Polyphenolic Composition and Antioxidant Activity of Galician Monovarietal Wines from Native and Experimental Non-Native White Grape Varieties

Abstract

Galicia region (Northwestern Spain) wine production is mainly focused to high-quality white wines. The polyphenolic composition and the antioxidant activity have been determined to characterize the wines obtained from white grape varieties (Albariño, Branco Lexitimo, Caiño blanco, Godello, Loureiro, Torrontes, and Treixadura) grown in Galician protected production areas, and the wines elaborated with non-native varieties experimentally grown in Galicia (Chardonnay, Gewürztraminer, Pinot blanc, Pinot gris, Riesling, and Sauvignon blanc). Since Albariños are the varietal wines with the highest production and commercialization extension, we have studied their polyphenolic composition as a function of the production subarea. Three vintage years (2010, 2011, and 2012) have been considered to account for climate variability. Liquid chromatography with diode array and mass spectrometry detection was used to obtain the phenolic profiles. Wines from native grapes could be fully discriminated by their phenolic composition, with only Albariño being partially confused with other varietal wines. Albariño wines produced in O Rosal and Ribera do Ulla could be clearly differentiated, whereas Condado do Tea and Val do Salnés wines were partially confused. The polyphenolic profile was enough to differentiate the wines obtained from non-native grapes. The total polyphenols content and the antioxidant activity of white wines elaborated in Galicia were comparable, although significant differences were found among varieties.

Keywords:

• White wines
• Polyphenolic composition
• Antioxidant activity
• High-performance liquid chromatography
• Galician wines characterization

INTRODUCTION

The consumption of foods rich in polyphenols, including wine, has been associated with the benefits of their antioxidant properties on the prevention of cardiovascular diseases, certain types of cancer, and other diseases related to aging.^[1,^2] Most of the research effort worldwide has been made to determine the phenolic composition of red wines,^[3] while little attention has been paid to the non-colored phenols in white varietal wines.^[4] Since in the winemaking process white wines are elaborated with a minimal contact of the skins, or even without any, white wines contain significantly lower amounts of total polyphenols (TPs) compared to red wines; however, white wines show a non-negligible antioxidant activity mainly due to their content in hydroxycinnamic and hydroxybenzoic acids and flavan-3-ols.^[5–7]

Galicia is a region in Northwest Spain with a long tradition in quality white wines production, which accounts for about 33% of the agriculture. Almost 500 wineries produce about 300,000 hectolitres per year, protected by five Denominations of Origin (DO): RiasBaixas, Ribeiro, Valdeorras, Ribeira Sacra, and Monterrei. RiasBaixas, the most extended DO, includes five subareas: Val doSalnés, Condado do Tea, O Rosal, Ribera do Ulla, and Soutomaior. All these areas are very different in climate, soil, and viticulture practices. Albariño is the most extended white grape variety, mainly produced in DO RiasBaixas, and it is usually processed to make monovarietal wines. Other grapes such as Treixadura and Godello can be elaborated as monovarietal wines. Other native grape varieties (Loureiro, Caiño, and Torrontés) have a limited production and are frequently used in mixtures. BrancoLexitimo wines are attracting a growing attention despite their limited production.

In spite of the importance and quality of the Galician white wines, there is still little knowledge on their polyphenolic composition and antioxidant activity, and also on the ability of these parameters to characterize and differentiate the wines by variety and origin. Previous work assessed the impact of phenolic indices and chromatic characteristics on the differentiation and characterization of red and white Galician wines,^[8] together with the presence of the two isomers of resveratrol.^[9] The high-performance liquid chromatography (HPLC) analysis of several polyphenols and their correlation with the antioxidant activity in Albariño commercial wines has also been reported.^[10]

In this study we aim to determine the polyphenolic composition and the antioxidant activity of varietal wines from seven native white grapes grown in protected production areas. The power of these variables to differenciate the wines by variety will be fully discussed. Since Albariños are currently the most relevant wines by their production and commercialization extension, we have classified them by their subarea of production. In addition, in the current study we have included the wines obtained with foreign grape varieties that are currently grown in Galicia experimentally (Chardonnay, Gewürztraminer, Pinot blanc, Pinot gris, Riesling, and Sauvignon blanc). Three vintage years (2010, 2011, and 2012) have been considered to account for climate variability.

