Abstract
Vitis vinifera cv. Carmenere, a member of the Cabernet family, was recently rediscovered in Chile where it had been misidentified as Merlot. The Carmenere variety produces complex wines, marked with herbaceous and “green” aromas. Results indicate that methoxypyrazines are present in high concentrations in this variety, probably responsible for the strong vegetative character found in Carmenere wines. Wines from Carmenere grapes had high 3-isobutyl-2-methoxypyrazine (IBMP) concentrations (5.0 to 44.4 ng/L), much higher than most Cabernet Sauvignon wines. The genotype dramatically affected the methoxypyrazine content of Carmenere wines, with some clones containing three-fold more than others. Climatic conditions, rather than maturity or terroir, had a critical effect upon IBMP concentration in grapes. 3-Isopropyl-2-methoxypyrazine evolved differently during grape maturation and was less influenced by harvesting year.
Vitis vinifera cv. Carmenere (Grande Vidure), now virtually extinct in its French homeland, was once widely planted in the Bordeaux regions of Médoc and Graves. The vineyards planted with this variety disappeared in the mid-19th century after the phylloxera plague. This aphid destroyed many European vineyards, drastically decreasing the diversity of V. vinifera (This et al. 2006). In the Bordeaux region, after the introduction of several disease-resistant, indigenous American rootstocks, the devastated vineyards were replanted with Merlot and Cabernet Sauvignon varieties, but not with Carmenere, which was considered less productive and difficult to graft.
The Carmenere variety was imported from Europe to Chile around 1850, together with other Bordeaux grapes (Cabernet Sauvignon and Merlot), before the phylloxera plague. For many years this variety was misidentified as Merlot in Chile. In 1994, the French ampelographer Jean Michel Boursiquot discovered that much of Chile’s Merlot was Carmenere (Pszczolkowski 2004). A similar situation existed in Italy where, until 1991, the variety was confused with the so-called Italian Cabernet franc (Calo et al. 1991). The largest established vineyards of this variety (13,000 ha) are currently in Chile, where vines are directly planted without grafting (Pszczolkowski 2004).
The first Chilean Carmenere wines that appeared in the market exhibited harsh and green aromas, resulting in notes of green bell pepper, celery, and green beans. The green aromas resulted from the fact that Carmenere grapes require three to four weeks more to reach ripeness than Merlot; and if Carmenere grapes are grown and vinified like Merlot, for example, then these characteristics can be overpowering and unappealing.
These sensory attributes are assimilated in other varieties to the presence of methoxypyrazines, mainly 3-isobutyl-2-methoxypyrazine (IBMP) (Bayonove et al. 1975), which is responsible for the characteristic green, herbaceous, or vegetative aromas found in wines of Sauvignon blanc (Allen et al. 1991, Lacey et al. 1991), Cabernet franc, and Cabernet Sauvignon (Kotseridis et al. 1999a,b, Roujou de Boubée et al. 2000, 2002). It has a high olfactory impact, resulting from its very low olfactive detection threshold (2 ppt in water). The green bell pepper note is clearly perceived in red wines containing IBMP concentrations higher than 15 ppt (Roujou de Boubée et al. 2000, 2002). High IBMP concentrations can also mask fruity sensory attributes in a wine (Falçao et al. 2006). A second methoxypyrazine, 3-isopropyl-2-methoxypyrazine (IPMP) has also been identified in several grape varieties, conferring more earthy, asparagus, and vegetative aromas to wines (Allen et al. 1995, Chapman et al. 2004), and with a similar olfactory detection threshold (2 ppt).
Methoxypyrazine concentration decreases in a sustained manner during ripening, reaching a basal level before harvest (Allen and Lacey 1999, Sala et al. 2004). High methoxypyrazine concentration in grapes at harvest is generally associated with unripeness (Chapman et al. 2004, Hashizume and Samuta 1999, Kotseridis et al. 1999a), negatively impacting the final wine quality. Methoxypyrazine concentration in grape berries depends on climate, sun exposure, and vine vegetative growth and yield (Allen 2001, Bogart and Bisson 2006, Chapman et al. 2004, Hashizume and Samuta 1999, Prouteau et al. 2004, Roujou de Boubée 2000, Sala et al. 2004, 2005, Falçao et al. 2007). IBMP is particularly sensitive to light and temperature (Prouteau et al. 2004). The more the berries are exposed to sunlight, the lower the final pyrazine levels (Allen and Lacey 1999, Hashizume and Samuta 1999, Sala et al. 2004). Therefore, sunnier and dryer years result in lower IBMP concentrations in winegrapes (Roujou de Boubée et al. 2000).
