Abstract
Petioles, rachis, and leaves are all matter other than grape (MOG), and although originating from vines, they potentially contaminate primary fermentations of red grape must. Fruit contaminated with high levels of MOG may be downgraded or rejected at the winery; however, management practices such as hand harvesting or fruit sorting may reduce MOG. Petioles are the most common form of MOG to be included in the must, yet little is known about their chemical composition and resulting effects on the sensory and chemical properties of red wines. A descriptive analysis panel (n = 12) examined the sensory profiles of Cabernet Sauvignon wines to which 0.5% or more of MOG (petioles or unripe berries) had been added. This sensory analysis showed that an addition of 10% petioles to the must had a perceived impact on sensory wine qualities, increasing floral aromas and reducing bitterness relative to the sensory qualities of control wines, and resulting in a fuller body than that of wines made with unripe (i.e., green) berries. These sensory results are related to increased terpene concentrations as petioles increase in concentration in the wine must. Methoxypyrazines can also be detrimental contaminants to Cabernet Sauvignon wines. Concentrations of the predominant methoxypyrazine in the wines, 3-isobutyl-2-methoxypyrazine, increased with increasing additions of unripe berries to the must. Wines made with 0.5% or more of unripe berries during fermentation were associated with increased perceived acidity in the sensory analysis and an aroma that was more leafy and vegetal green than the aromas of some wines made with added petioles. The results of this study provide winemakers with important information to better understand how grape-harvesting choices may affect the perceived quality and chemical composition of their wines.
- matter other than grape
- MOG
- sensory analysis
- wine composition
- gas chromatography-mass spectrometry
- GC-MS
- multivariate statistics
- grape sorting
In response to rising labor costs and lower staff availability, vineyard mechanization is becoming a common practice worldwide, which is certainly the case in Australia (Clingeleffer 2013). Since the introduction of the first machine harvesters to Australia in the early 1970s (May 1973, May et al. 1974), virtually all of its national crop is now mechanically harvested. The potential drawbacks of machine harvesting are increasing amounts of matter other than grape (MOG) and of unripe (i.e., green) fruit that may end up in the must. However, the presence of stems in fermentations is well known to be important for the quality and composition of red wine. Red wines made from nondestemmed grapes usually contain higher amounts of phenolic compounds (Sun et al. 2001) that can impact such properties as color, browning, bitterness, and astringency. These compounds can also be potentially modified by the yeast to produce certain volatile compounds. MOG is mostly composed of petioles and has been measured at a concentration of 4.7% in machine-harvested fruit (Huang et al. 1988). At such concentrations, it is predicted that the MOG will certainly have an impact on the sensory attributes of the wine (Huang et al. 1988), and at high levels, MOG may result in fruit being downgraded or rejected (Allan 2014). Furthermore, as all fruit is harvested, including any unripe bunches may introduce compounds into the ferment that also influence the sensory characters of the resulting wine. Although it has been recognized that MOG is commonly included in primary fermentation of grapes, few studies have investigated the effects of MOG or unripe fruit on the final sensory and chemical properties of the finished wine.
Although MOG and unripe berries may affect the quality of wine produced from all grape varieties, some varieties may be more prone to sensory taints introduced by these materials. One such cultivar is Cabernet Sauvignon, which produces methoxypyrazines that impart vegetal or herbaceous sensory characters to wines (Allen and Lacey 1998, Allen et al. 1996). Methoxypyrazines partly define Cabernet Sauvignon varietal character; however, as methoxypyrazine concentrations increase, sensory descriptors may begin to change to more green and earthy characters, which are seen as negative attributes (Lattey et al. 2010). Lattey et al. (2010) also reported that green flavor is important for discriminating acceptance scores among consumer clusters. Methoxypyrazine concentrations in the wine may particularly increase with the addition of MOG to a fermentation. For example, bunches of unripe fruit or vegetative tissues have levels of methoxypyrazines up to 10-fold higher than are observed in ripe Cabernet Sauvignon berries (Dunlevy et al. 2010). This material is more likely to enter fermentation when the fruit has been mechanically harvested. Cabernet Sauvignon is also predisposed to high numbers of live green ovaries (Dry et al. 2010), which even with hand harvesting have the potential to contaminate the fermentation and contribute to methoxypyrazine concentrations.
