Detailed Characterization of Proanthocyanidins in Skin, Seeds, and

Fractionation was conducted on an Agilent 1100 preparative HPLC system (Agilent, ... For phloroglucinolysis, an aliquot of each fraction (1 mL) was dr...
0 downloads 0 Views 861KB Size
ARTICLE pubs.acs.org/JAFC

Detailed Characterization of Proanthocyanidins in Skin, Seeds, and Wine of Shiraz and Cabernet Sauvignon Wine Grapes (Vitis vinifera) Rachel L. Hanlin,*,†,§ Mark A. Kelm,‡ Kerry L. Wilkinson,† and Mark O. Downey§ †

The University of Adelaide, School of Agriculture, Food and Wine, PMB 1, Glen Osmond, SA 5005, Australia Constellation Wines, U.S. R&D, 12667 Road 24, Madera, California 93637, United States § Department of Primary Industries Victoria, P.O. Box 905, Mildura, VIC 3502, Australia ‡

bS Supporting Information ABSTRACT: The distribution of proanthocyanidin (PA) polymer lengths, proanthocyanidin concentration at each polymer length, and polymer composition were determined in the seed, skin, and wine of Shiraz and Cabernet Sauvignon grape berries grown in southeast Australia. PA was fractionated by semipreparative high performance liquid chromatography (HPLC) and analyzed by phloroglucinolysis and HPLC to report the degree of polymerization (DP), concentration, and composition at 11 DP values in seed and wine and 21 DP values in skin. In skin, the highest PA concentration was observed at a DP of 31 in Shiraz and 29 in Cabernet Sauvignon representing 15% of the total PA in both varieties. The distribution of seed PA had the highest concentration at a DP of 7 in Shiraz and 6 in Cabernet Sauvignon representing around 30% of the total PA. In the wine PA distribution, the highest concentration was observed at a DP of 11 in Shiraz and 9 in Cabernet Sauvignon representing around 26 and 32% of the distribution, respectively. A second peak in wine PA concentration was observed at the largest DP of 18 in Shiraz and 15 in Cabernet Sauvignon representing around 20% of the distribution. The composition in wine did not vary at different DP, but the proportion of epicatechin gallate varied in seed PA less than 4 DP. The proportion of epigallocatechin increased with increasing DP in skin PA. Wine PA had a DP range and composition similar to the distribution of skin PA between DP 4 and 18 suggesting that larger skin PAs are not extracted into wine. This study provides information that could be used to target the important PA fractions in grapes that need to be measured to understand (or predict) PA extraction into wine and eventual mouthfeel. KEYWORDS: Vitis vinifera, proanthocyanidin distribution, polymer length, phloroglucinolysis, HPLC

’ INTRODUCTION Proanthocyanidins (PAs) are found in the seed and skin of wine grapes and are transferred into wine during vinification. In wine, PAs are responsible for astringency and long-term color stability.13 In grapes and wine, PAs are polymers composed of subunits analogous to the flavan-3-ols linked via interflavan bonds between the C-4 and C-8 carbon atoms and less commonly C-4 and C-6 atoms.4 The most common flavan-3-ols found in grape PA are catechin, epicatechin, epicatechin-gallate, and epigallocatechin.5,6 Differences in PA structures influence perceived astringency and have been partially characterized.7 An increase in astringency has been correlated with an increase in the degree of polymerization (DP) and galloylation, while the latter has also been associated with the level of “coarseness” and higher amounts of epigallocatechin influence the astringent perception of “smoothness”.7 However, as PA is extracted from grape into wine and as wine ages, the PA structure is modified by enzymatic and chemical processes and may polymerize, depolymerize, and form polymeric pigments.8 The structure of grape PA that is favored in these reactions and the influence it has on astringency and wine aging are relatively unknown. The different structures of grape PA will also influence its association with other material in grapes and wine including cell walls and proteins, as well as PA solubility and colloid formation. Some PA structures interact more strongly with cell walls, which influences the amount and type of PA extracted into r 2011 American Chemical Society

wine.9 For example, PA with larger DP and higher proportions of galloylation will preferentially bind to cell walls.10,11 However, the PA structures that interact with cell walls are yet to be thoroughly characterized. It is also yet to be established how different PA structures influence PA solubility in wine. For example, it is known that increasing ethanol concentration extracts more PA with increasing DP;12 however, it is not known if different PA structures are preferentially extracted by the fermentation matrix. PAs also form colloids in wine, associating with other material such as soluble grape and yeast polysaccharides, which can influence astringency. For example, the addition of some polysaccharides to wine can decrease astringency.13,14 The structure of PAs from grapes that form colloids are also largely unknown. Limiting our understanding of PAs and their involvement in these phenomena is a thorough characterization of PA distribution in grapes and wine. PA distribution is the concentration and composition of PA at different polymer lengths (DP). Determining PA distribution is the first step toward determining how PAs with different concentrations and composition in fruit are extracted into wine and the role they play in wine astringency,

Received: August 29, 2011 Accepted: November 15, 2011 Revised: November 8, 2011 Published: November 16, 2011 13265

dx.doi.org/10.1021/jf203466u | J. Agric. Food Chem. 2011, 59, 13265–13276

Journal of Agricultural and Food Chemistry wine color stability, and other chemical reactions and associations. One way that has been adopted to investigate PA distribution in grapes and wine is to separate PA by chromatography using C18 or silica columns and collecting PA fractions of different DP with a range of solvent mixtures that can be further characterized for concentration and composition.5,6,15,16 More recently, diol phase chromatography has been used to separate PA by increasing DP,17 and gel permeation chromatography, a size exclusion method, has been used to elute PAs of decreasing molecular weight to provide information on the distribution of DP.18 Fractionation of PA has enabled a DP range (smallest to largest DP) to be reported. Seed DP has been reported to range from 2 to 32 subunits, while skin and wine DP can range from 4 to 86 and 2 to 22 subunits, respectively.5,6,16,1822 The composition has been reported to change with the DP, with the proportion of epigallocatechin increasing with DP in skin and the proportion of epicatechin gallate increasing in seed.5,6,19,20,22 Despite a number of studies that report a DP range and the composition at different DP, there are few studies that report the amount of PA at each DP or distribution of PA.20,22 The only studies that have reported the PA distribution were in the skin of Cabernet Franc, Tourgia Nacional, Cabernet Sauvignon, and Castal~ao. In addition, the seed and wine distribution was reported for these varieties as well as for Trincadeira and Syrah (syn. Shiraz), although the skin distribution was not reported for Trincadeira and Syrah (syn. Shiraz).20,22 The two major red wine grape varieties grown in Australia are Shiraz and Cabernet Sauvignon. The DP range and distribution for Shiraz seed and wine and Cabernet Sauvignon seed, skin, and wine have previously been reported for grapes grown in Lisbon, Portugal. However, the DP range and distribution for Shiraz skin has not been reported, and the PA distribution for both varieties may differ for Australian growing conditions. The aim of this article was therefore to characterize the distribution of PA in terms of DP, concentration, and composition in Shiraz and Cabernet Sauvignon grown in Australian conditions in order to provide information on the PA distribution in grapes and its extraction into wine.

