The Extraction of Condensed Tannins in Red Wine Production

The winemaking portion of the experiment was completed on time on the commercial scale and on the small scale at UCD. Samples were collected and prepared for analysis. The various chemical analyses are still underway due to the large number of samples and the demand for analysis equipment by other projects. In summary, approximately 300 samples have been analyzed to date, all for phenolics by normal and reverse phase HPLC, as well as spectral measures of co-pigmentation. These have included all samples from the commercial wines, but not those from the small scale analyses. Since each of those procedures yields many results, the challenge now is to try an interpret the data to the point where it can be summarized in simple and understandable terms. Data collected so far shows some expected trends, but also some unexpected ones, in particular the increase in tannin levels during aging. With the large amount of data collected, additional analysis is needed to look for interactions and consistency. Additional data still must be collected and analyzed on the small scale wines. This data will be used to plan for the third year’s experiments during summer 1999.

Adaptation of a Simple Assay to Measure Tannins in Grapes and Wines

Premium red wine production is dependent upon proper extraction of pigments and tannins from grape berry skins and seeds during winemaking. While many methods are available to estimate the level of polymeric phenols in grapes and wines, none of the simple ones give an independent estimate of the amount of tannin present. Our aim was to provide a simple tannin assay that would be suitable for analysis of grape berries during maturation as well as for monitoring tannin extraction during fermentation and extended maceration. We have taken a tannin assay that was originally developed for measuring tannins in persimmon and adapted it so that it can be used for measuring tannins in grapes and wines. We studied many of the compounds in wine that could potentially interfere with this analysis and have found that while some can indeed inhibit the assay at very high concentrations, they do not interfere at the levels found in wines. Thus, the plate binding assay has been successfully adapted. In the course of working with the plate binding assay we realized that with some simple modifications we might be able to develop a much simpler solution assay that would not require a specialized piece of equipment like a microtiter plate reader. Although our initial research objectives were only related to adapting the plate binding assay, we were able to quickly demonstrate the feasibility of a solution assay, and we have now successfully developed a new assay method for tannins. The solution assay is simpler than the plate binding method we originally set out to adapt for grapes and wine. It is also much less variable than the plate-binding assay and is useful over a much wider range. Because it requires only a spectrophotometer and a small table-top centrifuge rather than a specialized piece of equipment like a microtiter-plate reader, this analysis will be suitable for use in many winery laboratories. One of our primary objectives in this project was to provide a simple tannin assay. The new analysis that we have developed meets that objective better than the plate binding assay ever could have.

The Extraction of Condensed Tannins in Red Wine Production

A commercial scale maceration trial was set up with two wineries, Wente Brothers and Sebastiani Vineyards. Cabernet Sauvignon grapes were harvested from four vineyards, two lots going to each of the wineries. In each case, five wines were made in 20 ton fermentations. The fruit was distributed between all fermentors. The treatments were a control, delayed fermentation, extended maceration, an oak tannin treatment, color extracting enzymes and rotary fermentors. The phenolic composition of the wines was tested using the Folin Ciocalteau, spectral tests, and both reverse and normal phase HPLC procedures. In general, the differences between vineyards were greater than those of the treatments. One of the most notable features was that many of the treatments did not cause consistent changes. The delayed fermentation increased phenol levels slightly, the extended maceration had very little effect on total phenol except for one vineyard which increased significantly, while the rotary and color extraction procedures had little effect. Only the oak tannin treatment significantly decreased the total phenols. Treatments did have trends for specific components. The extended maceration did increase catechin and epicatechin while it decreased unbleached anthocyanins as well as monomelic anthocyanins in most but not all wines. Catechin and epicatechin levels were very low in the wines with short maceration time, those in the oak tannin and color enzyme treatments, although the short but intensive rotary treatment had mixed results. The ratio of small to large tannin was not affected by any of the treatments. In conclusion, none of the treatments significantly altered any of the wines total phenol content consistently except to decrease when short contact times were used, and some individual compounds did respond to treatments.

Adaptation of a Simple Assay to Measure Tannins in Grapes and Wines

We have accomplished the objectives proposed for the 1996 season. We extracted seed tannin from Cabernet Sauvignon seeds and obtained enough to construct standard curves for several years. We will provide an aliquot this standard tannin preparation to those who wish to conduct similar assays. By assaying fractions from a seed tannin extract separated by HPLC, we determined that the microtiter assay detects the large polymeric tannins present in seed extracts. Simulated fining experiments showed that casein, gelatin and ovalbumin removed the tannin material detected by the assay. We determined that of the total amount of tannin detected by the plate-binding assay, 22%came from the skins and 78%came from the seeds. We have used the principal of the microtiter plate assay to develop a solution assay for tannin. This novel analysis is based on the observation that alkaline phosphatase activity, unlike many other enzymes, is not inhibited by tannin binding. The solution assay is less sensitive but shows less variation than the titer plate assay. The solution assay can be carried out using only a spectrophotometer and eliminates the need for a microtiter plate reader. Thus, the solution assay should prove useful in more winery laboratories than the plate assay.

