivatives being the main source (Guerrero et al., 2009). Resveratrol is found in the seed and skin of grapes (not in the flesh) and, hence, in grape juice and wine. Its concentration in red wine is higher than in white one because in red winemaking the must, grape skin and seed are in contact during the whole fermentation process. The amount of resveratrol in wines varies widely depending on many factors: grape variety, geographic region, agronomic factors, climate, plant stress conditions and oenological practices all have an influence (Bavaresco, 2003). Regarding optimum oenological practices for increasing resveratrol content, all the processes that maximize the extraction of phenols from skin are recommended (Vrhovsek, Wendelin, & Eder, 1997). It is difficult to predict the amount of resveratrol in wines because of many factors affecting resveratrol biosynthesis. Different strategies can be undertaken to increase the concentration of stilbenes in grape, since these compounds are phytoalexins and, therefore, can be induced via different kinds of stress. In particular, postharvest treatment of grapes by ultraviolet-C light (UVC) has been suggested as suitable stress-promoting technology to obtain stilbene-enriched wine under laboratory conditions (Cantos, Espín, & Tomás-Barberán, 2003); however, the validity of the process needs to be confirmed in a standard-scale experimental winery as a prior step to industrial-scale production. Stilbene-enriched wines potentially offer added value compared to traditional wines (Barreiro, Colombo, & Cantos, 2008). In this study, the winemaking process has been monitored both by oenological parameters and by content in stilbene compounds, to obtain a stilbene-enriched wine of standard quality, using grapes enriched by postharvest UVC treatment. In addition, the organoleptic characteristics of the resulting wine were also investigated using a tasting panel and GC–olfactometry. 2. Materials and Methods 2.1. Reagents trans-Resveratrol and piceatannol were purchased from Sigma– Aldrich (Madrid, Spain). The trans isomers were transformed to cis forms under UV light. Acetic acid, dichloromethane, formic acid and methanol, all of analytical grade, were supplied by Panreac (Barcelona, Spain). Ultrapure water from a Milli-Q system (Millipore Corp., Bedford, MA) was used throughout this research. 2.2. Grapevine The local red grape variety Jaen tinto (Vitis vinifera) was grown in Jerez de la Frontera (IFAPA, Rancho de la Merced). Ripeness of the grape (sugars, acids, pH and berry weight) was monitored to determine the optimum harvest date (data not shown). 2.3. UVC treatment and storage period The UVC was carried out according to the procedure described by the patent WO/2002/085137; ES 2177465. Briefly, grapes were irradiated inside a system of 34 UV-C lamps (G30T8; Silvania, Danvers, MA) – 17 each on top and bottom panels – with a theoretical power of 510 W and an average flow velocity of 14.72 mW/cm2 (Vilber Lourmat VLX 254 radiometer; Vilber Lourmat, Mame-laValle, France) for 60 s at 42 cm. The parameter ‘‘maximum day” (Dm) was defined as the number of days elapsed after UVC treatment to achieve the maximum trans-resveratrol concentration in grapes. Harvested grapes were divided into three batches (Fig. 1): the first batch (named CT–CT) was processed immediately after harvesting, starting winemaking without any prior treatment or

storage of grapes. The second batch (CT) was stored at 20 °C (without irradiation) after harvesting until Dm was reached. The third batch (UV) was irradiated with UVC after harvesting and stored at 20 °C until Dm. The temperature of 20 °C was found to be the optimum for obtaining a balance between the synthesis of stilbenes and the deterioration of grape quality (unpublished data). Each batch was processed in triplicate.

2.4. Red winemaking process The grapes were de-stemmed, crushed and placed in a 50-l steel vessel (the CT–CT batch immediately after harvesting and the other two batches after being stored for 3 days) Pectolytic enzymes (3 g/ 100 kg, Vinozym Vintage FCE, Novozymes, Bordeauz, France) and sulfur dioxide (70 mg/kg) were added to maximize extraction and to protect the must. One day later, fermentation was started after yeasting (Actiflore F5, Laffort, Spain), and temperature was maintained at 27 ± 1 °C during alcoholic fermentation (AF). As soon as the tumultuous fermentation had finished (density 999 g/l), the wine was vatted: the free-run wine was decanted and the solid parts were placed in a pneumatic press (Willmes, Germany) to obtain the press-run wine, which was mixed with the free-run wine to form the press wine. For the malolactic fermentation, lactic bacteria Oenococcus lacti (1 g/hl, Challenge Easy ML, Sepsa-Enartis, Spain) and nutrients (20 g/hl Nutriferm ML, Sepsa-Enartis, Spain) were used. When this stage of the process was finished, the wine was decanted into another vessel, eliminating the first lees (racking), and wines were rendered inert (using N2) and stored in a cold chamber (at 0 °C) until clarification (using 10 g/hl egg white albumin, Laffort, Bordeaux, France). Finally, the wine of each batch was bottled. A complete diagram of the process is shown in Fig. 1. During each step of the process, both the oenological parameters and the content in stilbenes were determined. In addition, samples were taken from both the skins (solid phase) and the must-wine (liquid phase) for analysis of the stilbene content on each day of the AF.

