Alex Fredrickson* and Misha Kwasniewski
*University of Missouri, 1453 Wikiup Dr, Santa Rosa, CA 95403 (ajfpvn@mail.missouri.edu)

Using a newly-optimized method to precipitate, quantify, and characterize grape and wine proteins, we identified >1200 unique proteins in juice and 900 in wine of Vitis interspecific hybrid cv. Chambourcin. To modify tannin concentrations, the following treatments were employed in the winery: large exogenous additions of 350mg/L cat- echin equivalents (CE) of tannin (Tannin+) at (i) crush or (ii) one month post press, or (iii) skin fragmentation (accentuated cut edges or ACE). Mean tannin concentrations at bottling were 36 mg/L CE in the control versus 91 mg/L in ACE. In-crush Tannin+ yielded 187 mg/L. The post-press Tannin+ was nearly 12-fold greater than the control (419 mg/L). Total protein content in wine correlated poorly with final tannin concen- tration, with an r2 of 0.2146. However, by modeling final tannin content with spectrum counts of individual proteins in juice and wine, we identified numerous proteins with predictive ability of final tannin content (2322 predictors). Overall, these yielded an r2 of 0.8031, indicating a strong inverse relationship between concentrations of certain proteins and tannin retention. Within the model using variable importance projection (VIP) scores, the top 10 individual proteins resulted in eight Vitis proteins, including vacuolar invertase, ß-fructofuranosidase, thaumatin-like, and osmotin-like. The other two proteins were yeast-derived. Of the individual proteins, all eight Vitis proteins had a negative logarithmic r2 of 0.8 or better, while the yeast proteins were lower. Proteomics showed many proteins that relate to tannin retention, and they were not limited to pathogenesis-related proteins. Although important, the direct relationship of protein and tannin in wine may not be a simple driving force in the extraction/retention of tannin in red wine.

Funding Support: Grape and Wine Institute (Missouri)