RESEARCH ON 2-METHOXY-3-ISOBUTYLPYRAZINE IN GRAPES AND WINES

Transcription

1 RESEARCH ON 2-METHOXY-3-ISOBUTYLPYRAZINE IN GRAPES AND WINES Dominique ROUJOU DE BOUBEE (School of Oenology, University of Bordeaux II) INTRODUCTION Pyrazines (1,4-diazines) are nitrogen-containing heterocyclic compounds that are quite widely distributed in nature in both the animal and plant kingdoms. The food industry is the area in which these compounds have been the most extensively studied. They are considered to be the heterocyclic compounds most widely represented in food aromas (Vernin and Vernin, 1982). They can be classified into three groups depending on their origins: those formed by heat treatment, those formed by micro-organisms and those present in the natural state in plants. Amongst the methoxypyrazines in the last category, the most important ones are 2- methoxy-3-isopropylpyrazine (IPMP), 2-methoxy-3-sec-butylpyrazine (s-bmp) and 2- methoxy-3-isobutylpyrazine (IBMP). IBMP was identified for the first time in green pepper (Capsicum annuum var. grossum) by Buttery et al. (1969). Its detection threshold in water is estimated to be 2 ng/l, and those authors consider it to be responsible for the characteristic aroma of green peppers. It was subsequently identified in several raw vegetables such as chili peppers (Capsicum frutescens), beans (Phaseolus vulgaris), broad beans (Vicia faba), lettuce (Lactuca sativa), spinach (Spinacea oleracea), etc. (Murray and Whitfield, 1975). Those authors note that one of the other methoxypyrazines is often dominant in some species. This is the case for IPMP in asparagus (Asparagus officinalis), peas (Pisum sativum), cucumber (Cucumis sativus), lettuce, potatoes (Solanum tuberosum) or even sow-thistle (Sonchus oleraceus) and for s-bmp in beets and carrots (Daucus carota sativa). In 1975, IBMP was identified for the first time in Cabernet Sauvignon grapes (Vitis vinifera L. cv. Cabernet Sauvignon) by Bayonove and Cordonnier, who claimed that it was responsible for the green pepper aroma that is characteristic of this variety. In 1982, Augustyn et al. identified IBMP in Sauvignon blanc grapes (Vitis vinifera L. cv. Sauvignon blanc). Over the last 10 to 15 years, its contribution to the vegetal and green pepper aromas of Cabernet Sauvignon, Merlot and Sauvignon blanc wines has been demonstrated (Harris et al., 1987; Maga, 1989; Allen et al., 1989; Allen et al., 1991; Allen et al., 1994; Kotseridis et al., 1998; Roujou de Boubée et al., 2000). These authors also show that s-bmp is rarely detected in wines, whereas IPMP is found at levels below its detection threshold in water (2 ng/l). As such, 2-methoxy-3-isobutylpyrazine would seem to be the key compound involved in the green pepper aroma of Cabernet Sauvignon, Sauvignon blanc and some Merlot wines. With these varietals, the IBMP concentration significantly exceeds the detection threshold. This compound is present in grapes, it has no known precursor and its concentration decreases during ripening (Allen et al., 1989; Lacey et al., 1991) under the influence of light (Heymann, 1986; Allen and Lacey, 1993; Hashizume and Samuta, 1999). Consequently, a high concentration in grapes at harvest is associated with a lack of ripeness and has a negative impact on wine aroma quality. Winemakers and oenologists commonly associate such aroma characteristics in grapes with low anthocyanin content and with mediocre tannin quality. It is therefore crucial to know what conditions influence IBMP concentration in grapes and wine. In this paper, we present the results of our research on the methoxypyrazine aroma of wines and on the conditions under which IBMP forms or breaks down in the grapevine. 1

2 1. The Methoxypyrazine Character of Wines and Its Evolution during Winemaking and Ageing 1.1. The Organoleptic Impact of 2-Methoxy-3-Isobutylpyrazine in Wines and Its Distribution in Wines of Different Varieties The quantification method that we developed (Roujou de Boubée et al., 2000) is based on that developed by Harris et al. in 1987 and then used by Allen et al. in By improving the ease and rapidity of extract preparation, we have been able to perform analyses with good repeatability on relatively large series of samples. In oenology, the value of interest is the threshold beyond which the green pepper aroma caused by IBMP becomes perceptible in wine. However, the determination of a detection threshold in wine does not yield an indicative value, owing to the great variations in composition from one wine to another and in descriptors used from one taster to another. As such, we sought to establish not a threshold value for a given wine, but rather a representative threshold for a given wine type, in our case a red Bordeaux-type wine. To do this, 50 red wines from Bordeaux and the Loire Valley (Cabernet Sauvignon, Cabernet franc and Merlot varieties from several different vintages for the Bordeaux wines; Cabernet franc from the 1991 and 1992 vintages for the Loire Valley) were evaluated by a panel of 10 persons at the Bordeaux School of Oenology. The wines were noted on a scale of 0 to 5 for the intensity of their green pepper aroma. The median of the scores for each wine was calculated and then correlated with the IBMP content determined by gas chromatography coupled with mass spectrometry (GC/MS). The relationship is linear between the median scores and the IBMP contents (r_ = 0.739, p>0.01%) (Figure 1). Wines with no green pepper aroma (median=0) contain 10 ng/l of IBMP on average. Those with only a slight green pepper aroma (median=1) have an average content of 15 ng/l, whereas beyond this level, the perception of the green pepper aroma is medium to strong. Median of tasting scores y = x R 2 = [IBMP] (ng/l) Figure 1: Correlation between the median of tasting scores obtained with 50 red wines from Bordeaux and the Loire Valley and the concentration of 2-methoxy-3-isobutylpyrazine as determined by GC/MS. We can therefore estimate that the detection threshold for the green pepper aroma, also known as the methoxypyrazine aroma, is 15 ng/l of IBMP in red Bordeaux and Loire Valley wines. In other words, this is the concentration at which the tasters identify a vegetal aroma in those wines. We then performed a statistical study on 96 red and white wines made from the Cabernet Sauvignon, Cabernet franc, Merlot and Sauvignon blanc varieties in order to determine those 2