MATERIALS AND METHODS

Chemicals

Pure standards of caftaric acid 98% (CAS 67879-58-7), protocatechuic acid 98% (CAS 121-33-5), caffeic acid 98% (CAS 331-39-5), gallic acid 99% (CAS 149‐91‐7), (+) catechin 99% (CAS 154‐23‐4), (-) epicatechin 97% (CAS 490‐46‐0), (-) epicatechin-galate 98% (CAS 1257-08-5), procyanidin B1 98% (CAS 20315-25-7), procyanidin B2 98% (CAS 29106-49-8), quercetin 98% (CAS 117-39-5), isoquercetin (quercetin‐3‐glucoside) 98% (CAS 482-35-9), rutin (quercetin‐3‐rutinoside) 98% (CAS 153-18-4), and quercetin‐3‐glucuronide 98% (CAS 22688-79-5) were all supplied by Sigma-Aldrich (Steinheim, Germany). Individual standard stock solutions of 2000–8000 µg mL^–1 were prepared in methanol. Working solutions in water containing the target analytes were obtained by appropriate dilution. Solutions were stored at 20ºC and protected from the light. The Folin–Ciocalteu (FC) reagent, and 2,2-diphenyil-1-picrylhydrazyl (DPPH) were obtained from Sigma-Aldrich (Steinheim, Germany), methanol HPLC grade from Panreac (Castellar del Vallès, Barcelona, Spain), acetone HPLC grade and formic acid (98–100%) from Merck (Darmstadt, Germany), and acetonitrile LC-MS Chromasolv from Fluka (Germany). Ultrapure water was produced in the laboratory with a Milli-Q gradient system (Millipore, Bedford, MA, USA).

Samples

Monovarietal wines elaborated with white grapes (Vitis vinifera sp) of native Galician varieties Albariño, Caiño, Godello, Loureiro, Torrontés, and Treixadura; were kindly supplied by 22 wineries of Galicia (NW Spain) belonging to five protected DO: Monterrei, Rias Baixas, Riberia Sacra, Ribeiro, and Valdeorras. The Rias Baixas DO samples include the recognized production sub-areas Condado de Tea, O Rosal, Ribeira do Ulla, and Val do Salnés. Wine samples of foreign white grape varieties (Chardonnay, Gewürztraminer, Pinot blanc, Pinot gris, Riesling, and Sauvignon blanc) were obtained at the Experimental Station for Viticulture and Winemaking of Ribadumia (Pontevedra). Also, wines of the variety Branco Lexitimo (also known as Albarin), elaborated at the Centre of Training and Experimentation in Agroforestry (Guisamo, A Coruña) have been considered. A total of 99 wine samples were obtained from three vintage years (2010, 2011, and 2012). Table 1 shows the samples distributed by variety and vintage. The number of samples of each variety and each geographical area was in accordance to its relevance on the total grape production. All samples were analyzed in duplicate or triplicate.

TABLE 1 White wine samples included in the study

	Vintage year
Variety	2010	2011	2012	Total
Albariño	16	20	16	52
Branco lexítimo	2	1	1	4
Caiño blanco	1	3	2	6
Godello	2	5	4	11
Loureiro	2	1	3	6
Torrontés	1	1	0	2
Treixadura	3	4	2	9
Chardonnay	1	1	1	3
Gewürtztraminer	1	0	0	1
Pinot blanc	1	0	0	1
Pinot gris	1	0	0	1
Riesling	1	1	0	2
Sauvignon blanc	1	0	0	1
Total	33	37	29	99

TPs

TPs content was determined according to the FC colorimetric method,^[11] using a calibration curve prepared with gallic acid standard solutions in concentrations ranging from 3 to 20 mg.L^–1 (R² = 0.9982), and expressed as gallic acid equivalents (GAE mg L^–1).

DPPH Radical Scavenging Activity

DPPH scavenging activity was determined using a modified method against Trolox^®.[12] DPPH 0.1 mM was dissolved in 100% methanol. Wine (0.1 mL) was added to 3.9 mL of the methanol DPPH solution. The mixture was shaken vigorously and allowed to stand at room temperature in the dark for 30 min. The decrease in absorbance of the resulting solution was monitored at 515 nm at 30 min. The antiradical activity (AA) was determined using the following equation (y = 0.5223x + 0.0276; R² = 0,999) obtained from linear regression after plotting the A₅₁₅ of known solutions of Trolox against concentration (0.08-1 mM). The radical stock solution was prepared fresh daily.