Automated grape sorting to replace people has become a more viable technology for removing MOG (Prade 2003) and unripe berries (Falconer et al. 2006) from machine-harvested fruit prior to crushing. Labor savings may justify the introduction of this type of technology to replace hand harvesting of fruit. If the fruit has previously been machine-harvested, an improvement in wine quality is required to justify investments in new processes and systems. To provide the wine industry with information to support decision making about alternative fruit processing, in particular, automated grape sorting, we sought to determine at which levels unripe berries and petioles in the must may impact sensory profiles of the resulting wines. Therefore, this study aimed to determine the effects of petioles and unripe berries on (1) the sensory properties, and (2) the chemical composition of Cabernet Sauvignon wine. This was achieved by controlled additions of varying levels of either unripe berries or petioles prefermentation to subsamples of a large parcel of grapes. Laboratory-scale fermentations were conducted, and the resulting wines were then assessed by a sensory panel. The concentrations of selected volatile and other chemical compounds were also examined.
Materials and Methods
Grape variety and production of experimental wines.
Cabernet Sauvignon (Vitis vinifera L.) grapes were hand-harvested in the 2010 to 2011 season at commercial maturity (24.3 Brix) from a vineyard located in the Coonawarra geographical indicator region in southeastern South Australia (see Wine Australia website at www.wineaustralia.com). Preveraison (i.e., unripe) berries were collected from the same region ~6 wks before fermentation and after addition of 50 mg/kg potassium metabisulfite (PMS; K2S2O5), were stored at 0°C in the dark. Petioles were picked two to three days before use and stored at 0°C in the dark.
Additions of either petioles or unripe berries to replicated, small-scale Cabernet Sauvignon musts were incrementally increased. Specifically, 10 must addition treatments were performed in triplicate: Control, 0.5, 1, 2, and 5% unripe berries (w/w %), and 0.5, 1, 2, 5, and 10% petioles (w/w %). The unripe berries and petioles were added to crushed and destemmed (ENO-15; Enoitalia, Florence, Italy) must, yielding 30 individual fermentations of 20 kg contained in 22-L stainless-steel pressure vessels.
Musts were inoculated with the active dry form of the aroma-neutral yeast strain PDM at 0.5 g/L (Maurivin, Sydney, Australia) and then with MBR LALVIN VP41 malolactic bacteria (Lallemand, Adelaide, Australia) the following day according to manufacturer instructions. Fermentation kinetics (i.e., changes in malic acid, glucose, and fructose concentrations) were monitored daily (data not presented), and the cap was hand-plunged twice daily. A preventative addition of 200 mg/L diammonium phosphate (DAP; (NH4)2HPO4) was added to all wines at 10.8 Brix. Once the primary fermentation reached ~3.6 Brix, it was pressed with a 20-L bladder press (Zambelli, Hydro 20, Camisano Vicentino, Italy) into 10-L glass demijohns. When the primary and malolactic fermentations had completed, a 60 mg/L PMS addition was made as the wine was racked off the yeast and bacteria lees. After racking off the lees, a 0.1 mg/L CuSO4 addition was made to all fermentations to remove any H2S, and 1 g/L tartaric acid was also added to aid with SO2 protection. After cold stabilization had completed, the wines were filtered using a pad filter (Colombo-Rover pump and six-pad filter; Polverara PD, Italy) with 0.8-μm Z6 cellulose filter pads (Ekwip; IMCD Australia Ltd., Mulgrave, Victoria, Australia). Further SO2 additions were performed as required, maintaining 40 mg/L of free SO2 until bottling under inert gas in 375-mL green glass bottles with metal rolls on tamper-evident screw caps. Wine bottles were stored horizontally at 15°C until further chemical and sensory analyses.
Sensory analysis.
A descriptive analysis (DA) was performed four months after bottling to quantitatively define any differences in the sensory profiles among the 30 wines. Wines were evaluated over August, September, and October 2011 by a panel of 12 people (five women and seven men). Panelists were University of Adelaide students enrolled in either postgraduate enology and viticulture programs or in Ph.D. research programs. One panelist was a qualified, practicing winemaker. Prior to formal DA training, the panelists underwent 20 hours of high-level training in aroma, taste, and trigeminal sensation and in detection, identification, evaluation, and ranking of wine samples over a five-week period. Panelists then underwent eight weekly two-hour training sessions and three two-hour formal evaluation sessions.
