Abstract
Effects of light exposure and vine microclimate on C13-norisoprenoid concentration in Cabernet Sauvignon grapes and wines were investigated by measuring the amounts of β-damascenone (megastigma-3,5,8-trien-7-one), TDN (1,1,6-trimethyl-1,2-dihydronaphthalene), and vitispirane (6,9-epoxy-3,5(13)-megastigmadiene) using gas chromatography-mass spectrometry. In grapes and wines, the most exposed treatment (all lateral and primary leaves removed) had the highest light intensity and temperature and showed the highest concentrations of TDN and vitispirane (p < 0.05). However, in the more shaded treatments, concentrations of all norisoprenoids were variable and dependent on the treatment conditions. When leaves were removed, C13-norisoprenoid concentrations were linearly (r > 0.90; p < 0.1) and positively correlated with increasing sunlight exposure. In contrast, in the most shaded treatments with no leaf removal there were high concentrations of norisoprenoids. β-Damascenone concentrations in particular were highest when no leaves were removed. Grapes and corresponding wines from the south side of the vine had higher levels of measured norisoprenoids than those from the north side. However, the leaf layer number was also correlated (r > 0.89; p < 0.1) with norisoprenoid concentration. Norisoprenoid concentrations in grapes were significantly correlated (p = 0.025) with those in wines. These results suggest that in addition to sunlight exposure, leaf removal and cluster microclimate can significantly alter norisoprenoid concentrations in grapes and wines.
C13-Norisoprenoids are important flavor constituents of many varieties of Vitis vinifera including Chardonnay, Cabernet Sauvignon, Syrah, Sauvignon blanc, Chenin blanc, Semillon, and white Riesling (Sefton et al. 1989, 1993, Abbott et al. 1990, Francis et al. 1992, Marais et al. 1992a,b). These compounds typically have low sensory thresholds and therefore can make important contributions to grape and wine aroma even at very low concentrations. For example, the norisoprenoid β-damascenone (mega-stigma-3,5,8-trien-7-one) is one of the most potent odorants known, with a reported aroma threshold in water of 9 ng/L and a flowery, slightly fruity character (Ohloff 1978). TDN (1,1,6-trimethyl-1,2-dihydronaphthalene) has a kerosene-like aroma and gives a characteristic bottle-aged aroma to many wines, particularly Rieslings, and has a reported aroma threshold of 20 μg/L (Belitz and Grosch 1992). Vitispirane (6,9-epoxy-3,5(13)-megastigmadiene) has a eucalpytus or camphoraceous odor with an aroma threshold in wine of 80 μg/L (Simpson 1978). Numerous other individual norisoprenoids have been identified that also contribute complex aromas to both red and white wines, including grassy, tea, lime, honey, oak, and pineapple characters (Francis et al. 1992, Sefton et al. 1994).
Norisoprenoids in grapes are thought to arise from photochemical and enzymatic degradation of carotenoids found in the skin and pulp. Their concentrations typically increase during grape ripening, and this increase, which starts at veraison, is accompanied by a decrease in carotenoid concentration (Razungles et al. 1988, 1993, Razungles and Bayonove 1996). The major carotenoids in grapes are β-carotene and lutein, present at concentrations of 1 to 2 mg/L (Marais et al. 1990, Razungles et al. 1988, 1998, Oliveira et al. 2004). Low levels (μg/L) of several other oxygenated carotenoids, including neoxanthin, flavoxanthin, 5,6-epoxylutein, luteoxanthin, and violaxanthin, have also been identified in grapes (Razungles et al. 1988, 1998, Mendes-Pinto et al. 2004, Oliveira et al. 2004). Each of these different carotenoid precursors yields different norisoprenoid products, which may influence grape and wine aroma and flavor.