HPLC-Diode Array Detector (DAD)

Wine samples were filtered through 0.22 µm PVDF filter (Simplepure, Membrane Solutions, TX) and analyzed using a Varian Prostar HPLC equipped with a DAD, and a 3.9 × 150 mm, 4 µm, 60 Å, Waters Nova-Pak C₁₈ column. The injection volume was 20 µL. The mobile phase solvents were (A) 1% formic acid/water, and (B) 1% formic acid/methanol. The mobile phase gradient program started with 5 % B, changed to 20% B at 20 min, and then changed to 100% B at 25 min. The entire HPLC run time was 25 min with a flow rate of 1.0 mL/min, and 50°C column temperature. Acids, catechin and epicatechin, were detected at 280 nm, and quercetins were detected at 350 nm. Polyphenols were identified by comparison of their retention times and ultraviolet (UV) spectra with those of pure standards. Calibration was performed in the range of concentrations 1–200 mg.L^–1 for the acids, and 5–700 mg.L^–1 for the flavonoids, with 5 concentration levels and two to three replicates by level. Good linearity (R² ≥ 0.9903) and reproducibility (relative standard deviation [RSD] ≤ 12%) were achieved.

Liquid Chromatography with Tandem Mass Spectrometry (LC-MS-MS)

The liquid chromatographic system used was a Finnigan Surveyor™ HPLC Thermo Fisher Scientific (Madrid, Spain) equipped with a TSP AS3000 autosampler. The experimental conditions are the same used with the HPLC-DAD. Electrospray mass spectrometry was performed with a TSQ Quantum Discovery triple stage quadrupole mass spectrometer from Thermo Fisher Scientific. Column effluent was monitored using the Selected Reaction Monitoring (SRM). Polyphenols were detected in the negative mode using Electrospray Ionization (ESI) and thus, producing mainly the [M-H]⁻ pseudo molecular ions, with the exception of quercetin, quercetin-glucuronide, and quercetin-glucoside, which were detected in the positive mode. The ESI-MS/MS was operated with a scanning range of m/z 100–600. The capillary voltage and temperature were set to 3.0 kV and 320°C, respectively. High purity nitrogen (99.9%) was used as sheath gas and auxiliary gas at 40 and 10 psi and 350°C, respectively. Argon was the collision gas at 30 psi. Identification was performed using ESI-negative ionization [NI] of precursor > product ion transitions. The m/z values for the parent/product ions pairs are shown in Table 2, as well as the corresponding collision energies. Table 2 also summarizes the linearity study for each compound.

TABLE 2 LC-MS/MS analytical parameters

Compound	Retention time (min)	Parent ion (m/z)	Product ions (m/z)^a	Collision Energy (eV)	Calibration range (µg mL^−^1)	Determination coefficient (R^2)
Gallic acid	2.19	169.0 [M-H]^−	125.0	26	25-5000	0.9991
Protocatechuic acid	3.56	152.9 [M-H]^−	108.0/109.0	26/17	5-1000	0.9982
Caftaric acid	3.91	310.9 [M-H]^−	148.9/174.9/178.9	30/19/26	25-5000	0.9998
Procyanidin B1	4.38	577.0 [M-H]^−	288.9/407.0/424.9	26	50-5000	0.9983
(+)Catechin	4.88	289.0 [M-H]^−	203.1/245.0	26/15	100-5000	0.9985
Procyanidin B2	5.38	577.0 [M-H]^−	288.9/407.0/424.9	26	50-5000	0.9982
Cafeic acid	5.83	178.9[M-H]^−	134.0/135.0	28/19	50-5000	0.9983
(-)Epicatechin	6.14	289.0 [M-H]^−	203.1/245.0	26/15	50-5000	0.9976
(-)Epicatechin gallate	7.34	441.0 [M-H]^−	125.0/169.0/289.0	26	50-2000	0.9962
Quercetin-3-glucuronide	11.14	479.0 [M+H]+	302.9/461.5	18/14	5-5000	0.9989
Quercetin-3-glucoside	11.30	465.0 [M+H]+	256.9/302.9	41/14	100-5000	0.9945
Quercetin-3-rutinoside	10.33	609.1 [M+H]-	178.8/270.9/300.0	44/56/37	5-2000	0.9993
Quercetin	11.47	303.1 [M+H]+	153.0/229.1	33/28	500-2000	0.9961

^aUnderlined product ions were used for quantification.