2.5. Stilbene extraction Stilbenes were extracted from solid samples (grape skin and pomace) according to the procedure previously described by authors Guerrero, Puertas, Fernández, Palma, and Cantos (2010). Regarding liquid samples (must, lees and wines) stilbene extraction was performed following other authors (Bavaresco, Pezzutto, Ragga, Ferrari, & Trevisan, 2001). All extractions were conducted in triplicate, in the dark and at low temperature to avoid the occurrence of oxidation and isomerisation processes.

2.6. UPLC and HPLC analysis Stilbenes were identified using an ACQUITY Ultra Performance LCTM system and were quantified by HPLC-DAD. The methods followed have recently been described in detail by Guerrero et al. (2010). Briefly, stilbenes were identified using an ACQUITY Ultra Performance LCTM system (Waters, Milford, MA) linked simultaneously to both a PDA 2996 diode array detector (Waters) and an ACQUITY triple quadrupole mass spectrometer (Waters MS Technologies, Manchester, UK), equipped with an electrospray ionisation (ESI) source operating in negative mode. MassLynxTM software (Version 4.0, Waters) was used to control the instruments, and for data acquisition and processing. In order to quantify stilbenes, a Waters HPLC system with a Model 1525 pump, a Waters 996 diode array detector and a Waters 474 fluorescence detector was used.

197

R.F. Guerrero et al. / Food Chemistry 122 (2010) 195–202

GRAPE CTCT-CT

CT

No storage

UV UVC treatment

Storage for three days

Enzymes + SO2

De-stemming and Crushing

Yeast inoculation + Cap punched down daily

Alcoholic Fermentation (AF)

Stems

Vatting and Pressing Bacteria inoculation + Nutrients

Albumin + Cold stabilisation

Pomace

Malolactic Fermentation Racking + Inertization

1ª Lees

Clarification

2ª Lees

Bottling

CT-CT wine

CT wine

UV wine

Fig. 1. Diagram of grape treatment and winemaking applied to the three batches of grapes.

2.7. Oenological parameters of grape and wines Ethanol, density, total and volatile acidity, pH, organic acids, dry extract, sugars, glycerine, metals (Cu, Fe, Zn, Mg, and K), volatile compounds, total polyphenols (IPT), colour intensity, hue, anthocyanins and tannins were determined in bottled wine according to the official methods (OIV, 1990). Colour parameters were measured by spectrophotometry and CIELab with a Konica-Minolta 3.600-D colourimeter. Acetic and gluconic acids were measured using enzymatic tests (Boehringer-Mannheim, Germany).

flow splitter to the column exit. The column was a DB-WAX from J&W (Folsom, CA), 30 m  0.32 mm I.D., with 0.5 lm film thickness and was preceded by a 3 m  0.32 mm I.D. uncoated (deactivated, intermediate polarity) precolumn from Supelco (Bellefonte, PA). The method of analysis is described by Culleré et al. (2009). The most significant odorants (Modified Frequency percentage, MF >30%) were identified by comparison of their odours and their retention index and, whenever possible, by comparison of their chromatographic retention properties and MS spectra with those of pure reference compounds.

2.8. Sensory characterization

2.10. Statistical analysis

An expert wine panel (7 judges) applied seven of the sensory descriptors that are most likely to be affected by the winemaking process: colour intensity, hue, aroma intensity quality, persistence, flavour intensity to the mouth and flavour quality. A score from 0 to 10 was given to each descriptor depending on the perception of each expert taster.

The results were subjected to analysis of variance, using Latin Square Design. The statistical analysis was carried out using the statistics programme STATISTIX 8.0. Significance of reported results was evaluated at p < 0.05.

