Methoxypyrazines in Sauvignon blanc Grapes and Wines

Transcription

1 Methoxypyrazines in Sauvignon blanc Grapes and Wines MICHAEL J. LACEY 1, MALCOLM S. ALLEN 2", ROGER L. N. HARRIS 3"', and W. VANCE BROWN 4 Three methoxypyrazines were identified and quantified by gas chromatography/mass spectrometry in wine and juice samples of Vitis vinifera L. cv. Sauvignon blanc. Twenty-two wines of Australian, New Zealand, and French origin were analyzed together with 16 juice samples from four Australian regions. 2-Methoxy-3-(2- methylpropyl)pyrazine was present in all wines ( ng/l) and all juice samples ( ng/l); it was invariably the major methoxypyrazine. 2-Methoxy-3-(1 -methylethyl)pyrazine was found in 11 wines ( ng/l) and almost all juice samples ( ng/l). In three wines and some juice samples, small quantities of 2-methoxy-3-(1 -methylpropyl)pyrazine were found (typically ng/l). Methoxypyrazine levels in the New Zealand wines were significantly higher than in the Australian wines (p < 0.001). Fruit grown under cool conditions gave higher grape methoxypyrazine levels than fruit grown under hot conditions. Grape methoxypyrazine levels were relatively high at veraison but decreased markedly with ripening. KEY WORDS" Sauvignon blanc, methoxypyrazine, grape, wine, aroma, analysis Wine flavor is crucial to wine quality, and grape varietal aroma is particularly important to wine flavor. Despite this, remarkably little is known about components responsible for grape varietal aroma in Vitis vinifera, except for the volatile monoterpenes that are important in "floral" varieties, particularly the Muscat varieties (13,16). Some factors that challenge investigation of grape aroma are the very low concentration of grape flavor components, the difficulty of relating odors of individual compounds to overall grape or wine aroma, and the chemical changes that may occur during grape processing, wine production, and wine aging (13,17,18). Wines from Cabernet Sauvignon and Sauvignon blanc grapes often have a characteristic aroma described as vegetative, herbaceous, grassy, or green. This aroma has been attributed to methoxypyrazine (MP) components. The occurrence of 2-methoxy-3-(2- methylpropyl)pyrazine (Fig. la, isobutyl MP, also known as 2-methoxy-3-isobutylpyrazine or 2-isobutyl-3- methoxypyrazine) has been reported in Cabernet Sauvignon grapes (3). Later, it was reported with other MP components in Sauvignon blanc grapes (2). The odor of isobutyl MP has been described as bell pepper-like. It contributes to the characteristic aroma of some vege- 'Senior Principal Research Scientist, CSIRO, Division of Entomology, GPO Box 1700, Canberra, ACT 2601, Australia; 2Lecturer in Wine Science, Ron Potter Centre, Charles Sturt University- Riverina, Wagga Wagga, NSW 2650, Australia; 3Principal Research Scientist, CSIRO, Division of Plant Industry, GPO Box1600, Canberra, ACT 2601, Australia; 4Senior Experimental Scientist, CSIRO, Division of Entomology, GPO Box 1700, Canberra, ACT 2601, Australia. Author to whom correspondence should be directed. 'Present address: Brindabella Hills Winery, Woodgrove Close, viahall, ACT 2618, Australia. The authors gratefully acknowledge the provision of wines and/or juice samples by Arrowfield Wines, Corbans Wines Ltd, Evans and Tate Vignerons and Winemakers, Hungerford Hill Wines Pty Ltd, Katnook Estate Pty Ltd, Lindemans Wines Pty Ltd, Montrose Winery, Selaks Wines, S. Smith & Son Pty Ltd, Valley Growers Cooperative Ltd, and the Committee of the Australian 1986 National Wine Show. This work was supported by a grant from the Australian Grape and Wine Research Council. This research was conducted at the Ron Potter Centre, Charles Sturt University-Riverina, Wagga Wagga, and at CSIRO, Division of Entomology, Canberra. The authors thank Mrs. H. McFadden for recording NMR spectral data. Manuscript submitted for publication 10 April Copyright 1991 by the American Society for Enology and Viticulture. All rights reserved. 