Stella Kassara

Relationship between Red Wine Grade and Phenolics. 2. Tannin Composition and Size

Commercial red wines ( Vitis vinifera L. cv. Shiraz) produced during the 2009 vintage underwent winemaker assessment for allocation grade soon after production. The wines were then subjected to phenolic analysis to measure wine color (total anthocyanin, SO(2) nonbleachable pigment, and wine color density) and tannins (concentration, composition, and average degree of polymerization). A positive relationship was found between wine phenolic concentration and projected bottle price. Tannin compositional analysis suggested that there was specifically a relationship between wine grade and skin-derived tannins. These results suggest that maximization of skin tannin concentration and/or proportion is related to an increase in projected wine bottle price.

Sensory Properties of Wine Tannin Fractions: Implications for In-Mouth Sensory Properties

Different molecular structures of grape tannins have been shown to influence astringency, however, the in-mouth sensory effects of different molecular structures in red wine tannins remains to be established. The objective of this research was to assess the impact of wine tannin structure on in-mouth sensory properties. Wine tannin was isolated from Cabernet Sauvignon wines of two vintages (3 and 7 years old) and separated into two structurally distinct subfractions with liquid-liquid fractionation using butanol and water. The aqueous subfractions had greater mean degree of polymerization (mDp) and contained a higher proportion of epigallocatechin subunits than the butanol-soluble subfractions, while the older wine tannin fractions showed fewer epicatechin gallate subunits than the younger tannin fractions. The red wine had approximately 3:1 mass ratio of the aqueous and butanol tannin subfractions which approximated an equimolar ratio of tannin in each subfraction. Descriptive sensory analysis of the tannin subfractions in model wine at equimolar concentrations revealed that the larger, more water-soluble wine tannin subfractions from both wines were perceived as more astringent than the smaller, more hydrophobic and more highly pigmented butanol-soluble subfractions, which were perceived as hotter and more bitter. Partial least squares analysis indicated that the greater hydrophobicity and color incorporation in the butanol fractions was negatively associated with astringency, and these characteristics are also associated with aged wine tannins. As the larger, water-soluble tannins had a greater impact on the overall wine astringency, winemaking processes that modulate concentrations of these are likely to most significantly influence astringency.

Phenolic Compositions of 50 and 30 Year Sequences of Australian Red Wines: The Impact of Wine Age

The phenolic composition of red wine impacts upon the color and mouthfeel and thus quality of the wine. Both of these characteristics differ depending on the age of a wine, with the purple of young wines changing to brick red and the puckering or aggressive astringency softening in older wines. This study investigated the color parameters, tannin concentrations and tannin composition of a 50 year series of Cabernet Sauvignon wines from a commercial label as well as 30 year series of Cabernet Sauvignon and Shiraz wines from a separate commercial label to assess the impact of wine age on phenolic composition and concentration. The wine color density in wines of 40 to 50 years old was around 5 AU compared with 16 AU of wine less than 12 months old, which correlated well with the concentration of non-bleachable pigments and pigmented polymers. Conversely, the anthocyanin concentrations in 10 year old wines were substantially lower than that of recently bottled wines (around 100 mg/L compared with 627 mg/L, respectively), adding further evidence that non-bleachable pigments including pigmented polymers play a much larger role in long-term wine color than anthocyanins. No age-related trend was observed for tannin concentration, indicating that the widely noted softer astringency of older red wines cannot necessarily be directly related to lower concentrations of soluble wine tannin and is potentially a consequence of changes in tannin structure. Wine tannins from older wines were generally larger than tannins from younger wines and showed structural changes consistent with oxidation.

Rapid measurement of methyl cellulose precipitable tannins using ultraviolet spectroscopy with chemometrics: application to red wine and inter-laboratory calibration transfer.