2.9. Gas chromatography–olfactometry (CG–O)

3.1. Stilbene content and oenological parameters in grape

The volatiles of the macerates were collected using a purgeand-trap system following the headspace strategy proposed by Culleré, Cacho, & Ferreira (2009). GC–O was carried out using a Thermo 8000 Series GC equipped with an FID and a sniffing port (ODO-1 from SGE, Ringwood, Victoria, Australia), connected by a

Grapes were sampled weekly, to monitor ripeness from veraison until harvesting (data not shown). Grapes were harvested at the appropriate stage of ripeness, while their capacity for the induction of resveratrol was preserved (Jeandet, Bessis, & Gautheron, 1991). As described in the Materials and Methods section

3. Results and discussion

198

R.F. Guerrero et al. / Food Chemistry 122 (2010) 195–202

(Fig 1), grapes were separated into three batches (CT–CT, CT and UV). The CT–CT grapes were de-stemmed and crushed immediately, while the CT and UV grapes were stored in tanks at 20 °C to preserve grape quality until the day of optimum induction (Dm), which was three days for the Jaen tinto variety. On the day of harvesting, CT–CT grapes contained a very low concentration of trans-resveratrol (0.53 mg/kg of fresh weight, fw) and trace amounts of viniferins (Table 1). Grapes grown under warm climate conditions usually possess a low stilbene concentration (Bavaresco, 2003). Differences found between CT–CT and CT grapes were not significant. In contrast, UV grapes presented significantly higher content of piceatannol, trans-resveratrol, e- and d-viniferin, identified by UPLC–DAD-MS–MS (Guerrero et al., 2010). In the UV grapes concentrations of piceatannol, trans-resveratrol and viniferins were 7.4, 9.6 and 4.7 times higher, respectively, than in the CT grapes (Table 1). Unlike the considerable differences found in stilbene content, no significant differences between the three batches were found in any of the oenological parameters analysed in the must when grapes were de-stemmed and crushed. Brix degree, total acidity, pH, tartaric and malic acids, potassium, anthocyanins, tannins and IPT showed correct values and non-significant differences between the batches (Table 1). Gluconic and acetic acids were measured daily to observe any possible deterioration in quality, since these acids are associated with infection by Botrytis cinerea and acetic bacteria, respectively. As observed in Table 1, all the grapes showed a healthy condition. The significance of these data is that, in this study, it has been demonstrated that grape quality was not adversely affected by postharvest UVC treatment (UV batch) nor by the period of pre-processing storage required (CT batch). As previous experiments have confirmed (unpublished data), good quality of such grapes enriched in stilbenes is a fundamental requirement for producing good quality ‘‘functional” wines. 3.2. Stilbene extraction during the alcoholic fermentation (AF) Since the samples of the CT–CT and CT batches were not significantly different, only those of the CT and UV batches have been compared in the following figures.

Table 1 Content of stilbene in grapes and oenological parameters in must after destemming and crushing (i.e., just after harvesting for CT–CT batch, and at maximum induction day for CT and UV batch).

Piceatannol (mg/kg fw) trans-Resveratrol (mg/kg fw) q-Viniferin (mg/kg fw) a-Viniferin (mg/kg fw) Total stilbenes (mg/kg fw) Brix degree TA (g/l TH2) pH Tartaric acid (g/l) Malic acid (g/l) K (mg/l) Anthocyanins (mg/l) Tannins (g/l) IPT Gluconic acid (g/l) Acetic acid (g/l)

CT–CT

CT

UV

Significance level

nda 0.53a 0.06a 0.02a 0.61a 20.87a 4.59a 3.70a 5.35a 1.80a 1996a 45a 1.88a 13.15a 0.15a 0.02a

0.40a 1.13a 0.19a 0.11a 1.82a 20.33a 4.56a 3.69a 5.19a 1.89a 1953a 44a 1.85a 12.22a 0.09a 0.04a

2.95b 10.89b 1.06b 0.34b 15.25b 20.13 a 4.53a 3.67a 4.70a 1.85a 1850a 43a 1.78a 12.16a 0.06a 0.03a

* ** ** ** **

– – – – – – – – – – –

Different superscript letters (a, b or c) for the same parameter denote significant differences (p < 0.05). Analyses of variance, levels of significance: – (ns, not significant). IPT, index polyphenols total; TA, Total acidity; nd, non-detected (LOQ = 0.04 ppm was used in the statistical treatment). * p < 0.05. ** p < 0.01.