Vegetal aromas are a recurrent aromatic attribute when describing Carmenere wines, suggesting that methoxypyrazines might be present in large quantities and be responsible for this particular aromatic note. The presence of significant amounts of IBMP in one Italian Carmenere wine has been reported (Calo et al. 1991). Results from our work demonstrate that both methoxypyrazines (IBMP and IPMP) were effectively present in Carmenere grapes and wines and are probably responsible for the vegetative descriptors of the variety. The influence of terroir, genotype, harvesting date, and year on methoxypyrazine concentration in Carmenere grapes and wines is also reported.
Materials and Methods
Analytical reagents.
The following reagents were used: NaCl and NaOH (Merck, Darmstadt, Germany), [2H3]-IBMP (Cambridge Isotope Laboratory, Andover, MA), IPMP (Interchim, Mountlucon, France), and IBMP (Sigma-Aldrich Chemical Co., St. Louis, MO). The solid-phase microextraction (SPME) manual holder and 65-μm polydimethylsiloxane divinylbenzene (PDMS-DVB) fibers were from Supelco (Bellefonte, PA). Water was purified with a MilliQ system from Millipore (Bedford, MA).
Grapes from commercial vineyards.
The experiments were conducted for three years (2003–2005) with Vitis vinifera L. cv. Carmenere grown in well-drained soils on vertical shoot-positioned systems in three valleys in Chile: Maipo Valley (lat. 33°40′, long. 71°15′; loam/sandy loam soil), Cachapoal Valley (lat. 34°20′, long. 71°05′; loam/clay loam soil), and Colchagua Valley (lat. 34°30′, long. 71°20′; sandy loam/clay loam soil). Grape clusters were sampled at four different dates, every two weeks, starting approximately 30 days after veraison (15 March). Technical maturity (M) was considered as the moment in which the grapes reached the maximum Brix content with a good sugar-acidity balance, rather than the actual harvesting date. Indeed, winemakers normally leave the Carmenere grapes longer for minimizing green aromas, which could result in overripe grapes. Three different rows of each parcel were randomly chosen, and 10 clusters were randomly collected from both expositions in each row, avoiding border effects. The grapes were immediately frozen and conserved at −20°C until analysis.
Experimental wines from selected clones.
Clonal selection of Carmenere vines was carried out as follows: during the 1999–2000 season, vines of Vitis vinifera cv. Carmenere were collected from the Central Valley of Chile. Plant identity and purity were determined by ampelographic and PCR-based molecular markers including microsatellite DNA markers (simple sequence repeats) (Hinrichsen et al. 2001). The clone stocks were planted directly in the soil in the COPEVAL S.A. Research Station located in San Fernando (Colchagua Valley, VI region, Chile) at 1.0 m × 2.0 m spacing and trained in unilateral Guyot. The vineyard system consisted of a randomized complete block design with three replications of each clone. Nutrition, irrigation, pest control, and other vineyard operations were consistent with accepted commercial vineyard practices.
Fourteen varietal wines from different Carmenere clones were obtained during the 2004 season. At maturity the grapes from each clone were manually harvested, destemmed, and crushed; 3g/hL SO2 was added. Fermentations were carried out in triplicates in 20-L microfermentation stainless-steel tanks using Saccharomyces cerevisiae yeast (20 g/hL) EC1118 (Lallemand, Montreal, Canada), at which point 0.2 g/L ammonium phosphate was added. Fermentation temperature and density were checked daily. Upon reaching d = 992 g/L, wines were separated from solid parts and spontaneous malolactic fermentation was achieved. Malic acid consumption was evaluated by paper chromatography. At the end of malolactic fermentation, 6 g/hL SO2 was added. After 3 weeks at 4°C, the free SO2 level was corrected to 25 mg/L and the wines were bottled and stored at 4°C until chemical analysis.
Commercial wines.
Varietal Carmenere (seven from the 2002 vintage and 23 from the 2003 vintage) and Cabernet Sauvignon (14 from 2003) wines were kindly donated by several Chilean wineries. All wines had completed malolactic fermentation; none were aged with oak treatments.
Sample preparation (SPME).
