Use of Anthocyanin Profi les When Differentiating Individual Varietal Wines and Terroirs

Phenolic compounds are signifi cant secondary plant metabolites from grape skin, seeds and pulp that are extracted into wine (2). They form a group of secondary metabolites with various chemical structures and functions, and are produced during physiological growth of plants or in response to diff erent forms of environmental stress (3). Phenolic compounds can be successfully used for the assessment of wine authenticity as they are characteristic for the type of wine and can provide information on its geographical origin (4). Two groups of fl avonoids, the anthocyanins and fl avanols, are particularly important for the quality of red wines.


Introduction
Grapevine (Vitis vinifera L.) is one of the most widespread cultural fruit crops in the world.Its planting locations are situated between 30° and 50° latitude in both northern and southern hemispheres.Wine, the main product made from grapes, is greatly infl uenced by its geographical origin.
As far as wine-making and wine marketing are concerned, the preservation of valuable properties and qualities of wine (e.g.genetic parameters, geographical origin, and/or unique wine-making technologies) is very important.The chemical composition of wine is therefore considered to be an essential part of its quality and authenticity (1).
Phenolic compounds are signifi cant secondary plant metabolites from grape skin, seeds and pulp that are ex-tracted into wine (2).They form a group of secondary metabolites with various chemical structures and functions, and are produced during physiological growth of plants or in response to diff erent forms of environmental stress (3).Phenolic compounds can be successfully used for the assessment of wine authenticity as they are characteristic for the type of wine and can provide information on its geographical origin (4).Two groups of fl avonoids, the anthocyanins and fl avanols, are particularly important for the quality of red wines.
Anthocyanins are very important compounds because they determine the colour of both grapes and red wine.As far as the red wine is concerned, colour can be regarded as one of the most important parameters.An thocyanins can be defi ned as water-soluble vacuolar pigments that belong to the parent class of molecules called fl avonoids.They are glycosides of anthocyanidins of which the most important are: pelargonidin, cyanidin, peonidin, delphinidin, petunidin and malvidin.Anthocyanins can conjugate with hydroxycinnamic acids (p-coumaric or caff eic acid) and also with some other organic acids (e.g.acetic acid).In red wine made from diff erent varieties of Vitis vinifera L. one of the most important groups of anthocyanins are monomeric 3-O-diglucosides, especially malvidin--3-O-diglucoside and its derivatives (5).
The genome of each grapevine variety predetermines its typical profi le of anthocyanins, by which individual grapevine cultivars can be diff erentiated (6)(7)(8)(9)(10)(11)(12) and thus the anthocyanin profi le can be used to identify the grape variety used for making varietal wines (13).Consequently, it can also be used for the authentication of red wines, especially in the cases when the grapevine variety is displayed on the label (14).
Characteristics of wine are very closely related to its terroir.The wine origin has been and still is considered an indicator of wine quality and consumers o en require information on the wine provenance (15).Generally, the differentiation and classifi cation of wine samples on the basis of their chemical composition, geographical origin, variety and/or quality represent the basic methods of application of chemometry in oenology (16).The content and composition of anthocyanins in grapes and wine may be infl uenced by both environmental and genetic factors.Climatic factors (e.g.temperature, solar radiation, availability of water and its content in soil) infl uence signifi cantly the content of anthocyanins in grapes (17)(18)(19).
Viticulture and wine-making represent an important part of our cultural heritage (20).The region of Dolní Kounice (Czech Republic) is a territory with a long-term tradition of growing grapevine varieties used for making red wine.
The aim of this study is to determine varietal authenticity and terroir of red wines based on their anthocyanin content and composition.Another objective of this study is to identify the anthocyanin pigments that characterize well not only the cultivated grapevine varieties but also the terroir of each wine sample.