103 tables (9,10,11), including bell peppers (4), and it has an extremely low sensory detection threshold in water of about 2 ng/l (2 parts per trillion) (4,14). Chemical determination of MP components has been frustrated by their low concentration in grapes and wine. Rigorous identification and quantification were not possible in the early reports of their detection (2,3), and they were not detected in a thorough investigation of a Cabernet Sauvignon wine (15). Furthermore, their odor detection threshold is more than two orders of magnitude below the detection limit of a quantitative method based upon HPLC techniques (7). Recently, a method was developed for rigorous identification and accurate quantification ofmp components (6). It allows investigation of their occurrence (8) and sensory role. This paper reports application of that method to a range of Sauvignon blanc wine and juice samples and to investigation of the influence of grape maturity and ripening temperature on grape MP concentrations. Materials and Methods Wine and juice samples: Commercially available Australian, New Zealand, and French Sauvignon blanc wines were used. Australian and New Zealand wines were less than three years old and represented a range of qualities (as assessed by performance in the 1985 and 1986 Australian National Wine Shows, Canberra). French wines were of the Sancerre/Pouilly-sur-Loire region and were four to ten years old. Juice samples were obtained from Vitis vinifera L. cv. Sauvignon blanc grapes collected either from the varietal collection at Charles Sturt University-Riverina, Wagga Wagga, or from vineyards in the Coonawarra, Great Western, and Adelaide Hills regions of Australia. Berries of fruit collected at Wagga Wagga were cooled to 4 C then passed through a juicer (National MJ65-N). The resulting juice was treated with sulfur dioxide (100 mg/l), centrifuged (3000 rpm, 10 min), the clarified juice treated with further SO 2 (100

2 104- METHOXYPYRAZINES IN SAUVIGNON BLANC mg/l) and stored at-20 C in the dark, under carbon dioxide. Samples for MP analysis (240 ml) were further stabilized by addition of sodium chloride (10% w/v) and sodium fluoride (0.25% w/v). Juice samples from other locations were obtained through each location's normal procedure for maturity assessment samples; they were stabilized with sulfur dioxide (250 mg/l), ascorbic acid (50 mg/l), and sorbic acid (500 mg/l), transported to Wagga Wagga (1-2 days), and stored at-20 C in the dark. Juice samples were warmed to redissolve potassium bitartrate prior to measurement of soluble solids, ph, titratable acidity, and individual acids and sugars. Soluble solids content was determined with an American Optical A hand-held refractometer after sample clarification by centrifugation (10000 rpm, 3 rain). Measurement of ph was carried out with a Radiometer PHM64 meter and Russell CMAW-TB combination electrode. Titratable acidity was determined with standardized sodium hydroxide solution, to a ph 8.3 end- point, after removal of dissolved carbon dioxide gas by gentle boiling (20 sec). Individual acids and sugars were analyzed by HPLC according to the method of Frayne (5) with a Waters Associates 510 pump, temperature controller, R401 differential refractometer, and Lambda-Max 481 LC spectrophotometer using peak height for quantification. Isolation of MP components: The method described previously (6) was modified as follows. Both 2- methoxy2h3-3-(2-methylpropyl)pyrazine (Fig. ld) (2.065 ng/lal in dichloromethane, 12.0 ~L) and 2-methoxy-2H3-3-(1-methylethyl)pyrazine (Fig. le) (1.503 ng/~l in dichloromethane, 12.0 lal) were added as internal standards to wine and juice samples (240 ml). Volatile material was isolated by distillation. The distillate was stirred with ion-exchange resin (200 mg) for one hour in the dark, and the filtered resin (glass sinter) washed with water (3 1 ml). A suspension of the washed resin in water (1 ml) was adjusted to ph 10 by dropwise addition of 10% to 15% KOH or NaOH solution. After five minutes, the supernatant solution was removed, the resin washed with water (0.5 ml), and the aqueous solutions combined. The resin was washed with dichloromethane (0.5 ml) and the aqueous solution extracted consecutively with the dichloromethane washings and a further portion of dichloromethane (0.5 ml). Combined dichloromethane extracts were sealed into glass ampoules and stored at -20 C in the dark until concentrated by controlled evaporation under nitrogen prior to assay by gas chromatography/mass spectrometry (GC/MS). During MP isolation, subdued illumination was used to minimize any photodecomposition (7). Deuterium-labeled MP standards: Synthesis of 2-methoxy-2H3-3-(2-methylpropyl)pyrazine (Fig. ld) has already been described (6). 