Several viticultural parameters, including sunlight exposure, have been shown to influence norisoprenoid concentration (Marais et al. 1992a, 1999, Gerdes et al. 2002). Concentrations of TDN, vitispirane, hydroxy-TDN, ionols, actinidols, grasshopper ketone, and vomifoliol were significantly higher in sun-exposed Chenin blanc and white Riesling grapes compared with shaded grapes of the same varieties; however, β-damascenone concentrations did not appear to be influenced by sunlight levels (Marais et al. 1992a). In that study, naturally sun-exposed grapes received direct sunlight in the morning on the south side of rows and in the afternoon on the north side of rows. Shaded grapes were naturally shielded by the leaves from direct sunlight exposure, although actual differences in photosynthetically active radiation (PAR) between the two treatments were not measured. Another study also showed that sunlight exposures greater than 20% of full sun exposure, beginning at veraison, increased levels of two norisoprenoids, TDN and Riesling acetal, in white Riesling grapes (Gerdes et al. 2002). Interestingly, these authors also observed that leaf removal at veraison decreased TDN in the berries, independent of sunlight exposure.
In Syrah grapes sunlight exposure before veraison increased carotenoid concentrations, while sunlight exposure after veraison increased norisoprenoid concentrations (Razungles et al. 1998). Light intensity of the sun-exposed berries was about 10 times higher than that of the shaded berries, while differences in temperatures of up to 5°C were also observed between the two treatments (Razungles et al. 1998). Increases in berry temperature associated with sunlight exposure have been shown to have significant effects on grape berry growth and composition (i.e., pH, titratable acidity, anthocyanin, and polyphenol concentration) (Bergqvist et al. 2001). Therefore, in these studies, temperature effects cannot be completely excluded when considering the effects of sunlight exposure on grape berry composition.
Concentration of some norisoprenoids (vitispirane and TDN) were lower in white Riesling wines from relatively cool regions compared with wines from warmer regions, although sunlight exposure also differed substantially between the regions (Marais et al. 1992b). Microclimates within vineyard sites and within the grapevine canopy can be quite variable, making it difficult to directly evaluate climatic differences between regions. Other variables such as clonal variation and differences in rainfall and crop yield may influence production of secondary metabolites, including carotenoids and norisoprenoids. These factors highlight just a few of the difficulties encountered when attempting to study viticultural variables that may influence grape and wine composition.
These previous studies have indicated that while sunlight can influence norisoprenoid concentrations in both red and white varieties, the vine microclimate required to elicit these effects is unclear. Consequently, there is a need for an increased understanding of the effects of viticultural variables on flavor production in grapes and wines. The objective of the current study was to evaluate the effects of vine microclimate on norisoprenoid concentrations in Cabernet Sauvignon grapes and in the corresponding wines made from the grapes.
Materials and Methods
Grapes.
The experiment was conducted during the 2002 growing season using Cabernet Sauvignon vines grown in a commercial vineyard in Gonzales, California (viticultural region I; lat: 36.5N; long: -121.4W). The site consisted of Cabernet Sauvignon vines that were bilaterally cordon trained, spur-pruned, and trellised on vertical shoot-positioned systems. Cabernet Sauvignon vines were spaced 1.5 m apart in east-west oriented rows that were 2.4 m apart. Vines were selected for use immediately after berry set and divided into six vine plots. Overall crop yield for this vineyard was 15.7 ± 0.7 t/ha and was not different among experimental treatments because vines were adjusted to similar crop loads prior to initiation of experiments. Vines did not exceed two clusters per shoot.
Vine microclimate treatments.
Six canopy densities and cluster exposure treatments were established on the north and south sides of the vine rows by performing varying amounts of primary and lateral leaf removal following fruit set. All treatments were maintained by leaf manipulation, as needed throughout the season, to create and preserve the integrity of the light exposure: (1) clusters were tucked into the interior of the canopy; (2) control cluster exposure, no manipulation of leaves/clusters; (3) clusters were exposed by removing lateral leaves in the fruit zone; (4) clusters were exposed by removing primary leaves in the fruit zone; (5) clusters were exposed by removing lateral and primary leaves on every other node in the fruit zone; and (6) clusters were exposed by removing all lateral and primary leaves in the fruit zone.
Experimental vines were selected on the basis of uniformity of vigor. A randomized complete block method was used to disperse treatments, and each fruit exposure treatment was replicated six times. One representative vine was selected from each replication of treatment for collection of light and temperature data. Data collection was separated based on the side of the vine the fruit was located and the treatment.
Light intensity and leaf layer measurements.
