2.1. Study of Sample Pre-Treatment and Storage Conditions
Grape canes are generally only available in the pruning season (winter) and it was necessary to store and condition samples for them to be available during the development of the extraction method. Additionally, samples must be dried to guarantee microbiological stability and also the stability of some compounds. As a consequence, a suitable treatment method for sample conservation was required.
Several alternatives have been reported for the pre-treatment of grape canes and these are focused mainly on thermal drying [9
] or freeze-drying [8
]. Thermal drying, although the simplest method, can lead to losses by evaporation or decomposition, especially for thermo-labile compounds such as phenolics. This approach also requires long drying times, especially at room temperature (months), in comparison to oven-drying (from days to weeks). Freeze-drying is an interesting alternative because it is faster than thermal drying and the extent of some degradation reactions will be reduced, mainly because of the low temperatures used in the process.
Both drying methods will lead to an increase in concentration of the compounds because water is removed in both cases. The aim of the assay described here was to compare freeze-drying and thermal drying (the two most widely reported drying techniques) to evaluate the influence of the drying process on the recovery of stilbenes. Similar studies (including both freeze-drying and thermal-drying processes) have been reported for other solid samples and compounds: Keinanen and Julkunen-Tiitto [22
] studied the effect of different drying methods on birch leaf phenolics and concluded that the highest concentrations of phenolics were obtained by freeze-drying; de Torres et al.
] reported similar conclusions on the volatile and phenolic composition of grape skins and found that the freeze-drying method was less aggressive than thermal drying (60 °C) and the phenolic composition was better preserved than that in the thermally dried skins; Garcia-Perez et al.
] reported that freeze-dried grape stalks presented the highest antioxidant concentration with the highest effective diffusivity and mass transfer coefficient and, furthermore, that the material was also more porous than thermally dried samples.
A total of 15 grape cane samples from several grape varieties (Sauvignon Blanc, Tempranillo, Malbec, Rome, Zinfandel, Tannat, Carmenere, Flame, Melissa, Red Globe, Moscatel Rosado, Victoria, Matilde, Crimson seedless and Tintilla de Rota) were cut into pieces, divided into two equal batches and the different drying procedures were carried out on the same scale. Samples were dried to constant weight using both techniques including an oven-dry at 40 °C for 15 days and a freeze-dried for 7 days. The average water loss for both procedures was 50% weight in which no significant differences were found for loss of water (data not shown). A typical chromatogram of grape cane and the structure of the different molecules studied are shown on Figure 1
The samples were milled and then extracted. Experiments were carried out with 0.8 g of dried sample, which was placed in a volumetric flask (100 mL) made up to 25 mL with 80% ethanol in water and the same extraction conditions were applied (15 min extraction time, 70% amplitude, cycle 0.7, and 7 mm diameter tip, 75 °C) according to Piñeiro et al
]. All of these assays were carried out in duplicate with protection from light (flasks were covered with aluminum foil) to prevent degradation and/or isomerization by light.
The values for total stilbenes obtained in the extractions for different dried samples (Figure 2
) were, in most cases, significantly higher (0.05 < p
< 0.001) for freeze-dried samples than for oven-dried samples. There are only three cases (Tempranillo, Zinfandel and Sauvignon Blanc varieties) in which significant differences were not observed. In general terms, piceatannol was the only stilbenoid with higher contents in oven-dried samples (data not shown). Other compounds, such as trans
-ε-viniferin, were present in similar quantities for both drying methods, whereas trans-
resveratrol had significantly higher contents in the freeze-dried samples. Therefore, freeze-drying was the pre-treatment method used for grape cane samples when developing the UAE method. Similar results have previously been reported for other samples [22
]. These results could be related to the mechanical effects on grape cane cells, as freeze-drying produce more intense cell wall degradation, thus allowing the solvent to penetrate deeper into the sample. Additionally, the use of high temperatures can promote degradation processes for some compounds.
Additional experiments were carried out with two grape cane samples in order to select the best storage conditions once the samples had been freeze-dried. Room temperature and freezer storage (−20 °C) were assessed after one month of storage. Significant differences were not observed for any of the quantified stilbenes (data not shown) and freeze-dried samples can therefore be stored at room temperature for easier handling without compound degradation.
2.2. Stability of Stilbenes
In order to evaluate the performance of different extraction conditions with accuracy, the stability of the stilbenes during the extraction was determined prior to method development. The stability study was performed using a stock standard solution of four commercially available stilbenes prepared in methanol.