2-Methoxy-2H3-3-(1- methylethyl)pyrazine (Fig. le) was prepared as described for 2-methoxy-~H3-3-(2-methylpropyl)pyrazine except that 2-chloro-3-(1-methylethyl)pyrazine (0.8 g) (11) was refluxed in dry tetrahydrofuran (10 ml) containing sodium methoxide-2h3 prepared from methanol- 2H. (Stohler Isotopic Chemicals, 99%, 1 ml) and sodium hy~tride-oil suspension (60%, 0.2 g). NMR spectrum (CDC13): d 1.24, 1.32, d, 6H, (CH3)2CH; 3.39, m, 1H, (CH3)2CH; 8.01, q, 2H, JHH 2.9 Hz, pyrazine ring protons. Absence of the methoxy signal in the NMR spectrum confirmed that the product was > 99% 2H 3- labeled. Determination of MP components: A strategy based on capillary GC/MS (6) was used to measure accurately the levels ofisobutyl MP (Fig. la), as well as 2-methoxy-3-(1-methylethyl)pyrazine (Fig. lb, isopropyl MP, also known as 2-methoxy-3-isopropylpyrazine or 2-isopropyl-3-methoxypyrazine) and 2-methoxy-3- (1-methylpropyl)pyrazine (Fig. lc, sec-butyl MP, also known as 2-methoxy-3-sec-butylpyrazine or 2-sec-butyl-3-methoxypyrazine). Detection limits were typically 0.15 ng/l. Results and Discussion MP analysis: The method developed for MP determination in wine (6) was equally applicable to grape juice analysis. Determination of low MP concentrations was more reliable for juice samples than wine samples because there were fewer components in the former. To improve the precision of GC/MS determination of isopropyl MP, a deuterium-labeled isopropyl MP, 2- methoxy-2h3-3-(1-methylethyl)pyrazine (Fig. le), was added as internal standard in addition to the deuterium-labeled isobutyl MP. Calibration curves using chemical ionization were linear (e.g., r = , 2HflHo v o molar ratm , isopropyl MP). Since sec-butyl MP and the deuterium-labeled isobutyl MP possess similar chemical properties and retention times, the latter was used as the internal standard for the former. The quantity of isolated MP components was usually sufficient to permit duplicate GC/MS analysis, allowing assessment of reproducibility. Duplicate GC/ MS analysis was performed on two wine and four juice samples, and the data (Table 1) show the differences between duplicates to have mean values of 0.53 ng/l (isobutyl MP) and 0.15 ng/l (isopropyl MP). Previous work has reported tentative detection of 2- ethyl-3-methoxypyrazine in Sauvignon blanc grapes i "y" OCX3 R a. CH2CH(CH3)2 H b. CH(CH3)2 H c. CH(CH3)CH2CH3 H d. CH2CH(CH3)~ D e. CH(CH3)2 D Fig. 1. MP components identified in Sauvignon blanc grapes and wine (X = H), and deuterium-labeled analogs used as internal standards (X = D = 2H). X

3 METHOXYPYRAZINES IN SAUVIGNON BLANC- 105 Table 1. Reproducibility of GC/MS analyses. Means and differences of duplicate determinations of methoxypyrazine (MP) concentrations (ng/l). Isobutyl MP Isopropyl MP 1 Mean Difference Mean Difference Wine Wine Juice Juice Juice Juice Co-eluting material prevented determination oi isopropyl MP levels in the two wine samples. (2), and it has also been claimed to occur in potato (12). However, this MP component cannot be biosynthesized readily from a common protein amino acid by the suggested route (10). 2-Ethyl-3-methoxypyrazine was not evident in the first few wine samples examined in this study, but a rigorous search with a deuteriumlabeled analog was not carried out. Its odor detection threshold in water has been reported as 425 ng/l (14); therefore, even if present, it seems unlikely that it would occur at a concentration sufficient to contribute to wine aroma. Biosynthesis of 2-methoxy-3-methylpyrazine by the suggested biosynthetic route (10) would be possible from alanine. However, analysis of a particularly MP-rich wine established an upper limit for 2-methoxy-3-methylpyrazine of 0.1 ng/l (6). Its odor detection threshold has been reported as 4000 ng/l (14); therefore, any contribution of it to wine aroma in Sauvignon blanc wines appears highly improbable. 