’ MATERIALS AND METHODS Chemicals. Methanol, acetonitrile, acetic acid (HPLC grade), and phloroglucinol were purchased from Merck (Melbourne, Australia). Acetone, sodium acetate, ascorbic acid, hydrochloric acid, catechin, epicatechin, and the Amberlite resin XAD7HP were purchased from Sigma Aldrich (Melbourne, Australia). Sample Collection. Shiraz and Cabernet Sauvignon grape berries were collected in 2009 from a single vineyard in the Sunraysia region of southeast Australia (34°270 S,142°140 E). Grape bunches were collected at commercial harvest with Shiraz harvested on the 4 March, 2009 at 23.5 oBrix and Cabernet Sauvignon on the 26 March, 2009 also at 23.5 oBrix. Approximately, 100 kg of whole bunches were collected across 2 rows (40 vines) and stored overnight at 4 °C to remove field heat. A 2.5 kg subsample of bunches was taken, and grape skins were collected by expulsion of the seeds and flesh. The seeds were removed from the remaining flesh and dried with paper towel. Skins and seeds were immediately frozen in liquid nitrogen and ground to a fine powder using an IKA grinder (All Basic grinder, IKA Works, Petaling Jaya, Malaysia). Samples were stored at 80 °C until analyzed. The remaining grape bunches were used for winemaking.

ARTICLE

Winemaking. Wine was made at the Provisor small scale winemaking facility in Merbein (Victoria, Australia) following their standard winemaking protocol.23 Shiraz and Cabernet Sauvignon fruit was crushed and divided into three fermentation vessels (i.e., 3 replicates of around 25 kg each). Wines were pressed after seven days of maceration and fermented to dryness. Following fermentation, wines were racked, cold stabilized, and bottled. Finished wines were stored at 18 °C until analyzed. Sample Preparation. Ground skin (100 g) was extracted in 600 mL of 70% aqueous acetone (v/v) containing 0.1% ascorbic acid with sonication at room temperature (∼23 °C) for 60 min. Solids were separated by vacuum filtration through Whatman #1 filter paper followed by evaporation of the extract under reduced pressure at 30 °C to remove acetone. The aqueous extract (∼150 mL) was then filtered through Whatman #2 filter paper to remove any precipitates. The filtrate was split into three aliquots (3  50 mL) for purification on an XAD7HP Amberlite resin column. Each aliquot (50 mL) was independently loaded onto an XAD7HP column (100 g in 30  500 mm sintered glass) preconditioned with water. The column was washed with water (600 mL) to remove flavan-3-ol monomers, sugars, and organic acids. Skin PAs were eluted with 65% methanol in water (450 mL, v/v). The three fractions eluted with 65% methanol were combined, then rotary evaporated under reduced pressure at 30 °C to remove methanol and freeze-dried to remove water. Ground seeds (25 g) were extracted with 70% aqueous acetone (v/v) containing 0.1% ascorbic acid (150 mL) with sonication at room temperature (∼23 °C, 60 min). Solids were separated by vacuum filtration through Whatman #1 filter paper followed by rotary evaporation of the extract under reduced pressure at 30 °C to remove acetone and freezedried to remove water. Wine PA was extracted by rotary evaporating 375 mL of wine under reduced pressure at 30 °C to remove ethanol. Once ethanol was removed, the remaining wine was loaded onto an XAD7HP column, and wine PA was eluted following the same protocol used for skin PA. Fractionation of Grape Seed, Skin, and Wine Proanthocyanidins. To analyze grape PA distribution, seed, skin, and wine PA were fractionated on a diol phase column by semipreparative liquid chromatography. Fractionation was conducted on an Agilent 1100 preparative HPLC system (Agilent, Melbourne, Australia) with a 250  20 mm ID, 100 μm Develosil diol column (Phenomenex, Melbourne, Australia) as previously described.17 Proanthocyanidin extracts (375 mg) were dissolved in a 3 mL mixture of 25:75 mobile phase A/B, centrifuged (16 100g, 10 min), and filtered through a 0.45 μm PTFE filter prior to injection (2000 μL). The binary mobile phase consisted of (A) acetonitrile/acetic acid (99:1, v/v) and (B) methanol/water/acetic acid (95:4:1, v/v/v) with a flow rate of 15 mL/min. Gradient conditions for fraction collection were 030% solvent B from 0 to 45 min, 30% isocratic solvent B from 45 to 65 min, 3085% solvent B from 65 to 68 min, and 85% isocratic B from 68 to 80 min. Column temperature was maintained at room temperature (∼23 °C) with eluent monitored at 280 nm using a diode array detector (DAD). Fractions were collected at one minute intervals from 0 to 80 min yielding 80  15 mL fractions. Injections (2000 μL) and fraction collection were conducted in triplicate.

Phloroglucinolysis and High Performance Liquid Chromatography. Subunit composition and the average degree of polymerization (DP) for each fraction were determined by acidcatalyzed cleavage in the presence of phloroglucinol24 followed by separation and quantification by reversed phase high performance liquid chromatography (RP-HPLC).25 Phloroglucinol adducts and terminal subunits in each fraction were identified by mass spectrometry as previously described.25 13266

dx.doi.org/10.1021/jf203466u |J. Agric. Food Chem. 2011, 59, 13265–13276

Journal of Agricultural and Food Chemistry

ARTICLE

Table 1. Distribution of PA Extracted from Shiraz Seeda Shiraz Seed PA concentration DP

b

c

% conversion d

(mg/L HPLC) (% by HPLC) (% by HPLC CE) (% by UVvis CE)

e

% of extension subunits

% of terminal subunits

yield

Cat

Ecat

Ecatgall

Cat

Ecat

Ecatgall

2

394.2

10.2

8.9

11.1

39.1

5.3

85.4

9.2

27.2

34.0

38.8

3

306.0

7.9

7.4

7.9

47.0

1.7

82.2

16.1

31.5

30.8

37.7

4

374.1

9.7

9.5

7.8

62.3

1.3

76.8

21.9

28.1

27.0

44.9

5

243.8

6.3

6.3

4.7

68.6

1.1

74.4

24.5

27.1

26.3

46.6

6 7

389.9 1209.4

10.1 31.3

10.1 32.4

7.6 37.4

63.9 60.0

1.2 1.3

74.2 73.3

24.5 25.5

27.9 27.6

25.4 26.1

46.6 46.2

12

445.3

11.5

12.2

11.7

52.6

1.0

68.0

31.0

22.1

28.5

49.3

14

210.6

5.5

5.6

4.9

57.2

0.9

71.6

27.6

24.7

29.1

46.1

15

258.9

6.7

6.9

6.3

60.1

1.0

72.5

26.5

25.7

28.5

45.8

17

29.0

0.8

0.7

0.6

68.3

1.0

76.4

22.6

27.3

25.2

47.5

total (mg/L)