Statistical Analysis

One-way analysis of variance (ANOVA) was performed to determine the statistical significance of the vintage year, and to study the ability of each variable to establish homogeneous groups of wines based on grape variety and production zone. Linear discriminant analysis (LDA) was performed to classify wines in the pre-established categories. As a supervised technique of pattern recognition it requires previous knowledge of the samples origin. It generates a number of linear discriminant functions which is equal to the number of categories minus one. LDA was applied to discriminate among varietal wines, and among Albariño wines belonging to different production zones. The software used was Statgraphics XV Centurion software package (Manugistics Inc, Rockville, MD).

RESULTS AND DISCUSSION

Polyphenolic Composition and Antioxidant Activity of the Galician Monovarietal White Wines

To account for variability due to the many factors affecting the final concentration of polyphenols in grape and thus, in the resulting wine, it is mandatory to consider several vintage years.^[13] In fact, the vintage,^[14] the maturity of the grapefruit, soil, and climate,^[15,^16] among other factors related to winemaking process are responsible of the high differences in phenolic composition through years. Wine samples considered in this study were obtained from three vintages (2010, 2011, and 2012). Results of an ANOVA showed that 2012 wines presented antioxidant activity significantly higher (p < 0.05, confidence level, 95%) than wines from vintages 2010 and 2011. Regarding TP content, wines from 2011 and 2012 vintages showed significantly higher values than 2010 wines.

Average values of AA, TP, and the individual polyphenols concentrations are shown in Table 3. To avoid the large dispersion between the different polyphenolic variables values, these variables have been normalized using the following formula: log (X) +1, where X is the value of each of the transformed variables. Caíño and Treixadura wines showed the highest values of AA and TP followed by Loureiro and Albariño wines, while Branco Lexitimo, Godello, and Torrontés were characterized by much lower AA and in general also TP content. Caiño and Loureiro wines showed a considerably high sum of individual polyphenols. This confirms that each polyphenol contributes differently to the wine antioxidant capacity, so this is not simply given by the sum of the antioxidant capacities of each component, but determined by the interaction between them, being possible synergic behaviors.^[15] On the other hand, the lower polyphenol content was found in Godello, Branco Lexitimo, and Torrontés wines.

TABLE 3 Average values of AA, TP, and individual polyphenols in Galician varietal white wines

	Albariño	Branco Lexítimo	Caíño	Godello	Loureiro	Torrontés	Treixadura
AA	1086^b±34	607^c±72	1247^ab±153	666^c ±45	1138^ab±79	695^c±43	1264^a±70
TP	295 ^b±5	202^c±5	335^a±10	221^c±5	300^ab±5	261^bc±22	325^a±11
Gallic acid	5465^b±444	1170^d±25	12580^a±2432	3187^c±175	10244^a±2044	1052^d±28	7145^ab±1281
Caffeic acid	8319^b±796	5983^b±1271	8257^b±1524	1620^c±197	20349^a±3208	6696^b±1745	7263^b±1712
Caftaric acid	36499^b±1458	23075^b±3110	51036^a±1804	6867^c±1183	48585^a±3111	7434^c±521	32063^b±1514
Protocatechuic acid	537^d±32	260^e±49	585^cd±93	647^c±48	1179^a±66	169^e±10	834^b±77
Catechin	7430^a±503	1260^c± 373	10886^a±2156	3922^b±453	11837^a±2547	423d±42	7180^ab±1215
Epicatechin	5803^a±468	818^c±144	8232^a±2093	2727^b±337	9509^a±2153	514^c±9	7667^a±1732
Procyanidin B1	8752^bc±643	1239^d±203	13382^a±3087	5323^c±410	12425^ab±2549	1396^d±38	10023^b±1070
Procyanidin B2	2943^c±263	796^d±20	4840^ab±1228	2454^c±295	6928^a±1737	748^d±12	5140^a±1039
Quercetin-3-glucuronide	523^c±61	149^d±56	829^b±176	233^d±55	2113^a±222	40^d±1	687^bc±121
Sum of compounds	76275^b±3119	33793^c±3613	110628^a±11658	26985^cd±2399	123173^a±16109	18476^d±2265	78004^b±4637

Data: mean ± standard deviation.

Units: TP: mg GAE/L; AA: µmol Trolox/L; compounds: µg/L.

Superscripts a, b, c, d, e, indicate the ANOVA homogeneous groups.

