Metabolic Profiling of Grape and Wine Aromas

Most aroma compounds exist in grapes as glycosidically bound precursors and the aglycones (“free” aroma compounds) are released by enzymatic or acidic hydrolysis during crushing, fermentation and wine aging. We evaluated procedures for measuring the glycosidically bound volatiles using both acid and enzyme hydrolysis conditions followed by analysis of the free volatiles by Head. -Phase-Microextraction coupled with gas chromatography-mass spectrometry (HS-SPME-GC-MS). Acid hydrolysis of a mixture of standard compounds at pH 1 and 100°C for 1 hour released 20-60{aed9a53339cdfc54d53cc0c4af03c96668ab007d9c364a7466e3349a91bf0a23} of the bound volatiles, however, these conditions also resulted in significant degradation (>50{aed9a53339cdfc54d53cc0c4af03c96668ab007d9c364a7466e3349a91bf0a23}) of the free volatiles. Enzyme hydrolysis was generally more effective at releasing glycosidically bound compounds with minimal artifactual changes in the concentrations of the free volatiles. However, esterase and oxidase activity was still observed resulting in artifactual degradation of the free volatiles. Most previous procedures for evaluating volatile glycosides employ an initial sample preparation step using column chromatography to isolate the glycoside fraction from the grapes prior to hydrolysis and analysis of the free volatiles. We observed that different types of sample preparation columns were not equally effective at retaining the glycosides and no column-type was effective for all the glycosides in our mixture. Using optimized conditions obtained from the previous studies with model systems, we compared direct enzyme hydrolysis of Cabernet Sauvignon and Chardonnay grape homogenates with and without prior isolation of the glycoside fraction. Both procedures resulted in new “free” volatiles. However, results varied depending on the type of treatment. This research is providing important insight into approaches for rapid estimation of “aroma potential” of grapes.

Judging Wine Quality: Do We Need Experts, Consumers or Trained Panelists?

The overall objective is to determine if quality can be reliably evaluated by wine consumers, wine aficionados and/or wine experts and what their quality scores actually mean relative to the intensities of the sensory attributes we get from trained panels, and the liking scores we get from wine consumers. To choose wines with different quality levels we used the judging scores assigned to 27 Cabernet sauvignon wines in the 2012 California State fair judging competition.

At this competition Cabernet wines are divided into 9 regions. Within each region we picked the wine with the highest score (usually a gold medal wines), the wine with the lowest score (a wine that did not medal) and then a wine with a score as close to the mean between the highest and lowest score. The 15 descriptive analysis panelists (10 males, 22-72 years old, average age 37 yrs, median 31 yrs, standard deviation 17 yrs) were recruited from the students, staff and faculty of the Departments of Viticulture & Technology, Food Science & Technology, and Aerospace Engineering. They participated in 6 1-hr training sessions to collect, generate and obtain consensus on 21 aroma, 3 taste and 3 mouthfeel attributes. Their evaluation of the wines, in triplicate will be complete on 2/7/13. The consumer hedonic evaluation, using the 9-point hedonic scale has been scheduled for February 2, 2013 – we are aiming to have about 192 participants and each participant would evaluate 6 wines. Thus each of the 27 wines would be evaluated by 42 or 43 consumers.

Oxidation of Wine: Control for Quality

This report presents, first, the electrophilic nature of Q4MeC toward wines antioxidants and polyfunctional odoriferous desirable volatile thiols. 1H, 13C and 2D NMR analyses were used to identify the reaction products between 4-methyl-1,2-benzoquinone (Q-4MeC) and wine relevant nucleophiles like SO2, glutathione, ascorbic acid, phloroglucinol and a key odoriferous volatile thiol the 3-sulfanylhexanol. The formation of these reaction products was studied by HPLC-UV-MS, in wine like medium where the different nucleophiles were present in to two to four compounds mixtures. SO2, ascorbic acid and glutathione appear to provide a protection against 3SH consumption by the quinone, by acting as sacrificial nucleophiles. The antioxidant powers of SO2, AA and GSH were very similar and no antioxidant synergic effect was observed between these compounds when they tested in equimolar concentration.