Initially, during the weekly two-hour training sessions over the first three weeks, the panelists assessed all 30 wines to define the attributes that the panel determined by consensus to be differentiating characteristics of the wines. The 18 descriptive terms the panel ultimately agreed on included eight aromas, two tastes, three flavors (where flavor is defined as aroma by mouth), three mouthfeel, and two aftertaste attributes (Table 1). In five further two-hour training sessions, the panel was familiarized with the sensory attributes, scales, and assessment of wines in computerized sensory booths. Color was determined by the panel not to be a differentiating characteristic of the wines, but to ensure there was no influence by any visual cues, the sensory assessment was carried out under red lights. Panelists practiced rating the wines using an unstructured 15-cm line scale with indented end anchor points placed at 10 and 90% of the scale, respectively, and with a midline anchor point. The practice was held in isolated booths in a 12-booth sensory laboratory under conditions identical to those used in the subsequent formal tasting sessions.
List of descriptive analysis aroma and palate attributes with agreed definitions, reference standards, and scale anchors.a
Intensity rating standards for aromas were provided at each session to aid in the rating of the intensity of all eight aromas. High and low aroma intensity standards were prepared using a “high” 1:10 raspberry cordial (Cottee’s, Tullamarine, Victoria, Australia) dilution in water and a “low” intensity standard of a 1:80 dilution. Eight reference standards were used for the following aroma attributes: fruit, spice, earthy, floral, leafy green, herbal green, vegetal green, and medicinal green (Table 1). The reference standards were presented in black INAO 215-mL glasses and diluted in 40 mL of Shiraz bag-in-box wine (Yalumba Wine Company, Angaston, South Australia). They were presented to the panel before the wine assessment, and panelists were encouraged to reevaluate these references as required during the wine assessment and during forced breaks between samples. The panelists did not receive reference standards for taste, aftertaste, and mouthfeel during formal evaluation but had extensive training in these attributes.
Panel performance was evaluated in the last two training sessions by having each panelist assess a subsample of the wines in duplicate. These data were analyzed using Panel-Check (Nofima Mat and DTU – Informatics and Mathematical Modeling, Tromsø, Norway) and SENPAQ (Version 5.01; Qi Statistics, Berkshire, UK). As no significant panelist by sample interactions were found, the panel was judged to be qualified for commencing final evaluation of the samples.
Formal assessment of the wines was conducted in two sessions of two hours each in the individual sensory booths. Each wine was evaluated in a randomized presentation order, balanced for carryover effects, and in duplicate over the course of the formal rating period. A series of 30-mL wine samples was presented in clear INAO 215-mL tasting glasses covered with small glass Petri dishes and coded with three-digit random numbers. Before the panelists entered the booths, they familiarized themselves with both the intensity and the reference standards, which were also available throughout the DA if needed. Distilled water and unsalted crackers were provided to cleanse the palate between samples. A one-minute break between each wine sample and a five-minute break after five wine samples was imposed to prevent palate fatigue.
Chemical wine analyses.
Residual glucose and fructose was measured with a commercial kit (CLINITEST Reagent Tablets; Bayer Corporation, Elkhart, IN). The following measurements were performed as described in Iland et al. (2004): wine pH was measured directly with a pH meter, titratable acidity (TA) was measured by titrating to pH 8.2 with sodium hydroxide, total SO2 was determined by the aspiration method, volatile acidity (VA) was measured by titration, and alcohol percentage was determined by distillation followed by hydrometry. Final concentrations of malic acid in the wines were measured with the Enzytec l-malic acid kit (R-Biopharm AG, Darmstadt, Germany) according to the manufacturer instructions. The wine samples were diluted 1:10 with Milli-Q water prior to analysis for concentrations of iron, magnesium, sodium, and potassium. These wine trace elements were analyzed using an inductively coupled plasma-optical emission spectrometer (Optima 2100DV; Perkin Elmer, Waltham, MA) by Waite Analytical Services (University of Adelaide, Australia), using the method of Wheal et al. (2011).