4O 30 =m of_ L Isobutyl MP ~ Isopropyl MP I sec- Butyl MP Australian French New Zealand =ig. 2. Concentrations of MP components in Sauvignon blanc wines of ~,ustralian, French, and New Zealand origin. MP components in wines: The MP concentrations of 22 wines of Australian, New Zealand, and French origin are shown in Figure 2. Accurate determination of the isopropyl MP level was not possible with some wines because of co-eluting components. Only determinations free of this interference are shown in Figure 2. Isobutyl MP was identified in all wines. In all but two cases, its concentration ( ng/l) was above the reported odor detection threshold in water of 2 ng/l (4,14), indicating its potential to influence wine aroma. In 11 wines, isopropyl MP was also identified, although at much lower concentration ( ng/l). In three wines, sec-butyl MP was identified, but only in trace quantities ( ng/l). For the 11 wines in which isopropyl MP could be quantified, the ratio of isobutyl MP to isopropyl MP was fairly constant (mean 7.0, SD 1.9, range ) and showed less variation than the overall MP levels (range ng/l). A report of our preliminary findings (8) indicated that Sauvignon blanc wines of New Zealand origin had consistently higher MP levels than those from Australia. This agrees with a general perception that New Zealand Sauvignon blanc wines often possess a more intense vegetative aroma. The data of Figure 2 confirm this for a wider range of wines, with significant differences for both isobutyl MP and isopropyl MP (Table 2). Table 2. Comparison of Australian and New Zealand wine methoxypyrazine (MP) concentrations by the t-test. Mean wine MP concentrations (ng/l) Degrees of Australian New Zealand freedom t 1 isobutyl MP "'" Isopropyl MP "" 1--, ""indicate significance at p < 0.01 and p < 0.001, respectively. MP components in juice: The three MP components found in Sauvignon blanc wines were also identified in Sauvignon blanc juice samples. Grapes at Wagga Wagga were sampled from veraison to normal harvesting maturity in two consecutive years and compared with samples from three cooler areas of Australia. Determined MP concentrations are presented in Table 3 together with other measurements of grape maturity and the climatic averages for the four regions. In all samples, isobutyl MP was identified ( ng/l). Its concentration was invariably higher than that of isopropyl MP ( ng/l) which was identified in all but one sample. Only a few samples contained detectable amounts of sec-butyl MP. Climate is generally considered to be of paramount importance to grape flavor. In particular, cool ripening conditions can lead to enhanced levels of vegetative aroma in Sauvignon blanc grapes. The comparison of different regions in Table 3 supports the role of MP components in producing this effect, although the contribution of factors other than climate cannot be excluded. In the three cool areas (Adelaide Hills, Great Western, and Coonawarra), isobutyl MP was present at higher levels ( ng/l) than in the hot.area Am. J. Enoi. Vitic., Vol. 42, No. 2, 1991

4 106 m METHOXYPYRAZINES IN SAUVIGNON BLANC Table 3. Methoxypyrazine (MP) concentrations and maturity data of juice samples of four Australian locations. Location 1 Climatic Year Sampling MP components (ng/l) Brix averages 2 date 3 ph Titratable acidity 4 MJT HDD Wagga Wagga, NSW Adelaide Hills, SA Great Western, VIC Coonawarra, SA Isobutyl Isopropyl sec-butyl Jan Feb m Feb Mar ~ Jan ~ Jan Feb m Feb Mar Mar m Mar Apr Mar Mar Mar NSW = New South Wales, SA = South Australia, VlC = Victoria. 2 MJT = Mean January temperature ( C); HDD = Heat summation ( C), using 10 C base temperature, Oct. - Apr. 3 The dates 30 Jan and 8 Jan corresponded to mid-veraison at Wagga Wagga. 4 Titratable acidity as g/l tartaric acid equivalent. (Wagga Wagga) ( ng/l) at comparable stages of sugar accumulation ( Brix). In the hot area, isobutyl MP levels fell below the reported odor detection threshold in water, 2 ng/l (4,14), by normal harvesting maturity. Grapes at Wagga Wagga were sampled over a wide range of grape maturity. Isobutyl MP levels were high at veraison (> 30 ng/l) but decreased markedly with ripening to less than 4% of the veraison level by the time the grapes had attained 21.3 to 21.4 Brix (Table 3) r-] o O~ a. >,.- 20 :3 JO O 10 l l....1 I... I I J.. J ~ Days (from 1 Jan.) Brix GDD Fig. 3. Variation of isobutyl MP concentrations in two years with markedly different ripening temperatures as a function of time (days from 1 Jan.), soluble solids content ( Brix), and growing degree days (GDD).