Stilbenes in solid and liquid phases were daily analysed to monitor the extraction rate of stilbenes during the AF. As expected, on crushing, stilbenes are mainly found in the solid phase (grape skin) rather than in the liquid phase (must) (Fig. 2, day 0). It was observed that stilbene content moved progressively from the solid to the liquid phase as the content of ethanol in the liquid increased during the AF (Fig. 2), as reported by other authors Gambuti, Strollo, Ugliano, Lecce, & Moio (2004). However, the extraction rate was different, depending on the initial amount of trans-resveratrol in grapes. It was observed that the higher the initial concentration of trans-resveratrol in the solid phase, the lower the extraction rate from the solid to the liquid phase. After one day of AF, the CT must contained almost 40% of the total trans-resveratrol contained in CT grapes, whereas, also after one day, the UV must contained only 15% of the total initial trans-resveratrol content of those UV grapes (Fig. 2A). On the second day of AF, the UV must contained 50% of the trans-resveratrol extracted, from solid to liquid phase (Fig. 2A). The extraction yield was also different for CT and UV wines. At the end of the AF, almost all the stilbene contained in CT grapes had been transferred into the wine made from this batch (the yield was 90%, taking into account that one kilogram of grape produces 0.8 l of wine). However, in the case of the UV wine, a yield of only over 60% was obtained (Fig. 2A). Thus, around 40% of the initial trans-resveratrol was lost from the UV grapes (i.e., it was not detected in either the solid or liquid phase after the AF) and two tentative hypotheses are suggested. First, some of the initial trans-resveratrol may be oxidised into a non-identified compound since it has previously been suggested that trans-resveratrol is highly sensitive to oxidation (Castellari, Spinabelli, Riponi, & Amati, 1998). In this line, the degradation and transformation of trans-resveratrol during AF have been described (Threlfall, Morris, & Mauromoustakos, 1999), and this could explain the fact that no significant correlation has been found between the level of transresveratrol in the grape skin and the level of trans-resveratrol in the corresponding red wine (Sun, Ribes, Leandro, Belchior, & Spranger, 2006). The second tentative hypothesis is the possibility of a co-pigmentation reaction between resveratrol and some other wine component, in a similar way to what has been described for p-coumaroyl acids (Rentzsch, Schwarz, Winterhalter, & Hermosín-Gutiérrez, 2007). At the end of the AF, viniferins were only present in solid parts (pomace). In fact, viniferins possess low solubility in must-wine, due to their hydrophobic character (Karacabey & Mazza, 2008), and they were not extracted during AF in this study. The extraction rate of total stilbenes during AF is marked mainly by the extraction rate of trans-resveratrol (Fig. 2A and B). Piceatannol was present at the end of AF only in the UV wine (2.61 mg/l) but not in the CT wine. It seems that the small amount of piceatannol found in CT grapes (Table 1) was lost during the AF. The UV wine possessed trans-resveratrol and total stilbenes contents that were, respectively, 4.7 and 6.4 times higher than those possessed by the corresponding control wine (of the CT batch) at the end of AF (Fig. 2). 3.3. Changes in stilbene content and oenological parameters during winemaking The stilbene content was determined at each stage in the winemaking process (Fig. 1) for the CT–CT, CT and UV batch assays. The extraction of trans-resveratrol from solid phase (skin) to liquid phase (wine) was completed after pressing. In accordance with previous data, pressing was carried out just at the end of the AF (when density was 999 mg/l), to avoid the precipitation of transresveratrol with yeast. After vatting, the trans-resveratrol content slightly decreased when the press-run wine was added to the free-run wine in the UV wine (Fig. 3), which is in agreement with findings of other authors (Vrhovsek et al., 1997).

199

R.F. Guerrero et al. / Food Chemistry 122 (2010) 195–202

20

15

mg/L mg/kg

A

CT Must UV Must CT Solid UV Solid

10

5

0

B

mg/L mg/Kg

20

15

10

5

0 0

1

2

3

End of AF

Days Fig. 2. Evolution of trans-resveratrol (A) and total stilbenes (B) during alcoholic fermentation in both liquid (mg/l) and solid phases (mg/kg).

7 CT UV

Resveratrol (mg/L)

6

5

4

3

2

1

3

Bo ttl ed M on th s ag 6 ed M on th s ag ed

at io n

+I ne rt

C la rif ic

M F

R ac k

s

wi ne

w in e Pr es

ru n Pr es s

Fr

ee

ru n

wi ne

0

Fig. 3. Evolution of trans-resveratrol during the various stages of winemaking and ageing in bottle. Press wine = free-run wine + press-run wine; MF, malolactic fermentation; Rack + Inert, racking plus inertization.

After pressing, differences in stilbene concentration between wines were at their highest (Fig. 3). In Table 2 (pressed wine) it can be observed that there were still no significant differences regarding stilbene content between the CT–CT and CT wines.

The UV wine was the only one which contained piceatannol, and its trans-resveratrol concentration at this stage was much higher than that of either of the control batches (CT–CT and CT).