3861.2

76.9

3.9

69.8

26.3

28.4

36.7

34.9

Total PA Extract 6 a

PA concentration, percent conversion yield, and the proportion of extension and terminal subunits are shown at individual DP. The average DP, percent conversion yield, and proportion of extension and terminal subunits are given for the total PA extract. Abbreviations: Epigall = epigallocatechin, Cat = catechin, Ecat = epicatechin, and Ecatgall = epicatechin gallate. b PA concentration at each DP following HPLC/phloroglucinolysis determined by the summation of individual subunit concentrations determined using published conversion factors.24 c The proportion of the total concentration for all DP calculated from the PA concentration determined following HPLC/phloroglucinolysis and the calculation of individual subunits using published conversion factors.24 d The proportion of the total concentration for all DP determined following HPLC/phloroglucinolysis and calculated as catechin equivalents. e The proportion of the total concentration for all DP determined by absorbance at 280 nm on a UVvis spectrophotometer before phloroglucinolysis. For phloroglucinolysis, an aliquot of each fraction (1 mL) was dried under reduced pressure, and phloroglucinol reagent (33 μL of 0.1 M hydrochloric acid in methanol, containing 50 g/L phloroglucinol and 10 g/L ascorbic acid) was added to each dried fraction and vortexed until the sample was resuspended. Samples were heated at 50 °C for 20 min. Aqueous sodium acetate (33 μL, 200 mM, pH 7.5) was then added to each fraction to stop the reaction. The sample was then centrifuged (16 100g, 5 min) and supernatant transferred to a HPLC vial for analysis. Following HPLC separation, subunit concentration was determined using a catechin standard curve and published conversion factors for each subunit relative to catechin.24 Total PA concentration was determined as both the sum of the individual subunit concentrations calculated using the conversion factors of each subunit relative to catechin and as the sum of each individual subunit calculated as catechin equivalents. DP was determined by dividing the sum of extension subunits and terminal subunits by the total of terminal subunits following determination of the concentration using conversion factors for each subunit relative to catechin. The average subunit composition and DP for the total extract prior to fractionation was also determined by phloroglucinolysis. Fractions of the same DP were pooled to report PA concentration, percent conversion yield, and the proportion of extension and terminal subunits at individual DP values.

Spectrophotometric Determination of Total PA Concentration and Percent Conversion Yield. In order to calculate

percent conversion yield for phloroglucinolysis, a 50 μL aliquot of each fraction prior to phloroglucinolysis was suspended in 450 μL of methanol (100%). The absorbance of each fraction diluted in methanol (100%) was recorded at 280 nm on a SpectraMax384 UVvis absorbance microplate reader (Molecular Devices, Australia) using polystyrene 96 well microplates (Greiner Bio-One, Australia). The total PA concentration of the fraction determined by UVvis spectrophotometry was determined as catechin equivalents against a

catechin standard curve. The percent conversion yield was determined as the proportion of total PA concentration determined by UVvis compared to the total PA concentration determined by HPLC following phloroglucinolysis and calculated as catechin equivalents.

’ RESULTS DP Range and Distribution. The DP range and distribution was determined by calculating the DP of each fraction and pooling the concentration of fractions with the same DP. The DP, concentration, percent conversion yield, and proportion of extension and terminal subunits of individual fractions are reported as Supporting Information. Individual fractions were numbered 1 to 80 to correspond with the time of elution (minutes). In seed PA, extension and terminal subunits were detected and identified by mass spectrometry from fractions 17 to 80 (Supporting Information Tables 1 and 2), while in skin PA, extension and terminal subunits were detected and identified from fractions 40 to 80 (Supporting Information Tables 3 and 4). Wine PA extension and terminal subunits were detected and identified from fractions 41 to 80 (Supporting Information Tables 5 and 6). In seed extracts, only flavan-3-ol monomers eluted prior to fraction 17, while in skin and wine extracts, PA was not detected prior to fraction 40 as monomeric and small oligomeric material were most likely removed by purification on the XAD7HP column. For Shiraz seeds, DP ranged from 2 to 17 subunits with distribution reported at 10 different DP values (Table 1). The concentration calculated by HPLC was highest at a DP of 7 representing 31.3% of the PA distribution followed by DP 12, 6 and 2, which represented 11.5, 10.1, and 10.2% of total Shiraz seed PA, respectively. The DP with the lowest proportion of PA calculated by HPLC was DP 17, which represented less than 13267

dx.doi.org/10.1021/jf203466u |J. Agric. Food Chem. 2011, 59, 13265–13276

Journal of Agricultural and Food Chemistry

ARTICLE

Table 2. Distribution of PA Extracted from Cabernet Sauvignon Seeda Cabernet Sauvignon Seed PA concentration DP

b

c

% conversion d

(mg/L HPLC) (% by HPLC) (% by HPLC CE) (% by UVvis CE)

e

% of extension subunits

% of terminal subunits

yield

Cat

Ecat

Ecatgall

Cat

Ecat

Ecatgall

2

237.0

6.7

5.9

5.8

39.3

4.3

86.1

9.6

30.1

37.5

32.5

3

381.3

10.8

10.0

8.6

39.3

2.0

84.1

13.9

37.4

33.0

29.6

4

372.5

10.5

10.3

8.1

59.5

1.3

81.4

17.3

32.9

28.7

38.4

5

226.1

6.4

6.5

5.1

64.8

1.1

77.9

21.0

30.2

26.7

43.1

6 7

1040.8 346.8

29.4 9.8

30.4 10.1

38.4 7.7

65.1 61.1

1.2 1.4

76.6 79.0

22.3 19.7

31.7 31.7

26.2 26.0

42.1 42.4

8

64.1

1.8

1.8

1.1

69.7

1.2

79.5

19.3

28.4

24.5

47.0

9

30.4

0.9

0.8

0.7

55.9

1.2

79.9

19.0

26.0

23.5

50.5

10

352.4

10.0

10.4

12.3

52.6

1.5

73.3

25.3

33.0

28.4

38.6

13

229.1

6.5

6.6

5.9

58.3

1.0

76.7

22.2

28.4

30.3

41.3

14

211.6

6.0

6.1

5.7

59.3

0.9

77.5

21.6

27.4

28.8

43.8

15

43.1

1.2

1.0

0.8

68.3

1.0

79.7

19.3

31.5

26.0

42.5

total (mg/L)

3534.9

80.9

5.4

71.1

23.5

35.6

33.7

30.7

Total PA Extract 6 a

PA concentration, percent conversion yield, and the proportion of extension and terminal subunits are shown at individual DP. The average DP, percent conversion yield, and proportion of extension and terminal subunits are given for the total PA extract. Abbreviations: Epigall = epigallocatechin, Cat = catechin, Ecat = epicatechin, and Ecatgall = epicatechin gallate. b PA concentration at each DP following HPLC/phloroglucinolysis determined by the summation of individual subunit concentrations determined using published conversion factors.24 c The proportion of the total concentration for all DP calculated from the PA concentration determined following HPLC/phloroglucinolysis and the calculation of individual subunits using published conversion factors.24 d The proportion of the total concentration for all DP determined following HPLC/phloroglucinolysis and calculated as catechin equivalents. e The proportion of the total concentration for all DP determined by absorbance at 280 nm on a UVvis spectrophotometer before phloroglucinolysis.