5 METHOXYPYRAZINES IN SAUVIGNON BLANC ~ 107 l.,so u,.. I Malate Fructose i 5! O Tartrate : ]i I 30 4 / 20 o t~ 3 ~ o. ~ 1 t... t ] 8 Jan. 22 Jan. 1 Feb. 12 Feb. Fig. 4. Sauvignon blanc grape juice components, on a per berry basis, during ripening. Data for 1988, on fruit sampled at Wagga Wagga. The influence of ripening temperature is supported by comparison of fruit sampled from the same vines at Wagga Wagga in 1987 and 1988 (Table 3). Cooler ripening conditions prevailed in 1987, with grape maturity approximately four weeks later than in Mean January temperatures for 1987 and 1988 were 22.2 C and 27.0 C, respectively, and heat summations in degree days (calculated on a daily basis, October to April, base temperature 10 C) were 1835 and 2340 for years 1987 and 1988, respectively. In the climatic terms of Amerine and Winkler (1), this region IV area behaved like a region III and a region V, respectively, in these two years. Figure 3 shows that the isobutyl MP concentration difference between these two years, during the main period of MP component decrease, was less when the isobutyl MP levels were expressed in terms of growing degree days (GDD) than in terms of soluble solids content or calendar days. This is consistent with ripening temperature having an influence on levels of isobutyl MP, with high temperature decreasing the level ofisobutyl MP and having a greater effect upon the isobutyl MP level than on sugar accumulation, leading to a higher isobutyl MP concentration in the cooler year at comparable stages of sugar accumulation. To avoid the dilution effect of berry enlargement during ripening, the concentrations ofisobutyl MP and other substrates were also determined on a per berry basis in 1988 (Fig. 4). The per berry quantity ofisobutyl MP fell from 4.54 pg (4.54 X 10 -~ g) at veraison to 3.4% of this level in 35 days. Loss ofisobutyl MP in percentage.-, e- terms was greater than that ofmalate which fell to 6.9% of its veraison level. Comparison with glucose and fructose showed that isobutyl MP loss had largely occurred before sugar accumulation had attained half its final value. Conclusions Three MP components have been quantified in Sauvignon blanc wine and grape juice samples. A wide variation in their concentrations was found, but the relative proportion of the three components was fairly constant. The most abundant component was isobutyl MP, which occurred at concentrations of 0.6 to 38 ng/l in wines. Lesser quantities of isopropyl MP could be detected in many cases. Small quantities of sec-butyl MP were identified in a few samples. The concentration of these components in grape juice declined markedly as ripening progressed. The extent of this decline was dramatic even on a per berry basis: more than 96% of the veraison level ofisobutyl MP was lost by normal harvesting maturity. Isobutyl MP levels from juice samples of different sites, and of one site in two years of very different ripening temperatures, were consistent with ripening temperature having a strong influence. Higher MP levels were present under cooler ripening conditions. Literature Cited 1. Amerine, M. A., and A. J. Winkler. Composition and quality of musts and wines of California grapes. Hilgardia 15: (1944). 2. Augustyn, O. P. H., A. Rapp, and C. J. van Wyk. Some volatile aroma components of Vitis vinifera L. cv. Sauvignon blanc. S. Afr. J. Enoi. Vitic. 