To measure the light intensity, photosynthetically active radiation (PAR) incident to random grapes was determined at the following stages of fruit development: fruit set, veraison, and before harvest. Light intensity (PAR) at each cluster was evaluated three times throughout the day (900, 1300, and 1600 Pacific Daylight Savings Time [PDT]) using a handheld Ll-189 quantum sensor (LI-COR, Inc., Lincoln, NE). The sensor was placed at the top, center, and bottom of each cluster, parallel to cluster orientation, oriented perpendicular to the ground, facing the center of the vineyard, and along the edge of the cluster.
Leaf area was measured using a LI-COR area meter and leaf layer number determined by point quadrant measurements using a previously described method (Smart 1988b). The sharpened tip of a 1-m rod (3 mm diam) was positioned perpendicularly to the canopy surface at the height of the fruit zone. The rod was inserted into the canopy interior at an angle of 90° with respect to the canopy exterior, and the number of leaves, clusters, and gaps intercepted by the tip of the rod recorded. Since canopy treatments were separated by side, the leaves, clusters, and gaps were measured up to the middle of the cordon from each side and shoots were ignored. Twenty insertions per replicate were made at 5-cm intervals. Leaf layer number represents the mean number of leaf contacts per insertion, while the percentage of canopy gaps refers to the percentage of insertions in which no leaf contacts were made. Interior clusters and leaves represent those organs located within the canopy, beneath one or more leaf layers.
Temperature measurements.
Berry temperature was measured at the same fruit development stages and times as described above using a handheld Omega HH 23 temperature monitor with dual hypodermic thermocouples (Omega Engineering, Stamford, CT). Berry temperature was determined by inserting the probe into the berry center. A shielded probe was placed next to the berry to monitor ambient air temperature. Isolated clusters were selected for temperature measurements at three different positions within the canopy: clusters completely exposed, clusters covered by one leaf layer, and clusters located within the center of the canopy covered by two or more leaf layers. Three berries from the top, center, and bottom of each cluster were used for temperature measurement.
Harvest.
Grapes were harvested on the same date at 23 ± 1 Brix (Table 1⇓). Random samples of ~15 kg of berry clusters (~25 clusters) were collected separately from both the north and the south sides of the vine row. From each of these samples, ~1 kg of individual berries were randomly picked from the interior and exterior of the clusters and used for norisoprenoid analysis. These berries were stored on ice during transport to the lab. Within 24 hr after harvest, the berries were crushed by hand and pressed through cheesecloth to obtain the juice. The juice was stored frozen (<−20°C) until the isolation and analysis of C13-norisoprenoids. Winemaking was initiated on the remaining grapes, also within 24 hr of harvest.
Wines.
Wines were made from each treatment according to uniform winemaking practices. Twelve lots were made (six treatments by two sides of the vine), each lot consisting of 45 ± 4.5 kg hand-harvested fruit collected from the field replicates. For each lot, the fruit was refrigerated at 15°C upon arrival at the UC Davis research winery and held at that temperature overnight. The following morning, the fruit was sent through a Valley Foundry and Machine Works (Fresno, CA) crusher-destemmer into 64-L cylindrical plastic containers. Each fermentation contained and ~38 L of must. The must was dosed with 50 mg/L SO2 punched down to mix. Premier Cuvee (Universal Foods, Milwaukee, WI) was thoroughly dissolved according to the manufacturer’s suggested procedure and used as the inoculum for each fermentor at 10 g dry yeast/38 L must. Brix readings using hydrometers and manual punchdowns with a stainless-steel punch were completed twice per day on each fermentor. The wines were pressed off the skins with a stainless-steel hydraulic basket press when the fermentation reached 0 Brix as determined by hydrometer reading. After the initial press, the press cake was broken up by hand and pressed a second time (and combined with the wine from the first pressing). The pressed wine was subsequently inoculated with actively growing malolactic starter culture (MCW strain; Vinquiry, Healdsburg, CA) and malolactic fermentation was carried out at room temperature (23 ± 2°C). After completion of malolactic fermentation, as determined by paper chromatography (Zoecklein et al. 1995), wines were racked, sulfured (30 mg/ L), and cold-stabilized before bottling.
Juice/must composition was determined using standard procedures (Zoecklein et al. 1995) (Table 1⇑). Brix was determined by hydrometer, titratable acidity (TA) was determined by titration to pH 8.2 and reported as g tartaric acid/L, and pH was measured using an Orion 420A pH meter and electrode (ThermoElectron, Philadelphia, PA).