3 in which IBMP plays a role in the vegetal aroma. This study shows that IBMP is the main contributor to the vegetal aroma in Cabernet Sauvignon, Cabernet franc and Sauvignon blanc wines. However, this compound is perceptible in only a minority of Merlot wines. Once the methoxypyrazine character had been defined, its threshold determined and the grape varieties identified, we focused on how the IBMP content changes during the winemaking process Extractability of 2-Methoxy-3-Isobutylpyrazine during Winemaking. Observations on the Evolution of IBMP Content in Wines during Bottle Ageing. Whole clusters of manually harvested Sauvignon blanc grapes were placed in a pneumatic press (Bücher, 70 hl). The juice samples obtained by simple crushing of the berries during press filling and at the beginning of the press cycle (0.2 bar; 30 min) had the highest IBMP contents (Table 1). Table 1: IBMP concentration (ng/l) of Sauvignon blanc musts sampled at different press levels during the 1998 harvest. Pressure Level Batch 1 Batch 2 Filling Free run 0.2 bar (30 min) Free run 0.2 bar (60 min) bar (90 min) bar (120 min) bar (140 min) bar (180 min) 7 10 The musts extracted subsequently contain less IBMP, and as the volume of liquid is lower, these fractions contribute less to the final IBMP content. The final IBMP concentration differs little from that of the first free-run juice obtained during press filling. IBMP is easily extracted from grape clusters during crushing and at the beginning of pressing, at least in the case of pneumatic pressing of whole clusters. Table 2 shows the effect of settling on the IBMP content of musts. Table 2 : Effect of settling on the IBMP concentration (ng/l) of Sauvignon blanc musts (1998). Batch 1 Batch 2 Before settling 9 13 After settling (200 NTU) 4 6 The clarified musts (200 NTU) contain about half as much methoxypyrazine as the unsettled musts. A part of the IBMP seems to interact with the grape solids and is thus eliminated by must clarification. It has previously been observed that settling limits the grassy aroma of white musts by lowering the contents of C6 aldehydes and alcohols (Dubourdieu et al., 1986) and of methionol (Lavigne, 1996). The effect of settling on IBMP content in musts has never been reported before. Under real conditions, the IBMP contents of Cabernet Sauvignon 24 hours after tanking-down and at the end of maceration are not very different (Figure 2). Most of the 3

4 IBMP in the free-run wine is therefore extracted into the aqueous phase prior to alcohol fermentation. 35 CS 97 CS1 98 CS IBMP (ng/l) Days after tanking down Figure 2: Evolution of the IBMP concentration during winemaking with Cabernet Sauvignon grapes in 1997 (CS 97) and 1998 (CS1 98 and CS2 98). To monitor the kinetics of IBMP extraction as precisely as possible with Cabernet Sauvignon grapes, a micro-batch of wine was made in the laboratory (7 kg of grapes in 15- litre stainless steel drums). 10 IBMP (ng/l) End AF Racking End MLF Time (h) Figure 3: Evolution of the IBMP concentration during micro-batch winemaking from Cabernet Sauvignon grapes in 1999 (CS 99). The extraction of IBMP from the grapes into the must is even quicker in this case (Figure 3). Within 24 hours, before the alcohol fermentation even begins, all of the IBMP found in the wine after racking has already been extracted from the grapes. The IBMP content is not increased by the successive punching-down operations performed during fermentation nor by post-fermentation maceration. Finally, the final concentration of the wine after racking does not seem to be influenced much by the frequency of pump-overs or by the skin contact time. However, as Kotséridis et al. (1999) observed, the IBMP content of the press wine can be greater than that of the free-run wine (Table 3). As such, IBMP certainly participates in the 4

5 increased vegetal character of many press wines. A certain fraction of the IBMP associated with the solid parts of the grapes can therefore be extracted during the mechanical operations of pressing. Table 3: Evolution of the IBMP concentration (ng/l) in Cabernet Sauvignon press wines (1998) during the press cycle. Batches 1 and 2. Batch 1 Batch 2 Start of pressing (0.2 bar) Pressing after 1 hr. (0.8 bar) Pressing after 2 hrs. (2 bar) Quite often, it can be observed that the thermovinification of red grapes (heating of the grapes to between 60 and 80 C for a short period to promote extraction of phenolic compounds and to destroy oxidases) leads to a decrease in the vegetal character of some wines. Of the 5 examples presented (Table 4), thermovinification systematically decreases the IBMP concentration to below the 15 ng/l threshold, and as such the vegetal character is no longer perceptible in those wines. This decreases varies from 29 to 67%, depending on the case. Table 4: Influence of thermovinification on the IBMP content (ng/l) of 5 Cabernet Sauvignon wines. Control Thermovinification Wine Wine Wine Wine Wine We thus sought to provide an analytical interpretation for an empirical observation. In the laboratory, a Cabernet Sauvignon must doped with IBMP was heated by means of a water bath to 60 C in a flask connected to a rotary vacuum evaporator in order to recover the volatilised fraction. It was shown that all of the IBMP that disappeared from the must was found in the evaporated fraction, which shows that this compound is volatilised during heating (boiling point of IBMP = 50 C). Thermovinification can be of interest in cases where the grapes have not reached optimal ripeness for different reasons (difficult weather conditions, unfavourable soil/exposure, excessive yield, etc.) or where the grapes are mouldy. This technique can lead to highly coloured, supple, fruity (ester-type) and less vegetal wines. The evolution of IBMP content for one Cabernet Sauvignon wine and one Sauvignon blanc wine during bottle ageing was also monitored. After three years of ageing in a dark cellar, no significant change was recorded. The chemical stability of IBMP explains this result. Thus, one should not count on time to diminish this olfactory defect in the bottle. The aroma of Sauvignon blanc or Muscat wines is determined by the grape ripening conditions and also, in large part, by the winemaking conditions (Peyrot des Gachons, 2000; Günata, 1984). For the Bordeaux varieties, we show here that obtaining fruity wines with no vegetal character depends mostly on the grape ripening conditions. This means that to get a low IBMP concentration in the wines, it is imperative to understand what happens prior to the harvest. What is the metabolism for IBMP in the grapevine? In what parts of the grapevine is 5