1% of the total. Each of the remaining DP represented between 5.5 and 9.7%. The total concentration of PA in Shiraz seed calculated by UVvis spectrophotometry and expressed as catechin equivalents was also highest at a DP of 7, representing 37% of the total PA concentration. DP 12 and 2 had the second highest concentrations representing 11.7 and 11.1% of the distribution, respectively. The DP of 17 also had the lowest concentration representing less than 1% of the distribution with the remaining DP representing between 4.7 and 7.9%. For Cabernet Sauvignon seed, DP ranged from 2 to 15 subunits with 12 different DP values observed (Table 2). The concentration calculated by HPLC was highest at a DP of 6 representing 29.4% of the PA distribution. By HPLC, the distribution represented around 10% for DP values of 3, 4, 7, and 10. The DP values of 8, 9, and 15 had the lowest distribution representing less than 2% of the total distribution each. The remaining DP values were each around 6% of the distribution. The determination of total PA concentration by UVvis spectrophotometry for Cabernet Sauvignon seed showed that the DP with the highest concentration, DP 6, represented 38% of the total distribution. The proportion at DP 10 was slightly higher than the proportion calculated by HPLC, representing 12% and was around 1% lower than the values calculated by HPLC at DP values of 3, 4, and 5. For Shiraz skin, DP ranged from 4 to 63 subunits with distribution reported at 21 DP values (Table 3). Calculated by HPLC, each DP value represented between 15.2 and less than 1% of the total concentration of PA extracted from the skin. The DP with the highest concentration calculated by HPLC was DP 31,

representing 15.2%, followed by DP 52 and 56, representing 12.8 and 10.7% of the distribution, respectively. Above DP 56, there were three DP values representing between 4.1 and 7.9% of the total concentration. Below DP 31, DP 14 had the highest concentration representing 9.5% with the remaining DP representing between 5.7 and less than 1% of the total Shiraz skin PA extract. The DP with the lowest concentration was DP 4 representing 0.4% of the distribution. The Shiraz skin PA distribution calculated by UVvis spectrophotometry differed from the results obtained by HPLC. By UVvis spectrophotometry, we observed that the DP with the highest PA concentration was DP 14, representing 14.5%, followed by DP 31 and 52, representing 12.7 and 10.5%, respectively. The remaining DP represented between 0.5 and 8.4% with values between DP 4 and 10 up to 5% higher than the proportion calculated by HPLC. In comparison, the proportion of the concentration calculated by UVvis spectrophotometry was slightly lower than those calculated by HPLC for DP above 31 subunits. For Cabernet Sauvignon skin, DP ranged from 4 to 76 subunits with distribution reported at 20 different DP values (Table 4). The DP with the highest PA concentration calculated by HPLC was DP 29, representing 14.6% of the distribution, followed by DP 42, representing 13.5%. The DP values next highest in proportion were DP 15, 48, and 53 representing 8.6, 9.1, and 10.6%, respectively. The remaining DP values represented between 0.5 and 7.3%. The DP with the lowest concentration of PA was DP 5 representing 0.5%. Similar to Shiraz skin, the distribution of Cabernet Sauvignon skin calculated by UV vis spectrophotometry was slightly different. The DP with the highest proportion was 15, representing 13.3%, followed by DP 13268

dx.doi.org/10.1021/jf203466u |J. Agric. Food Chem. 2011, 59, 13265–13276

Journal of Agricultural and Food Chemistry

ARTICLE

Table 3. Distribution of PA Extracted from Shiraz Skina Shiraz Skin PA concentration DP

b

c

% conversion d

e

(mg/L HPLC) (% by HPLC) (% by HPLC CE) (% by UVvis CE)

yield

% of extension subunits Epigall Cat Ecat Ecatgall

% of terminal subunits Cat

Ecat

Ecatgall

4

6.3

0.4

0.4

2.0

5.9

22.1

2.3 71.3

4.3

63.8

26.2

9.9

5

13.1

0.8

0.9

4.1

6.4

25.0

1.6 63.8

9.7

59.9

28.1

12.0

6

26.3

1.6

1.8

5.3

10.2

22.6

1.3 65.8

10.3

69.2

20.6

10.2

7

50.5

3.1

3.4

8.4

12.4

24.0

1.6 64.9

9.5

69.8

19.3

10.9

8 9

54.3 10.5

3.3 0.6

3.6 0.7

5.5 0.9

20.0 24.1

24.9 26.1

2.6 62.9 2.9 61.0

9.7 9.9

70.0 71.0

19.7 18.8

10.3 10.2

10

27.7

1.7

1.8

2.1

25.6

26.1

3.1 60.7

10.1

68.9

20.3

10.8

11

7.7

0.5

0.5

0.5

31.2

24.8

3.1 62.8

9.2

65.5

23.9

10.6

12

23.4

1.4

1.5

1.6

29.6

25.7

3.2 61.7

9.5

63.6

24.6

11.8

13

32.6

2.0

2.1

1.9

33.7

25.9

3.0 61.5

9.6

62.7

24.9

12.3

14

154.3

9.5

10.0

14.5

27.6

33.3

3.4 51.4

11.9

65.4

24.7

9.8

15

92.5

5.7

6.3

5.2

40.5

27.7

3.3 58.6

10.4

66.5

23.2

10.3

17 31

15.9 247.8

1.0 15.2

1.1 15.5

0.8 12.7

38.0 36.3

25.9 40.1

2.9 61.4 5.4 43.4

9.8 11.1

74.9 68.4

14.5 25.0

10.7 6.6

52

209.2

12.8

12.5

10.5

36.2

40.9

5.6 44.5

9.0

67.2

31.0

1.8

54

97.7

6.0

5.6

3.2

51.7

42.2

4.3 46.2

7.3

63.3

36.7

n.d.

55

114.9

7.0

6.5

3.9

50.2

43.3

4.8 44.7

7.3

64.8

35.2

n.d.

56

174.4

10.7

10.3

6.9

43.9

41.1

5.5 45.3

8.1

62.2

36.2

1.6

59

75.8

4.7

4.3

2.4

52.7

42.5

4.1 46.2

7.2

61.3

38.7

n.d.

62

128.6

7.9

7.5

5.3

41.5

41.8

5.0 45.6

7.7

63.4

36.6

n.d.

63 total (mg/L)

66.4 1629.6

4.1

3.7

2.1

51.5

43.5

4.2 45.6

6.7

61.0

39.0

n.d.

35.7

44.8

3.4 44.6

7.3

77.4

22.6

n.d.