200

R.F. Guerrero et al. / Food Chemistry 122 (2010) 195–202

Table 2 Stilbene content and oenological parameters in pressed and bottled wines. Pressed Wine

Piceatannol (mg/l) trans-Resveratrol (mg/l) Total stilbenes (mg/l) TA (g/l TH2) pH Tartaric acid (g/l) Malic acid(g/l) K (mg/l) Anthocyanins (mg/l) Tannins (g/l) IPT Colour intensity Hue L* a* b* Ethanol (%vol.) VA (g/l AcH) Acetic acid (g/l) Citric acid (g/l) Lactic acid (g/l) Succinic acid (g/l) Glycerine (g/l) Dry extract (g/l) Fe (mg/l) Cu (mg/l) Zn (mg/l)) Sugars (g/l) R Higher alcohols (mg/l)

Bottled Wine

CT–CT

CT

UV

Significance level

CT–CT

CT

UV

Significance level

nda 0.60a 0.60a 6.36a 3.69a 2.89a 1.75a 1725a 748b 3.65a 52.3a 1.26b 0.44a 52.6a 61.0b 3.7b

nda 1.45a 1.45a 6.57a 3.63a 3.20a 1.81b 1805a 885a 3.47a 49.2a 1.44a 0.41a 41.1a 65.7a 5.1a

1.90b 4.62b 6.52b 6.49a 3.63a 3.63a 1.64b 1783a 916a 3.54a 50.1a 1.44a 0.41a 41.1a 66.8a 6.2a

**

nda 0.91a 1.10a 4.55a 3.72a 1.72a 0.14a 1259a 499a 3.17a 45.1a 1.32a 0.52a 44.8 a 57.2a 6.9a 11.4a 0.30a 0.37b 0.09a 1.65a 1.03a 7.96a 26.7a 1.01a 0.11a 0.46a 2.6a 285a

nda 0.96a 0.96a 4.83a 3.68a 1.71a 0.12a 1268a 483b 2.76b 39.2b 1.31a 0.52a 44.6a 56.6a 5.6b 11.0a 0.44b 0.61a 0.15a 1.80a 1.19a 8.19a 24.6b 1.10a 0.08a 0.50a 2.4a 264a

1.05b 3.07b 4.12b 4.68a 3.73a 1.78a 0.15a 1254a 504a 2.99a 41.8ab 1.38a 0.52a 42.0a 57.6a 8.2c 11.2a 0.32c 0.55ab 0.15a 1.78a 1.14a 8.09a 27.8c 0.93a 0.07a 0.60a 2.4a 255a

**

** **

– – – – – *

– – *

– – * *

** **

– – – – – * * *

– – – – *

– * *

– – – – *

– – – – –

Different superscript letters (a, b or c) for the same parameter denote significant differences (p < 0.05). Analyses of variance, levels of significance: – (ns, not significant). TA, total acidity; IPT, index polyphenol total; VA, volatile acidity; nd, non-detected (LOQ = 0.04 ppm was used in the statistical treatment). * p < 0.05. ** p < 0.01.

Malolactic fermentation did not affect the trans-resveratrol content (Fig. 3) because Oenococcus oeni, which does not modify the stilbene profile in wines, was used (Hernández et al., 2007). On the contrary, racking and inertization did affect the trans-resveratrol content in the wine (Fig. 3). Thus, it seems that the tartaric acid precipitated the trans-resveratrol through tartrate crystals. In fact, when tartrates were analysed, the trans-resveratrol concentration was high (9.92 mg/kg of tartrate). Albumin was selected for the clarification, since, as it has been previously described, albumin does not fine resveratrol (Vrhovsek et al., 1997), and consequently resveratrol content in the second lees (Fig. 1) was low. The wine was not filtered, in order to preserve maximum trans-resveratrol content, in agreement with previous experiments. Throughout the entire process, piceatannol behaved in a similar way to trans-resveratrol (data not shown). No viniferins were detected in wine of any of the batches. As commented before, the initial viniferin content of the grapes remained 100% in the pomace, which in the UV batch contained 2.16 mg/kg; thus pomace could be used as a source of viniferins. Finally, the UV bottled wine contained 3.07 mg/l of trans-resveratrol (3.2 times the content of the CT wines) and 1.05 mg/l of piceatannol (26 times the content of the CT wine, taking the limit of quantification (LOQ) as 0.04 ppm) (Table 2, bottled wine). During ageing of the wine in bottle, a slight decrease in trans-resveratrol content was observed (Fig. 3), which has also been reported by other authors Bravo et al. (2008). In order to verify if grape treatment (storage and UV exposure) had affected the quality of the wine, oenological parameters were measured at each stage in the process (Fig. 1). In particular, oenological parameters in pressing are reported, since maximum differences in stilbenes were shown at this point (Table 2). All wines presented similar values for total acidity, pH, tartaric and malic