We have compared our data with those previously published. Due to the broader extension of their cultivars, Albariño, Godello, and Treixadura grapes can be currently processed as monovarietal wines in the DO protected regions Rias Baixas, Valdeorras, and Ribeiro, respectively. Caiño, Loureiro, and Torrontés, are cultivated in minor extension and so, they are more frequently used in the elaboration of multi-varietal wines. Thus, regarding the varietal wines considered in this study, only data for Albariño, Godello, and Caiño were previously referenced (although Godello was produced outside Galicia in the neighboring region of Bierzo, León). Previous published TP content for Caiño wines produced in O Rosal was much lower (97 mg GAE/L, vintage 2009 data)^[17] than the average TP we found (335 mg GAE/L, 2010–2012 vintages data). Since 2009 production was very similar to that of 2010 included in this study, differences in TP may be attributed to different winemaking procedures in the sampled wineries.

Analyzing each compound individually, results also indicate that Loureiro, Caiño, and Treixadura wines present in general the highest concentrations of the target polyphenols, whereas Torrontés and Branco Lexitimo wines show the lowest concentrations. Regarding Albariños, obtained results were comparable to those previously found in samples from 2006 vintage,^[10] with the exception of procyanidines that we found at lower concentrations.

All varieties with the exception of Godello presented similar phenolic distribution profiles, with caftaric acid being the predominant compound. In Torrontés wines, caftaric and caffeic acids were at similar concentrations. In the case of Godellos, their profile is more uniform for all compounds, as also described for Godello wines produced in Bierzo region, although in this case slightly lower values were obtained for both TP and catechin.^[18]

The Pearson correlation coefficients between the AA and the concentration of the individual polyphenols have been calculated, showing the highest correlations of AA with catechin, epicatechin, and PB1 (0.67, 0.64, and 0.67, respectively). Lower correlation coefficients have been found with PB2 (0.53), gallic acid (0.53), and quercetin-3-glucuronide (0.43). Caftaric, protocatechuic, and caffeic acids showed the lowest correlation with AA (0.35, 0.35, and 0.24, respectively). Similar results were previously reported in a study on the polyphenolic changes in white wines during storage^[19] with high correlations of AA with epicatechin (0.78) or TP (0.856) in freshly bottled white wines. Other authors found a much lower correlation between AA and TP in white wines (0.55).^[20] High correlations of AA with the polyphenols content in the studied wines have also been found in the corresponding grape marc generated in their winemaking process.^[21]

Differentiation between Varietal Wines from Galicia Native White Grape Varieties

Discriminant analysis was applied to the wines produced with the autochthonous varieties Albariño, Godello, Treixadura, Caíño, and Loureiro. Torrontés and Branco Lexitimo wines were excluded from the analysis since they were elaborated in little wineries and we could not obtain samples from the three vintages. The three first discriminant functions explained 95.2% of variability. The eigenvalues, explained and cumulative variance (%), and canonical correlations for the functions are shown in Table 4. The coefficients of the variables in the discriminant functions indicated which variables had a greater influence on the discrimination (Table 5). Variables showing the highest discriminatory power in the first function were PB1, and also epicatechin and quercetin-3-glucuronide; in the second function, PB2, epicatechin, and PB1; and catechin in the third function. Using all variables a 92% of correct classification of the wines in their respective groups was observed. It is worthy to note that misclassified wines (12%) were only Albariño that were confused in similar proportions (about 4%) with Caiño, Godello, and Treixadura wines. The distribution of the wine samples in the two first discrimination functions graph (Fig. 1) shows that Loureiro wines appears noticeably separated from the other groups, mainly due to their high content of caffeic and protocatechuic acids, and quercetin-3-glucuronide.

TABLE 4 Eigenvalues, explained variance, cumulative variance, and canonical correlation of the discriminant functions

Galician varietal wines
Discriminant function	Eigenvalue	Explained variance, %	Cumulative variance, %	Canonical correlation
1	3.147	61.22	61.22	0.871
2	1.308	25.46	86.68	0.753
3	0.438	8.53	95.21	0.552
Albariño wines
1	4.705	82.63	82.63	0.90814
2	0.604	10.61	93.24	0.61369
Experimental wines
1	5187.62	80.52	80.52	0.99990
2	987.468	15.33	95.85	0.99949