Isolation and analysis of C13-norisoprenoids.
All solvents, buffer salts, and internal standards were purchased from Sigma-Aldrich (St. Louis, MO) in the highest purity available. β-Damascenone was donated by Haarman and Reimer, Inc. (Holzminden, Germany) and TDN and vitispirane were donated by Peter Winterhalter (Technische Universität, Braunschweig, Germany). Free norisoprenoids were isolated from 150 mL juice and wine by liquid/liquid extraction (20 hr extraction) using pentane/diethyl ether (1:1, v:v) as the extracting solvent (Gerdes et al. 2002). Internal standard (2-octanol; 5 mg/L) was added to the juice and wine before extraction.
Using the juice and wine remaining from the liquid/liquid extraction, glycosidically bound volatiles were isolated with simultaneous distillation and extraction (Gerdes et al. 2002). Any pentane/ether and ethanol remaining in the juice was first removed by distillation under vacuum at room temperature with rotary evaporator. A citrate/phosphate buffer (pH 2.1) was added to the juice and 1-heptanol (5 mg/L) added as an internal standard. The solution was extracted with pentane/diethyl ether (1:1, v:v) by simultaneous distillation and extraction for 1 hr. After extractions were completed, the organic phases were dried over anhydrous sodium sulfate and concentrated to 1.5 mL on a Vigreux column. The concentrated extract was stored at <10°C until gas chromatography (GC) analysis. Immediately before GC analysis, the extract was further concentrated to 100 μL under a gentle nitrogen stream.
All extracts were analyzed by GC with mass spectrometric detection (GC-MSD) using a Hewlett-Packard 6890 GC with a 5972 MSD and ChemStation software (Agilent Technologies, Santa Clara, CA). Helium was the carrier gas at a flow rate of 0.8 mL/min and the MSD interface temperature was 260°C. A 30 m x 0.25 mm i.d. x 0.25 μm film thickness DB-Wax (polyethylene glycol) column (J&W Scientific, Folsom, CA) was used for all analyses. The oven temperature was programmed from an initial temperature of 50°C to a final temperature of 240°C at a rate of 4°C/min. The oven was held at the final temperature for 25 min. The injector temperature was 230°C, and 1 μL of concentrated extract was injected in the splitless mode.
Peak identity of individual peaks was confirmed by comparison of retention times and mass spectra with those of authentic standards injected under identical conditions. For quantitation, the GC-MSD was operated in selected ion monitoring mode. Quantitation of damascenone, TDN, and vitispirane was based on peak areas of the ions at m/z 121, 157, and 192, respectively. Concentration of each peak was calculated relative to the peak area of the internal standard, 2-octanol (free norisoprenoids) or 1-heptanol (bound norisoprenoids). No differences in trends among treatments were observed when norisoprenoid concentrations were determined on a per berry basis or as a function of concentration per volume of juice; therefore, all data are reported as mg/L juice. All extractions were duplicated and the average of the two values is reported.
Statistical analysis.
The effects of vine microclimate on norisoprenoid concentrations in grapes and wines were analyzed by one-way analysis of variance. Means comparisons were performed with Student Newman-Keuls test using SAS statistical software (version 8.02; SAS Institute, Cary, NC). Student t-tests and linear regression analyses were performed using Excel 98 (Microsoft, Redmond, WA).
Results
Light intensity and temperature of grapes.
The greatest differences in light exposures for the various treatments were observed at ~1600 hr on the south side of the vines (Figure 1⇓). Vine manipulations that did not remove leaves—control (treatment 2) and clusters tucked into the canopy (treatment 1)—allowed the least amount of light to reach the berry clusters. When leaves were removed (treatments 3, 4, 5, 6), the total light intensity for exposed clusters increased as leaf layer number in the fruit zone decreased on both the north and south sides of vines (Figure 1⇓). In addition, as leaf layer number decreased in the fruit zone, the percentage of total clusters located on the canopy exterior increased (data not shown). Overall, clusters on the south side of the vine received more direct sunlight exposure than those on the north side.