Total PA Extract 52 a

PA concentration, percent conversion yield, and the proportion of extension and terminal subunits are shown at individual DP. The average DP, percent conversion yield, and proportion of extension and terminal subunits are given for the total PA extract. Abbreviations: Epigall = epigallocatechin, Cat = catechin, Ecat = epicatechin, and Ecatgall = epicatechin gallate, and n.d. = not detected. b PA concentration at each DP following HPLC/ phloroglucinolysis determined by the summation of individual subunit concentrations determined using published conversion factors.24 c The proportion of the total concentration for all DP calculated from the PA concentration determined following HPLC/phloroglucinolysis and the calculation of individual subunits using published conversion factors.24 d The proportion of the total concentration for all DP determined following HPLC/phloroglucinolysis and calculated as catechin equivalents. e The proportion of the total concentration for all DP determined by absorbance at 280 nm on a UVvis spectrophotometer before phloroglucinolysis.

values of 29 and 42, representing 12.2 and 10.3%, respectively. Similar to Shiraz skin, the proportion was slightly higher when calculated by UVvis spectrophotometry from DP 2 to 9, while it was slightly lower at the higher DP values of 29 to 76. The DP range, composition, and distribution were also analyzed in wine. For Shiraz wine, DP ranged from 6 to 18 subunits with distribution reported at 11 DP values (Table 5). Each DP value represented between around 1 and 26% of the total concentration. The DP with the highest proportion was DP 11 representing around 26% of the total PA when calculated by both HPLC and UVvis spectrophotometry. When calculated by HPLC, the DP with the second highest proportion was DP 18 representing 20.5%; however, when calculated by UVvis spectrophotometry, the proportion of PA at DP 18 was only 12.3%. The proportion of PA calculated by UVvis spectrophotometry was higher for DP values below DP 11 compared to that calculated by HPLC following phloroglucinolysis. The proportion of PA calculated by UVvis spectrophotometry was lower for high DP values compared to those by HPLC.

For Cabernet Sauvignon wine, DP ranged from 5 to 15 subunits with distribution reported at 11 DP values (Table 6). The distribution was similar when calculated by both HPLC and UVvis spectrophotometry. The DP value with the highest concentration was DP 9 representing 31.9% when calculated by HPLC and 34.3 when calculated by UVvis spectrophotometry. The DP value with the second highest proportion of the total PA distribution was DP 15 representing 20.9% when calculated by HPLC and 14.8% when calculated by UVvis spectrophotometry. DP 6 and DP 7 were the next highest in proportion representing around 10% each when calculated by HPLC and around 11% when calculated by UVvis spectrophotometry. In addition to the concentration calculated by HPLC and UVvis spectrophotometry, the percent conversion yield is also reported at each DP. For Shiraz seed, the percent conversion yield ranged from 39.1 to 68.6%. Cabernet Sauvignon seed had a similar conversion yield ranging from 49.3 to 69.7%. The seed conversion yield was lowest in both Shiraz and Cabernet Sauvignon at DP 2 and 3. The conversion yield for Shiraz skin 13269

dx.doi.org/10.1021/jf203466u |J. Agric. Food Chem. 2011, 59, 13265–13276

Journal of Agricultural and Food Chemistry

ARTICLE

Table 4. Distribution of PA Extracted from Cabernet Sauvignon Skina Cabernet Sauvignon Skin PA concentration DP

b

c

% conversion d

e

(mg/L HPLC) (% by HPLC) (% by HPLC CE) (% by UVvis CE)

yield

% of extension subunits Epigall Cat Ecat Ecatgall

% of terminal subunits Cat

Ecat

Ecatgall

4

14.1

0.8

0.9

2.8

9.2

30.5

3.3 61.3

4.9

74.5

19.6

5.9

5

8.2

0.5

0.5

1.4

10.3

31.1

3.1 61.1

4.8

76.2

18.0

5.8

6

25.3

1.4

1.6

4.4

8.4

29.7

2.4 65.2

2.8

73.0

20.2

6.8

7

48.9

2.7

3.2

5.6

13.0

29.6

2.6 63.2

4.6

77.8

17.0

5.2

8 9

36.9 24.4

2.0 1.4

2.4 1.6

3.3 2.1

18.0 18.7

32.0 31.8

2.9 60.4 2.7 61.0

4.7 4.6

78.8 76.7

16.1 18.3

5.1 5.1

10

33.4

1.9

2.1

2.6

18.3

34.7

3.0 57.8

4.4

70.8

21.7

7.5

11

44.2

2.5

2.7

3.5

19.3

34.8

2.9 58.0

4.3

72.7

22.0

5.3

12

24.2

1.3

1.5

2.2

15.0

34.5

2.9 58.1

4.4

71.3

22.8

5.8

13

16.3

0.9

1.0

1.1

19.6

34.5

2.9 58.6

4.0

69.1

24.6

6.3

14

110.1

6.1

6.9

7.1

23.5

36.2

3.4 55.9

4.5

76.1

18.6

5.4

15

154.9

8.6

8.6

13.3

18.5

55.8

4.0 33.0

7.3

79.7

16.2

4.1

29 42

263.0 243.3

14.6 13.5

14.5 13.0

12.2 10.3

30.9 36.1

50.2 51.8

5.6 39.6 5.1 39.3

4.5 3.7

82.3 77.6

16.5 20.5

1.2 1.9

48

162.9

9.1

8.7

6.1

37.1

50.4

4.7 41.9

3.1

68.8

31.2

n.d.

53

191.3

10.6

10.1

8.1

31.9

51.6

4.1 40.9

3.3

71.1

28.9

n.d.

55

76.1

4.2

4.0

2.9

40.1

51.5

3.0 42.9

2.6

60.1

39.9

n.d.

64

80.5

4.5

4.2

2.9

46.0

52.3

2.8 42.0

2.9

57.4

42.6

n.d.

66

110.8

6.2

5.8

3.6

44.0

51.0

3.6 42.5

2.9

58.8

41.2

n.d.

76

130.9

7.3

6.9

4.5

40.0

51.0

3.8 42.2

3.0

65.3

34.7

n.d.

total (mg/L)

1799.8

46.5

54.5

2.3 40.3

2.9

69.0

31.0

n.d.

Total PA Extract 48 a

PA concentration, percent conversion yield, and the proportion of extension and terminal subunits are shown at individual DP. The average DP, percent conversion yield, and proportion of extension and terminal subunits are given for the total PA extract. Abbreviations: Epigall = epigallocatechin, Cat = catechin, Ecat = epicatechin, Ecatgall = epicatechin gallate, and n.d. = not detected. b PA concentration at each DP following HPLC/ phloroglucinolysis determined by the summation of individual subunit concentrations determined using published conversion factors.24 c The proportion of the total concentration for all DP calculated from the PA concentration determined following HPLC/phloroglucinolysis and the calculation of individual subunits using published conversion factors.24 d The proportion of the total concentration for all DP determined following HPLC/phloroglucinolysis and calculated as catechin equivalents. e The proportion of the total concentration for all DP determined by absorbance at 280 nm on a UVvis spectrophotometer before phloroglucinolysis.