acids, potassium, tannins, IPT, hue and L*. The only differences were found in the CT–CT wine, which showed less colour intensity, anthocyanins, red (a*) and blue (b*) components than the CT and UV wines. It has been suggested that, during the storage of grapes until Dm was reached, a soft intracellular maceration process could have taken place. However, the differences found were of low significance (p < 0.05) and all parameters were within the range established for quality wines (Ribereau-Gayon, Glories, Maujean, & Dubordieu, 2000). To ensure final product quality, oenological parameters were also analysed in bottled wine, (Table 2). No differences were found in ethanol content, total acidity, pH, organic acids (apart from acetic), glycerine, sugars, higher alcohols and metals. Volatile acidity and acetic acid showed differences of low significance among samples (p < 0.05) but, in both parameters, all values were under the legal limit for red wine (CE 479/2008). Some differences of low significance were found in dry extract, tannins, anthocyanins, b* and IPT; which could be attributed to the intrinsic variability of winemaking. It is worth to mention alcohol content was not high, as is required for a healthy wine. 3.4. Sensory analysis Three basic methods were used by the panel of expert tasters to assess the wines: visual aspect, aroma and flavour. No significant differences were found between CT–CT, CT and UV wines on any of the three assessments (data not shown). All wines showed a very good visual aspect. However, some herbaceous notes were detected by nose and by mouth, in all wines equally. To determine the compound/s responsible for this unwanted characteristic, wines were analysed by GC–olfactometry. Overall, the panel

201

R.F. Guerrero et al. / Food Chemistry 122 (2010) 195–202 Table 3 Odorants found in wines: gas chromatographic data, olfactory description, chemical identity and modified frequency percentage (MF). RI

Descriptor

Identity

MF (%)

1218–1220 1244–1249 1025–1033 1123–1128 950–959 1096–1103 931–937 1684–1688 1822–1828 1064–1069 1459–1462 1884–1890 1921–1924 1731–1740 1051–1053 1304–1306 1713–1720 1509–1513 1563–1565 1868–1870 1395–1398 1643–1647

Fusel, rancid, cheese Anise, fruity, liquorice Fruity, strawberry, hazelnuts Banana Strawberry, cream, sweet Fusel, rancid, solvent Alcoholic Rancid cheese, feet, Dry plum, cooked apple Anise, fruity Vinegar Pollen, floral, phenolic Roses Honey, tropical fruity Anise, sweet strawberry Mushroom Synthetic, rubber Rancid, moisture (humidity) Lemon, floral Phenolic, spicy Herbal, vegetal Cheese, vomit

Isoamyl alcohol Ethyl hexanoate Ethyl butyrate Isoamyl acetate Diacetyl Isobutanol Propyl acetate Isovaleric acid b-Damascenone 3-Methybutyl acetate Acetic acid Ethyl dihydrocinnamate b-Phenylethanol unknown Ethyl 2-methylbutirate 1-Octen-3-one Methionol 2-Nonenal Linalool Guaiacol 3-Hexenol Butyric acid

91 87 82 75 71 71 70 66 66 63 63 54 51 46 38 36 36 35 33 33 31 31

RI, retention index; MF, modified frequency.

reported 108 separate perceptions. To reduce this large number, we set an arbitrary limit at modified frequency (MF) > 30% (Table 3). The most potent aroma on the polar column was isoamyl alcohol, which has a fusel character. Similar results were obtained by Petka, Ferreira, Gonzalez-Vinas, and Cacho (2006) in Slovenian wines. This result is in apparent contrast to the finding that the fusel aroma notes were rather weak in the wine tasting, but it could be attributed to the good solubility of this alcohol in wine, and to the fact that this compound is a fixed constituent of wine aroma. Something similar happened to odorants such as ethyl hexanoate, diacetyl and to many other compounds in Table 3. In contrast, no component was particularly notable as responsible for the herbaceous notes. In fact, the herbaceous notes (3-hexen-1-ol) presented an MF = 31%, close to the limit to be considered an unimportant contributory component (MF < 30%). Thus, no conclusion on responsibility can be drawn from the olfactometric analysis. It seems that the factor responsible for the herbaceous notes is not one single but a complex mix of several compounds. We suggest that this aroma was originated from the ripeness stage of grapes and not from the processing method applied, since it was also found in wines made from non-treated grapes. Even though this could not be avoided in the conditions of this study, it could be possible to mask the aroma by certain oenological techniques such as micro-oxygenation or use of wood chips. Further studies need to be conducted on this aspect of the topic.