TABLE 5 Coefficients of the standardized canonical discrimination functions

	Galician varietal wines			Albariño wines		Experimental wines
Compound	1	2	3	1	2	1	2
AA	0.969	–0.519	–1.065	–0,960	0,458	6,607	–9,328
TP	–0.668	0.459	0.880	1,233	–0,122	–3,323	1,884
Caffeic acid	1.009	0.243	–0.827	–0,717	–0,430	12,116	0,579
Caftaric acid	1.178	0.584	–0.246	–0,149	–0,683	18,082	6,239
Catechin	0.303	–0.576	1.922	–3,173	–2,139	–1,569	–1,768
Epicatechin	–1.324	2.079	–1.274	3,9793	3,411	2,576	–1,323
Gallic acid	0.209	0.028	0.777	–0,759	0,121	0,022	2,193
PB1	–2.122	1.628	–1.470	0,519	1,585	–11,958	–4,765
PB2	0.685	–2.862	0.927	–2,048	–2,248	2,398	–6,990
Protocatechuic acid	0.505	–0.874	–0.382	1,2948	–0,631	–1,663	0,667
Quercetin-3 glucuronide	1.315	–0.016	0.367	–0,1598	–0,036	—	—

FIGURE 1 Classification of Galician wines by varietal origin.

This classification indicates that the selected nine polyphenolic compounds in addition to the AA and TP indices as discriminating variables may be adequate to classify the wines by grape variety in the five groups considered. These results are consistent with the thesis that the differences related to genetic (varietal) and environmental influences are more deeply associated with the major polyphenols,^[22] with mainly wine flavanols PB1 and PB2, and caftaric acid, showing the largest influence to differentiate wines by both the geographic origin and variety; similar results regarding the varietal discrimination of white wines were obtained,^[23] although in this case the phenolic profile was not effective for differentiating the samples by their geographical origin.

To check for the robustness of the classification, we included in the analysis five additional samples: a wine elaborated in a little winery of O Rosal zone with a mixture of Albariño and Caiño grapes, being Albariño predominant; a wine of Torrontés variety elaborated in the Ribadumia Experimental Station for Viticulture and Winemaking; a wine of Branco lexítimo elaborated in the Guísamo Centre for Training and Experimentation Agroforestry; and two wines produced in the major Albariño producing zone (Salnés), acquired in a supermarket as Albariño varietal wines but at a price of less than 3 euro (price for this kind of wines is about 6–12 euro). Results of the analysis showed that the Albariño-Caiño wine was classified as Albariño (97%) indicating that the proportion of Caiño variety modifies the phenolic composition of the resulting wine. The two low cost Albariño samples were included in the Albariño group, without confusion with the other varieties, but with a probability of 70 and 94%, thus indicating that in this case the predominant variety used in the elaboration was Albariño but it has been blended with one or more varieties not considered in this study. Regarding the experimental wines of Torrontés and Branco lexítimo, they were both confused with Albariño/Godello wines (55%/43% and 79%/15%, respectively).

Differentiation of Albariño Wines Produced in the Subareas of DO Rías Baixas

DO Rias Baixas, with Albariño grape as the predominant variety, includes five production subareas (Val do Salnés, Condado do Tea, O Rosal, Ribeira do Ulla, and Soutomaior) that were sampled keeping the proportion with their wine production. In this way, 43 wine samples belonging to the DO were obtained of which 26 were from Val do Salnés, ten from Condado do Tea, four from O Rosal, and three from Ribeira do Ulla. Soutomaior area only represents 0.1% of the total DO production and was not considered in the study.

The average concentrations of polyphenols in the wines produced in the subareas of DO Rias Baixas are detailed in Table 6, as well as the homogeneous groups resulting from the ANOVA (confidence level, 95%). In general, O Rosal wines showed significantly higher values for all variables except for TP that was similar in Ribeira do Ulla wines. Condado do Tea wines showed the lowest values for the studied parameters. In addition to the geographical origin, the particular practices in winemaking might influence the results obtained for O Rosal wines. In fact, these wines were elaborated after the maceration of grapes during several hours, which favours the transfer of skin polyphenols among many other components to the must previously to wine fermentation. This would be consistent with the much lower TP reported for Albariño wines from O Rosal elaborated in a different winery to those participating in this study (107 mg GAE/L).^[17]

TABLE 6 Average values of AA, TP, and individual polyphenols in Albariño wines from Rías Baixas subzones

