DETERMINATION OF MAJOR ANTHOCYANIN PIGMENTS AND FLAVONOLS IN RED GRAPE SKIN OF SOME TABLE GRAPE VARIETIES (VITIS VINIFERA SP.) BY HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY–PHOTODIODE ARRAY DETECTION (HPLC-DAD)

1: Department of Agronomy, University of El-Tarf, PB 73, El-Tarf 36000, Algeria 2: Institut de Recherche en Horticulture et Semences UMR1345 (INRA / Agrocampus-Ouest / Université d’Angers), 42 rue Georges Morel BP 60057, 49071 Beaucouzé Cedex, France 3: Department of Biology, University of Badji-Mokhtar, Annaba, PB 12, Annaba 23000, Algeria 4: University A. Mira of Bejaia, Faculty of Sciences of Nature and Life, Department of Food Science, 3BS Laboratory, Bejaia 06000, Algeria


INTRODUCTION
Grape (Vitis vinifera) is one of the world's largest fruit crops, and its composition and properties have been extensively investigated, with several reports of the presence of large amounts of phenolic compounds. Most phenolic compounds found in grape can act as antioxidants. Grape skin is a source of natural pigments (anthocyanins and flavonols), which are known to possess broad pharmacological activities and therapeutic potentials (Bagchi et al., 2004;Leifert and Abeywardena, 2008). Anthocyanins, the major polyphenols found in red grape skins, have been reported to show cardioprotective effects against ischemic reperfusion injury and to possess other diverse biological properties and therefore are considered as secondary metabolites with potential nutritional value (Kallithraka et al., 2005;Das et al., 2007). Anthocyanins, being located in the berry skin, are the main flavonoids responsible for the red color of grape cultivars. It is well known that the anthocyanin concentration can vary widely among different vintages of a given cultivar, due to both environmental (seasonal conditions) and agronomical factors. Therefore, the anthocyanin profile has been used as a chemotaxonomic parameter for the classification of red Vitis vinifera varieties (Revilla et al., 2001;Mattivi et al., 2006). Grape anthocyanins are monoglucosides of five anthocyanidins, namely delphinidin, cyanidin, petunidin, peonidin and malvidin. The acylated anthocyanins are esters of the glucose moiety of the free anthocyanins with acetic, p-coumaric or caffeic acids. Flavonols are one of the most studied classes of polyphenolic phytochemicals, because of the importance pertaining to their antioxidant potency and other biological activities (O'Byrne et al., 2002). Flavonols constitute a group of flavonoids that vary in color from white to yellow and are closely related in structure to the flavones. Derivatives of the most commonly encountered aglycones, including quercetin, myricetin, kaempferol and isorhamnetin, have been found in grapes (Vitis vinifera sp.). The conjugates are exclusively 3-O-glycosides, whereas sugar attachment on other positions of the flavonol skeleton has never been reported. For isorhamnetin, only glucose derivatives have been identified, but myricetin, quercetin and kaempferol may also occur as glucuronides (Makris et al., 2006).
The objective of the present study is to identify phenolic compounds in skin of red grape varieties grown in Mediterranean climate. The grapes of three varieties commonly used in Algeria (Cardinal, Gros noir and Muscat noir) were analyzed and compared for their anthocyanin and flavonol content. To the best of our knowledge, no research has examined the anthocyanin and flavonol content in these grape varieties grown in El-Tarf (Algeria).

Plant material
Skins from three grape cultivars, namely Cardinal, Gros noir and Muscat noir, were examined. Grape samples were grown in the region of El-Tarf located in the north-east of Algeria (36° 45' 00» N; 81° 10' 00» E) and collected at maturity. The region has a Mediterranean climate, which is divided into a hot season from June to October and a rainy season from November to April. The rest of the year the region enjoys a mild and pleasant climate.

Sample preparation
Before extraction, skins were manually separated from the whole berries and dried in oven at 50°C until constant weight. Dried grape skins were crushed in a grinder for 2 min and then used for extractions.

Extraction of anthocyanins and flavonols
The extraction procedures was done according to Cadot et al. (2012) and Brossaud et al. (1999). Briefly, dried skin powder (2 g) was extracted successively with 80 mL of methyl alcohol/water/TFA (80:20:0.05) and 50 mL of acetone/water (60:40). The extraction was repeated twice: 25°C, 15 min, 250 rpm and the extract was centrifuged (10°C/10 min/10 000 rpm). The supernatant was then filtered through a glass microfiber filter (GF/A Whatman® 1.6 µm) before drying under vacuum at 30°C using a rotary evaporator (Buchi®) and dissolved in 5 mL of methanol to yield a crude skin flavonoid extract.