Headspace volatile analysis solid-phase microextraction-gas chromatography-mass spectrometry (SPME GC-MS).
Two SPME GC-MS methods were used to analyze the volatile compounds produced during the fermentation of the 10 different treatments. All wines underwent a stable isotope-dilution analysis method coupled with headspace volatile analysis SPME GC-MS to measure the levels of 3-isobutyl-2-methoxypyrazine (IBMP), as this compound has been implicated in vegetal and green sensory characters in wines made from Cabernet Sauvignon grapes. IBMP concentrations were measured according to the method described by Dunlevy et al. (2010). Furthermore, SPME GC-MS using a nontargeted approach was conducted to determine whether any other volatile components were changing in response to the different additions to the musts.
For the nontargeted analysis of the volatile profiles of the wines, sample preparation and chromatographic conditions were identical to those described by Keyzers and Boss (2010). In this analysis, compounds for more detailed analyses were first identified by overlaying the chromatographs and identifying components consistently changing in response to MOG additions. These volatile compounds were identified through a library search of an in-house mass spectral library of 346 compounds, the United States National Institute of Standards and Technology 2008, and the Wiley Registry 9th edition mass spectral libraries, and the identities were confirmed by comparison of linear retention indices with those of authentic standards. The linear retention indices were calculated from a compound retention time relative to the retention of a series of n-alkanes (C8 to C20). The selected compounds were then quantified by running standard curves (containing six data points) with authentic standards in model wine (12% ethanol, pH 3.7) along with a d13-hexanol internal standard. These curves were then used to determine the amount of target analyte in the experimental wines.
Statistical analyses.
The data on wine chemical composition were analyzed by one-way ANOVA with Tukey’s honestly significant difference (HSD) post hoc test in XLSTAT (v. 2013.5.03; Addinsoft SARL, Paris, France). For the DA, a mixed-model, two-way ANOVA with panelists as random effects and samples as fixed factor effects was used, with Fisher’s least significant SD post hoc test where p < 0.05 was considered statistically significant using SENPAQ. A principal component analysis (PCA) was also performed in XLSTAT. The presentation orders of the wine samples and random three-digit codes for each sample in the DA were generated in Design Express Version 1.6 (Qi Statistics).
Results and Discussion
Wine chemical composition.
The results of standard chemical composition analyses indicated that the wines with the different petiole and berry additions had largely similar physicochemical and chemical characteristics (Table 2), indicating that these characteristics would not strongly, if at all, contribute to any sensory differences among the wines. Because all wines had residual sugar (RS; glucose + fructose) concentrations of equal or less than 0.62 g/L, technically, they were considered dry. The wines ranged in pH from 3.46 to 3.58. Only wines with petioles added at 0.5 or 2% had a significantly higher pH than wines with unripe berries added at 2 or 5%.
Results of the chemical analysis of the 10 Cabernet Sauvignon wines.a
Statistically significant (p < 0.05) differences in alcohol percentages were also observed among some of the wines. Specifically, the wine with the highest unripe berry addition and therefore lowest fermentable sugar concentration, i.e., the 5% unripe berry-addition wine, had the lowest alcohol percentage. However, these differences, although statistically significant, were generally not large (e.g., the largest difference in alcohol content was only 0.4 percentage points) and was therefore unlikely to have had a major impact on sensory perception (Yu and Pickering 2008).
DA of MOG-added Cabernet Sauvignon wines.
The experimental Cabernet Sauvignon wines were analyzed by the DA panel to determine any sensory effects the different prefermentation additions of either petioles or unripe berries had on the wines. Of the 18 sensory attributes evaluated, seven were statistically significantly different (p < 0.05) among the wines: four aroma attributes (i.e., vegetal, leafy, floral, and earthy), two taste descriptors (bitterness and acid), and one mouthfeel attribute (body; Table 3). There was a petiole concentration–dependent increase in perceived floral wine aroma, and notably the wine with 10% added petioles had significantly higher perceived floral aroma than all of the other wines, except for that with the 5% petioles. The opposite pattern was observed for vegetal and leafy aromas; the wines with 2, 5, and 10% petioles were rated as having less vegetal and leafy aromas than the other wines, and whether this reduction was the result of a suppressive effect due to a mixture of odors (Ache 2010, Cashion et al. 2006) needs to be further investigated. Generally, petiole-containing wines tended to have greater body and less perceivable acid taste. Although crossmodal interactions between the gustatory and olfactory senses have been demonstrated (Boakes and Hemberger 2012), whether these effects accounted for the suppressed acid taste perception due to increased floral or decreased green flavor in our wines remains to be tested.