also contained piceatannol (1.05 mg/l), which was not detected in any of the control wines (made from CT–CT and CT batches). In contrast, few significant differences were found in the oenological parameters of these three wines, measured during the course of the winemaking process. All the wines (from the CT– CT, CT and UV batches) presented satisfactory values for alcohol content, total acidity, pH, individual organic acids, metals, superior alcohols, anthocyanins, tannins and colour parameters. However, in the sensory analysis, herbaceous notes were found in all the wines. It was not possible to determine the particular compound responsible for this characteristic from the GC–olfactometry performed. Wine enriched in trans-resveratrol and piceatannol may provide an additional source of anti-oxidants. It is known that wine regularly consumed in moderate amounts has various beneficial health properties. These wines present added value to the consumer since, with the same ingestion of ethanol, the intake of stilbenes is significantly increased. Additionally, the pomace obtained as a by-product from UVC-treated grapes could be a source of viniferins for the manufacture of nutraceutical products. Postharvest UVC treatment of grapes is already being used to produce nutraceuticals (http://revidox.com/) and stilbene-enriched juice (González-Barrio, Vidal-Guevara, Tomás-Barberán, & Espín, 2009). The use of this technology for enriched wine still needs to be optimised, but it is a useful and promising technology for innovative companies in the highly competitive wine market.

4. Conclusion Acknowledgements Although warm climates are associated with a low content of stilbenes in grape and wine, the stilbene content in grapes notably increased with postharvest UVC treatment. Grapes of the UV batch showed increased content in piceatannol, trans-resveratrol and viniferins, respectively, to around 7.4, 9.6 and 4.7 times the content measured in the non-treated controls.trans-Resveratrol concentration decreased progressively during winemaking, especially during AF; this is a finding which requires further research to determine by what mechanisms the trans-resveratrol is lost. The final bottled wine made from UV grapes contained 3.07 mg/l of trans-resveratrol (3.2 times the content measured in wine made from control grapes of the CT batch). Moreover, bottled wines of the UV batch

The authors would like to thank the INIA for financing this study via the Project RTA2008-00014, and to the CSIC for permitting the use of Patent WO/2002/085137; ES 2177465.

References Barreiro, J., Colombo, S., & Cantos, E. (2008). Is there a market for functional wines? Consumer preferences and willingness-to-pay for resveratrol enriched quality red wines. Food Quality and Preference, 19, 360–371. Baur, J. A., Pearson, K. J., Price, N. L., Jamieson, H. A., Lerin, C., Kalra, A., et al. (2006). Resveratrol improves health and survival of mice on a high-calorie diet. Nature, 444, 337–342.

202

R.F. Guerrero et al. / Food Chemistry 122 (2010) 195–202

Bavaresco, L. (2003). Role of viticultural factors on stilbene concentrations of grapes and wine. Drugs Under Experimental and Clinical Research, 29(5/6), 181–187. Bavaresco, L., Pezzutto, S., Ragga, A., Ferrari, F., & Trevisan, M. (2001). Effect of nitrogen supply on trans-resveratrol concentration in berries of Vitis vinifera L cv. Cabernet Sauvignon. Vitis, 40(4), 229–230. Bertelli, A. A. E. (2007). Wine, research and cardiovascular disease: instructions for use. Atherosclerosis, 195(2), 242–247. Bertelli, A., Bertelli, A. A. E., Gozzini, A., & Giovannini, L. (1998). Plasma and tissue resveratrol concentrations and pharmacological activity. Drugs Under Experimental and Clinical Research, 24(3), 133–138. Bravo, M. N., Feliciano, R., Silva, S., Coelho, A. V., Vilas-Boas, L., & Bronze, M. R. (2008). Analysis of trans-resveratrol: Comparison of methods and contents in Muscatel fortified wines from Setubal region in Portugal. Journal of Food Composition and Analysis, 21(8), 634–643. Cantos, E., Espín, J. C., & Tomás-Barberán, F. A. (2003). UV–C irradiated grapes as a potential source for producing stilbene-enriched red wines pulses. Journal of Agricultural and Food Chemistry, 51(5), 1208–1214. Castellari, M., Spinabelli, U., Riponi, C., & Amati, A. (1998). Influence of some technological practices on the quantity of resveratrol in wine. Zeitschrift für Lebensmittel-Untersuchung und Forschung, 206, 151–155. Culleré, L., Cacho, J., & Ferreira, V. (2009). Comparative study of the aromatic profile of different kinds of wine cork stoppers. Food Chemistry, 112(2), 381–387. Gambuti, A., Strollo, D., Ugliano, M., Lecce, L., & Moio, L. (2004). transResveratrol, quercetin, catechin, and epicatechin content in south Italian monovarietal wines: relationship with maceration time and marc pressing during winemaking. Journal of Agricultural and Food Chemistry, 52(18), 5747–5754. González-Barrio, R., Vidal-Guevara, M. L., Tomás-Barberán, F. A., & Espín, J. C. (2009). Preparation of a resveratrol-enriched grape juice based on ultraviolet C-treated berries. Innovative Food Science and Emerging Technologies, 10, 374–382. Guerrero, R. F., García-Parrilla, M. C., Puertas, B., & Cantos-Villar, E. (2009). Wine, resveratrol and health: a review. Natural Products Communication, 4(5), 635–658. Guerrero, R. F., Puertas, B., Fernández, M. I., Palma, M., & Cantos, E. (2010). Induction of stilbenes in grapes by UV–C: comparison of behaviour of different subspecies of Vitis. Innovative Food Science and Emerging Technologies, 11, 231–238. Hernández, T., Estrella, I., Pérez-Gordo, M., Alegría, E. G., Tenorio, C., Ruiz-Larrea, F., et al. (2007). Contribution of malolactic fermentation by Oenococcus oeni and Lactobacillus plantarum to the changes in the nonanthocyanin polyphenolic composition of red wine. Journal of Agricultural and Food Chemistry, 55(13), 5260–5266.