	Condado do Tea (n = 10)	O Rosal (n = 4)	Val do Salnés (n = 26)	Ribeira do Ulla (n = 3)
AA	982^c±53	1984^a±113	1052^c±42	1438^b±37
TP	266^c±7	367^a±18	308^b±9	327a^b±5
Gallic acid	4448^b±359	14979^a±3497	4110^b±251	5097^b±372
Caffeic acid	7921^b±598	19000^a±3091	7143^c±1318	1867^d±129
Caftaric acid	29897^c±3270	57192^a±5168	38650^b±1787	39427^bc±2261
Protocatechuic acid	441^c±18	779^a±157	614a^b±55	331^bc±30
Catechin	5352^c±655	20672^a±2211	7299^bc±565	10278^b±231
Epicatechin	4253^c ±594	17733^a±3368	5123^c±466	11225^b±581
Procyanidin B1	7239^b±533	24495^a±4082	8474^b±818	10447^b±474
Procyanidin B2	2155^c±194	10115^a±2174	2550^c±254	4662^b±279
Quercetin-3-glucuronide	276^c±40	1239^a±227	603^b±102	797^abc±197
Sum of compounds	61982^b±4714	166207^a±20713	74567^b±2881	84132^b±3702

Data: mean ± standard deviation.

Units: TP: mg GAE/L; AA: µmol Trolox/L; compounds: µg/L.

Superscripts a, b, c, indicate the ANOVA homogeneous groups.

Discriminant analysis was applied to the Albariño wines. Table 4 shows the eigenvalues, the explained and cumulative variance, and the canonical correlation of the discriminant functions. It can be seen that with the two first functions, 93.2% of the variability can be explained. The coefficients of the variables in the discriminant functions indicated that epicatechin was the variable with a greater influence on the discrimination in both the first two functions (Table 5). A correct classification of the O Rosal and Ribeira do Ulla in their respective groups was achieved, highlighting the clear separation of the O Rosal wines from those produced in the other subareas (Fig. 2). Samples of Val do Salnés and Condado do Tea were partially confused and misclassified; indicating that a correct classification of Albariño wines by their geographical origin would require additional variables.

FIGURE 2 Classification of Albariño varietal wines by production area.

Characterization of White Wines Elaborated with Experimental Foreign Grape Varieties Grown In Galicia

Wines from six non-native grape varieties grown in the DO Rias Baixas subzone Val of Salnés were considered. Varietal wines of Pinot blanc, Pinot gris, Gewurtzträminer, and Sauvignon blanc, from vintage year 2010; and Riesling and Chardonnay (2010, 2011, and 2012 vintage years; see Table 1) were provided by the Ribadumia Experimental Station for Viticulture and Winemaking.

Table 7 shows the average values of AA, TP, and the individual polyphenols. ANOVA was used to compare each wine composition (confidence level of 95%), allowing classifying them in homogeneous groups. Riesling wines AA was significantly higher than the other wines. Rieslings are also characterized by high TP and caftaric acid concentration. In general, reported data on wines from different geographical origin showed lower values than those obtained in our study, as is the case of Riesling.^[24,^25] We have found that Pinot blanc, Sauvignon blanc and Gewürztraminer show a polyphenolic profile characterized by the abundance of procyanidin B, which differentiate these wines from the other ones characterized by a higher content of caftaric acid. These results are different of those obtained for the varietal wines grown in France and Germany, indicating that the geographical origin has a critical influence in the polyphenol composition. For Sauvignon blanc wines from France, it was reported that gallic acid and caftaric acid were the most abundant phenols.^[26] Gewürztraminer wines elaborated in Germany^[25] showed a similar profile to that of Galician wines. Galician Chardonnay is characterized by its high levels of catechin, epicatechin, PB1, and PB2, although found concentrations are lower than those reported for these wines in other countries.^[23–27] However, caftaric and caffeic acids are at similar concentrations than those previously reported in Chardonnays.^[23]

TABLE 7 Average values of AA, TP, and individual polyphenols in experimental wines