6 this compound found? How does its concentration change during the reproductive cycle and what factors affect it? 2. Evolution and Location of 2-Methoxy-3-Isobutylpyrazine in Various Grapevine Organs during the Reproductive Cycle 2.1. From Fruit Set to Veraison The determination of IBMP content in grape berries at an early stage (prior to veraison) enables us to observe a synthesis phase preceding the start of veraison (Figure 4), as previously reported by Hashizume and Samuta (1999). Mid-veraison [IBMP] (ng/l) Jul 17-Jul 22-Jul 27-Jul 01-Aug 06-Aug 11-Aug 16-Aug 21-Aug Figure 4 : Evolution of IBMP content in Cabernet Sauvignon grapes from berry touch to veraison (1999). Normally, the breakdown of IBMP is rapid initially and then it slows down as harvest approaches. However, in 1999, we observed in this vineyard block that the IBMP content increased in the berries between 31 July and 12 August (mid-veraison). This phenomenon, which had never been observed previously, is related to specific weather conditions. Between 27 July and 10 August, approximately 180 mm of rain fell on the block. The soil water reserves were thus restored, and the grapevines, which had been close to the end of their growth cycle, began to grow again. These vineyard observations do not enable us to draw any definitive conclusions. However, it would seem that IBMP synthesis is related to the vegetative growth of the grapevine. This could explain why vigorous vines that stop growing relatively late in the season produce grapes that generally have high IBMP levels. This result should, however, be compared with those we obtained in the following experiment set up by researchers at INRA Bordeaux (Tandonnet et al., 1996). A block of Cabernet Sauvignon vines was subjected to three soil-water statuses: normal (vintage conditions), irrigated (the vines received 4.6 mm of water/day from late June to late August) and dry (the soil was covered with a tarp from late June to harvest time). The grapes were vinified in small batches, and the IBMP concentrations were determined in the wines in 1994, 1995 and 1996 (Figure 5). We can note that irrigation leads to a significant increase in IBMP concentration in the wines 6

7 (+79% in 1994 and +39% in 1996 with respect to the normal wines). In 1996, tarping-over at soil level led to a 57% decrease in the wine IBMP levels with respect to the normal wines. Perhaps by inducing high vine vigour, irrigation leads to greater IBMP synthesis. [IBMP] (ng/l) Normal Irrigated Dry Figure 5 : Effect of different soil water statuses on IBMP concentration for Cabernet Sauvignon wines in 1994, 1995 and 1996 (normal: vintage conditions; irrigated: 4.6 mm of water per day from late June until veraison in late August; dry: tarping-over at soil level under the vines from late June until ripeness). Regardless of the cause, IBMP synthesis seems to occur between fruit set and two to three weeks prior to the onset of veraison. At this stage, the stems contain a large proportion of IBMP (Figure 6). Inside the berry, IBMP is found mainly in the skin (72%) and also in the seeds (23.8%). The pulp contains very little IBMP (4.2%). At this stage, IBMP is found in the berries, but we cannot determine if it is synthesised in situ. IBMP has in fact been identified in Cabernet Sauvignon leaves at the time of grape harvest (Hashizume et al., 1997). This finding suggests that IBMP could be synthesised in the leaves. We have therefore examined whether IBMP is present in the leaves upon berry touch. During this analysis, the clusters were grouped as a function of their insertion point on the shoot. On the primary shoot, we distinguished between leaves in the basal area (the first three to four leaves from the base), the leaves in the intermediate zone and the leaves in the apical zone where growth has not finished. 14.8% 0.9% 5.1% 79.2% Skins Pulp Seeds Stems Figure 6: Location of IBMP in different components of Cabernet Sauvignon grape clusters prior to veraison (4/08) in