Sunlight exposure generally plays a pivotal role in influencing grape berry temperatures in the canopy (Millar 1972, Bergqvist et al. 2001) and berry temperatures typically parallel diurnal solar radiation (Spayd et al. 2002). The effects of vine microclimate on the temperature of Cabernet Sauvignon grapes in this study are shown in Figure 2⇓. The highest berry temperatures were generally observed when all lateral and primary leaves were removed (treatment 6), which also corresponded to the treatment with the highest sunlight exposure. The maximum temperature on both sides of the vine was reached at ~1600 PDT. The vineyard site is classified as region I, and coastal influences minimize high temperatures in late afternoons.
Light exposure and berry temperature were correlated in some cases, and the strongest, positive correlations between light and temperature were observed during the late afternoon, especially for grapes on the south side of the vine (Table 2⇓). For grape clusters on the south side, light exposure in the afternoon was 2 to 3 times greater when leaves were removed and temperatures increased by ~9°C, compared with the control. Except for the morning time period (0900 PDT), clusters on the south side of the vine had higher average temperatures than those on the north (Figure 2⇑).
C13-Norisoprenoid concentrations.
In the grapes and wines, concentrations of bound norisoprenoids were always higher than free norisoprenoids (data not shown). That is consistent with literature reports indicating that norisoprenoids are generally present as glycoconjugates or polyol precursors, which are easily degraded under acidic conditions (Sefton et al. 1989, Roscher and Winterhalter 1993, Winterhalter and Schreier 1994, Winterhalter and Skouroumounis 1997). For comparison purposes, all data are reported here as the sum of free and bound (i.e., total) norisoprenoids. Total norisoprenoids ranged from ~0.05 to 0.35 μg/mL, consistent with other observed ranges (Strauss et al. 1987). Norisoprenoid concentrations were always 2 to 3 times higher in grapes than in wines (Figure 3⇓); however, there was a significant, positive correlation (p = 0.025) between concentrations in the grapes and in the corresponding wines.
North vs south exposure.
With few exceptions, grapes and corresponding wines from the south side of the vine had higher levels of measured norisoprenoids (sum of damascenone, TDN, and vitispirane) than those from the north side (Figure 3⇑). Exposed clusters on the south side of the vine (treatments 4, 5, 6) also received more direct sunlight exposure (and were exposed to higher temperatures) than the north side, as discussed previously.
Vine microclimate.
Although the concentrations of individual norioprenoids were typically higher in samples from the south side of the vine as discussed above, similar trends among treatments were observed regardless of whether samples were from the north or south side. For both grapes and wines, the most exposed berry clusters (all lateral and primary leaves removed, treatment 6) had the highest concentrations of TDN and vitispirane (p < 0.05); however, norisoprenoid concentrations were variable in the more shaded treatments, depending on whether or not leaves were removed to alter the cluster microclimate (Figure 4⇓). Generally, as leaves were removed and fruit was more exposed to sunlight (treatments 3, 4, 5, 6), a significant positive, linear correlation (r > 0.90, p < 0.1) between light exposure and C13-norisoprenoids levels was observed (Table 3⇓). However, significantly higher concentrations (p < 0.05) of β-damascenone and vitispirane were also found in the most shaded treatments with no leaf removal (treatments 1 and 2), resulting in poor correlation between light exposure and norisoprenoid concentration when all treatments were evaluated together (Table 3⇓). Interestingly, damascenone was generally highest in treatments with no leaf removal (treatments 1 and 2), even though sunlight exposures (and temperature) were lowest for these samples. These results indicate that leaf removal and other microclimate effects may influence norisoprenoid concentrations independent of sunlight exposure.
Individual norisoprenoid concentrations for treatments 3, 4, 5, and 6 (treatments where leaves were removed) were also correlated with temperatures; as leaves were removed and temperatures increased, a positive correlation between temperature and C13-norisoprenoid concentrations was observed (Table 3⇑). Most correlations were weaker than those observed with light exposure, particularly on the north side of the vine where temperature differences among treatments were not great (<3°C) and maximum temperatures were ~9°C lower than on the south side.