A 15-mL aliquot of wine and 10 mL of MilliQ water were transferred to a 40-mL amber vial equipped with a magnetic stirring bar and 7.5 g NaCl was added. For grape must samples, a 25-mL aliquot was used; no water was added. 55 ng [2H3]-IBMP (10 μL of a 55 μg/L solution in ethanol) and 500 μL NaOH 1N were then added. The vial was sealed with a screwcap with a Teflon-faced septum. Solution was equilibrated by stirring at 750 rpm for 5 min at room temperature. A SPME needle was then inserted through the septum and the fiber (PDMS-DVB) extended into the headspace, and allowed to equilibrate for 1 hour at room temperature. The fiber was then removed and immediately desorbed into a gas chromatograph. All solutions were maintained in amber flasks at −20°C.
Instrumental analysis.
The samples were analyzed with a Hewlett Packard 6890 gas chromatograph equipped with a HP 5973 mass detector (Agilent Technologies, Santa Clara, CA). Splitless injection (1 min) was carried out using a 0.75-mm SPME inlet at 260°C. The column was a HP-FFAP (60 m × 0.25 mm, 0.25-μm film). The carrier gas was helium with a nominal flow rate of 1.5 mL/min. The oven was kept at 70°C for 3 min; temperatures were then increased 3°C/min up to 115°C, 1°C/min up to 120°C, and finally 10°C/min up to 230°C, before holding for 10 min. Total running time was 44 min. Detection of trace quantities of IBMP and IPMP was carried out using selected ion monitoring (SIM). The mass selective detector transfer line was heated to 250°C. The electron impact energy was 70 eV. Mass selective source and quadrupole temperatures were set at 250 and 120°C, respectively. The following ions were used for quantification: m/z 124, 127, and 137 for IBMP, [2H3]-IBMP, and IPMP, respectively; while m/z 99, 56, and 77 were used as qualifier ions for IBMP, [2H3]-IBMP, and IPMP. All samples were analyzed in triplicate.
Method linearity, accuracy, and precision.
The SPME GC-MS method for methoxypyrazine quantification allowed for simultaneous analysis of IBMP and IPMP in the same injection, with similar recoveries and repeatability to other published methods (Chapman et al. 2004, Sala et al. 2002, Wampfler and Howell 2004). Calibration curves were carried out using increasing amounts of commercial IBMP and IPMP (10 to 100 ng/L), spiked in a model wine solution (3.5 g/L tartaric acid, 120 mL/L ethanol, pH 3.5).
Excellent linearity was obtained for both analytes: IBMP (y = 1.21 x, r2 = 0.997), and IPMP (y = 1.965 x, r2 = 0.998). The limit of detection (LOD) was estimated from the minimum concentration of IBMP and IPMP found in wines. On a basis of a signal to noise ratio of 3, LODIBMP = 1 ppt and LODIPMP = 0.7 ppt were calculated. Relative standard deviations of six replicates were 5.0% for IBMP and 9.8% for IPMP. Finally, for recovery studies, five replicates were carried out of the same wine, spiked with 9.8 ng/L IBMP and 9.9 ng/L IPMP. Excellent recovery yields were obtained: 95.4 ± 2.0% for IBMP and 91.4 ± 1.9% for IPMP.
Olfactometric identification.
The sensory impact of the compounds was confirmed by GC-olfactometric analysis with a HP 6890 gas chromatograph hyphenated to a FID detector and a Sniffer 9000 port (Brechbuhler Inc., Spring, TX), with the same column and conditions as above. Odor detection and description were also checked using pure commercial compounds (IBMP and IPMP, from Sigma-Aldrich and Interchim, respectively).
Climate data.
Data used in this study were obtained from meteorological stations situated close to the vineyards. Data were collected by the Chilean Meteorological Agency and consisted of daily observations of minimum and maximum temperatures and rainfall.
Statistical analysis.
Analyses of variance were performed using StatGraphics Plus version 4.0 (Manugistics, Rockville, MD). Significant differences were determined using Duncan’s test at p < 0.05.
Results and Discussion
Methoxypyrazines in Carmenere wines.
Olfactometric analysis (GC-O) was used to detect the olfactive zones and molecules involved in the vegetal character of Carmenere wine. Five odorant zones were found with vegetative descriptors of herbs, green beans, green peppers, cooked celery, and dry parsley. The two most intense odorant zones correlated with green bell pepper and green pea/asparagus aroma, respectively, and corresponded to the retention times of pure IBMP and IPMP compounds. These results show that the latter were responsible for the main vegetative characteristics of the wines. Further work needs to be done to determine the identity of the other three odorant zones, with lower olfactory impact.