Information relating to tannin concentration in grapes and wine is not currently available simply and rapidly enough to inform decision-making by grape growers, winemakers, and wine researchers. Spectroscopy and chemometrics have been implemented for the analysis of critical grape and wine parameters and offer a possible solution for rapid tannin analysis. We report here the development and validation of an ultraviolet (UV) spectral calibration for the prediction of tannin concentration in red wines. Such spectral calibrations reduce the time and resource requirements involved in measuring tannins. A diverse calibration set (n = 204) was prepared with samples of Australian wines of five varieties (Cabernet Sauvignon, Shiraz, Merlot, Pinot Noir, and Durif), from regions spanning the wine grape growing areas of Australia, with varying climate and soils, and with vintages ranging from 1991 to 2007. The relationship between tannin measured by the methyl cellulose precipitation (MCP) reference method at 280 nm and tannin predicted with a multiple linear regression (MLR) calibration, using ultraviolet (UV) absorbance at 250, 270, 280, 290, and 315 nm, was strong (r 2 val=0.92; SECV=0.20 g/L). An independent validation set (n = 85) was predicted using the MLR algorithm developed with the calibration set and gave confidence in the ability to predict new samples, independent of the samples used to prepare the calibration (r 2 val = 0.94; SEP = 0.18 g/L). The MLR algorithm could also predict tannin in fermenting wines (r 2 val=0.76; SEP=0.18 g/L), but worked best from the second day of ferment on. This study also explored instrument-to-instrument transfer of a spectral calibration for MCP tannin. After slope and bias adjustments of the calibration, efficient calibration transfer to other laboratories was clearly demonstrated, with all instruments in the study effectively giving identical results on a transfer set.

Effect of Wine pH and Bottle Closure on Tannins

The impact of wine pH and closure type on color, tannin concentration, and composition was investigated. A single vintage of Cabernet Sauvignon wine was divided into three batches, the pH was adjusted to 3.2, 3.5 or 3.8, and the wines were bottled under screw caps with either SaranTin (ST) or Saranex (Sx) liners. After 24 months, the tannin concentration, tannin percent yield (relating to the proportion of acid-labile interflavan bonds), and the mean degree of polymerization (mDp) had decreased significantly, all of which can contribute to the softening of wine astringency with aging. The higher pH wines contained less percent (-)-epicatechin 3-O-gallate subunits, whereas the Sx pH 3.2 wines were significantly lower in percent yield and mDp than the other wines. Overall, the tannin structure and wine color of the lower pH wines (pH 3.2) bottled under Sx screw caps changed more rapidly with aging than those of the higher pH wines (pH 3.8) bottled under ST screw caps.

Comparison of extraction protocols to determine differences in wine-extractable tannin and anthocyanin in Vitis vinifera L. cv. Shiraz and Cabernet Sauvignon grapes.

Cabernet Sauvignon and Shiraz grapes were sourced from different regions within Australia, and microvinified with a skin contact period of 6 days. Grape samples were extracted using two protocols: a 15% v/v ethanol, 10 g/L tartaric acid extract of gently crushed berries (wine-like, WL) and a 50% v/v ethanol, pH 2 extract of grape berry homogenate. It was found that in WL extracts, grape tannin and anthocyanin concentrations were strongly related to wine tannin, anthocyanin and color density achieved during the skin contact period. No relationship was observed for grape tannin concentration analyzed in homogenate extracts and wine tannin, but a strong, positive relationship was found for anthocyanin concentration. When the data obtained from homogenate extraction was treated separately by grape variety, a stronger relationship between grape and wine tannin concentration was observed. Tannin compositional analysis in wines indicated that higher tannin concentrations were due to the extraction of tannin of higher molecular mass during fermentation, most likely from grape skins.

Ethanol Concentration Influences the Mechanisms of Wine Tannin Interactions with Poly(L-proline) in Model Wine.

Changes in ethanol concentration influence red wine astringency, and yet the effect of ethanol on wine tannin-salivary protein interactions is not well understood. Isothermal titration calorimetry (ITC) was used to measure the binding strength between the model salivary protein, poly(L-proline) (PLP) and a range of wine tannins (tannin fractions from a 3- and a 7-year old Cabernet Sauvignon wine) across different ethanol concentrations (5, 10, 15, and 40% v/v). Tannin-PLP interactions were stronger at 5% ethanol than at 40% ethanol. The mechanism of interaction changed for most tannin samples across the wine-like ethanol range (10-15%) from a combination of hydrophobic and hydrogen binding at 10% ethanol to only hydrogen binding at 15% ethanol. These results indicate that ethanol concentration can influence the mechanisms of wine tannin-protein interactions and that the previously reported decrease in wine astringency with increasing alcohol may, in part, relate to a decrease tannin-protein interaction strength.