	Chardonnay	Pinot Blanc	Pinot Gris	Gewürtztraminer	Riesling	Sauvignon blanc
AA	676^c±42	720^bc±7	757^bc±4	816^b±3	1247^a±58	716^bc ±4
TP	254^ab±21	214,5^b±0,1	247^ab±3	239^ab±2	280^a±3	230^b±1
Gallic acid	1241^ab±107	563^c±40	991^b±5	1015^b±20	1345 ^a±78	1323^a±83
Caffeic acid	14996^a±107	786^b±19	15157^a±132	986^b±32	12303^a±3623	1654^b±12
Caftaric acid	23839^b±8273	421^c±6	18471^b±182	202^c±2	62637^a±4952	511^c±9
Protocatechuic acid	173^b±10	131^b±10	367^a±7	380^a±10	314^a±42	108^b±2
Catechin	1643^a±112	582^bc±57	346^c±10	742^b±53	583^bc±54	663^b±27
Epicatechin	1943^a±199	773^b±22	515^b±10	503^b±11	575^b±32	591^b±25
Procyanidin B1	3520^a±372	1299^c±4	1558^c±28	2690^b±32	1558^c±18	1723^c±30
Procyanidin B2	1585^a±109	726^b±3	342^c±8	380^c±9	735^b±33	782^b±20
Sum of compounds	48940^b±907	5280^c±44	37746^b±252	6899^c±82	80051^a±1370	7355^c±187

Data: mean ± standard deviation.

Units: TP: mg GAE/L; AA: µmol Trolox/L; compounds: µg/L.

Superscripts a, b, c, indicate the ANOVA homogeneous groups.

LDA was applied to the foreign varieties wines. Due precisely to the experimental character of their elaboration, the number of samples was limited. We analyzed three different samples of each kind of wine from each available vintage, so 24 cases were included in the LDA. The results of the analysis should be considered in this context. Five discriminant functions with p-values less than 0.05 were statistically significant at the 95.0% confidence level. With the first two, almost 96% of variability could be explained (Table 4), being caftaric acid the compound showing the greater discriminating power in these both functions (Table 5). Figure 3 shows that differences in polyphenolic composition are sufficient for clearly classify the wines in their respective varietal groups (percent correct classification = 100%). It should be accounted that all wine samples come from grapes cultivated in the same geographical area and, consequently, both terrain features and climatic characteristics were the same in each vintage. Then, the varietal character represented here by the polyphenolic profile would be enough to distinguish between wines.

FIGURE 3 Discrimination functions for the wines elaborated with foreign grape varieties.

Comparison of Galician Wines from Native and Experimental Foreign Varieties

Figure 4 shows the comparison of the varietal wines considering the sum of the individual polyphenols, TP, and AA values. Despite the caution with which we must draw conclusions due to the experimental character of the wines from foreign varieties considered in this work, it can be seen that regarding polyphenolic composition, both native and non-native varieties compare well, with Riesling, Chardonnay and Pinot gris wines placed among those with the highest concentration of polyphenols. This conclusion is further reinforced when considering the AA and TP, for which there are not differences between wines from a source or another.

FIGURE 4 Comparison of varietal wines elaborated with Galician native and non-native grapes, a: sum of individual compounds concentrations, b: TP and AA values.

Conclusions

The phenolic composition and the antioxidant activity of wines obtained from white grape varieties cultivated in Galicia (Spain) were determined. The study includes 99 wine samples elaborated with 13 grape varieties from 3 vintage years in 22 wineries and 2 experimental centres. Wines from native and foreign varieties grown in the region were characterized by their polyphenolic profile. Results indicate that wines from native grapes can be fully differentiated by their phenolic composition, with the exception of Albariños that are partially confused with other wines. Also, classification of Albariño wines by their area of production was studied, finding that wines from O Rosal and Ribera do Ulla could be clearly differentiated, whereas those of Condado do Tea and Val do Salnés were partially confused. Wines elaborated with foreign grapes cultivated in Galicia have also been studied, and results indicate that their polyphenol profile was enough to differentiate the corresponding monovarietal wines. The total content of polyphenols and the antioxidant activity of all wines elaborated in Galicia were comparable, although significant differences have been found between varieties. Among the native wines, Caiño and Treixadura showed the highest TP and AA, whereas Riesling highlights by those parameters among the non-native wines.

ACKNOWLEDGMENTS

The authors are also very grateful to the collaborating wineries, as well as the Ribadumia Experimental Station for Viticulture and Winemaking, and the Guisamo Centre of Training and Experimentation in Agroforestry.

FUNDING

This research was supported by European Regional Development Fund 2007-2013 (FEDER), and projects 09TAL012209PR and GPC2014/035 (Consolidated Research Groups Program), both from Galician Government (Xunta de Galicia).

Additional information

Funding

This research was supported by European Regional Development Fund 2007-2013 (FEDER), and projects 09TAL012209PR and GPC2014/035 (Consolidated Research Groups Program), both from Galician Government (Xunta de Galicia).