Methoxypyrazine concentration was determined in 30 Chilean Carmenere wines from 2002 and 2003 (Table 1⇓). IBMP concentration ranged from 5.0 to 44.4 ng/L, far above the sensory detection threshold (2 ng/L) for this molecule in young red wines (Romero et al. 2006). Furthermore, 21 of these wines (70%) had IBMP concentrations above the olfactory recognition limit (15 ng/L); and six (20%) had IBMP concentrations over 30 ng/L.
When comparing IBMP concentration in Carmenere wines with other red winegrape varieties, the former had the highest mean IBMP content, 21 ppt, much higher than Chilean Cabernet Sauvignon for the same year (Table 2⇓). Similar concentrations have been reported, however, for some French Cabernet Sauvignon wines (Roujou de Boubée et al. 2000) and some New Zealand and Australian Cabernet wines (Allen et al. 1994).
Influence of vine genotype on methoxypyrazine.
IBMP concentration in wines from four 4-year-old Carmenere clones was studied during the 2004 season. Wide differences among the clones were determined, ranging from 45 to 161 ng/L IBMP (Table 3⇓), all exceeding the recognition olfactory threshold (Roujou de Boubée et al. 2000). IPMP concentration ranged from 0 to 8.6 ng/ L (Table 3⇓). Similar results on clonal heterogeneity of methoxypyrazine concentration of several Cabernet sauvignon clones were recently reported (Battistutta et al. 2000). These results strongly suggest that clonal selection of Carmenere and Cabernet Sauvignon vines for IBMP is critical in managing the vegetal character of resulting wines. The extremely high IBMP values obtained for the Carmenere clones in this study could have resulted from the climatic conditions observed in 2004 (see below).
Methoxypyrazines in Carmenere grapes.
The influence of season, climatic conditions, and harvesting date on methoxypyrazine concentration in Carmenere grapes focused on the three different valleys. Samples were taken every two weeks from mid-March until the beginning of May. The evolution of IBMP and IPMP was different for the three years under study (Table 4⇓). The highest IBMP concentration was in 2004, irrespective of the valley. The methoxypyrazine concentration significantly dropped after technical maturity (M) only in the Maipo Valley and for the 2004 season.
Results showed three evolution patterns of IBMP during the final six weeks of berry development. First, there was a sharp decrease in IBMP upon maturity. This behavior was found for the 2005 season and has been previously noted for Cabernet Sauvignon and Sauvignon blanc (Allen and Lacey 1999, Roujou de Boubée et al. 2000, Sala et al. 2005).
Second, IBMP levels remained stable upon reaching maturity. This pattern was observed in our Carmenere grapes in 2003 and in French Cabernet Sauvignon grapes (Kotseridis et al. 1999a). During this vintage, low IBMP concentrations were found in immature grapes.
Third, high initial IBMP concentrations consistently decreased, or remained constant, during the maturation process. This trend for methoxypyrazine evolution has not been reported until now. This behavior was observed during berry ripening for the 2004 vintage, with very high IBMP concentrations, which remained over the season, resulting in ripened grapes with very high methoxypyrazine concentrations.
These three patterns occurred in different seasons but were similar for the three different valleys studied, suggesting that climatic conditions, rather than terroir or maturity, are determinant of IBMP in Carmenere grapes, confirming the results of van Leeuwen et al. 2004. These authors recently studied the importance of climate, soil, and cultivar over berry composition and showed that the effect of climate was greatest on most parameters, followed by the effects of soil and cultivar.
Specific climatic conditions (Table 5⇓), such as monthly rainfall and average minimum and maximum temperatures (Figure 1⇓; Figure 2⇓), help to explain the main differences among the three seasons under study. The lowest methoxypyrazine concentrations were obtained at technical maturity (M) for the 2003 and 2005 vintages. The latter was marked by high summer temperatures, with a significant number of days over 30°C, and higher than in the 2004 season. For the Colchagua Valley, the 2003 season had fewer days over 30°C but had higher IBMP levels (86 ng/kg) two weeks before technical maturity. The year with lowest methoxypyrazine concentrations, 2003, was marked by heavier winter rains, the highest number of days below 0°C in winter, and the highest summer temperatures (January), contradicting previous results (Kotseridis et al. 1999b, Roujou de Boubée et al. 2000). The key point may not be the amount of rainfall, but the period in which the rain falls and its intensity in relation with berry developmental stage and plant status.
For the 2004 vintage, higher minimum and maximum temperatures were observed in winter (June) compared with the 2003 and 2005 vintages for all valleys (Figure 1⇑). A recent work reported differences in IBMP concentrations in Cabernet Sauvignon correlated with seasonal climatic data and altitude (Falçao et al. 2007) and also found a reasonable relationship between bell pepper aroma and lower winter temperatures. We observed the contrary, with the highest IBMP concentrations in 2004 (Figure 1⇑).