Second we observed that the presence of carbonyl compounds in the matrix appear to decrease the SO2 antioxidant activity in a compound dependent manner. Acetaldehyde and pyruvate are very strong and fast SO2 binders, resulting in limited 3SH protection via their addition reaction with quinones. Ketoglutarate and galacturonate appeared less reactive with SO2, but their SO2 binding power needs further investigation under enological conditions. Finally, our NMR metabolomic approach of measuring the nucleophilic potential of wines reveled the importance of both SO2 and glutathione in controlling oxidation in wines, and opened new horizons for the development of powerful analytical tools for controlling oxidation during the vinification process and aging.

The Fate of Anthocyanins Under Warm Growing Conditions

In the proposal for the 2012, we outlined three objectives: 1. Compare the rate of anthocyanin degradation in warm climate grapes vs cool climate grapes 2. Determine if other phenolic compounds also experience degradation in warm climates 3. Identify by-products of degradation and potential mechanisms for their formation We have continued to make progress on developing the methods needed to make the desired measurements and those results are outlined in the attached proposal. We have clarified the accumulation of anthocyanins in Cabernet Sauvignon over two vintages and believe we have discovered a hidden pool of metabolites in the phenolic synthesis pathway in Vitis vinifera.

Towards objectives 1 and 2, we have collected all the grape samples from two vineyards, one at the UC Davis Oakville vineyard and the other at the UC Kearney Agricultural Center between July and October 2012. We have cluster temperature data from each site, collected hourly. Six berries were collected from 3 vines treated with tracer and 2 control vines at each site approximately 8 times during the growing season. The berries will be extracted during February, and student assistants are currently being trained to undertake this exercise. We extract the skin of each berry separately to improve the statistical significance of our data. We have the instrumentation ready to analyze the extracts once they are prepared. Towards objective 3, we have collected data from berries incubated in the lab, at 25?C versus 45?C, with high levels of the 13C6 phenylalanine (Phe) with an M+6 mass. This provided anthocyanin metabolites with approximately 35{aed9a53339cdfc54d53cc0c4af03c96668ab007d9c364a7466e3349a91bf0a23} containing the M+6 label. This allows for facile detection of the anthocyanins and other metabolites. Berry skins from the 8 biological replicate pairs of samples were all extracted and the extracts analyzed by Chip-LC-MS, using a QToF detector. This detector has a very high resolution and mass accuracy, so that detected substances can be identified with more certainty.

The large and complex dataset was transferred to Austrian collaborators we identified at the Grape Research Coordination Network meeting. Using this technique, lists of potentially labeled molecular features were found in positive and negative mode analyses. Of these candidate ions, 42 positive and 11 negative molecular features were found to be statistically different between 25°C and 45°C treatments. Only a select few molecular features were found to be significantly higher in the 45°C treatment; these compounds present our best candidates for the degradation products of anthocyanins. From the list of molecular features believed to be derived from Phe13 metabolism, a pair-comparison t-test was performed over the 8 biological replicates to determine significant differences in the concentrations of labeled features. As expected, many phenolic compounds are susceptible to degradation under high temperatures. At least one of each anthocyanidin moiety appeared to be degraded under 45?C temperatures, although every single anthocyanin did not vary significantly between treatments.