3:53-60 (1982). 3. Bayonove, C., R. Cordonnier, and P. Dubois. Etude d'une fraction caract~ristique de I'ar6me du raisin de la vari~t~ Cabernet-Sauvignon mise en evidence de la 2-m~thoxy-3-isobutylpyrazine. C.R. Acad. Sc. Paris., Ser. D 281:75-8 (1975). 4. Buttery, R. G., R. M. Seifert, D. G. Guadagni, and L. C. Ling. Characterization of some volatile constituents of bell peppers. J. Agric. Food Chem. 17: (1969). 5. Frayne, R. F. Direct analysis ofthe major organic components in grape must and wine using high performance liquid chromatography. Am. J. Enol. Vitic. 37:281-7 (1986). 6. Harris, R. L. N., M. J. Lacey, W. V. Brown, and M. S. Allen. Determination of 2-methoxy-3-alkylpyrazines in wine by gas chromatography/mass spectrometry. Vitis 26:201-7 (1987). 7. Heymann, H., A. C. Noble, and R. B. Boulton. Analysis of methoxypyrazines in wines. 1. Development of a quantitative procedure. J. Agric. Food Chem. 34: (1986). 8. Lacey, M. J., W. V. Brown, M. S. Allen, and R. L. N. Harris. Alkyl methoxypyrazines and Sauvignon blanc character. In: Proceedings of the Second International Cool Climate Viticulture and Oenology Symposium. R. E. Smart, R. J. Thornton, S. B. Rodriguez, and J. E. Young (Eds.). pp New Zealand Society for Viticulture and Oenology, Auckland (1988). 9. Maga, J. A., and C. E. Sizer. Pyrazines in foods. A review. J. Agric. Food Chem. 21:22-30 (1973). 10. Murray, K. E., J. Shipton, and F. B. Whitfield. 2-Methoxypyrazines and the flavour of green peas (Pisum sativum). Chem. Ind (1970). 11. Murray, K. E., and Whitfield, F. B. The occurrence of 3-aikyl-2- methoxypyrazines in raw vegetables. J. Sci. Food Agric. 26: (1975). 12. Nursten, H. E., and M. R. Sheen. Volatile flavour components of cooked potato. J. Sci. Food Agric. 25: (1974). 13. Rapp, A. Wine aroma substances from gas chromatographic analysis.

6 108 m METHOXYPYRAZINES IN SAUVIGNON BLANC In: Modern Methods of Plant Analysis, New Series Vol. 6. Wine Analysis. H. F. Linskens and J. F. Jackson (Eds.). pp Springer-Verlag, Berlin (1988). 14. Seifert, R. M., R. G. Buttery, D. G. Guadagni, D. R. Black, and J. G. Harris. Synthesis of some 2-methoxy-3-alkylpyrazines with strong bell pepper-like odors. J. Agric. Food Chem. 18:246-9 (1970). 15. Slingsby, R. W., R. E. Kepner, C. J. Muller, and A. D. Webb. Some volatile components of Vitis vinifera variety Cabernet Sauvignon wine. Am. J. Enol. Vitic. 31:360-3 (1980). 16. Strauss, C. R., B. Wilson, P. R. Gooley, and P. J. Williams. The role of monoterpenes in grape and wine flavor - a review. In: Biogeneration of aroma compounds. T. H. Parliment and R. B. Croteau (Eds.). pp ACS Symposium Series 317, American Chemical Society, Washington, DC (1986). 17. Strauss, C. R., P.J. Williams, B. Wilson, and E. Dimitriadis. Formation and identification of aroma compounds from non-volatile precursors in grapes and wine. in: Proceedings of the Alko Symposium on Flavour Research of Alcoholic Beverages. L. Nykanen and P. Lehtonen (Eds.). pp Foundation for Biotechnical and Industrial Fermentation Research 3 (1984). 18. Williams, P.J., C. R. Strauss, and B. Wilson. Developments in flavour research on premium varieties. In: Proceedings ofthe Second International Cool Climate Viticulture and Oenology Symposium. R. E. Smart, R. J. Thornton, S. B. Rodriguez and J. E. Young (Eds.). pp New Zealand Society for Viticulture and Oenology, Auckland (1988).