The perception of astringency (“drying, rough, puckering”) in wine is largely credited to condensed tannins (proanthocyanidins). As described in Chapter 14, condensed tannins in grapes and other plants are polymers of flavanoid monomers, and possess innumerable structures due to variations in subunit number, subunit structure/stereochemistry, and connection points. Even assuming only two subunit choices, no branching, and a single initiator, a tannin 30‐mer could have 230 →1 billion different structures. This chemistry is further complicated in wines where proanthocyanidin linkages undergo hydrolysis and subsequently rearrange or else react with a large number of wine components, for example, anthocyanins or aldehydes (Chapters 24 and 25). Consequently, methods to characterize tannins typically fall into one of three categories:
Of existing analytical methods, those based on precipitation of tannins by proteins or other macromolecules have been particularly successful in modeling astringency in wine, as determined by trained sensory panels (Chapter 14). Harbertson and Adams adapted the 1978 method of Hagerman and Butler [4] for wine, using bovine serum albumin as the binding substrate [5]. Sarneckis et al. utilized methyl cellulose as a precipitant to develop a comparable method [6]. Following precipitation, the phenolic content of the pellet or supernatant can be measured. These precipitation methods can achieve very high correlations between “astringency intensity” and “tannin” (r 2 > 0.8), likely because they mimic reactions that would occur between tannins and lubricating salivary proteins in the mouth [7].
While precipitation based assays are useful, they only model the overall intensity of perceived astringency. However, winemakers and others commonly use terms to discriminate types of astringency, and several reports have shown that the wines can be distinguished by trained panels based on either temporal behavior (e.g., how astringency intensity decays once the wine is expectorated) or subterms (“grainy,” “velvety,” “puckering,” and more) [8]. At this point, it is unclear if these subterms arise from structural differences among tannins or the presence or absence of other wine components [9]. A challenge for chemists today is to look for analytical methods that could produce results that relate to these variations in perception.
Traditional protein‐precipitation methods use an excess of protein, so that strong and moderate binding tannins are fully precipitated and consequently not well distinguished. McRae et al. have hypothesized that the strength of tannin–protein binding, as measured by isothermal titration calorimetry (ITC) [10], could be related to the qualities of astringency perception, particularly its time‐intensity properties [11]. The enthalpy of binding of isolated wine tannin to polyproline, a synthetic model protein, drops by over 30% in young versus old wines [11]. While ITC methods are difficult to execute, comparable data can be achieved through HPLC analyses at different temperatures [12].
Relating the astringent subterms to specific chemical components is also an ongoing area of research, in wine and other fields. For example, some flavonol glycosides (Chapter 13) are reported to have “velvety astringent” sensory properties [13]. This astringent characteristic may have a different mechanism than that described for the “puckering astringency” of condensed tannins. Instead of affecting astringency indirectly through protein precipitation, these compounds may act directly on tactile receptors [14]. While intriguing, this hypothesis has not been validated with sensory studies on multiple wines. Other subterms may arise from the combined presence of other flavor compounds and condensed tannins; for example, high acid and high tannin may be responsible for the perception of “green tannin” [15]. Finally, future advances in relating tannin chemistry to sensory properties may benefit from using novel analytical chemistry approaches, such as detailed mass spectral data on tannin composition [16].