One milliliter of stock solution was directly added to 25 mL of ethanol and submitted to the extraction conditions to check the stability of the stilbenes at 75 °C up to 35 min according to previous studies in our lab. Higher temperatures were not assayed because of the solvent will be evaporated after some minutes. Longer times were not assayed because of long extraction methods are not interesting for regular applications. Final composition in the resulting extract was compared to the starting composition of the stock standard dilution. The average recovery of all of the compounds in the sample was over 90%, e.g., piceatannol (96.7% ± 2.9%), trans-resveratrol (96.0% ± 0.4%) and trans-ε-viniferin (97.6% ± 1.0%). A temperature of 75 °C was therefore selected as the maximum temperature to be used when developing the UAE method, as in the extraction of real samples the use of a higher the temperature would lead to a faster extraction process.
2.3. Solvent Selection
Initial extractions were performed to determine the best solvent choice and these were carried out using a 1:30 sample-solvent ratio (0.83 grams of grape cane sample in 25 mL of each solvent) for 15 min according to previous studies. Palomino Fino 2011 samples were used throughout all of the method optimization assays. The assayed solvents, selected according to literature data [12
], were as follows: ethanol, 80% ethanol in water (v
), 80% methanol in water (v
), 60% ethanol in water and 60% acetone in water. Selected solvents covered a wide range of boiling points (from 56 for acetone to 78 °C for ethanol) and three extraction temperatures were therefore assessed (25, 65 and 75 °C) to avoid artifacts related to solvent loss. The total stilbenes extracted with the different solvents and temperatures are shown in Table 1
At 25 °C, the solvent that extracted the highest amount of stilbenes was 60% acetone in water, although significant differences were not observed with other solvents except for 100% ethanol, which extracted the lowest amount of stilbenes. Indeed, the use of 100% ethanol gave the lowest amounts of each stilbene at most of the temperatures assayed (except for 75 °C).
At 65 °C (a sufficiently high temperature to avoid total evaporation of each assayed solvent during the extraction time), the best solvent was 60% ethanol in water, the use of which led to the highest amounts of trans-ε-viniferin along with total stilbenes. In contrast, the methanol/water mixture was the solvent producing the lowest recoveries.
At 75 °C (the maximum extraction temperature used to develop the UAE method on the basis of stability results) the acetone/water mixture evaporated rapidly and it was therefore impossible to assess this solvent system. Ethanol/water mixtures were the most efficient solvents for all compounds and there was no significant difference between 60% and 80% ethanol for the extraction of stilbenes. It can be seen from the results that the extraction of samples at 75 °C led to a considerable increase in the stilbene content when compared to 25 °C: the total stilbenes concentration in the ethanol/water mixtures extracts increased to 89% (1362.9 vs. 721.9 mg·kg−1 dry weight) and 81% (1365.8 vs. 753.2 mg·kg−1 dry weight) achieved with 60% and 80% ethanol in water, respectively, in comparison to 35% (732.4 vs. 540.7 mg·kg−1 dry weight) in the ethanol extract. Therefore, it can be concluded that the use of ethanol/water mixtures at temperatures close to its boiling point is a better system for ultrasound transmission, thus allowing the ultrasound energy to reach the sample in a more effective way.
The use of 60% ethanol in water has several advantages over the use of 80% and these include environmental compatibility, lower toxicity and lower cost. These factors suggested the use of this solvent in the extraction method being developed. Thus 60% ethanol/water mixture was employed as the extraction solvent for subsequent optimization of the extraction conditions.
2.5. Extraction Time
With regard to the determination of the time required for the extraction, the recovery of the studied compounds was determined on applying extraction times from 5 to 35 min. The results are shown in Table 2
. Significant differences were found for trans
-ε-viniferin and total stilbene content on comparing times of 5 and 10 min, but this was not the case for trans
-resveratrol. It can be seen from the results that the majority of the total stilbenes present in the sample were extracted within 5 min (approximately 92%) and an increase in the extraction time from 5 to 10 min led to an increase of 7.6 of the total amount of trans
-ε-viniferin extracted. The use of longer extraction times (20–35 min) led to the same levels of recovery as 10 min, indicating that the quantitative extraction of stilbenes from the sample had occurred after this time. Therefore an extraction time of 10 min was employed.
In an effort to verify recovery, the number of extraction steps needed for total recovery was investigated. Thus, two successive extraction steps were carried out with fresh solvent, including a step that involved rinsing the solid between the first and second extractions. After first extraction, the sample was filtered under vacuum and the extract was collected and filled up to 25 mL. The solid was rinsed with 5 mL of fresh solvent (collected separately) and dried for 5 min. The sample was subsequently re-extracted with fresh solvent by the same procedure to provide a third fraction. All fractions were analyzed separately. The recoveries of the total stilbenes in each fraction were 926.9 (extract), 10.7 (rinse) and 25.0 (re-extract) mg·kg−1 dry weight, respectively. Most compounds were almost totally recovered in the first fraction (data not shown); however, evidence for the presence of all of the stilbenes was also found in the rinsing fraction. Therefore, a rinsing step is required in order to obtain quantitative recoveries (higher than 95%) for all the assayed compounds.