8 The leaves on the secondary shoots (or summer laterals) were also divided into groups. Firstly, we show that the leaves contain IBMP at this stage (Figure 7). They are therefore capable of synthesising this compound. Secondly, we can note that the basal leaves have a very high IBMP content, i.e. much greater than in the other leaves or in the clusters Clusters Basal l. Middle l. Apical l. Lateral l. Figure 7: IBMP content in clusters and leaves (l.) at different levels of Cabernet Sauvignon shoots (30/07/1999). Knowing that many products synthesised in the leaves are then transported to the berries, we decided that it would be interesting to determine whether IBMP is transported from the leaves to the clusters. To do this, we used fruit-bearing cuttings of Sauvignon blanc obtained in accordance with the protocol described by Ollat et al. (1998a) and Gény et al. (1998). Twelve fruit-bearing cuttings of Sauvignon blanc (4 cuttings _ 3 repetitions A, B and C), each with one cluster, were selected between the small-pea stage and the beginning of veraison for their homogeneity (physiological stage, cluster size and compactness). On each of the cuttings, eight leaves were selected from all along the shoot. Each leaf was placed in a plastic container and then treated with a solution of 2( 2 H 3 )methoxy-3-isobutylpyrazine (1 mg/l), which is the deuteriated analogue of IBMP and is used as an internal standard for GC/MS. The deuteriated analogue solution of IBMP was deposited on the leaf every morning and evening for three days. On the fourth day, the clusters were harvested. For the leaves treated with the deuteriated IBMP solution, the leaf blade was removed (to avoid any contamination risk). Only the petiole was sampled. The leaves not treated with deuteriated IBMP (i.e. the young leaves near the apex) as well as the apices were sampled and gathered together. Finally, the clusters were picked. The distribution of deuteriated IBMP in the shoot was then measured (Figure 8). Firstly, we can note that the deuteriated IBMP is detected in the petioles, which shows that it penetrated the leaf blade and that it was transported. The great quantity of deuteriated methoxypyrazine found in the petioles for repetition C (80%) seems to indicate that the compound deposited on the leaf had not yet fully migrated towards the other plant organs. For the three repetitions, we can also observe a low level of redistribution to the growing parts of the vine (i.e. young leaves and apex, which always contain less than 10%). This corroborates the fact that at this stage, the cluster is the organ to which metabolites are preferentially routed. 8

9 Repetitions C B A % 20% 40% 60% 80% 100% Figure 8: Distribution (in %) for three repetitions (A, B, C) of deuteriated IBMP treatment on Sauvignon blanc shoots after deposition on the leaves and migration. 1 repetition = 4 cuttings. Finally, we found deuteriated IBMP in the stems and then in the berries. We therefore demonstrate that this compound is transported by the phloem from the leaves to the berries. Knowing that leaves contain high IBMP levels during ripening and that this compound can migrate from the leaves to the clusters, we can imagine that the leaves form IBMP reserves that can supply the clusters. We can also postulate that the berries are also capable of synthesising methoxypyrazine. Before veraison, synthesis seems to occur faster than breakdown. Perhaps at this stage, the berries have not yet acquired the capacity to break down IBMP (or only at very low levels). The IBMP content in the berries would thus be the result of transport from the leaves and of synthesis in situ. Unfortunately, no advances have been made in the field of IBMP biosynthesis in plants, and even the origin of this compound remains unknown. Murray et al. (1970) put forth the hypothesis of a condensation between glyoxal (which is also involved in the formation of other methoxypyrazines, according to these authors) and leucine (after having accepted an amide group). In line with this hypothesis, a study on the biosynthesis of 2-methoxy-3- isobutylpyrazine was undertaken using cell cultures of Cabernet Sauvignon. For the first time, we demonstrate that an undifferentiated callus of Cabernet Sauvignon is capable of synthesising IBMP. Moreover, adding leucine, i.e. the supposed precursor of IBMP, to the culture medium increases IBMP production by the cells. However, in one experiment, the addition of stable isotopes (L-leucine-d 10 and 15 NH 4 Cl) did not lead to isotopic enrichment of the IBMP produced by the cells. From fruit set up through about two to three weeks before mid-veraison, IBMP is synthesised and accumulates in the leaves and/or berries, perhaps from leucine. Thereafter, the IBMP content drops up through harvest From Veraison to Ripeness Apical zone Petiole Stems Berries The shape of IBMP content curves during ripening does not change as a function of whether this content is expressed as ng/l, ng/kg of fresh matter or pg/berry (Figure 9). 9

10 140 [IBMP] ng/l ng/kg of FM pg/berry Jul 22-Jul 01-Aug 11-Aug 21-Aug 31-Aug 10-Sep 20-Sep 30-Sep Figure 9: Evolution of IBMP content (expressed in ng/l, in ng/kg of fresh matter or in pg/berry) in Cabernet Sauvignon grapes from berry touch till harvest (1999). Mid-veraison occurred on 11 August. As such, and in contrast with what occurs with tartaric acid, for example, the decrease in IBMP content expressed in ng/l is independent of the dilution that occurs through the increase in berry volume during this period. This compound is thus truly broken down during ripening. Several studies have revealed a strong relationship between exposure of the cluster to light and the decrease in IBMP concentration (Allen and Lacey, 1993; Noble et al., 1995; Hashizume and Samuta, 1999). They thus confirm the first studies performed by Heymann (1986) and then by Maga (1989), which showed that methoxypyrazines are broken down by light. A photo-degradation study of IBMP in a solvent (methanol, 10% v/v) and in wine (white and red) enabled us to confirm that it breaks down when it is exposed to normal daytime sunlight. We then showed that the breakdown products (including 2-methoxy-3- methylpyrazine, which has been identified as a reaction intermediate) are present in very low quantities and do not seem to have an organoleptic impact. Until now, it was possible to suppose that an IBMP breakdown product could be involved in the aroma of Cabernet Sauvignon wines. While this variety has a strong green pepper aroma when unripe, this vegetal character disappears when conditions permit it and only then can a great Cabernet Sauvignon wine with fruity and toasty aromas be made. Our work seems to show that the origin of these aromas is not directly related to IBMP breakdown. 10