HPLC analysis of anthocyanins and flavonols
Analysis of monomeric phenolics was performed following a method described by Cadot et al. (2012) with some modifications. Prior to analysis, skin extracts were filtered using a HVLP 0.45-µm filter (Millipore). Quercetin and malvidin-3-O-glucoside were used as external standards for flavonols and anthocyanins, respectively (Extrasynthese, Genay, France).
The analyses were done on extract as described previously by Roggero et al. (1992) using a Waters Millennium HPLC-DAD system (Milford, MA). The column consisted of a reversed-phase (Phenomenex   Table 2. according to their retention time and UV-visible (UVvis) spectra. Quantification was carried out from peak areas at 360 and 520 nm using quercetin and malvidin-3-O-glucoside as external standards. For quercetin-O-glucuronide, a co-injection of a standard with the sample was done in order to validate the identification.
As extractions and injections were done in duplicate, the final result was the arithmetic average of four analyses.

Statistical analysis
Results are expressed as mean ± standard deviation (SD). Statistical analysis was carried out using the Statistica software version 5.0 (StatSoft, France). Differences between means were first analyzed using the ANOVA test, and the least significant differences (Fisher's LSD) were calculated following significant F test (P≤0.05).

RESULTS
Polyphenol analysis was carried out on the grape skins of red varieties, because this is the part of the berry that contains the two main classes of polyphenols (anthocyanins and flavonols) (Guerrero et al., 2009).

Anthocyanins
Retention time (min)  Table 2. Retention time of identified anthocyanins and flavonols in skins of red grape varieties.

Anthocyanin profile
The presence of anthocyanidin 3-O-monoglucosides along with their acyl derivatives in the berry skin has been demonstrated in different grape cultivars  Table 2.
including Tempranillo, Palomino negro and Cabernet sauvignon (Guerrero et al., 2009), and Syrah (Boss et al., 1996;Marquez et al., 2013). On the other hand, Pinot noir completely lacks acylated anthocyanins (Mattivi et al., 2006). However, not all acylated derivatives were detected in all varieties. In two of the three varieties studied (Cardinal and Muscat noir), malvidin-3-acetylglucoside, petunidin-3-pcoumaroylglucoside and malvidin aglycone were not detected, whereas peonidin-3-O-acetylglucoside was present in Cardinal but not in Gros noir and Muscat noir. Table 3, in the two black cultivars (Gros noir and Muscat noir), malvidin-3-O-glucoside was the major anthocyanin component (72.73 and 41.11%, respectively), which is in agreement with the general perception that malvidin-3-O-glucoside is the main anthocyanin in grapes (Revilla and Ryan, 2000;Revilla et al., 2001). In agreement with the findings of Brar et al. (2008) and Castillo-Munoz et al. (2009), peonidin-3-O-glucoside was the major anthocyanin found in Cardinal berries, which accounted for 59.27% of total anthocyanins. It has been reported that the cyanidin and peonidin levels are higher in red grape cultivars, whereas delphinidin and malvidin levels are higher in black grape cultivars (Mizuno et al., 2006). Our data indicate that the color of grape berries may be related to the relative concentration of peonidin-3-O-glucoside and malvidin-3-O-glucoside in the grape berry skin. In the biosynthetic pathway it is known that cyanidin is a precursor to the other anthocyanidins and is converted into peonidin by the action of 3´-Omethyltransferase or into delphinidin by the action of 3´-hydroxylase. Delphinidin is further converted via petunidin to malvidin by the action of 3´5´-Omethyltransferase. The activity of these enzymes increased during grape ripening compared to other activities upstream in the flavonoid pathway (Boss et al., 1996;Pomar et al., 2005;Fournand et al., 2006). So, delphinidin-3-O-glucoside is among the minor anthocyanins in the varieties studied, and according to Roggero et al. (1986) the concentration of delphinidin-3-O-glucoside in grape berry increases during the early stages of berry ripening to reach a maximum value, and subsequently decreases till harvest due to its conversion into petunidin-3-Oglucoside and malvidin-3-O-glucoside by the action of methyl transferases.