Intensity rating scores of sensory attributes in Cabernet Sauvignon wines produced from fermentations with or without varying amounts of unripe berries or petioles.
A correlation matrix, generated with the mean rating values from all the panelists of each wine across all of the seven significant attributes from the three sessions, was analyzed by PCA. The first and second principal components (PCs) accounted for 83.4% of the variance in the sensory data of the 10 Cabernet Sauvignon wines (Figure 1); the third PC explained a further 7.7% of the variation (data not shown), and earthy aroma was strongly loaded on this PC. PC1 differentiated the wines according to the scores for floral aroma as opposed to the scores for vegetal, leafy, and, to a lesser extent, earthy aromas. PC2 separated the wines either by bitterness and body or by acid taste. The control wine and those made with 0.5 or 1% petiole additions and with 1% berry additions clustered in the upper left-hand quadrant of the graph. This clustering indicated that these wines all had similar intensities of the differentiating attributes, being lowest in floral aromas, having more vegetal and leafy aromas, and being moderately to highly bitter.
PCA biplot of the means of sensory attributes (solid lines) and 3-isobutyl-2-methoxypyrazine and terpene concentrations as additional data (dashed lines) for the 10 different Cabernet Sauvignon wines (solid circles). Wines were made from subsamples of the same starting must, but with different petiole and unripe berry matter other than grape (MOG) additions. Control: no MOG added to must; BER 0.5, 1, 2, and 5: 0.5, 1, 2, and 5% unripe berry additions to the must; PET 0.5, 1, 2, 5, and 10: 0.5, 1, 2, 5, and 10% petiole additions to the must.
The scores for floral sensory characters increased with the percentage of petiole content in the must. Most notably, the wines produced with 2 or 5% petiole additions had an increase in both body and floral aromas but were not perceived as acidic or leafy (Figure 1); the wine made with 10% petioles had the greatest floral aroma of all wines and had the lowest bitterness of the other petiole-added wines and the least vegetal or leafy aroma overall.
The wines produced from musts with the additions of 0.5, 2, and 5% of unripe berries were separated on PC2, which accounted for 15.3% of the variation among the wine sensory attributes. These wines had increased intensities of leafy aroma and acid taste and lower bitterness and body (Figure 1). The chemical analysis showed that the wines did not significantly differ in TA or malic acid (Table 2), yet the pH values were significantly lower in the wines with 2 and 5% unripe berry additions. It is possible that the perceived increase in acidity was associated with the increase in unripe leafy aromas in the 0.5, 1, and 2% unripe berry-addition wines, but when more than 5% unripe berries are added, acid perception may dominate and reduce perception of the leafy and vegetal aromas. The panelists may have perceived higher acidity with increased green aromas, as both of these attributes typically are associated with unripe fruit (Roujou de Boubee et al. 2000). Notwithstanding any crossmodal effects on perceptions by the DA panel, it was clear that the addition of unripe berries at ~2% and the addition of petioles of between 5 and 10% both affect the sensory profiles of the resulting wines such that they influence perceived wine quality of Cabernet Sauvignon.
Wine elemental analysis.
An elemental analysis indicated no notable trends in elemental composition among the wines with the unripe berry additions (Table 4). In contrast, the wines produced from musts with added petioles exhibited significant increases in some elemental ions (Table 4). The wine produced after a 10% petiole addition had significantly higher concentrations of iron and sodium than the other wines. Higher iron concentrations may negatively impact the finished wine quality because they may increase transition metal-catalyzed oxidation reactions that prematurely age wines (Danilewicz and Wallbridge 2010). We observed significant increases also in magnesium in the wines with 2, 5, and 10% petiole additions and in potassium in the wines with the 5 and 10% petiole additions (Table 4). Although high potassium levels can precipitate tartaric acid in wine, we found no evidence for this in our wine series, as the TA and pH remained unchanged among these treatments and the other wines (Table 4). Potassium supplementation has been also shown to improve fermentation outcomes when its concentrations are low in musts (Schmidt et al. 2011); however, there were no issues with sluggish fermentations during wine production.