The unusual rains in November 2003 and March–April 2004 may have contributed to the high methoxypyrazine concentration that season (Figure 2⇑). Roujou de Boubée (2000) reported that IBMP increase correlated with summer rains. One possible explanation for the elevated methoxypyrazine for 2004 is that rainfalls occurred when Carmenere grapevines were close to the end of their growth cycle, resulting in a second vegetative growth, which promoted de novo IBMP biosynthesis. Methoxypyrazine could have been transported by the phloem to berries (Bogart and Bisson 2006). Alternatively, the vegetative growth increased the fruit shading, diminishing methoxypyrazine photodegradation. Indeed, highly irrigated vines generally lead to high IBMP in wines (Sala et al. 2005). Moreover, the 2004 production yield of Carmenere grapes was particularly low, as a consequence of the high incidence of coulure, a blossom disorder resulting in asynchronous ripening caused by weather alterations; uneven ripening has been suggested as a prevailing condition that promotes high IBMP concentration (Zoecklein 2006).
The 2005 season was mixed, with high initial IBMP, like 2004, but with a significant decrease at maturity, reaching ~30 ng/kg in the three locations under study. Climatic conditions were also different (Table 5⇑). Winter rains were similar to 2004, with lower minimum winter temperatures and days over 30°C in summer, approaching those during the 2003 season. There were also some rains in November (i.e., before grape setting) for the 2004 and 2005 vintages, which could explain the high initial IBMP concentrations.
It has been suggested that IBMP concentration in ripe fruit is mostly related to the prevailing weather conditions and that its berry decrease is not photo- but temperature-dependent (thermo-degradation) (Zoecklein 2006). The latter suggests that IBMP decrease could be faster during warmer nighttime temperatures, as it occurs with malic acid. This could be related to the higher proportion of days over 30°C for the vintages where the lowest methoxypyrazine concentrations were obtained (2003 and 2005).
IBMP is formed early in fruit and breaks down following veraison. Some researchers have suggested that lower temperatures in the period previous to veraison more significantly influence IBMP concentration in grapes than after maturation (Lacey et al. 1991). In this work, samples were analyzed from the final stages of berry development. Future work will focus also on the early stages of grape berry formation, that is, after setting and before véraison, to determine how climatic conditions affect IBMP biosynthesis and degradation.
Finally, while some authors have observed that lower Brix at harvest resulted in higher IBMP (Chapman et al. 2004, Hazhimuda and Umeda 1996, Kotseridis et al. 1999a), we could not reproduce this observation. For example, in 2004, similar Brix levels were measured in grape berries than for the other vintages, which accumulated much higher IBMP concentrations (Table 4⇑). Reaching maturity is important, and it pertains not only to acceptable Brix and AT but also to the absence of marked vegetal characters (enological ripeness).
IPMP showed a different behavior than IBMP. Other researchers indicated that IPMP evolved differently during grape ripening, suggesting that its turnover (biosynthesis and degradation) is different than that of IBMP (Sala et al. 2004). IPMP concentration seemed to be affected by both climatic factors and location (terroir). Indeed, in 2005, the lowest IPMP was found in grapes from the three valleys. The highest IPMP was found in the 2003 season for Maipo and Colchagua Valleys and in 2004 for Cachapoal Valley. This compound may have influenced the final aroma of wine in 2003 and 2004 vintages, because the concentrations found were above the detection threshold for IPMP in red wine (2 ppt) (Sala et al. 2004).
Conclusions
Carmenere grape berries were found to contain very high methoxypyrazine concentrations that could explain the strong vegetative character found in these particular wines. Wines from Carmenere grapes had high concentrations of IBMP, much higher than most Cabernet Sauvignon wines. The genotype dramatically affected the methoxypyrazine of Carmenere wines, with some clones containing three-fold more methoxypyrazine than others. Climatic conditions, rather than maturity or terroir, might have a critical effect upon IBMP concentration in grapes. IPMP evolved differently during grape maturation and was less influenced by harvest year.
Footnotes
Acknowledgments: The authors are grateful to Dr. Jorge Perez for providing the Carmenere wines obtained from different clones. The authors also thank Juan Pablo Maldonado and Lenka Torres for their skillful technical assistance. This work was supported by FONDECYT grant 1030484.
- Received August 2006.
- Revision received May 2007.
- Copyright © 2007 by the American Society for Enology and Viticulture