Studies on the Interaction of Flavor Compounds With Non-Volatile

During the previous year of this proposal we focused on developing sensitive gas chromatographic headspace procedures for quantifying odorant/polyphenol interactions in model solutions. We are currently using these procedures to evaluate the effects of polyphenol structure on the strength of odorant interactions. The main focus this year has been on developing and evaluating sensory procedures for measuring these interactions. Although previous studies of flavor interactions in other foods and beverages have largely focused on measuring aroma (nasal) intensity only, we wanted to also be able to quantify oral (retronasal) aroma stimulation, similar to that experienced when wine is consumed. We have used a time-intensity procedure and model solutions to show that retronasal aroma perception is significantly affected by the chemical/physical nature of the aroma compound, by the nature of the matrix, and by individual judge factors such as salivary flow rate. We feel that this procedure now provides us with a promising tool for the sensory evaluation of matrix effects. This, in combination with the analytical headspace procedures, will provide extensive information about the importance of odorant/polyphenol interactions in wine.

Studies on the Interaction of Flavor Compounds With Non-Volatile

During the previous year of this proposal we focused on developing sensitive gas chromatographic headspace procedures for quantifying odorant/polyphenol interactions in model solutions. We are currently using these procedures to evaluate the effects of polyphenol structure on the strength of odorant interactions. The main focus this year has been on developing and evaluating sensory procedures for measuring these interactions. Although previous studies of flavor interactions in other foods and beverages have largely focused on measuring aroma (nasal) intensity only, we wanted to also be able to quantify oral (retronasal) aroma stimulation, similar to that experienced when wine is consumed. We have used a time-intensity procedure and model solutions to show that retronasal aroma perception is significantly affected by the chemical/physical nature of the aroma compound, by the nature of the matrix, and by individual judge factors such as salivary flow rate. We feel that this procedure now provides us with a promising tool for the sensory evaluation of matrix effects. This, in combination with the analytical headspace procedures, will provide extensive information about the importance of odorant/polyphenol interactions in wine.

Factors affecting bitterness, astringency and mouthfeel in wine

The effect of varying sourness on perceived astringency has been studied using different astringents and different acids. Astringency of phenolic compounds (grape seed tannin, catechin, tannic acid or gallic acid) increases with increasing citric acid. Alum, in contrast, shows a very strong decrease in astringency with the addition of acid. The results of this experiment call into question basic sensory studies of astringency which have interchangeably used alum and other astringent compounds such as tannic acid. Astringency of alum was decreased equivalently by addition of lactic acid, citric acid or hydrochloric (HC1) acid. The acids themselves have different astringencies at equally sour concentrations. Citric acid was least sour, while lactic and HC1 were equal in astringency to the alum solution. In previous experiments, increasing sucrose concentration in red wine decreased both maximum intensity and total duration of astringency (Ishikawa and Noble, 1995). The role of the sweetness versus the contribution of viscosity of the added sucrose in reducing astringency was unknown. In the present experiments using the sweetener aspartame, increasing sweetness with no increase in viscosity resulted in no change in astringency, while the bitterness of the GST was slightly decreased. In contrast, as viscosity was increased, both maximum intensity (Imax) and total duration (TOT) of astringency of grape seed tannin (GST), citric acid and alum were significantly decreased. Sour Imax and TOT of citric acid and alum were also decreased as the viscosity was raised. In contrast, no change in bitter Imax and TOT of GST was perceived as viscosity was raised. These results suggest that perception of the astringency is influenced by tactile factors such as viscosity, more than by tastes. Continuing studies of the role of saliva in sensory perception, consistent with results found for thickened sweet solutions (AVF 94 final report), there was no difference in perception of astringency of thickened solutions of grape seed tannin between high- and low-flow judges. In contrast, perception of unthickened grape seed tannin solutions which contained no and increasing levels of aspartame, astringency was rated more intensely by the low flow subjects.