Elemental analysis of iron, magnesium, sodium, and potassium concentrations in the 10 finished wines.a
It is possible that the elemental differences in the wines, which appeared to be due to the addition of the petioles, had altered yeast metabolism and had thus changed the profile of the metabolites responsible for the wine sensory attributes. Nevertheless, it is more likely that the presence of certain metabolites in the unripe berries or in the petioles had more directly contributed to the volatile profiles of the wine. To investigate this possibility, the differences in volatile components among the wines were examined with SPME GC-MS.
SPME GC-MS analysis of the wines.
We ran two SPME GC-MS analyses to measure volatile compounds in the wines produced after the unripe berry or petiole additions. First, a SIDA method was employed to measure the levels of IBMP in the samples because this compound has been implicated in vegetal and green sensory characters in wines made from Cabernet Sauvignon (Preston et al. 2008). Although three methoxypyrazines have been previously detected in wines (Allen and Lacey 1998), neither 3-isopropyl-2-methoxypyrazine nor 3-sec-butyl-2-methoxypyrazine were observed in the wine samples produced in this study. We did note trends toward lower IBMP concentrations as petiole additions increased, yet these trends were not statistically significant (Table 5). This may mean petioles were absorbing IBMP from the fermentations and that IBMP was then removed during the racking off the wine from the berry skins and MOG. This observation supports the findings of Dunlevy et al. (2010) who showed that Cabernet Sauvignon leaves (and petioles) contain relatively low levels of IBMP.
3-Isobutyl-2-methoxypyrazine (IBMP) concentrations in the experimental and control wines.
The addition of 5% unripe berries significantly increased IBMP concentrations in the wines compared with the control wine (Table 5). As the unripe berries were picked just preveraison, and research indicates that IBMP levels in the berries steadily decrease during the ripening phase (Dunlevy et al. 2010, Ryona et al. 2009), this result was not unexpected. However, the addition of lower amounts of unripe berries did not significantly increase IBMP levels in the wines (Table 5). This lack of an increase in IBMP may explain why increasing the amounts of unripe berries in the fermentations did not increase the sensory scores for the vegetal aroma, which is often associated with higher IBMP levels. Instead, the increase in petiole additions was associated with a decrease in vegetal aroma (Figure 1). IBMP concentrations in the petiole-addition wines were lower than in the unripe berry-addition samples, but only the difference with the 5% unripe berry-addition sample was statistically significant, suggesting that the differences in IBMP concentration did not explain the sensory differences between the unripe berry-addition and petiole-addition wines. However, it has been shown that other volatile compounds can mask the sensory effect of IBMP, which may be the reason for the sensory differences in the wines produced after petiole additions (Hein et al. 2009). In addition, the vegetal sensory character may originate from another volatile component of the wine (Preston et al. 2008).
With these considerations in mind, we also analyzed the wines by SPME GC-MS in a nontargeted manner to determine if volatile components were changing in response to the different additions to the musts. Although no statistically significant differences were observed in the total ion chromatograms of the wines made after additions of unripe berries (data not shown), a number of compounds appeared to increase as an increasing percentage of petioles were added to the musts. These compounds were putatively identified as several monoterpenes and as the aromatic compounds eugenol, methyl salicylate, and ethyl salicylate. These compound identities were confirmed with authentic standards. In total, we measured 13 compounds in the wines produced after petiole addition (Table 6).
Terpene concentrations in wines made with varying amounts of petioles.
The concentrations of these compounds in the wine samples with the 10% petiole addition were higher than those in all the other wine samples. This observation was consistent with that of the PCA analysis (Figure 1) which suggests that this sample is most different from the control wines. The wine produced with 5% petiole additions had significantly higher levels of 10 of these compounds than the control wine, and the 2% petiole-addition wine significantly differed from the control wine in the concentrations of two of the quantified compounds, reflecting the separation of these samples on PC1 (Figure 1). The 2, 5, and 10% petiole-addition wines also had higher scores for floral aroma and lower scores for vegetal aroma (Figure 1).