As shown in
It has been suggested that the anthocyanin profile of berry skin may be directly related to the intensity of color, as evidenced from the higher concentrations of anthocyanins produced at the end of the biosynthetic pathway, namely peonidin-3-O-glucoside, malvidin-3-O-glucoside and petunidin-3-O-glucoside, in the higher pigmented cultivars (Carreno et al., 1997).
The highest quantity of free anthocyanins (not bounded to tannins) was present in the Gros noir skin extract (214.10 mg/kg of berries), a variety that belongs to the so-called Teinturier varieties. For that reason, it was not surprising that its grapes also showed the highest average content of pigments (P<0.05). Lower anthocyanin concentrations found in skins from clusters studied appear to be due to elevated temperature, either through degradation, inhibition of synthesis, or, more likely, both (Spayd et al., 2002). Non-acylated anthocyanins were the main class of grape anthocyanins, and coumaroylated derivatives were the most important within the minor acylated anthocyanins. The percentage of coumaroylglucoside, acylated and non-acylated anthocyanins in each cultivar was calculated based on the peak area. All the varieties had higher percentages of nonacylated anthocyanins over acetic acid acylated and coumarate anthocyanins. The ratios of these different anthocyanins are also important for grape variety characterization: for instance, Pinot noir has no acylated anthocyanins (Cortell et al., 2007). Nonacylated derivatives were predominant (>70%) but coumaroylglucoside derivatives were also present in quantities >5%, compared to the percentage of acylated anthocyanins which was very low (<2%). Muscat noir presented the highest non-acylated anthocyanin content, followed by Cardinal and Gros noir with 92.83, 88.02 and 70.74%, respectively. The lack of acylation in red grapes is not well understood, and two different hypotheses should be considered: the lack of genes related to the synthesis of acyltransferases or the lack of regulatory genes that module the expression of the first set of genes (Revilla et al., 2012). Further, according to Fournand et al. (2006), the degradation rate of coumaroylglucoside derivatives is higher than that of non-acylatedglucoside derivatives or acetylglucoside derivatives at high sugar content in pulp (over 200 g/L). Muscat noir showed the highest content of sugar followed by Cardinal and Gros noir with 21, 17.30 and 17°Brix, respectively (data not shown in this study). Coumaroylglucoside derivatives might be hydrolyzed to restore the non-acylated forms, or the cinnamoyltransferase activity might decrease compared to the other activities involved in anthocyanin biosynthesis. Coumaroylglucoside derivatives may also be more reactive and consequently more involved in the formation of derived pigments than other anthocyanins. Their higher reactivity has been previously observed during fermentation, where the rate of formation of p-coumaroylvitisins was estimated to be higher than that of non-acylatedvitisins (Morata et al., 2006). These results are consistent with previous studies performed on berries from different varieties and on the corresponding wines in which the authors systematically observed a higher ratio of nonacylatedglucosides to coumaroylglucosides (He et al., 2010;Zhao et al., 2010).
As stated previously, the presence of the enzyme flavonoid 3-O-glucosyltransferase is necessary for anthocyanin biosynthesis. However, the biosynthesis of the different anthocyanin precursors is driven upstream of the enzyme UFGT by the activity of flavonoid-3´-hydroxylase and flavonoid-3´5´hydroxylase enzymes, which add either a single hydroxyl group or two to dihydrokaempferol. Once converted to dihydroquercetin or dihydromyricetin, these intermediates flow through common downstream enzymes to form disubstituted (cyanidin, peonidin) and trisubstituted anthocyanins (malvidin, delphinidin and petunidin), when UFGT is expressed, and to form other polyphenols (flavanols, flavonols) at different developmental stages (Hernández-Jiménez et al., 2013). It has been demonstrated that the distribution of the ratio of disubstituted/ trisubstituted anthocyanins is associated to the expression of F3´5´H and F3´H, responsible for the hydroxylation of the flavonoid Bring. Castellarin and Di Gaspero (2007) showed that the F3´5´H/UFGT and F3´5´H/F3´H ratios are strongly correlated to the proportion of trisubstituted and disubstituted anthocyanins in several grape varieties. Gros noir exhibited a higher disubstituted/ trisubstituted ratio than the other varieties. A possible explanation for this is that the enzymes cited above are highly activated in this variety.
The anthocyanin profile is under a strong genetic control, therefore, the profiles of anthocyanins for each variety are relatively stable, while absolute concentrations can vary widely between different vintages, due to both environmental and agronomical factors. The anthocyanin profile can therefore be used as a chemotaxonomic parameter for the classification of red Vitis vinifera varieties (Mattivi et al., 1990;Mattivi et al., 2006). For this reason, red berry grapevine cultivars can be classified into groups: -Cultivars with a higher proportion of disubstituted anthocyanins (Cardinal); -Cultivars with a higher proportion of trisubstituted anthocyanins (Gros noir); and -Cultivars with high non-acylated anthocyanins (Muscat noir).
From the description presented above, anthocyanins could be considered useful markers for distinguishing grape varieties, although this characteristic should be used with care since anthocyanin content is heavily influenced not only by agronomical factors such as soil composition, irrigation, light intensity, etc., but also by the year's climatic conditions (Adams, 2006;Conde et al., 2007).