Factors affecting bitterness, astringency and mouthfeel in wine

Astringency and sweetness were rated using time-intensity methods(T-I) in red wine containing two levels of phenolics, and five levels of sucrose. Maximum intensity and total duration for astringency decreased significantly with increasing sucrose concentrations. In contrast, sweetness was not affected by varying astringency levels. No differences in perception of astringency or sweetness were found as a function of PROP (6-n-propylthiouracil) taster status, although there was significant difference in intensity and persistence of astringency as a function of salivary flow. Consistent with our previous work, judges with low salivary flow rates rated astringency higher than high-flow subjects, but no difference for sweetness was found. Five attributes were rated by T-I in six hydroxy benzoic acid derivatives, differing only in number and position of hydroxy groups. Maximum intensity varied significantly for astringency, bitterness, prickling, sourness and sweetness. Gentisic acid was the most sour and bitter, salicylic and gentisic acids were highest in astringency, and m-hydroxybenzoic acid was sweetest. Benzoic acid had the highest prickling feeling which lasted 20 sec longer than salicylic acid and 40 sec longer than the other samples which elicited lowest prickling sensation. To understand factors influencing temporal perception, sweetness of solutions of 20, 80 or 140 g/L glucose was rated by T-I. Saliva expectorated by these same judges initially at 7 sec and at 30 sec intervals up to two min was analyzed for glucose concentration. Saliva glucose concentration was higher in the low flow judges at 4 of the 6 intervals. For each subject, the decrease in sweetness intensity was correlated with the oral glucose concentration. However, there was no significant difference in sweetness perception between the saliva flow-groups. Sweetness and viscosity were rated by T-I in solutions of glucose which were thickened with a tasteless gum. In contrast to the strong salivary response to acid or tannin found previously, which increases viscosity from that of water (1 cp) to the high gum level (50 cp) failed to increasing the salivary flow over that observed in response to water. Small significant differences in salivary flow were observed when 9 and 14%(w/v) glucose were tasted. No perceptual interaction occurred between viscosity and sweetness. Unlike previous work in which sweetness was increased by physical viscosity, no difference in sweetness was found when the viscosity was raised by 50 cp. Correspondingly, no increase in viscosity were perceived when sugar level was raised.

AN APPRAISAL OF LIQUID CHROMATOGRAPHY-MASS SPECTROMETRY FOR THE ANALYSIS AND TRACKING OF WINE AND GRAPE COMPONENTS

Objectives and Results: Our initial assessment was that this technique was not well developed enough at this point to be genuinely useful for wine analysis. It appears to lack the sensitivity to provide chemically useful information on the minor components in wines, and it is these components that are the interesting ones, since the major components are already well known. We are continuing to work with the UC Davis Facility for Advanced Instrumentation which is installing a new LC/MS system, to evaluate its applicability to wine analysis. Our evaluation focussed on the analysis of phenolic compounds.

Reducing Bitterness in Wines

As reported previously (Noble, final report, AVF 91), astringency and bitterness of 18 wines, varying in ethanol (1, 8, 14%), pH (3.0 and 3.6) and added phenolics (none or 1500 mg/L catechin or tannic acid) were evaluated by time-intensity (Tl) methods. Using this procedure, the wine is ingested and perceived intensity is rated continuously Increasing ethanol concentration dramatically increased both the intensity and duration of bitterness and area under the Tl curve, while having very little effect on any astringency. Addition of tannic acid dramatically increased bitterness and astringency intensity and duration, more than addition of an equivalent amount of catechin. Raising pH from 3.0 to 3.6 had virtually no effect on bitterness but decreased astringency intensity and duration. As reported herein, salivary flow was collected in response to the same wines. The largest increase in production of saliva was elicited by decrease in pH (increase in TA); tannic acid, which elicited intense astringency, had the next largest effect, followed by increase in ethanol. Examining the taste responses of the judges, grouped by salivary flow rate, showed that (pooled over all samples) low-flow judges perceived bitterness and astringency more intensely and longer than high-flow individuals. The effect of method of sample evaluation on sourness, bitterness and astringency was preliminarily explored in water and in wine. The temporal responses using “sip and spit” procedure (the normal lab protocol) was compared to the Tl responses when samples were sipped and swallowed, more similar to actual consumption method. With the exception of astringency in wine, no differences between the two methods were found. For evaluation of astringency, when samples were swallowed, astringency intensity was slightly lower and the total duration was slightly shorter, than when the samples were expectorated. However, considerably variation among judges occurred , thus this effect is not be concluded to significantly affect perception based on this preliminary study. To explore the effect of rate of salivary flow on Tl perception of these three attributes in water and in wine, data for the expectorated samples was analyzed for low, medium and high flow subjects for each samples. Inconsistent differences were seen across flow groups, further, by analysis of variance problems with reproducibility were encountered. In contrast to the previous study, here high flow judges tended to have a longer persistence of bitterness or astringency than low flow judges. These low-flow-judges rated the samples higher in both attributes, consistent with the above study. Because of the inconsistent responses of judges, we have repeated the experiment in wine to determine if any real effects of salivary flow can be determined. Twenty judges rated the intensity of sourness, bitterness and astringency in wines by time-intensity methods following extended training. These results, while not demonstrating a large influence of salivary flow on perception were more consistent with the first study. Although the results are still being analyzed statistically, preliminary analysis across all judges indicates that the data are reproducible across replications and that significant differences among samples for each attribute were found for all Tl parameters. While neither concnetration nor salivary flow rate affected the time-to maximum, maximum intensity and total duration increased significantly between samples. Although very small differences in maximum intensities were found among flow groups, the subjects with the highest flow rates, perceived all three attributes for the shortest duration.