Properties of Wine Polymeric Pigments Formed from Anthocyanin and Tannins Differing in Size Distribution and Subunit Composition

To explore the effect of tannin composition on pigment formation, model ferments of purified 3-O-monoglucoside anthocyanins (ACN) were conducted either alone or in the presence of two different tannins. Tannins were isolated from grape seeds (Sd) or skins (Sk) following exhaustive extraction in 70% v/v acetone. The Sd and Sk tannin fractions had a mean degree of polymerization of 5.2 and 25.6, respectively. The Sd fraction was highly galloylated, at 22%, but galloylation was <2% in the Sk fraction. The Sk fraction was distinguished by a high proportion of prodelphinidin, at 58%. After a 6 month aging period, polymeric pigments were quantified and their color properties determined following isolation by solid-phase extraction. Wine color and polymeric pigment were highest in the treatment containing ACN+Sd and similar in the ACN+Sk and ACN treatments. The same trend between treatments was observed for total and polymeric nonbleachable pigments. Only minor changes in tannin subunit composition were found following ACN incorporation, but the size distribution of polymeric pigments determined by gel permeation chromatography decreased, in particular for the ACN+Sk treatment. Color incorporation in the higher molecular mass range was lower for ACN+Sk wines than for ACN+Sd wines. Compositional differences between the two tannin fractions may therefore limit the incorporation of ACNs in the colored form. The results suggest that in the ACN+Sk and ACN treatments, the formation of lower molecular mass oligomeric pigments was favored. In polymeric pigments derived from ACNs, the presence of ethyl- and vinyl-linked ACNs to the level of trimers was identified using mass spectrometry.

Characterization of macromolecular complexes in red wine: Composition, molecular mass distribution and particle size.

Precipitates were prepared from two compositionally different Pinot noir wines with addition of excess ethanol, and contained primarily polysaccharide, tannin and protein. The ethanol-soluble material was further fractionated into polymeric (tannin) and monomeric phenolics. Tannin associated with precipitates was of a higher molecular mass than that remaining in ethanolic solution. Wine fractions were reconstituted at the ratios of the original wine and analyzed using nanoparticle tracking analysis. The average particle size of the tannin fraction was 75-89 nm, and increased when combined with the precipitate (≅ 200 nm). Addition of the monomeric fraction to the tannin-precipitate complex increased both the incidence and concentration of smaller particles, reducing the average particle size. The formation of aggregates occurred in all fractions and only minor differences in particle size distribution were found between wines. Differences in particle concentration between wines appear to be due to differences in the total concentration of macromolecules rather than compositional differences.

Measuring the molecular dimensions of wine tannins: comparison of small-angle X-ray scattering, gel-permeation chromatography and mean degree of polymerization.

The molecular size of wine tannins can influence astringency, and yet it has been unclear as to whether the standard methods for determining average tannin molecular weight (MW), including gel-permeation chromatography (GPC) and depolymerization reactions, are actually related to the size of the tannin in wine-like conditions. Small-angle X-ray scattering (SAXS) was therefore used to determine the molecular sizes and corresponding MWs of wine tannin samples from 3 and 7 year old Cabernet Sauvignon wine in a variety of wine-like matrixes: 5-15% and 100% ethanol; 0-200 mM NaCl and pH 3.0-4.0, and compared to those measured using the standard methods. The SAXS results indicated that the tannin samples from the older wine were larger than those of the younger wine and that wine composition did not greatly impact on tannin molecular size. The average tannin MWs as determined by GPC correlated strongly with the SAXS results, suggesting that this method does give a good indication of tannin molecular size in wine-like conditions. The MW as determined from the depolymerization reactions did not correlate as strongly with the SAXS results. To our knowledge, SAXS measurements have not previously been attempted for wine tannins.