Access features, news and views from the latest issue and from our chemistry archives.

September/October 2019

Astringency and tannin size

Some readers of this column have sought my thoughts on a paper recently published in the Journal of Agricultural and Food Chemistry ( on the relationship between tannin size and dryness perception. The paper from the University of California, Davis is headed ‘Red Wine Dryness Perception Related to Physicochemistry’, a somewhat more rigorous title than in the online Food & Wine magazine of ‘Dry Red Wines “Big Tannins” Are Literally Bigger, Study Says’ ( The researchers – Aude Watrelot, Hildegarde Heymann and Andrew Waterhouse – examined the chemistry and sensory effects of condensed tannins in relation to the perception of mouthfeel dryness. Condensed tannins are found in grape skins, seeds and flesh and consist of repeating units of flavan-3-ol units; for example, (+)-catechin.

Two wines were used in the study: a Cabernet Sauvignon and a Pinot Noir. Using plots of intensity of sensory response versus time in mouth (commonly called time/intensity plots), the group of trained tasters found that the Cabernet Sauvignon showed a higher overall intensity and a longer time in mouth response than the Pinot Noir. The tannins were extracted from each wine, the mean degree of polymerisation (mDP; the number of repeating flavan-3-ol units) determined and the extent of their interaction with protein assessed through turbidity measurement. The same measurements were performed on the original wines and on wines to which the extracted tannins had been added.

The tannins in the Cabernet Sauvignon were ‘bigger’ with a mDP value of 6.3 compared with 2.4 for the Pinot Noir. This difference was even more pronounced with the tannin extracts: 9.4 versus 3.6. There were also structural differences between the two extracts, particularly with respect to substitution on the B-ring. Both wines came from very ripe grapes showing high alcohol concentration (15.4% and 14.5%) and a relatively high pH for red wine (3.97 and 3.77).

Dryness is a common mouthfeel descriptor for astringency. A major contributor to dryness is the consequence of salivary protein interaction with tannin. Proline-rich proteins are known to be significantly involved in generating the astringent response. So, the authors examined the interaction between the two tannin extracts as well as the two wines themselves with poly-L-proline, a commonly used model compound for proline-rich proteins. Mucin and human saliva were also used with turbidity as a measure of the interaction.

Limited turbidity was observed with mucin, perhaps not surprisingly because it is a glycoprotein, a class not generally known for its contribution to astringency. Significant turbidity was found when poly-L-proline was used, but the most extensive turbidity was found when saliva was used – the saliva was actually a pooled sample from 4 (two female and two male) participants in the study. In essence, the Cabernet Sauvignon wine with added tannin extract from either wine showed the highest turbidity, with the somewhat quirky result that the Pinot Noir wine with added tannin from the same wine had a higher turbidity than when the Cabernet Sauvignon extract was used.

In these experiments, turbidity measurements are based on the formation of aggregates of sufficient size to cause particle formation. The turbidity results here agree with those for the sensory assessment of astringency, thereby reinforcing the idea that larger tannin molecules are the basis for ‘big tannin’ taste.

The actual mechanism of astringency is extremely complex and our understanding is still evolving. An extensive review published in 2014 summarises the state of knowledge at that time (Trends Food Sci. Technol. 2014, vol. 40, pp. 6–14). It is well known that there can be marked differences in individual saliva composition and this alone can result in differences in astringency assessment. To add to this complexity, it has been shown that salivary protein profile can change with food ingestion and with a tentative suggestion that the profile may be influenced by circadian rhythm (Brandão et al. Food Res. Int. 2014, vol. 64, pp. 508–13).

There is increasing evidence from research at the University of Porto, Portugal, that oral cells contribute to astringency by binding tannins to their surface. These tannins then interact with and bind salivary proteins, leading to the formation a network of oral cell–tannin–salivary protein complexes (Soares et al. J. Agric. Food Chem. 2016, vol. 64, pp. 7823−8). This opens up a new level of complexity in astringency assessment.

One of the frustrations in obtaining a molecular handle on astringency has been the inability to assess interactions in actual solutions. Rather, surrogates such as turbidity have been used. Here again the University of Porto group has taken the lead, using fluorescence quenching and saturation transfer difference-nuclear magnetic resonance (STD-NMR) (see J. Agric. Food Chem. 2017, vol. 65, pp. 6415–25). STD-NMR allowed the binding sites of various tannins (the ‘epitopes’ in the authors’ terminology) to be mapped while discrimination of tannin–salivary protein binding efficiencies could be achieved with fluorescence quenching.

The role of polysaccharides in influencing the tannin–salivary protein interaction is another complicating factor and still the subject of extensive research, as is the difference in salivary protein structure (disordered versus structured) and its effect on astringency. Some rather exciting research still needs to be carried out. Something to contemplate when next consuming a red wine or two.

Geoffrey R. Scollary FRACI CChem ( has been associated with the wine industry in production, teaching and research for the last 40 years. He now continues his wine research and writing at the University of Melbourne and the National Wine and Grape Industry Centre at Charles Sturt University.

Book and software reviews

To offer your services as a book or software reviewer for Chemistry in Australia, please contact Damien Blackwell at