Although increasing light exposure was associated with increasing norisoprenoid concentrations when leaves were removed (treatments 3, 4, 5, 6), the lower light (and temperature) exposures associated with the north side of the vine still resulted in relatively high norisoprenoid concentrations overall (Figure 5⇓), indicating that light (or temperature) alone cannot account for the observed effects. When leaf layer number was considered, a significant negative, linear correlation (p < 0.1) was observed and norisoprenoids decreased as leaf layer number increased (Figure 6⇓). A slightly greater slope (change in norisoprenoid concentration as a function of leaf layer number) was observed for treatments on the south side of the vine compared with the north. Norisoprenoid concentration in treatments with no leaves removed (treatments 1 and 2) were not correlated with leaf layer number (Figure 6⇓). For simplicity, only results for grapes are shown, although there were similar results for wines.
Discussion
Many viticultural practices have been developed and applied in order to manipulate the grapevine canopy and to improve fruit exposure to sunlight (Reynolds et al. 1986, Dokoozlian and Kliewer 1995, Hunter et. al. 1995, Bureau et al. 2000). While effects of these canopy manipulations on basic fruit parameters such as Brix, pH, and TA have frequently been measured, few studies have evaluated the effects of canopy microclimate on fruit and wine flavor composition. In this study, the canopy of Cabernet Sauvignon vines was manipulated by leaf removal, and total light intensity of grape clusters increased as the leaf layer number in the fruiting zone decreased. Temperature also increased as light intensity increased, so it is difficult to separate fully the effects of sunlight and temperature on observed changes in berry composition. Higher berry temperatures generally increase fruit transpiration rates and berry dehydration, while previous studies have indicated that some metabolites (e.g., anthocyanins) increase with sunlight exposure, independent of berry temperatures (Kliewer 1970, 1977, Kliewer and Torres 1972, Bergqvist et al. 2001, Spayd et al. 2002). Although we observed significant effects on light exposure and temperature as a result of the canopy manipulations in this study, other important microclimate effects may have also occurred but were not measured (e.g., differences in moisture/relative humidity and in foliage exposure to sunlight). Many other factors also contribute to the amount of sun exposure clusters receive as well as to cluster temperature, such as climate, row orientation and spacing, topography, time of day, time of year, altitude, cultivar selection, rootstock selection, trellising, and wind speed (Kliewer 1970, Crippen and Morrison 1986, Dokoozlian 1990, Morrison and Noble 1990, Bergqvist et al. 2001). Norisoprenoid concentrations were higher in grapes than in wines, for reasons that remain unclear, although the concentration in grapes was positively correlated (p = 0.025) with that in wines. Similar relationships between norisoprenoid concentrations in grapes and wines have been observed (Marais et al. 1992a).
Consistent with other studies (Marais et al. 1992a, 1999, Gerdes et al. 2002), we observed increasing levels of norisoprenoid concentrations in grapes (and wines made from the grapes) as sunlight exposure in the fruiting zone increased. However, the method by which sunlight exposure was altered also had a significant effect. When no leaves were removed from the vine, high concentrations of some norisoprenoids were observed even though light intensity at the exposed clusters was low. The highest concentrations of β-damascenone were in the control samples and when clusters were tucked into the canopy (treatment 1). Similar effects have been observed with white Riesling grapes (Gerdes et al. 2002). Reasons for these microclimate effects are unclear. It is possible that a critical variable other than light or temperature of the cluster was affected, resulting in alterations in norisoprenoid synthesis or degradation. Foliage exposure to sunlight may regulate fruit composition (e.g., pH) (Bergqvist et al. 2001), so it is possible that differences in sunlight exposure to the vine foliage among the different treatments may have occurred, resulting in either a direct or indirect effect on norisoprenoid concentrations.
Our results are consistent with a study in Chenin blanc and white Riesling grapes that indicated that sunlight exposure affected β-damascenone levels differently compared to other norisoprenoids (Marais et al. 1992a). Reasons for differences in the responses of individual norisoprenoids are unclear. Norisoprenoids are formed from carotenoid precursors and different carotenoids will yield different norisoprenoid products (Figure 7⇓). Sunlight has been shown to influence carotenoid formation and composition in grapes (Marais et al. 1991, Bureau et al. 1998, Razungles et al. 1998, Oliveira et al. 2004); therefore, our experimental treatments may have resulted in changes in individual carotenoid precursor concentrations. In addition, observed differences in responses of the individual norisoprenoids may be related to the pathways involved in their formation. Synthesis of both β-damascenone and vitispirane may require an enzymatic reduction of a ketone precursor, while TDN can form directly from acid catalyzed rearrangement of a ketone precursor (Winterhalter 1994) (Figure 7⇓). TDN contributes an important bottle-aged character to aged wines, and high levels result in a strong kerosene-like off aroma. These results indicate it may be possible to identify viticultural practices that enhance formation of β-damascenone and its corresponding floral, fruity character while minimizing TDN off-flavor formation. However, sensory studies are necessary to fully characterize the effects of these microclimates on final wine flavor. In addition, further studies evaluating the effects of microclimate on concentrations of the carotenoids and other precursors may help identify factors that influence the formation of norisoprenoids.