11 As we have seen, prior to veraison, IBMP begins to break down in the berries. However, the distribution of this compound in the clusters remains the same throughout ripening (Figure 10) Before veraison After veraison Harvest % 20% 40% 60% 80% 100% Stems Skins Seeds Pulp Figure 10: Distribution (in %) of IBMP in different components of Cabernet Sauvignon clusters during ripening in Regardless of the phenological stage, the pulp contains little IBMP and the stems contain a lot. However, from veraison to harvest, the proportion of IBMP decreases in the stems and increases in the skins. It also decreases slightly in the seeds during this period. Upon harvest, IBMP is found mainly in the stems. As such, we can understand why the green pepper character in a wine can be greatly influenced by destemming quality. Between 11 August and 23 September, the IBMP content increases most in the basal and intermediate leaves (Figure 11). In fact, the IBMP concentration in the adult leaves evolves in the opposite direction from that of the grapes. This result may appear paradoxical. Under identical environmental conditions, IBMP accumulates in the leaves, whereas it is broken down in the grapes during ripening. If IBMP is broken down by light, why is it that the basal leaves, which receive as much if not more light than the clusters, display increasing levels of IBMP during the ripening phase? Everything suggests that the metabolism of this compound is different in the fruit and in the leaves. 11

12 [IBMP] (ng/kg of FM) Jul 23-Sep Clusters Adult leaves Middle leaves Young leaves Figure 11: IBMP in the clusters and leaves (l.) at different insertion levels on the shoot during ripening of Cabernet Sauvignon (1999). These results confirm the advantages of leaf removal from the grapevine to decrease the vegetal/green pepper aromas of grapes. By exposing the grapes to greater sunlight, leaf removal increases IBMP breakdown, and it might also decrease the IBMP content in the berries by removing organs that could be a source of IBMP supply for the clusters Effect of Vineyard Practices and Conditions on IBMP Content in Grapes First, we measured the cumulative effect of all summer pruning and thinning work on the IBMP content of Cabernet Sauvignon and Merlot grapes. This work includes debudding (between budbreak and bloom), removal of summer laterals in the cluster zones on the east side of the vine row (at the end of fruit set), leaf thinning in the fruiting zone on the east side (at berry touch) and cluster thinning to limit the yield to around 50 hl/ha (at the start of veraison). Table 5: Influence of summer pruning and thinning (SPT) work on the composition of Merlot (M) and Cabernet-Sauvignon (CS) wines, Alcohol Total acidity Yield Anthocyanins IBMP TPI % vol. (g/l H 2 SO 4 ) (hl/ha) (mg/l) (ng/l) M control M SPT CS control CS SPT TPI: Total Polyphenolic Index Summer pruning and thinning has a direct effect on decreasing the IBMP content of Cabernet Sauvignon and Merlot grapes, from veraison till harvest. The photolabile nature of 12

13 IBMP explains this result. Grapes from the control group and from the thinned group were vinified on a large scale (311-hL tanks) and then analysed at the start of ageing (Table 5). The differences are more marked for Cabernet Sauvignon. The wine made from the test vines (i.e. the thinned vines) has more alcohol, a higher phenolic content (+40% anthocyanins) and its IBMP content is much lower than that of the wine made from the control vines (-39%). The control wine is clearly marked by the green pepper aroma of IBMP, whereas IBMP is not perceptible in the test wine. It is widely acknowledged that the grapevine reacts differently depending on the date on which summer vine work is performed. An experiment was conducted in 1998 on Cabernet Sauvignon and Sauvignon blanc vines in order to determine how the timing of summer vine work (summer lateral removal and leaf thinning in the cluster zone on the east side) affects the IBMP content of grapes at harvest. This experiment led to the following conclusions. It is generally important to perform lateral removal and leaf thinning early, i.e. between fruit set and berry touch. If this is done, the grapes have a higher sugar content at harvest, they are smaller in size and they contain less IBMP (Table 6). Late leaf thinning, while it does increase the sugar content in the grapes, does not lead to a sufficiently great decrease in IBMP content (this content is 65% greater than for a vine on which the leaf thinning and lateral removal were done early). Table 6: Influence of the timing of summer pruning and thinning (difference with respect to a control group) on the composition of Cabernet Sauvignon grapes upon harvest Berry weight Total acidity Reducing sugars Summer lateral removal Lateral removal at fruit set Lateral removal and leaf & leaf thinning at fruit set & leaf thinning after veraison thinning after veraison -7.4% -4.4% +2% Not significant Not significant Not significant +8.5% +6.7% +3% IBMP -68.4% -10.5% = There are no significant differences in grape composition at harvest if this work is performed at fruit set or when the berries are the size of small peas. In both cases, the work improves ripening. From a practical point of view, this offers greater flexibility in scheduling summer vine work. There is a day period in which one can perform this work without affecting final grape quality. Once this period is over, the risk becomes greater of harvesting grapes with marked green pepper aromas. These results were obtained in the case of low-trained vines with high planting density (8,550 vines/ha) and also with high-trained vines with low density (3,300 vines/ha). Analogous results were found with Sauvignon blanc. However, these results obtained in 1998 were not confirmed in 1999, when the IBMP contents in the grapes at harvest were close to zero, regardless of when the summer vine work was performed. The 1999 vintage was an atypical year, with heavy rainfall during veraison, which lengthened the veraison period and led to heterogeneity in berry ripeness. Thereafter, the weather was variable, with storms being intermixed with sunny, hot periods. These weather conditions led to unusual vine function. We can observe that IBMP breakdown was slower in 1999 than in The main difference between these two vintages lies in the 13