Flavonol profile
The total amount of flavonols found in the grape extracts ranged from 20.37 to 49.13 mg/kg of berries (Table 4). According to Mattivi et al. (2006), the flavonol content in the skin of 64 Italian red grape varieties and 27 Italian white grape varieties varied between 3.81 and 80.37 mg/kg of berries and between 1.36 and 30.21 mg/kg of berries, respectively. Guerrero et al. (2009) showed flavonol content between 221 and 538 mg/kg of berries in five red grape varieties grown in Andalusia (Spain). In another study, Mélo et al. (2006) noted that the levels of flavonols in Italia and Patrícia varieties from Brazil were respectively 2.38 and 8.85 mg/100 g fresh weight.
Differences in flavonol levels compared to those obtained in this study are probably due to several factors, including the extraction technique and/or the sensitivity of the method, varietal differences and geographical origin of the samples analyzed.
Looking at the pattern of the flavonols, the quercetin-3-O-glucoside was the major flavonol in grape skin at harvest, accounting for 86.77% of total flavonol content, which is in agreement with previous work on numerous grape tissues in many cultivars (Mattivi et al., 2006;Castillo-Munoz et al., 2007;Hernández-Jiménez et al., 2013). The isorhamnetin-3-Oglucoside flavonol was the second in importance (4.90%), followed by kaempferol-3-O-glucoside (4.32%), usually occurring as minor compounds and accounting for no more than 5%, and myricetin-3-Oglucoside (4%). Kaempferol-3 type flavonols have been found as minor compounds contributing to the flavonol profiles of red grape cultivars (Mattivi et al., 2006;Castillo-Munoz et al., 2007). This flavonol was absent in the Gros noir variety; the same observation was made by Mattivi et al. (2006) for the red grape variety Tannat. Myricetin-3-O-glucoside was not detected in the samples analyzed, with the exception of the black variety Gros noir (5.42 mg/kg). This result does not coincide with that of Mattivi et al. (2006), who detected myricetin only in white varieties. By contrast, Tarara et al. (2008) did not found myricetin-3-O-glucoside in Merlot (black grape), which is consistent with the results of this study where this flavonol was not recorded in the red (Cardinal) nor in the black variety (Muscat noir); the same flavonol was found in trace amounts in pink varieties (Mattivi et al., 2006). This could probably be due to the fact that these varieties contain low amounts of anthocyanins. According to these authors, the varieties having mostly cyanidin-3-glucoside and peonidin-3-glucoside derivatives, such as Cardinal and Muscat noir, are expected to derive from white varieties.
In summary, the red grape cultivars showed remarkable differences in their flavonols, which can be considered as a cultivar characteristic. Gros noir had the highest proportions of myricetin, quercetin and isorhamnetin-3-O-glucoside but not kaempferol flavonols. However, the aforementioned results are only preliminary, and confirmation is needed with further analysis. Although the grape samples analyzed were grown in the same viticultural  conditions, it is obvious that several unstudied factors (e.g., differences in viticultural practices, soil, or ripening degree) could markedly change the flavonol profile of a red grape cultivar. Other authors have obtained a good differentiation of several red grape cultivars from La Mancha (central southern Spain) on the basis of their total contents of quercetin, kaempferol, and dihydroflavonols (Castillo-Munoz et al., 2007;. The role of light on flavonol synthesis in grape berries seems well established and it is consistent with the UV-protection function exerted by these compounds. Merlot berries exposed to sunlight showed up to 10-fold more flavonols compared to shaded berries (Spayd et al., 2002;Tarara et al., 2008). Cluster shading caused a significant reduction in the synthesis of flavonols in Shiraz (Downey et al., 2003) and Cabernet sauvignon grapes (Matus et al., 2009). UV radiation also had a positive effect on flavonoid accumulation (Spayd et al., 2002). Moreover, sunlight exposure specifically increased the quantity of quercetin-3-O-glucoside in Merlot grapes, while kaempferol-3-O-glucoside remained unchanged; however, the authors did not detect myricetin-3-O-glucoside (Tarara et al., 2008). Taken together, these results point out that light is the most important factor determining the levels of flavonols in grape berries.

CONCLUSION
The current study clearly shows that healthpromoting compounds exhibit differences, confirming their importance in varietal characterization. The Gros noir variety is characterized by the highest levels of anthocyanins and flavonols. Based on these results, varieties can be classified as Gros noir > Muscat noir > Cardinal. Although the differences in anthocyanin and flavonol content between the varieties analyzed were significant, the levels of these compounds are such that extracting polyphenols from grape skin might be economically viable. Grape skin could then be used (i) as a natural alternative to the synthetic antioxidants used in the food industry to prolong the shelf life of food or (ii) as a good antioxidant source in dietary supplements.