Numerous epidemiological studies since the 90s have shown that a very moderate wine consumption may be beneficial to health (Renaud et al., 1999). The presence of polyphenols and particularly stilbenes in wine may explain these beneficial effects. Among stilbenes, resveratrol has demonstrated a number of promising biological activities, particularly in regards to disease prevention and anti-aging activities (Vang et al., 2011 ; Catalgol et al., 2012). Many studies have examined the resveratrol content in wine (Goldberg et al., 1995; Mattivi and Nicolini, 1997). In contrast, few studies have investigated stilbene oligomers in wine. However, some studies indicate that wine contains many other stilbenes (Pawlus et al., 2013 ; Moss et al., 2013).

The aim of the present work was to develop a method to characterize stilbene content in Tunisian red wines. While many wines, especially French wines, have been studied for their contents in resveratrol, to our knowledge, no Tunisian wine has yet been explored in these regards. The presence of various stilbenes has been studied using UPLC-MS analysis on wine extracts. These compounds have been quantified by HPLC-DAD and different stilbenes have been identified, from the simplest monomers, such as resveratrol and piceid, to the more complex oligomers, such as ε- and α-viniferin, hopeaphenol and isohopeaphenol. Concerning standards, if the simpler molecules may be obtained commercially, the complex stilbenes need to be purified from different plant sources. In this study, the standards were purified from vine parts and by-products from Carex species (Arraki et al., 2013 ; Papastamoulis et al., 2014).

In this study, seven Tunisian red wines mainly from the Mornag appellation were investigated. Seven stilbenes were identified in these wines: three monomers (resveratrol, piceid and piceatannol), one dimer (ε-viniferin), one trimer (α-viniferin) and two tetramers (hopeaphenol and isohopeaphenol). Regarding the presence of resveratrol derivatives, we found that one of the wines, Sidi Zahia, was the richest qualitatively. These first results are very encouraging, and the study should be extended to a larger number of wines and appellations.

Methanol and acetonitrile were obtained from Scharlab® S.L. (Sentmenat, Spain), trifluoroacetic acid (TFA) from Sigma-Aldrich® and deuterated solvents from Euriso-Top. Water was purified using an Elga water purification system (Bucks, UK) with a resistivity of no less than 18 MΩ/cm.

Wine samples

Seven Tunisian red wines from the Mornag appellation (Clipea, Distinto, Mornag, Selian, Sidi Brahim, Sidi Rais and Sidi Zahia) were examined for their stilbene content by using UPLC-MS and HPLC-DAD methods. The wines were purchased from local stores and kept in the dark until analyzed.

Red wine is a complex mixture of molecules. Among these, stilbenoids are in very low concentration and are often masked by other compounds, which sometimes makes it difficult to determine and quantify the stilbenes by techniques such as HPLC-DAD. Thus, in order to remove compounds such as sugars, anthocyanins or proanthocyanidins for example, we had to carry out a pre-purification step. The different wines were evaporated and purified on XAD16 and DOWEX columns (figure 1). These methods for the pre-purification of polyphenols have been developed and adapted from techniques already used in our laboratory and previously described concerning DOWEX columns (Decendit and Mérillon, 1996 ; Waffo-Téguo et al., 1996) or XAD16 columns (Pawlus et al., 2013). Each 750-mL bottle of wine was concentrated down to 650 mL in vacuo at 30 °C and diluted with 100 mL of water. The concentrated wine was poured over 500 g of Amberlite XAD-16 resin (Sigma-Aldrich), in an open column (5.1 cm id × 35 cm), and rinsed with 5 L of water to remove sugars, small organic acids and other non-polyphenolic compounds. The polyphenols were then eluted with 3 L of methanol and 1 L of acetone. The solvents were removed in vacuo and the samples were lyophilized. The concentrated XAD16 residue was then chromatographed over a cation-exchange resin column (DOWEX, Sigma-Aldrich), rinsed with distilled water (4 L), and eluted with 75% (v/v) aqueous methanol (2 L) to yield a solid powder after lyophilization. In order to evaluate the accuracy of the purification method, we proceeded to verify the chromatographic steps with mini-columns (0.8 mm id × 5 cm) and two pure stilbenoid standards ((E)-resveratrol and (E)-piceid); the proportions between the amount of deposited stilbenes, the amount of resins, and the volumes of water and eluents for each step were conserved. For example, 10 µg of each stilbene were dissolved in 2 mL water and submitted to these two chromatographic steps in these conditions. The various steps were followed by HPLC as described below. We were not able to detect stilbene in the rinsing waters of the columns, which means that there is apparently no loss of stilbenoids at this level. Elution with 75% methanol at the last step with the DOWEX column allowed the recover 96.9 ± 0.9% and 92.9 ± 1.6% of piceid and resveratrol, respectively. These experiments were confirmed a second time. It should also be noted that, when the amount of stilbenes deposited on the columns was 10 times greater, we observed a loss of resveratrol in the rinsing waters during the step of the XAD16 column (approximately one third). Under these conditions, the limit of solubility of resveratrol in water is exceeded and we have a precipitate, which cannot be retained by the XAD-16 resin. These stilbene concentrations here (equivalent to 100 mg/L) are also much too high compared to those found in wine.

Analyses were performed by analytical HPLC-DAD using a HPLC system model 1100 (Agilent technologies) coupled with a diode array detector (DAD). 20 µL of each sample was injected for each chromatographic analysis. Polyphenolic extracts were separated on a Bischoff® (Stuttgart, Germany) Prontosil® reverse-phase C₁₈ column (5 µm packing, 4 mm id × 250 mm) protected with a guard column of the same material. The solvent system used was ultrapure water acidified with 0.1% TFA (solvent A), and acetonitrile acidified with 0.1% TFA (solvent B). The elution program at 1 mL/min was 10% B (0-5 min), 10-60% B (5-35 min), 60% B (35-40 min) followed by a 5 min wash with 100% B and a 5 min reequilibration step. The chromatograms were monitored at 290 and 306 nm and the spectra (200-600 nm) continuously recorded.

Monitoring of the collected Tunisian red wines was achieved by UPLC-MS. The UPLC analyses were carried out using an UPLC system model 1290 Infinity (Agilent Technologies) coupled to an Esquire 3000+ mass spectrometer (Bruker Daltonics GmbH) with electrospray ionization and ion trap analyzer. The column used was a C₁₈ reversed-phase (100 × 2.1 mm, 1.8 µm, Agilent). Water 0.1% formic acid (solvent A) and acetonitrile 0.1% formic acid (solvent B) were used as mobile phases. 2.1 µL of each sample was injected and analyzed at 25 °C. The elution program at 0.21 mL/min was 10% B (0-2 min), 10-60% B (2-14 min), 60% B (14-16 min) followed by a 2 min wash with 100% B and a 5 min reequilibration step. The detection wavelengths were set at 290 and 306 nm. Electrospray ionization in positive and negative mode was used.

Table

In this study, we were able to obtain seven Tunisian wines mainly from the Mornag appellation. We were able to determine in these wines the content of seven stilbenoids: (E)-resveratrol, (E)-piceid, piceatannol, ε-viniferin, α-viniferin, hopeaphenol and isohopeaphenol. From the point of view of the presence of resveratrol derivatives, we found that one of the wines, Sidi Zahia, was the richest qualitatively. These first results are very encouraging. The study should now be extended to a larger number of wines and appellations.

Acknowledgements : The authors wish to thank the Ministry of Research and Higher Education of Tunisia for financial support of this research.