Most interestingly, 3 features were found to be in greater quantity in the 45?C than in 25?C grape treatments. These features had m/z values of 433.113, 347.076, and 585.170. The molecular feature of mass 433.113 was tentatively identified as a benzyl alcohol dihexose. A similar molecular feature was found when researchers probed the degradation of anthocyanins in flower petals (Bar-Akiva et al. 2010). The feature with m/z 347.076 was tentatively identified as a syringetin aglycone. Syringetin, while present in grapes at low levels, is most likely an artifact of malvidin-3-glycoside created during ESI, since their molecular structures only vary in the oxidation states of the flavonoid ?C? ring and their retention times are nearly identical. This could easily be caused in conditions in which the pH is above 2, when anthocyanins undergo structural transformations. The final molecular feature identified 585.170 has not yet been identified, although if the molecular feature represents a product of anthocyanin degradation, it would have to be an adduct of some kind since the m/z is higher than many of the degrading anthocyanins. We were expecting more compounds that would increase at higher temps, but have found background noise is limiting our sensitivity, most likely due to the type of LC separation used (Chip-LC). At the present time, we are trying various methods to reduce that noise, through complicated data analyses, in hopes that the signals of the temperature sensitive and rising metabolites will be more evident.

Improvement of Wine Quality: Tannin and Polymeric Pigment Chemistry

Accomplished Objectives (06/2012-01/2013)

Upon receipt of funding experimental design was finalized and a strategy for implementation was refined. Appropriate grape varieties were selected for winemaking. Meanwhile, instrumentation needs were satisfied and preliminary cocoa, preveraison skin, grapeseed extracts, and fractionated young wine samples were obtained. We are 4 months into the practical and laboratory experimentation portion of the project which began with September’s harvest. At the end of September and into the beginning of October fruit was harvested from the UC Davis Oakville Research vineyard. Production progressed through December with bottling January 28, 2013. Upon completion of secondary fermentation, instrument training, method development, and data acquisition took place.

Metabolic Profiling of Grape and Wine Aromas

Most aroma compounds exist in grapes as glycosidically bound precursors and the aglycones (“free” aroma compounds) are released by enzymatic or acidic hydrolysis during crushing, fermentation and wine aging. We evaluated procedures for measuring the glycosidically bound volatiles using both acid and enzyme hydrolysis conditions followed by analysis of the free volatiles by Head. -Phase-Microextraction coupled with gas chromatography-mass spectrometry (HS-SPME-GC-MS). Acid hydrolysis of a mixture of standard compounds at pH 1 and 100°C for 1 hour released 20-60{aed9a53339cdfc54d53cc0c4af03c96668ab007d9c364a7466e3349a91bf0a23} of the bound volatiles, however, these conditions also resulted in significant degradation (>50{aed9a53339cdfc54d53cc0c4af03c96668ab007d9c364a7466e3349a91bf0a23}) of the free volatiles. Enzyme hydrolysis was generally more effective at releasing glycosidically bound compounds with minimal artifactual changes in the concentrations of the free volatiles. However, esterase and oxidase activity was still observed resulting in artifactual degradation of the free volatiles.

Most previous procedures for evaluating volatile glycosides employ an initial sample preparation step using column chromatography to isolate the glycoside fraction from the grapes prior to hydrolysis and analysis of the free volatiles. We observed that different types of sample preparation columns were not equally effective at retaining the glycosides and no column-type was effective for all the glycosides in our mixture. Using optimized conditions obtained from the previous studies with model systems, we compared direct enzyme hydrolysis of Cabernet Sauvignon and Chardonnay grape homogenates with and without prior isolation of the glycoside fraction. Both procedures resulted in new “free” volatiles. However, results varied depending on the type of treatment. This research is providing important insight into approaches for rapid estimation of “aroma potential” of grapes.

Barrel Maturation, Oak Alternatives and Micro-Oxygenation: Influence on Red Wine Aging and Quality

The impact of micro-oxygenation (MOX) in conjunction with a variety of oak alternatives on phenolic composition and red wine aging was investigated and compared with traditional barrel aging.  Although several studies concluded that MOX give similar results to barrel aging, few have compared them directly and none directly compared MOX with and without wood alternatives and barrel aging.  Results confirmed that MOX had a positive effect on color density, similar to barrel aging, even after 5 months of bottle aging.  This is supported by an increase in polymeric phenol and pigment content not only with aging but in the MOX compared to non-MOX wine treatments.  Descriptive analysis results indicate that the wood treatment dominated differences among treatments.  This study showed that MOX in combination with wood alternatives such as oak chips and staves could mimic short term (six months) barrel aging in new American and French oak barrels.