Samples and wine technology
Evaluated wine samples were made from grapes of the following varieties: Blaufränkisch (seven samples), Blauer Portugieser (eight samples) and Saint Laurent (two samples).These red wines were produced in three diff erent terroirs situated in the cadastre of the village Dolní Kounice, viz.Karlov (ten samples), Šibeniční vrch (fi ve samples) and Na Nivách (two samples).The list of the evaluated wine samples is shown in Table 1.The village Dolní Kounice is situated in the wine-growing region Moravia, sub-region Znojmo (Czech Republic).Grapes were produced by two growers in Dolní Kounice and red wine was made using an identical technology.Grapes were hand-picked into plastic boxes and immediately after the harvest, they were transported to a winery for processing.
The mash was macerated in stainless steel vinifi cators and regularly stirred.Within the fi rst two days of maceration, the temperature was maintained between 28 and 30 °C and therea er it was decreased to 25 °C.The malolactic fermentation took place a er the end of alcoholic fermentation.A er the end of fermentation, the wine was stored and aged in stainless steel containers for a period of 6 months.Wine aged for 6 months was analysed continuously.All analyses were done in triplicate.
The reproducibility of the method was verifi ed by spik ing one sample from each series with malvidin-3-glu- Concentrations of anthocyanins for which standards were not available were determined by the calibration curve of MvGl.For the purpose of this work, the ratio of their peak areas is crucial, rather than their exact concentration.A typical chromatogram of anthocyanins is shown in Fig. 1.

HPLC analysis
The wine samples were centrifuged at 3000×g for 6 min and then diluted with 100 mmol/L of HClO (ratio 1:1).The chromatographic system Shimadzu LC-10A (Shimadzu, Kyoto, Japan) was used, which consisted of two pumps LC-10ADvp, a column thermostat with manual injection valve, a DAD detector SPD-M10Avp, and a PC running the chromatographic so ware.The chromatographic separations were performed using a column Alltech ® Alltima™ C18 (3 µm, 3×150 mm, Grace, Deerfi eld, IL, USA) equipped with a guard column (3 mm×7.5 mm i.d.) fi lled with the same sorbent.
The temperature of separation was 60 °C.The gradient elution began with the fl ow rate of 0.6 mL/min in the mobile phase A (15 mM of perchloric acid).The mobile phase B (5 mM perchloric acid, 10 % methanol, and 50 % of acetonitrile) was used as follows: 2 % B for 0 min, 26 % B for 20 min, 45 % B for 30 min, 70 % B for 35 min, 100 % B for 37 min, 0 % B for 38 min, 2 % B for 40 min and 2 % B for 45 min.Acetonitrile (ACN) and methanol (MeOH) were of HPLC super gradient grade and perchloric acid was produced by Sigma-Aldrich (Prague, Czech Republic).

Statistical data treatment
The obtained data were processed according to the grapevine varieties and terroirs, and were expressed as mean values±standard deviations.The infl uence of a grapevine variety and terroir was determined using one--way analysis of variance (ANOVA) and the least significant diff erence (LSD) test at p<0.05.The multivariate chemometric methods were used as a supervised learning technique for diff erentiation of wines to groups on the basis of the wine-growing region and fi nding markers which showed a signifi cant discrimination value.When using a multivariate statistical analysis, it is necessary to designate suitable variables for classifi cation of the samples fi rst.In order to achieve this, the elimination of redundant variables using principal component analysis (PCA) was done.Principal components are orthogonal, and each principal component is a linear combination of the original variables.Canonical discriminant analysis (CDA) was subsequently performed using variables which showed higher signifi cance in the PCA assessment.

Concentrations of individual anthocyanins in varietal wines
Concentrations of individual anthocyanins in varietal wines are presented in Table 2.The highest average values were determined in red wines made from the variety Blauer Portugieser, lower in Blaufränkisch wine and the lowest in Saint Laurent.In Austrian red wine, however, the highest content of MvGl was detected in Blaufränkisch (23).In red wines, concentrations of monoglucosides were generally the highest.Increased concentrations of PtGl were found in varieties Saint Laurent (7.0 mg/L) and Blauer Portugieser (6.3 mg/L), while higher concentrations of PnGl were detected in Blaufränkisch As far as other anthocyanins were concerned, concentrations of malvidin derivatives were also the highest.The highest concentrations of MvGlAc and MvGlCm were found in the variety Blauer Portugieser (6.2 and 5.0 mg/L, respectively).On the other hand, the lowest concentrations of both aforementioned anthocyanins were estimated in the variety Blaufränkisch (1.9 and 3.6 mg/L, respectively).A lower content of acylated glucosides in the variety Blaufränkisch was determined also by Papoušková et al. (24) in their study on Czech red wines.