ranged from 5.9 to 52.7% and increased with DP. Percent conversion yield values also increased with DP for Cabernet Sauvignon skin ranging from 8.4 to 46%. There was also a slight increase in conversion yield with increasing DP for wine with the conversion yield ranging from 4.5 to 12.1% for Shiraz wine and 10.1 to 13.8% for Cabernet Sauvignon wine. Extension Subunit Composition. The proportion of extension subunits was determined at each DP. For Shiraz and Cabernet Sauvignon seed, epicatechin was present at the highest proportion representing between 68.0 and 86.1% of extension subunits across the DP range (Tables 1 and 2). Epicatechin was highest at the lowest DP values of 2 and 3, but remained relatively constant over the rest of the DP range. Catechin extension subunits in seed represented less than 2% of extension subunits at all DP except for DP 2. At DP 2, catechin extension subunits represented around 5% of total extension subunits in Shiraz and 4% in Cabernet Sauvignon. The proportion of epicatechin gallate as an extension subunit in seed increased in both Shiraz and Cabernet Sauvignon from DP 2 to 5 ranging from around 9 to 24%. The proportion of epicatechin gallate remained relatively constant across the rest of the DP range representing

between 24.5 and 31% in Shiraz and 19 to 25.3% in Cabernet Sauvignon. For Shiraz and Cabernet Sauvignon skin, epigallocatechin and epicatechin were the extension subunits present in the highest proportion (Tables 3 and 4). Epigallocatechin increased with DP in both Shiraz and Cabernet Sauvignon, ranging from 22 to 43.5% in Shiraz and 29.6 to 55.8% in Cabernet Sauvignon. Epicatechin decreased with DP ranging from 73.3 to 43.4% in Shiraz skin and 65.2 to 33% in Cabernet Sauvignon skin. Catechin and epicatechin gallate were also present as extension subunits in skin but in much lower proportions. In both Shiraz and Cabernet Sauvignon skin, catechin ranged from around 1 to 5%, while in Shiraz, epicatechin gallate ranged from 4.3 to 11.9% but was slightly lower in proportion in Cabernet Sauvignon skin representing between 2.6 and 7.3%. In wine, epicatechin was the extension subunit representing the largest proportion in both Shiraz and Cabernet Sauvignon wine ranging from around 54.7 and 64.2% (Tables 5 and 6). Epigallocatechin was the extension subunit representing the second largest proportion ranging from 28.6 and 35.5% in Shiraz and 32.3 and 39.6% in Cabernet Sauvignon wine. Catechin and 13270

dx.doi.org/10.1021/jf203466u |J. Agric. Food Chem. 2011, 59, 13265–13276

Journal of Agricultural and Food Chemistry

ARTICLE

Table 5. Distribution of PA Extracted from Shiraz Winea Shiraz Wine PA concentration DP

b

c

% conversion d

(mg/L HPLC) (% by HPLC) (% by HPLC CE) (% by UVvis CE)

e

yield

% of extension subunits Epigall Cat Ecat Ecatgall

% of terminal subunits Cat

Ecat

6

13.3

3.1

3.1

4.9

4.8

34.7

3.0 59.2

3.1

77.6

22.4

7

24.6

5.8

6.0

10.1

4.5

28.6

2.7 64.2

4.5

83.7

16.3

8

42.3

9.9

10.0

13.4

5.8

29.5

2.8 63.5

4.2

78.6

21.4

9

40.0

9.4

9.0

9.7

7.3

31.8

3.5 60.1

4.7

68.7

31.3

10 11

55.7 113.5

13.1 26.6

13.2 26.8

12.7 26.1

7.8 8.4

30.1 30.9

2.7 62.7 2.5 62.3

4.5 4.3

76.2 77.6

23.8 22.4

12

6.9

1.6

1.6

2.2

7.2

31.5

2.2 62.2

4.0

71.2

28.8

14

8.4

2.0

2.0

1.4

10.0

28.7

2.7 63.5

5.1

72.9

27.1

15

12.2

2.9

2.7

3.1

6.5

33.2

2.1 61.0

3.6

65.5

34.5

17

22.2

5.2

5.1

4.1

9.1

34.0

1.8 60.4

3.8

70.9

29.1

18

87.5

20.5

20.5

12.3

12.1

35.5

1.6 57.7

5.2

75.5

24.5

total (mg/L)

426.6

29.4

2.2 64.5

3.9

86.0

14.0

Total PA Extract 6 a

PA concentration, percent conversion yield, and the proportion of extension and terminal subunits are shown at individual DP. The average DP and proportion of extension and terminal subunits are given for the total PA extract. Abbreviations: Epigall = epigallocatechin, Cat = catechin, Ecat = epicatechin, and Ecatgall = epicatechin gallate. b PA concentration at each DP following HPLC/phloroglucinolysis determined by the summation of individual subunit concentrations determined using published conversion factors.24 c The proportion of the total concentration for all DP calculated from the PA concentration determined following HPLC/phloroglucinolysis and the calculation of individual subunits using published conversion factors.24 d The proportion of the total concentration for all DP determined following HPLC/phloroglucinolysis and calculated as catechin equivalents. e The proportion of the total concentration for all DP determined by absorbance at 280 nm on a UVvis spectrophotometer before phloroglucinolysis.

epicatechin gallate were also present as extension subunits in wine, representing 1 to 4% and 2 to 5% in both wines, respectively. Terminal Subunit Composition. The terminal subunit composition was determined at each DP. For Shiraz and Cabernet Sauvignon seed PA, the terminal subunits, catechin, epicatechin, and epicatechin gallate were present in similar proportions (Tables 1 and 2). Catechin ranged from 22.1 to 31.5% across the DP range in Shiraz seed and 27.4 to 37.4% across the DP range in Cabernet Sauvignon seed. Epicatechin ranged from 25.2 to 34.0% in Shiraz seed and 23.5 to 37.5% in Cabernet Sauvignon seed, while epicatechin gallate ranged from 37.7 to 49.3% in Shiraz seed and 29.6 to 50.5% in Cabernet Sauvignon seed. In skin PA, the terminal subunit composition was primarily composed of catechin, which represented between 57.4 and 76.9% in both Shiraz and Cabernet Sauvignon (Tables 3 and 4). Epicatechin was also present as a terminal subunit in skin, representing between 14.5 and 42.6% at various DP in both varieties. As a skin terminal subunit, epicatechin gallate was present in both varieties from DP 4 to DP 52 in Shiraz and DP 4 to DP 48 in Shiraz but only represented 1.8 to 12.3% of the total PA in Shiraz and 1.2 to 7.5% in Cabernet Sauvignon. In wine PA, catechin was the primary terminal subunit followed by epicatechin in both Shiraz and Cabernet Sauvignon (Tables 5 and 6). Epicatechin gallate was not detected as a terminal subunit in wine. In Shiraz wine, catechin represented between 65.5 and 83.7% of terminal subunits, while epicatechin represented between 16.3 and 31.3%. In Cabernet Sauvignon wine, catechin and epicatechin represented 75.0 to 82.3% and 17.7 and 25.0% of terminal subunits, respectively. Average DP and Composition of the Total Extract. The total PA extracts for seed, skin, and wine were analyzed by