Evaluating the Effects of Sterile Membrane and Other Filtration on the Sensory and Chemical Properties of Wine

The goal of our project is to evaluate the effects of sterile membrane and other filtration on the sensory and chemical characteristics of wine. To do this, we have filtered two red wines, a Cabernet Sauvignon and a Merlot, and one white wine blend through 0.45 µm PVDF and PES membrane filters and compared the sensory and chemical characteristics of these wines to unfiltered control wines. Treatments were expanded with the Merlot and white blend to also examine the effects of a pad filter and cartridge depth filter used as prefilters. We have examined changes in dissolved oxygen content, tannin content, and color during the course of filtration and found only minor changes. Sensory panels were trained for each of these wines and each of the treatments evaluated immediately after filtering and then on a regular basis for 9 weeks (for the Cabernet), 24 weeks (for the Merlot), and 20 weeks (for the white blend). While all three wines changed significantly over time in the bottle, very few significant differences were observed in aroma or mouthfeel between filtration treatments. In other words, our results thus far indicate limited impact of sterile filtration on the sensory or chemical properties of the wine, regardless of the type of filter material used. We did observe a small decrease in tannin and astringency with the pad filter for the Merlot wine, but we are continuing to investigate the significance of this change. In addition, we are preparing to evaluate the effects of filtration of both a red and white wine using a cross-flow filter, as well as evaluating the effects of the associated pumps, during the next three months.

Influence of Grape and Wine Production Practices on Tannin Extractability and Activity.

The objectives of this proposal are to do the following: I. Develop optimized analytical methods for the measurement of: A. Grape skin tannin extractability and concentration potential in red wine. B. Predicted concentration-independent interaction of red wine tannin with protein. II. Explore the influence of grape and wine production practices on developed analytical methods These objectives are consistent with the highest priority research objective as outlined by the American Vineyard Foundation as well as the National Grape and Wine Initiative Part A of the first objective addresses observations that higher quality wines are associated with an increase in skin tannin concentration. Presumably this is a vineyard-derived aspect of wine quality in that the data collected to date were designed to understand vineyard management practices. Historically it has been difficult to predict tannin extraction into wine from grape-based data. Part of this difficulty can be attributed to the lack of accounting for the dependence of extraction on cellular integrity. In order to take into consideration the cellular integrity of the grape berry, an analytical approach which exposes grape berry samples to mild mechanical force is being employed. It is expected that incorporating an analytical component that accounts for cell integrity will provide an improved ability to predict red wine tannin extraction and quality, from fruit samples. Part B of the first objective is designed to improve our understanding of how tannins are perceived and address the generally held perception that tannin quality has a concentration-independent component. Recently published research has found that the enthalpy change of a protein-tannin interaction experiment is dependent in part on the “age” of the tannin in that apparent interaction between tannin and protein declines as tannin structure modification increases. This research was performed using Isothermal Titration Calorimetry on previously purified tannins. This portion of the project will focus on the development of an analytical method that can predict protein-tannin interaction from wines directly. In the second objective, optimized methods will be used to understand experiments in progress by other research programs, vineyards, and wineries. IV. Objective(s) and Experiments Conducted to Meet Stated Objective(s): Part A, Objective #1 The objective is to develop an extraction method that is mild yet predictive of potential skin tannin extraction. A successful method will compare favorably with wines made consistently from the same fruit. A further criterion for success will be that the analytical approach is as rapid as practically possible. Fruit samples have been collected from the Fresno State vineyard and have been subjected to readily-available equipment designed to introduce various degrees of sheer to the sample. The specific target was the introduction of sufficient sheer to differentiate between berries having variable cellular integrity. We initially investigated a Kitchenaid kitchen mixer equipped with various implements. After exploring this mixer with and without added ethanol, a decision was made to test a new instrument due to insufficient extraction. We next transitioned to a Cuisinart food processor. The food processor with a dough kneader worked well for our purposes and all extracts prepared were found to be sufficiently extracted. E&J Gallo was approached and they expressed interest in cooperating on this project. To provide a meaningful data set, they offered to provide us with berry samples and wines from Cabernet Sauvignon vineyards from warm interior valley sites, to cool north coast sites. Wines were also prepared from this fruit to provide comparison. To date, extracts have been prepared and have been analyzed. Also to be analyzed by May 2012 will be exhaustive skin extracts and the red wines produced from samples. Extracts prepared with the food processor and preliminary comments from E&J Gallo indicate that the results are positive. Extracts were analyzed by acid-catalyzed degradation in the presence of excess phloroglucinol. This analytical approach was used because it could provide information on the composition of tannins in the extracts and hence the tissue of origin. Of specific interest was determining the prodelphinidin content which is skin tissue-derived.