Differences in norisoprenoid concentrations on the north and south sides of the vine were not as great as expected given the differences in sunlight exposure (Figure 5⇑). When norisoprenoid concentrations on both sides were plotted together as a function of sunlight exposure, sunlight level alone was not correlated with norisoprenoid concentrations (similar results were obtained for temperature). In this study, the treatments were manipulated so that leaf layer numbers on the north and south sides of the vines were similar; when norisoprenoid concentrations were plotted as a function of leaf layer number (Figure 6⇑) a strong correlation (r > 0.89; p < 0.10) was observed. The change in norisoprenoid concentration as a function of leaf layer number was slightly greater for grapes on the south side compared to the north side, indicating that light (and temperature) does play an important role. However, leaf layer number appeared to be the dominant effect on norisoprenoid concentration, at least when leaves were removed. Interestingly, similar results were observed for other secondary metabolites in these samples, including anthocyanins and total polyphenols (J. Cotta, personal observation). For treatments where no leaves were removed (1 and 2), norisoprenoid concentrations appeared to be independent of those observed when leaves were trimmed (treatments 3, 4, 5, 6), as discussed previously.
Reasons for the effect of leaf layer number on observed norisoprenoid concentrations are not clear. The percent change in norisoprenoids did not appear to be associated with the relative percent change in light exposure (e.g., an 8 to 10% increase in light at 1600 PDT on the north of the vine resulted in an increase in β-damascenone concentration of ~120%, whereas a similar increase in light on the south side resulted in an increase of ~30 %) so there does not appear to be a simple attenuation of response associated with relative (as opposed to absolute) changes in light exposure. However, the percentage of exposed clusters increased at the surface of the canopy as leaf layer number decreased. Therefore, leaf layer number may provide a better overall integration of the effects of light exposure and temperature for both exposed and nonexposed clusters compared with either light or temperature measurements alone. In addition, other effects to foliage and/or light quality that were not measured in this study may also affect the observed responses (Smart et al. 1988a, Dokoozlian 1990, Schultz et al. 1998, Shabala and Wilson 2001, Núñez-Olivera et al. 2006, Reynolds et al. 2007).
Conclusions
While increased sunlight exposure to grape clusters generally increased the concentration of norisoprenoids in grapes and wines, removing the leaves decreased the concentration, even when cluster sunlight exposures were similar to those of vines with no leaf removal. Leaf layer number also strongly influenced norisoprenoid concentrations, independent of sunlight exposure on the north and the south sides of the vine. Individual norisoprenoids responded differentially to microclimate effects, possibly as a result of differences in chemical and enzymatic mechanisms involved in their synthesis. In this study we evaluated microclimate effects on concentrations of only one class of flavor compounds; however, other flavors (e.g., 2-methoxy-3-isobutylpyrazine) and secondary metabolites are also impacted by sunlight and microclimate in complex ways. Therefore, the overall impact of these microclimate effects on sensory properties of the final wines is unknown. Additional studies are needed to fully understand how canopy manipulations alter light, temperature, and other microclimate variables in order to influence grape and wine composition and sensory properties.
Footnotes
Acknowledgments: Financial support of the American Vineyard Foundation and USDA/Viticulture Consortium for NKD and JPC is gratefully acknowledged.
The authors thank McFarland Vineyards for the grapes, Marcy Webb and Jim Duane for making the wines, Doug Adams for helpful discussions, and Haarman and Reimer, Inc. and Peter Winterhalter for donation of norisoprenoid standards.
- Received May 2005.
- Revision received April 2007.
- Copyright © 2007 by the American Society for Enology and Viticulture