14 IBMP content at mid-veraison: it was three times greater in 1998 for the same vineyard block. In other words, for vintages like 1998, it is important to intervene early so that the IBMP in the grapes breaks down as quickly as possible. However, in 1999, early leaf thinning accelerated IBMP breakdown at the beginning, but at harvest no difference was observed between the different lots ( test and control ), owing to the low levels of IBMP present from the start. This proves that it is not so much the weather conditions during the ripening period that matter but rather those just prior to that period. It is probable that an early diagnosis (i.e. in late July) of the grape IBMP content would help in determining whether or not lateral removal or leaf thinning are needed in a vineyard block. While 1999 did not confirm the observations made in 1998, lateral removal and leaf thinning are major vineyard techniques enabling production of high-quality wine grapes. These practices are not widely used (since they cannot be mechanised and are thus costly), but they can be advantageous if performed early (at fruit set, the laterals are small and can be removed easily and quickly). Such practices facilitate further summer vine work by eliminating some of the vegetation in the fruiting zone; they improve the healthiness of the crop and lead to better aeration around the clusters; they remove some of the organs requiring influx of nutrients and thus they promote better redistribution of photosynthesis products for improved grape ripeness. CONCLUSION The application of an original protocol enabled us to define a upper limit for the IBMP concentration beyond which the methoxypyrazine character is perceptible in red Bordeauxtype wines. This level is on average 15 ng/l. We have been able to determine the contribution of IBMP to the vegetal aroma of Bordeaux varieties. It plays a major role for Cabernet Sauvignon, Cabernet franc and Sauvignon blanc and a minor one for Merlot. We have also shown that the IBMP content of wine depends mainly on that of the corresponding grapes and that it is only marginally affected by winemaking techniques. In the case of traditional winemaking, IBMP is highly extractable, independently of pressing conditions for white wines and of maceration time and the number of pump-overs for red wines. Only the settling of Sauvignon blanc musts and a careful selection of press wines for Cabernet Sauvignon can be used to limit the IBMP content of wines. In the case of thermovinification, heating the grapes leads to a significant drop in IBMP concentration, owing to volatilisation. We have observed no change in IBMP concentration in wines during bottle ageing. IBMP concentration increases in grapes from fruit set until about two to three weeks before mid-veraison. This phenomenon seems to be influenced by the grapevine water status before veraison. The use of Cabernet Sauvignon cell cultures has not enabled us to explain the mechanisms of IBMP biosynthesis. Nevertheless, we have shown that the addition of leucine, which is the supposed precursor of IBMP, leads to increased IBMP production. During this period, IBMP is synthesised in the leaves, which is where it is mainly located. It is also in the stems, skins and seeds (in order of decreasing importance). The pulp contains almost no IBMP. We have revealed for the first time that IBMP is transported from the leaves to the clusters during this stage. The maximum IBMP content in the grape is reached before veraison. After this, the compound begins to break down in the berries. The order of its distribution in the leaves and clusters remains the same throughout the ripening period. This breakdown is the result of IBMP s sensitivity to light. However, none of the photo-degradation products, including 2- methoxy-3-methylpyrazine, which we have identified here, appear to have any organoleptic 14

15 impact. Paradoxically, while it is broken down in the grapes, IBMP continues to accumulate in the leaves. The metabolism of methoxypyrazine does not seem to be the same in leaves as in clusters. The IBMP content of grapes at harvest can be controlled by summer vine work. Operations such as debudding, summer lateral removal, leaf thinning and cluster thinning lead to a large decrease in IBMP content in Merlot and Cabernet Sauvignon grapes during ripening. It is generally important to perform lateral removal and leaf thinning early, i.e. between fruit set and berry touch. After that, the risks are greater of harvesting grapes with marked green pepper aromas. The comparison of the results obtained in 1998 and 1999 enables us to postulate that the key to the methoxypyrazine character of ripe grapes is related to the weather conditions that prevail at an early stage and which lead to a certain initial IBMP content in the grapes prior to veraison. 15

16 BIBLIOGRAPHY AERNY J., Composés azotés des moûts et des vins. Revue Suisse Arboric. Hortic., 28, ALLEN M.S., LACEY M.J., BROWN W. V. and HARRIS R.L.N., Occurrence of methoxypyrazine in grapes of Vitis vinifera cv. cabernet-cauvignon and sauvignon blanc. In Actualités Œnologiques 89, Compte Rendu du IV e Symposium d Œnologie de Bordeaux, Dunod, Paris, ALLEN M.S., LACEY M.J., HARRIS R.L.N. and BROWN W.V., Contribution of methoxypyrazines to Sauvignon blanc wine aroma. Am. J. Enol. Vitic., 42, ALLEN M.S. and LACEY M.J., Methoxypyrazine grape flavour : influence of climate, cultivar and viticulture. Vitic. Enol. Sci., 48, ALLEN M.S., LACEY M.J. and BOYD S., Determination of methoxypyrazines in red wines by stable isotope dilution gas chromatography mass spectrometry. J. Agric. Food Chem., 42, AMRANI-HEMAIMI M., CERNY C. and FAY L. B., Mechanisms of formation of alkylpyrazines in the Maillard reaction. J. Agric. Food Chem., 43, ARNOLD R.A. and BLEDSOE A.M., The effect of various leaf removal treatments on the aroma and flavor of Sauvignon blanc wine. Am. J. Enol. Vitic., 41, AUGUSTYN O. P. H., RAPP A. and VAN WYK C. J., Some volatile aroma component of Vitis vinifera L. cv. Sauvignon blanc. S. Afr. J. Enol. Vitic., 3, BAYONOVE C., CORDONNIER R.A. and DUBOIS P., 1975 Etude d une fraction caractéristique de l arôme du raisin de la variété Cabernet-sauvignon ; mise en évidence de la 2-méthoxy-3-isobutylpyrazine. C. R. Acad. Sci. (Paris), Série D, 281, BLANK I., SEN A. and GROSCH W., Potent odorants of the roasted powder and brew of Arabica coffee. Z. Lebensm. Unters. Forsch., 195, BOIDRON J.N., Essai d identification des constituants de l arôme des vins de Vitis vinifera L. Premiers résultats. Thèse de doctorat. Faculté des Sciences de l Université de Bordeaux. BOIDRON J.N., CHATONNET P. and PONS M., Influence du bois sur certaines substances odorantes des vins. Conn. Vigne et Vins, 22, BOISON J.O.K., and TOMLINSON R.H., New sensitive method for the examination of the volatile flavour fraction of Cabernet-sauvignon wines. J. Chromatogr., 522, BOISSENOT E., Incidence du climat, des sols et du comportement de la vigne sur les caractères analytiques et organoleptiques des vins rouges du Haut-Médoc. Relation avec la maturation du raisin. Thèse de doctorat de l Université de Bordeaux II. BRANAS J., Viticulture. Imprimerie Déhan, Montpellier. BROCHET F., Dosage de la 2-méthoxy-3-isobutylpyrazine dans les vins de Bordeaux. Mémoire de DEA. Université de Bordeaux II. BROCHET F., La dégustation. Etude des représentations des objets chimiques dans le champ de la conscience. Thèse de doctorat de l Université de Bordeaux II. BUTTERY R.G., SEIFFERT R.M., GUADAGNI, D.G. and LING L.C., Characterization of some volatile constituents of bell peppers. J. Agric. Food Chem., 17, BUTTERY R.G., Vegetable and fruit flavors. In Flavor research Recent advances. Marcel Dekker ed., New York, CALO A., DI STEFANO R., COSTACURTA A. and CALO G., Caracterizzazione di Cabernet franc e Carmenère (Vitis sp.) e chiarimenti sulla loro coltura in Italia. Riv. Vitic. Enol., 3,