Oxidation of Wine: Control for Quality

This report presents first evidence of managing wine flavor evolution during oxidation (including aging), by altering the reactions of quinones. O-quinone of 4-methylcatechol has been produced in acetonitrile by periodate resin. Michael addition reactions of o-quinone 4-methylcatechol to various nucleophiles, including varietal volatile thiols, hydrogen sulfide, glutathione, sulfur dioxide, ascorbic acid and amino acids like methionine and phenylalanine have been carried out, obtaining alkylated adducts in fairly good to quantitative yields. The reaction rate and selectivity of o-quinone 4-methylcatechol toward the studied nucleophiles have been investigated by UV-Vis spectrometry by following the loss of quinone chromophore. The observed reactivity spans 3 orders of magnitude on passing from amino acid (methionine and phenylalanine) (KNu = 0.0002 s-1) to the most reactive nucleophiles, the hydrogen sulfide (KNu = 0.4188 s-1). These are the first direct reaction rate measurements of nucleophilic addition to the oxidized catechol. The classification of the studied nucleophiles revealed the existence of three categories. The first group consisted of amino acids (methionine and phenylalanine) having rates of essentially zero. Next, phloroglucinol, has a very low rate (K = 0.0064). Despite the slow rate of this reaction, this data suggests phenolics may be able provide a protection for varietal thiol aroma under some circumstances. The next group of compounds includes the volatile thiols having increasing reactions rates K as steric inhibition declines. The tertiary thiol, 4MSP, was much less reactive with quinones than the secondary (3SH) and primary thiols (2FMT), and quite close to the rate of phloroglucinol. The odoriferous volatile thiols (4MSP, 3SH, 2FMT) showed lower K values than those of the third group of the wines antioxidant compounds (SO2, GSH, AA) and H2S. The result indicates that both the antioxidant compounds (SO2, GSH, AA) and H2S would react preferentially with oxidation induced quinones, resulting the preservation of varietal aromas. The characterization of the reaction products between the studied nucleophiles and ortho-quinones derived 4-methylcatechol, was performed by using HPLC-MS analyse. Three o-quinone 4-methylcatechol/H2S; phloroglucinol adducts, two o-quinone 4-methylcatechol/3SH; 2FMT; 4MSP adducts, and one o-quinone 4-methylcatechol/SO2 were detected. No adduct was detected into the reaction medium where o-quinone 4-methylcatechol was incubated with both methionine and phenylalanine explaining why there observed reaction rates were essentially zero. Competition experiments provided strong support to involvement of antioxidants such as such as SO2, glutathione and ascorbic acid should to the preservation of wine?s varietal flavors.