2.8. Determination of Stilbenes in Grape Cane Samples
The optimized procedure was successfully applied to the determination of stilbene levels in 20 grape cane samples and the results are shown in Table 4
. Most of the grape cane varieties assayed had not been analyzed previously and only results for Sauvignon Blanc, Carmenere, Zinfandel and Malbec cultivars were found in the literature.
Among the red grape varieties, the Tintilla de Rota grape cane sample (an autochthonous Andalusian grape variety) contained the highest amount of total stilbenes and also the highest concentrations of trans
-piceid, whereas Zinfandel had the maximum trans
-ε-viniferin content (Table 4
Regarding white grape cane samples, Melissa was the variety that had the highest contents of trans-resveratrol, trans-piceid and piceatannol.
All of the studied stilbenes were found in all of the analyzed red grape cane samples. The most abundant stilbene in all grape canes was trans-ε-viniferin, followed by trans-resveratrol and trans-piceid.
The abundance of stilbenes differed depending on the cultivar. Tintilla de Rota showed a higher trans-ε-viniferin content (1964.8 ± 24.8 mg·kg−1 dry weight) than trans-piceid (952.8 ± 23.7 mg·kg−1 dry weight) and trans-resveratrol (575.1 ± 16.5 mg·kg−1 dry weight). In the Rome cultivar instead the most abundant stilbene was trans-ε-viniferin (1494.1 ± 18.8 mg·kg−1 dry weight) but followed by trans-piceid (348.4 ± 8.7 mg·kg−1 dry weight), being both higher than piceatannol (111.6 ± 4.2 mg·kg−1 dry weight) and trans-resveratrol (107.1 ± 3.1 mg·kg−1 dry weight) content. Surprisingly, low differences were found for the Palomino fino samples analyzed in different vintages, where the trans-resveratrol and trans-ε-viniferin contents remained quite similar. It is well known that stilbene content is influenced not only by cultivar, but also by climatic and cultural conditions in specific vintages.
-ε-viniferin concentrations measured in our samples were in the range described by other authors for different varieties such us Pinot Noir: 433–1500 mg·kg−1
dry weight [26
], Gewürztraminer, Cabernet Sauvignon, Cinsault, Moscatel Alejandría: 266–824 mg·kg−1
dry weight [24
], several Estonian varieties (Hasaine Sladki, Zilga, Yubilei Novgoroda): 700–1700 mg·kg−1
dry weight [26
], other major Vitis vinifera
cultivars (such as Chardonnay, Sauvignon Blanc, Semillon, Syrah or Merlot): 967–3737 mg·kg−1
dry weight [19
], or some wild-type species: 728–5739 mg·kg−1
dry weight [21
With regard to trans
-resveratrol content, the results reported in this paper are consistent with those described in the literature: 190–1526 mg·kg−1
dry weight [23
], 100–3200 mg·kg−1
dry weight [26
] and 383–6533 mg·kg−1
dry weight [20
] in different major Vitis vinifera
species, Estonian cultivars or several grape varieties grown in Chile, respectively. These concentrations were markedly higher than those reported recently by Houillé et al.
] in different Vitis vinifera
varieties, with values between 41 and 158 mg·kg−1
The high variability in trans
-resveratrol content between the same cultivars and different growing regions has also been reported previously, e.g., values of 723, 1526 and 3400 mg·kg−1
dry weight for the Pinot noir variety in Chilean, French or Canadian vineyards, respectively [20
]; or 529.2 and 1085 mg·kg−1
dry weight for the Zinfandel variety in a Spanish (in this study) and Chilean vineyards [20
]. Therefore, the higher values reported here could be explained in terms of cultivar practices and climatic conditions in Southern Spain.
Piceatannol concentrations were in the range reported by Vergara et al.
] for several cultivars, but are somewhat lower than those reported by other authors [19
In general terms, the average total stilbene contents found in red grape cane samples were higher than those in the white grape samples, a finding that is consistent with the results reported by Lambert et al.
Our findings were not unexpected since a number of abiotic or biotic stress factors, such as UV radiation, heavy metal ions or infection by fungi, are known to affect stilbene biosynthesis during grapevine growth. The differences observed between stilbene contents in Vitis
samples and genotype have been reported previously [19