Concentrations of acylated anthocyanins in varieties
Saint Laurent and Blaufränkisch were signifi cantly lower.Similar results were obtained also in a study of Austrian red wine from varieties Blaufränkisch and Saint Laurent (23).In the variety Blaufränkisch, the content of acylated anthocyanins was markedly lower than that of glycosidic anthocyanins.On the other hand, in varieties Saint Laurent and Blauer Portugieser the sums of acetylated, coumarylated and total anthocyanins were nearly the same.

Concentrations of individual anthocyanins in wine samples with diff erent terroirs
Concentrations of individual anthocyanins in varietal wines according to diff erent terroirs are presented in Table 3. Similarly as in the case of individual varieties, MvGl was the dominant anthocyanin also in this study.As far as the concentration of MvGl was concerned, there were signifi cant diff erences among individual terroirs.The highest concentration of MvGl was found in the terroir Na Nivách.Of the group of monoglucosides, PtGl and DpGl occurred in signifi cant amounts in the terroir Karlov.In contradistinction to varieties, the concentration of DpGl can also be signifi cantly infl uenced by the terroir.As far as other anthocyanins were concerned, the most frequent were malvidin derivatives (similarly as in individual varieties).A signifi cantly higher concentration of MvGlAc was recorded in the wine originating from the terroir Šibeniční vrch, while the highest concentration of MvGlCm was observed in the terroir Karlov.
The highest sums of monoglucoside anthocyanins, ace tylated anthocyanins and coumarylated anthocyanins were detected in wines made from grapes produced in terroirs Na Nivách, Šibeniční vrch and Karlov, respectively.

Chemometric diff erentiation of wine samples on the basis of individual varieties and terroirs
The principal component analysis (PCA) is a multivariate method that quantitatively analyses the dependent variables.It was used to analyse all variables of all anthocyanins.The fi rst two components represented 58.6 % of total variance (PC1=35.8%, PC2=22.8%).On the basis of the PCA method, variables for the canonical discriminant analysis were selected.The following anthocyanins were eliminated from further assessment: PnGl, CyGlAc and MvGlCm.These compounds turned out to be insignificant for wine diff erentiation using the PCA method.In order to increase the discrimination value obtained by means of the PCA method, the geographical origin of wine and the varietal authenticity were verifi ed by means of canonical discriminant analysis (CDA).This method included only those variables which proved to have a high signifi cance for the diff erentiation of geographical and varietal origin using the PCA method.
Canonical discriminant analysis is a dimension-reduction technique related to PCA and canonical correlation.Given a nominal classifi cation of variables together with several interval variables, the CDA method derives those canonical variables that explain the inter-class variation in the same way as the principal components summarize total variation (25).The CDA was therefore used for the determination of anthocyanins which were also the most important for the wine diff erentiation on the basis of terroir and of grapevine variety in this study.
The variance of the diff erentiation of grapevine variety was 96.1 % for CVA1 and 3.9 % for CVA2 (i.e.100.0 % of total variation).The diff erences between grape varieties were signifi cant.The standardized canonical discriminant functions for varietal diff erentiation are presented in Table 4.
As it can be seen in Fig. 2, the obtained results indicate that CVA1 was positively correlated with DpGl, DpGlAc and PnGlCm, and negatively with PtGlAc, PnGlAc, CyGl- The results are expressed in mg/L of malvidin-3-O-glucoside (MvGl).The values followed by diff erent le ers in the same row are signifi cantly diff erent by LSD test at p<0.05 Cm and DpGlCm.CVA2 was positively correlated with DpGl and MvGl, and negatively with PtGl and PnGlAc.CVA1 enabled good separation of the variety Blaufränkisch from Blauer Portugieser and Saint Laurent.This means that the variety Blaufränkisch can be well characterised on the basis of the presence of anthocyanins DpGl, DpGlAc and PnGlCm.On the other hand, CVA2 separated the variety Blauer Portugieser from varieties Blaufrän-kisch and Saint Laurent.This means that the variety Blauer Portugieser can be characterised very well on the basis of the profi le of anthocyanin pigments DpGl and MvGl.
As far as the separation of varietal red wines was concerned, concentrations of anthocyanins DpGl, MvGl, PtGl, DpGlAc, PnGlCm and PnGlAc represented the most important discrimination parameters.MvGl showed to be an important discrimination parameter in samples of red wine originating from Austria (23), Canary Islands (26) and Greece (7).DpGL also belonged to discrimination compounds that enabled the identifi cation of not only Hungarian red wines (24), but also of red wines originating from some other European countries (7,27,28).Among  the group of acylated anthocyanins, DpGlAc was a good discrimination parameter of individual grapevine varieties in a study on Spanish wines (29).It was proven in the studies evaluating the varietal authenticity that PnGl can also be used as a discrimination parameter (23,26).In our study, this anthocyanin was eliminated already a er the application of PCA.
The explained canonical variance used for the diff erentiation of terroir represents 98.5 % for CVA1 and 1.5 % for CVA2.The diff erences between individual regions were signifi cant (Wilks' lambda value was 0.0072 at the value of p=0.01).The Wilks' lambda value near zero indicates a good discrimination function of the individual CVAs.Standardized canonical discriminant functions for terroir-based diff erentiation are presented in Table 5.
As it can be seen in Fig. 3, the obtained results indicate that CVA1 was positively correlated with DpGl, MvGl and DpGlCm, and negatively with CyGl, PtGl and Dp-GlAc.CVA2 was positively correlated with DpGlAc and CyGl, and negatively with PtGlAc and CyGlCm.CVA1 separated very distinctively the vineyard site Karlov from other two vineyard sites (i.e.Šibeniční vrch and Na Nivách).Wines originating from the vineyard site Karlov were characterised very well on the basis of the concentrations of anthocyanins DpGl, MvGl and DpGlCm.When using CVA2, it was not possible to discriminate wine samples originating from vineyards Šibeniční vrch and Na Nivách.Monoglucosides DpGl, MvGl and a derivative of p-coumaric acid, DpGlCm, could therefore be taken into account as the main discrimination parameters of red wine samples originating from diff erent vineyard sites on the basis of the profi le of anthocyanin pigments.
In the wine samples from the village of Dolní Kounice it is therefore possible to consider anthocyanins MvGl, DpG and DpGlCm as the main discrimination factors applicable for the diff erentiation of individual terroirs within a certain region.The suitability of using MvGl as discriminant factor was also corroborated in the studies of terroirs in Greece (7), Argentina (30) and Brazil (31).Similarly as in our study, the suitability of using DpGl as a discriminant factor was demonstrated only in a study dealing with the evaluation of Greek red wines (7).It is possible to conclude that phenolic compounds may be considered as chemical markers in studies on the authenticity of the variety and the geographic origin of wine (30).This was also corroborated by the results obtained within the framework of this study on red wine.
In the previous studies of the phenylpropanoid pathway (7,(32)(33)(34), it was demonstrated that environmental factors also infl uence the biosynthesis of anthocyanins.Monoglucosidic forms of anthocyanins represent the fi nal product of the phenylpropanoid pathway (35).This means that terroir is a contributing factor to the formation of anthocyanins.