phloroglucinolysis and HPLC to determine the average DP and composition prior to fractionation. For seed total PA extracts, the average DP was 6 for both Shiraz and Cabernet Sauvignon with a percent conversion yield of 76.9 and 80.9%, respectively (Tables 1 and 2). The proportion of extension subunits was similar in both Shiraz and Cabernet Sauvignon with epicatechin, catechin, and epicatechin gallate representing around 70, 4, and 24%, respectively. In Shiraz seed, the terminal subunit composition was composed of catechin, epicatechin, and epicatechin gallate representing 28.4, 36.7, and 34.9%, respectively. In Cabernet Sauvignon seed, the terminal subunit composition was similar representing 35.6, 33.7, and 30.7% of catechin, epicatechin, and epicatechin gallate, respectively. For skin total PA extracts, the average DP was 52 in Shiraz and 48 in Cabernet Sauvignon with percent conversion yields of 35.7 and 46.5%, respectively (Tables 3 and 4). The extension subunit composition of Shiraz skin was composed of epigallocatechin, catechin, epicatechin, and epicatechin gallate representing 44.8, 3.4, 44.6, and 7.3%, respectively. In Cabernet Sauvignon skin, the proportion of the extension subunit epigallocatechin was slightly higher than that of Shiraz representing 54.5%, while catechin, epicatechin, and epicatechin gallate were slightly lower representing 2.3, 40.3, and 2.9%, respectively. In the skin total PA extract, the terminal subunits were catechin and epicatechin with catechin representing 77.4 and 69.0% in Shiraz and Cabernet Sauvignon, respectively, while epicatechin represented 22.6 and 31.0%, respectively. Epicatechin gallate was not detected as a terminal subunit in the skin total PA extract. The average DP for the wine total PA extracts was 6 for Shiraz and 5 for Cabernet Sauvignon (Tables 5 and 6). The composition of extension subunits was similar for both Shiraz and Cabernet Sauvignon representing around 29, 2, 64, and 4% for 13271

dx.doi.org/10.1021/jf203466u |J. Agric. Food Chem. 2011, 59, 13265–13276

Journal of Agricultural and Food Chemistry

ARTICLE

Table 6. Distribution of PA Extracted from Cabernet Sauvignon Winea Cabernet Sauvignon Wine PA concentration DP

b

c

% conversion d

(mg/L HPLC) (% by HPLC) (% by HPLC CE) (% by UVvis CE)

e

yield

% of extension subunits Epigall Cat Ecat Ecatgall

% of terminal subunits Cat

Ecat

5

37.6

6.3

6.6

7.9

10.8

33.0

4.0 60.4

2.6

81.6

18.4

6

60.1

10.1

10.4

11.1

10.1

32.5

4.0 61.1

2.5

81.3

18.7

7

60.7

10.2

10.6

11.6

10.3

32.3

4.2 60.9

2.7

82.0

18.0

8

39.0

6.6

6.6

6.4

11.0

32.4

3.4 61.5

2.8

82.3

17.7

9 10

189.3 18.9

31.9 3.2

31.9 3.1

34.3 3.5

11.4 11.7

36.0 36.6

2.9 58.4 3.0 57.6

2.8 2.8

80.0 79.3

20.0 20.7

11

14.7

2.5

2.7

2.7

10.3

35.6

3.3 58.0

3.2

76.2

23.8

12

6.4

1.1

1.1

0.9

11.9

34.2

2.8 60.1

2.9

75.0

25.0

13

26.8

4.5

4.3

5.1

11.1

34.0

2.9 60.0

3.1

75.7

24.3

14

16.0

2.7

2.6

1.8

13.1

37.6

2.9 56.6

2.9

77.5

22.5

15

124.0

20.9

20.0

14.8

13.8

39.6

2.4 54.7

3.3

81.0

19.0

total (mg/L)

593.5

28.7

2.6 64.6

4.2

84.2

15.8

Total PA Extract 5 a

PA concentration, percent conversion yield, and the proportion of extension and terminal subunits are shown at individual DP. The average DP and proportion of extension and terminal subunits are given for the total PA extract. Abbreviations: Epigall = epigallocatechin, Cat = catechin, Ecat = epicatechin, and Ecatgall = epicatechin gallate. b PA concentration at each DP following HPLC/phloroglucinolysis determined by the summation of individual subunit concentrations determined using published conversion factors.24 c The proportion of the total concentration for all DP calculated from the PA concentration determined following HPLC/phloroglucinolysis and the calculation of individual subunits using published conversion factors.24 d The proportion of the total concentration for all DP determined following HPLC/phloroglucinolysis and calculated as catechin equivalents. e The proportion of the total concentration for all DP determined by absorbance at 280 nm on a UVvis spectrophotometer before phloroglucinolysis.

epigallocatechin, catechin, epicatechin, and epicatechin gallate, respectively, in both varieties. The terminal subunit composition was also similar in wine total PA extracts representing around 85% for catechin and 15% for epicatechin.

’ DISCUSSION The aim of this article was to characterize the distribution of PA in the two major red wine grape varieties grown in Australia, Shiraz and Cabernet Sauvignon. This was done by determining the concentration and composition of PA across the DP range for each variety from 80 fractions collected from a semipreparative HPLC separation. For seed PA, DP ranged between 2 and 17 subunits in Shiraz and 2 and 15 subunits in Cabernet Sauvignon. Previously, the DP range has been reported in Cabernet Sauvignon as high as 30 subunits and as low as 6 subunits,21 while the range for Shiraz seeds has been reported up to 12 subunits.22 For skin PA, DP ranged between 4 and 63 subunits in Shiraz, while Cabernet Sauvignon ranged from 4 to 76 subunits. The DP range of Shiraz has not previously been reported; however, the DP range of Cabernet Sauvignon has been reported as high as 81 and 83 subunits.16,22 In the present study, the DP range of wine PA was 6 to 18 subunits in Shiraz and 5 to 15 in Cabernet Sauvignon. The range reported in this study is larger than what has previously been reported in Shiraz and Cabernet Sauvignon. Previous reports have ranged between 2 and 8 subunits in Shiraz and 3 and 6 subunits in Cabernet Sauvignon.16,22 While the DP range has been reported in a number of varieties, there are few studies that report the distribution or concentration at different DP. In this study, the advantage of analyzing and consolidating a large number of fractions was that we were able to report the DP distribution across a larger range of DP values.