17 CANTALEJO M.J., Analysis of volatile components derived from raw and roasted earth-almond (Cyperus esculentus L.). J. Agric. Food Chem., 45, CASTELL C.H., GREENOUGH M.F. and JENKIN N.L., The action on Pseudomonas on fish muscle : 2. Musty and potato-like odours. J. Fish. Res. Bd. Canada, 14, CHAMPAGNOL F., Eléments de physiologie de la vigne et de viticulture générale. Champagnol F. Ed., Saint Gély du Fesc. CHENG T.B., REINECCIUS G.A., BIORKLUND J.A. and LEETE E., Biosynthesis of 2-methoxy-3-isopropylpyrazine in Pseudomonas perolens. J. Agric. Food Chem., 39, COOMBE B.G., Research on development and ripening of the grape berry. Am. J. Enol. Vitic., 43, CRIPPEN D.D. JR. and MORRISON J.C., 1986a. The effects of sun exposure on the compositional development of Cabernet Sauvignon berries. Am. J. Enol. Vitic., 37, CRIPPEN D.D. JR. and MORRISON J.C., 1986b. The effects of sun exposure on the phenolic content of Cabernet Sauvignon berries during development. Am. J. Enol. Vitic., 37, CROUZET J., Les enzymes et les arômes des vins. Rev. Fr. Œnol., 102, CZERNY M., WAGNER R. and GROSCH W., Detection of odor-active ethenylalkylpyrazines in roasted coffee. J. Agric. Food Chem., 39, DECENDIT A. and MERILLON J.M., Condensed tannin and anthocyanin production in Vitis vinifera cell suspension cultures. Plant Cell Reports, 15, DECENDIT A., RAMAWAT K.G., WAFFO TEGUO P., DEFFIEUX G., BADOC A. and MERILLON J.M., Anthocyanins, catechins, condensed tannins and piceid production in Vitis vinifera cell bioreactor cultures. Biotechnol. Lett., 18, DI STEFANO R. and MAGGIOROTTO G., Différence entre la composition terpénique des cépages aromatiques. In Connaissance aromatique des cépages et qualité des vins, Actes du Symposium international, Montpellier, 9-10 février 1993, Rev. Fr. Œnologie Ed., DUBOURDIEU, D. ; OLLIVIER, CH. and BOIDRON, J.N., Incidence des opérations préfermentaires sur la composition chimique et les qualités organoleptiques des vins blancs secs. Conn. Vigne Vin., 20, DUMONT J.P., ROGER S. and ADDA J., 1974a. Volatile composition of whole and grated Parmesan cheese. Le Lait, 54, DUMONT J.P., ROGER S., CERF R. and ADDA J., 1974b. Etude de composés volatils neutres présents dans le Vacherin. Le Lait, 54, DUMONT J.P., ROGER S. and ADDA J., Mise en évidence d un composé à hétérocycle azoté responsable d un défaut d arôme dans le Gruyère et le Comté. Le Lait, 55, DUMONT J.P., MOURGUES R. and ADDA J., Potato-like off flavor in smear coated cheese : a defect induced by bacteria. In Sensory quality in foods and beverages : definition, measurement and control, WILLIAMS A.R., Atkins R.K. Eds. Ellis Horwood Limited, DUTEAU J., Alimentation en eau de la vigne dans le Bordelais en période estivale sèche. Exemple de l année 1980 à Saint Emilion et Pomerol. Sci. Sol, 1, EMDE K.M.E., BEST N. and HRUDEY S.E., Production of the potent odour agent, ispropyl methoxypyrazine, by Lysobacter enzymogenes. Environmental Technol., 13, ENGEN T., Psychophysics. In Experimental Psychology, KLING J.W. and RIGGS L.A. Eds., Holt, Rinehart ans Winston, New York. FIRMENICH and CIE, Brevet français n (25 janvier 1965). 17