Of the compounds whose levels were significantly higher in the 10% petiole-addition wine, linalool, α-terpineol, β-citronellol, nerol, ethyl salicylate, and geraniol have floral as one of their descriptors. Moreover, the concentrations of linalool, β-citronellol, and geraniol were all above the reported aroma detection thresholds in the 5 and 10% petiole-addition wines (and in the 2% petiole-addition wine, the concentration of β-citronellol was only slightly lower than its threshold concentration), which suggests that their presence will be detected by consumers.
The increase in floral aromas when greater than 5% of the must was petiole tissue appeared to be due to increased terpenoid concentrations in these wines. The terpenoids probably originated from the petioles because grapevine leaves are a rich source of these compounds (Gunata et al. 1986). Terpenes play a large role in the aromas of white wines such as Gewürztraminer, Riesling, and Muscat, which are well known for their floral and citrus notes (Dziadas and Jeleń 2010). Typically, Cabernet Sauvignon wines are considered terpene-neutral, and this class of compounds typically contributes little to the finished aroma of this wine (Kalua and Boss 2009).
The increase in floral sensory scores in the petiole-addition wines was surprising because we had hypothesized that the vegetative tissue would increase sensory scores associated with green attributes. Although floral attributes in Cabernet Sauvignon wines may not be a dominant characteristic of this variety, previous sensory studies, e.g., Bramley et al. (2011) and Robinson et al. (2011), have reported that they are present in these wines, and they may therefore not be noted as unusual by consumers.
Relationships between volatile IBMP and terpene flavor chemistries with wine sensory attributes.
When the IBMP and terpene concentration data were overlaid on the PCA biplot, acid taste and leafy aroma were both positively correlated with IBMP concentration, whereas body was negatively correlated with IBMP. Notably, the wines which had excessive unripe berries in the fermentation were more strongly associated with leafy aroma, acid taste, and lower body than the petiole-addition wines. Interestingly, the wines with higher perceived body also contained significantly higher concentrations of terpenes.
Conclusions
The impact of MOG on the sensory properties and chemical composition of wines is an understudied area of research. Using small-scale wine fermentations, we have shown here that inclusion of petioles can alter the sensory and chemical profiles of wines, but addition of up to 5% unripe berries had little impact on the sensory qualities of the wines. For petioles, the threshold amount required to produce significant sensory and chemical changes appeared to be above 5% of petiole material in the must fermentation. From a viticultural standpoint, petiole terpene concentrations may be manipulated with various vineyard practices. Furthermore, petiole additions offer a way to make wine with attributes that will suit some wine consumer segments. Nevertheless, the results in this study suggest that if unripe berries or petiole levels in a must are kept below 5%, the resulting wine should not be markedly sensorially different from a wine produced without these potential contaminants.
Because this was a preliminary trial, it would be worthwhile to conduct a study with a larger number of samples from varying sites, with different grape varieties, from regions using different yeast strains, and with a full-factorial design across a minimum of two to three seasons. Further work is also needed to clearly identify the MOG-related molecules affecting wine aroma via wine fractionation associated with sensory analysis, reconstitution experiments, GC olfactometry, and other avenues of research, such as crossmodal interactions.
Acknowledgments
Acknowledgments: This research was supported by Treasury Wine Estates, the University of Adelaide, and the Commonwealth Scientific and Industrial Research Organization; the latter two organizations being members of the Wine Innovation Cluster, Waite Campus. The authors appreciate the assistance of Waite Analytical Services for elemental analysis, Emily Nicholson for GCMS analyses, and Samuel Brook and vineyard staff in Coonawarra for harvesting the green and ripe fruit, respectively. The authors thank those students at the University of Adelaide who donated their time to participate in the DA panel.
- Received February 2015.
- Revision received June 2015.
- Accepted June 2015.
- Published online October 2015
- ©2015 by the American Society for Enology and Viticulture
Literature Cited
Vol 66 Issue 4
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