This gives a more accurate picture of PA distribution in grape seeds and skin, and subsequent wines. The distribution was reported at 10 different DP values in Shiraz seed and 12 in Cabernet Sauvignon, while earlier studies have reported between 4 and 8 different DP values in seed.20,22 In skin, the DP distribution has previously been reported between 7 and 11 DP values,20,22 while 21 DP values were reported in Shiraz for this study and 20 for Cabernet Sauvignon. The number of DP values reported in wine for this study was also greater than those in earlier studies with the distribution reported at 11 DP values in both Shiraz and Cabernet Sauvignon in the present study compared to 6 in an earlier study of wine PA distribution.22 Reporting distribution at a larger range of DP showed that PA is more broadly distributed across the DP range than previously reported. An earlier study of Shiraz seed distribution reported only 4 DP values ranging from 3 to 12 subunits with DP 5 representing the largest proportion at 46%. DP 3 and 12 also represented 23% each of the total distribution.22 In comparison, for Shiraz seed in the present study, the DP with the highest proportion represented only 30% of the concentration (calculated by HPLC) compared to 46% in the earlier study. While the largest DP of 17 in this study, contained less than 1%, the remaining DP all represented between 5 and 11% of the distribution, which was much lower than the 23% reported in the earlier study and indicates that DP is more likely to be evenly distributed across a broader range of DP. For Cabernet Sauvignon seed, the distribution in the present study was highest at DP 6 representing around 30% of the distribution (calculated by HPLC). The remaining DP ranged between one and 10%. An earlier study of Cabernet Sauvignon seed reported 5 DP values with a DP of 3, 4, and 6 subunits representing between 20 and 30% of the distribution with the remaining DP values of 2 and 11 representing 13 and 9% of total PA.22 13272

dx.doi.org/10.1021/jf203466u |J. Agric. Food Chem. 2011, 59, 13265–13276

Journal of Agricultural and Food Chemistry While the previous report of Cabernet Sauvignon seed PA had a higher concentration at the reported DP compared to the present study, the DP range was smaller. Despite differences in the proportions and range of seed DP between the present study and earlier reports of Shiraz and Cabernet Sauvignon seeds, in all of the studies, the DP with the highest proportion appears to be a good representation of the median value of the DP range. However, this is not always the case. A number of different distribution patterns have been reported in four other varieties (Cabernet Franc, Touriga Nacional, Trincadeira, and Castal~ao).20,22 While the distribution reported in Shiraz and Cabernet Sauvignon seed had a peak at the median DP value, the distribution reported in Castal~ao seed was highest at the largest DP of 11 representing 34%.22 For the other varieties (Cabernet Franc, Touriga Nacional, and Trincadeira), the lowest DP value was present in the highest proportion representing more than 26%.20,22 While Cabernet Franc seeds had one maximum peak at the lowest DP with the remaining DP values representing between 2 and 12%, Tourgia Nacional and Trincadeira had a second large peak at a DP of 7, which also represented more than 23% of the total PA.15,22 For skin PA, the distribution was generally low in concentration at small DP with a large peak at high DP.16,18,20,22 This pattern was observed in the present study, with the distribution of skin PA highest at a DP of 31 and 52 in Shiraz and 29 and 42 in Cabernet Sauvignon representing around 13 to 14% of the distribution (calculated by HPLC). At the highest DP values of 63 in Shiraz and 76 in Cabernet Sauvignon, the proportion was 4 and 7%, while at DP values below 13 subunits, the proportion was less than 3% in both varieties. Similar to seed distribution, the greater number of DP values reported have showed that PA was more evenly distributed across the DP range than earlier studies suggest. While the distribution of Shiraz skin PA has not previously been reported, Cabernet Sauvignon has been reported to have a distribution with the highest PA concentration representing 30 and 25% of the total PA at a DP of 26 and 38.22 Only 7 DP values were reported in the earlier study, with the largest DP value of 81 also representing a large proportion of the distribution at 23%, while the lower DP values of 6 to 11 subunits represented between 2 and 8% each. The distribution of skin PA has also been determined in Touriga Nacional and Castal~ao, but only 3 to 4 DP values were reported above a DP of 12 with each representing between 11 and 32%.22 In Touriga Nacional, DP was reported at values of 14, 15, 24, and 64 representing 11, 12, 20, and 27%, while Castal~ao was reported at DP 13, 17, and 49 representing 32, 26, and 21%. As only 3 to 4 DP values were reported at higher DP in the earlier study, it would appear that there was a large proportion of PA at each of these DP values. However, the skin PA distribution in the present study was more evenly spread across a large DP range. In contrast to the PA distribution in seed and skin, the distribution of wine PA showed two peaks across the DP range. The concentration of wine PA was highest around the middle of the DP range representing 26% of total PA at a DP of 11 in Shiraz and 32% of the distribution at a DP of 9 in Cabernet Sauvignon. Both varieties had a second peak representing around 20% of the distribution at the highest DP value of 18 in Shiraz and 15 in Cabernet Sauvignon. In Shiraz, DP below 11 subunits represented between 3 and 13% of the distribution but was less than 5% between DP 12 to 17. A similar distribution was observed in Cabernet Sauvignon wine.

ARTICLE

An earlier study of wine PA distribution also reported two peaks, although the DP range was smaller.22 In Touriga Nacional, Trincadeira, Castal~ao, Shiraz, and Cabernet Sauvignon wine, the DP range was between 2 and 9 subunits, but both peaks were generally observed between DP 2 and 6 with the largest peak representing between 21 and 35% of the distribution. Generally, the second peak did not occur at the largest DP value as was observed in the present study but still represented between 15 and 20% of the distribution.22 This observation in two independent studies suggests that a bimodal distribution may be a feature of wine PA. For a novel distribution of PA to be observed in wine, some rearrangement of the seed and/or skin PA must occur during vinification. Following this thought, one might hypothesize that the larger DP of skin PA could be cleaved during winemaking releasing smaller, more soluble polymers. What is observed as a second peak in the distribution of wine PA is some of these fragments near the upper solubility limit for PA in wine, which looks to be around DP 20 or less. An alternate hypothesis would be aggregation of the pool of smaller DP material into larger polymers, some of which are insoluble and are lost from the fermentation, and some of which are observed as a cluster near their solubility limit. The solubility of higher DP PA has long been speculated on, but limited studies exist on solubility of PA. An interesting observation from the data for both Shiraz and Cabernet Sauvignon was that wine PA had DP less than 20 subunits, while grape skin from both varieties had DP considerably greater than 20 subunits. This suggests that the higher DP material in grape skin is not extracted into the wine. While this is consistent with previous reports of PA extractability in ethanol mixtures,12 it raises a number of questions that may have implications for estimating PA extraction into wine based on grape skin or whole berry PA measures. If PA with DP greater than 20 subunits is not extracted into wine, some 6070% of the PA in the samples analyzed here, and likely a similar amount in other samples (e.g., 47% of Touriga Nacional, 21% of Castal~ao, and 55% in Cabernet Sauvignon22), may play no role in winemaking and wine sensory character. If long polymers are not extracted into wine, what accounts for the perception of highly astringent wines? This must be due to some previously unrecognized property of smaller polymers or the interaction of those PA with DP greater than 20 subunits with one another or with other elements in the wine. If the larger DP material is extracted during winemaking, why is it not observed in the subsequent wine? The answers to these questions will be a critical part of developing appropriate analytical tools for predicting PA extraction from grapes into wine. In this study, the distribution has been reported by HPLC and UVvis spectrophotometry to demonstrate any differences influenced by the conversion yield of PA during phloroglucinolysis. While the distribution determined by UVvis spectrophotometry prior to phloroglucinolysis was generally similar for seeds, there were some discrepancies for skin and wine PA distribution. For skin and wine, the distribution pattern calculated by HPLC and UVvis spectrophotometry was generally similar; however, the distribution based on HPLC data was slightly skewed toward longer polymers compared to the distribution based on the UVvis spectrophotometric data. This difference is likely to be a result of the percent conversion yield. In skin, the percent conversion yield increases with DP up to 52% and is generally low across the wine DP range (