Conclusions
The results of this study of anthocyanin fi ngerprinting of wine originating from the village Dolní Kounice show that it was possible to identify and separate not only grapevine varieties but also the terroir of wine.Very good results were obtained when trying to characterise the varieties Blaufränkisch and Blauer Portugieser on the basis of anthocyanins DpGl, DpGlAc, PnGlCm and DpGl, MvGl, respectively.It was found that the applicability of anthocyanin fi ngerprinting in the studies of the diff erentiation of wines from the viewpoint of terroir was very good.The terroir Karlov can be quite explicitly identifi ed on the basis of DpGl, MvGl and DpGlCm.This study on Moravian red wines therefore demonstrates the suitability of application of anthocyanin profi les in diff erentiation and identifi cation of individual varieties and terroirs.
(3.8 g/L) and Saint Laurent (3.4 g/L).Papoušková et al. (24) also mentioned similar contents of PtGl and PnGl in red wines originating from the wine-growing region Bohemia.In individual varieties, concentrations of DpGl were nearly the same.

Table 1 .
Grapevine varieties used in the study for the production of varietal wines coside.Recovery of the standards was in the range of (98-102) %.Retention times ranged from 26.23 to 26.41 min for MvGl.UV-VIS spectra were used to distinguish between p-coumarylglucoside and caff eoylglucoside derivatives of the anthocyanins.

Table 2 .
Concentrations of individual anthocyanins in varietal wines

Table 3 .
Concentrations of individual anthocyanins in samples of varietal wine originating from individual terroirs

Table 5 .
Standardized canonical discriminant functions used for terroir-based diff erentiation