Jang, M. S., Cai, E. N., Udeani, G. O., Slowing, K. V., Thomas, C. F., Beecher, C. W. W., et al. (1997). Cancer chemopreventive activity of resveratrol, a natural product derived from grapes. Science, 275(5997), 218–220. Jeandet, P., Bessis, R., & Gautheron, B. (1991). The production of resveratrol (3,5,40 trihydroxystilbene) by grape berries in different developmental stage. American Journal of Enology and Viticulture, 42(1), 41–46. Karacabey, E., & Mazza, G. (2008). Optimization of solid–liquid extraction of resveratrol and other phenolic compounds from milled grape canes (Vitis vinifera). Journal of Agricultural and Food Chemistry, 56(15), 6318–6325. Kitanaka, S., Ikezawa, K., Yasukawa, K., Yamanouchi, S., Takido, M., Sung, H. K., et al. (1990). (+)-a-Viniferin, an anti-inflamatory compound from Caragana chamlagu root. Chemical and Pharmaceutical Bulletin, 38(2), 432–435. Murias, M., Jaeger, W., Handler, N., Erker, T., Horvath, Z., Szekeres, T., et al. (2005). Antioxidant, prooxidant and cytotoxic activity of hydroxylated resveratrol analogues: structure-activity relationship. Biochemical Pharmacology, 69(6), 903–912. OIV (1990). Recueil des Méthodes Internationals d’Analyse des vins. France, Paris: Office International de la Vigne et du Vin. Petka, J., Ferreira, V., Gonzalez-Vinas, M. A., & Cacho, J. (2006). Sensory and chemical characterization of the aroma of a white wine made with Devin grapes. Journal of Agricultural and Food Chemistry, 54(3), 909–915. Pignatelli, P., Ghiselli, A., Buchetti, B., Carnevale, R., Natella, F., Germano, G., et al. (2006). Polyphenols synergistically inhibit oxidative stress in subjects given red and white wine. Atherosclerosis, 188(1), 77–83. Renaud, S., & De Lorgeril, M. (1992). Wine, alcohol, platelets, and the French paradox for coronary heart disease. Lancet, 339(8808), 1523–1526. Rentzsch, M., Schwarz, M., Winterhalter, P., & Hermosín-Gutiérrez, I. (2007). Formation of hydroxyphenyl-pyranoanthocyanins in Grenache wines: precursor levels and evolution during aging. Journal of Agricultural and Food Chemistry, 55(12), 4883–4888. Ribereau-Gayon, P., Glories, Y., Maujean, A., & Dubordieu, D. (2000). Handbook of enology. Chichester, UK: Jon Wiley and Sons. Sun, B., Ribes, A. M., Leandro, M. C., Belchior, A. P., & Spranger, M. I. (2006). Stilbenes: Quantitative extraction from grape skins, contribution of grape solids to wine and variation during wine maturation. Analytica Chimica Acta, 563, 382–390. Threlfall, R. T., Morris, J. R., & Mauromoustakos, A. (1999). Effect of variety, ultraviolet light exposure, and enological methods on the trans-resveratrol level of wine. American Journal of Enology and Viticulture, 50(1), 57–64. Vrhovsek, U., Wendelin, S., & Eder, R. (1997). Effects of various vinification techniques on the concentration of cis- and trans-resveratrol and resveratrol glucoside isomers in wine. American Journal of Enology and Viticulture, 48(2), 214–219.