18 FORS S., Alkylpyrazines Flavour compounds in foods. Odour characteristics of pure alkylpyrazines and their role in treated malt. PhD Thesis, Departement of Food Science, Chalmers University of Technology, Göteborg, Sweden. GALET P., Précis de Viticulture. 6 ème édition. Imprimerie Déhan, Montpellier. GALLOIS A., Les pyrazines présentes dans les aliments état actuel de nos connaissances. Sci. Aliments, 4, GALLOIS A., KERGOMARD A. and ADDA J., Study of the biosynthesis of 2- methoxy-3-isopropylpyrazine produced by Pseudomonas taetrolens. Food Chem., 28, GALLOIS A. and GRIMONT P.A.D., Pyrazines responsible for the potatolike odor produced by some Serratia and Cedecea strains. Appl. Environ. Microbiol., 50, GAMBORG O.L., MILLER R.A. and OJIMA K., Nutrient requirements of suspension cultures of soybean root cells. Exp. Cell Res., 50, GENY L., OLLAT N. and SOYER J.P., Les boutures fructifères de vigne : validation d un modèle d étude de la physiologie de la vigne. II. Etude du développement de la grappe. J. Int. Sci. Vigne Vin, 32, GERBER N.N., Three highly odorous metabolites from an actinomycete : 2-methoxy- 3-isopropylpyrazine, methylisoborneol and geosmin. J. Chem. Ecol., 3, GONZALO S.R. and KLIEWER W.M., Estimation of leaf area of two grapewine cultivars (Vitis vinifera L.) using laminea linear measurements and fresh weight. Am. J. Enol. Vitic., 34, GÜNATA Z.Y., Recherche sur la fraction liée de nature glycosidique de l arôme du raisin : importance des terpénylglycosides, action des glycosidases. Thèse Docteur-Ingénieur. Université des Sciences et Techniques du Languedoc. HALE C.J. and WEAVER R.J., The effect of development stages on direction of translocation of photosynthata in Vitis vinifera. Hilgardia, 35, HARRIS R.L.N., LACEY M.J., BROWN W.V. and ALLEN M.S., 1987; Determination of 2- methoxy-3-alkylpyrazines in wine by gas chromatography/mass spectrometry. Vitis, 26, HASHIZUME K. and UMEDA N., Methoxypyrazine content of japanese red wines. Biosci. Biotech. Biochem., 60, HASHIZUME K. and SAMUTA T., Green odorant cluster and their ability to cause a wine stemmy flavor. J. Agric. Food Chem., 45, HASHIZUME K., KIDA S. and SAMUTA T., Effect of stream treatment of grape cluster stems on the methoxypyrazine, phenolic, acid and mineral content of red wines fermented with stems. J. Agric. Food Chem., 46, HASHIZUME K. and SAMUTA T., Grape maturity and light exposure affect berry methoxypyrazine concentration. Am. J. Enol. Vitic., 50, HEYMANN H., Studies of methoxypyrazines and vegetative flavor od Cabernet Sauvignon wines. PhD Thesis, University of California, Davis. HUFFMAN V.L., SCHADLE E.R., VILLALON B. and BURNS E.E., Volatile components and pungency in fresh and processed jalapeno peppers. J. Food Sci., 43, IACONO F., BERTAMINI, M., SCIENZA A. and COOMBE B.G., Differential effects canopy manipulation and shading of Vitis vinifera L. cv. Cabernet-Sauvignon. Leaf gas exchange, photosynthetic electron transport rate and sugar accumulation in berries. Vitis, 34, JOSLIN W.S. and OUGH C.S., Cause and fate of certain C 6 compounds formed enzymatically in macerated grapes leaves during harvest and wine fermentation. Am. J. Enol. Vitic., 29,

19 KARAHADIAN C., JOSEPHSON D.B. and LINDSAY R.C., Volatile compounds from Penicillium sp. Contributing musty-earthy notes to Brie and Camembert cheese flavors. J. Agric. Food Chem., 33, KLIEWER W., Methods for determining the nitrogen status of vineyard. In Proceedings of the Int. Symposium on nitrogen in grapes and wines, RANTZ J. Ed. Am. Soc. Enol. Vitic., Davis, KOBLET W., Translocation des produits assimilés dans des pousses de vigne et influence de la surface de la feuille sur la qualité et la quantité de raisin. Die Wein- Wissenschaft, 8-9, KOBLET W., Translocation des produits d assimilation des différentes feuilles de la vigne pendant la maturation des raisins. Die Wein-Wissenschaft, 5, KOTSERIDIS Y., ANOCIBAR BELOQUI A., BERTRAND A. and DOAZAN J.P., An analytical method for studying the volatile compounds of merlot noir clone wines. Am. J. Enol. Vitic., 49, KOTSERIDIS Y., ANOCIBAR BELOQUI A., BAYONOVE C., BAUMES R. L. and BERTRAND A., Effect of selected viticultural and enological factors on levels of 2- methoxy-3-isobutylpyrazine in wines. J. Int. Sci. Vigne Vin., 33, KRISA S., 1999a. Production d anthocyanes et de stilbènes par culture cellulaire de Vitis vinifera. Marquage au 13 C et études biologiques. Thèse de doctorat de l Université de Bordeaux II. KRISA S., LARRONDE F., BUDZINSKI H., DECENDIT A., DEFFIEUX G. and MERILLON J.M., 1999b. stilbene production by Vitis vinifera cell suspension cultures : methyl jasmonate induction and 13 C biolabelling. J. Nat. Prod., 62, LACEY M.J., ALLEN M.S., HARRIS R.L.N. and BROWN W.V., Methoxypyrazines in Sauvignon blanc grapes and wines. Am. J. Enol. Vitic., 42, LAVIGNE-CRUEGE V., Recherches sur les composés soufrés volatils formés par la levure au cours de la vinification et de l élevage des vins blancs secs. Thèse de doctorat de l Université de Bordeaux II. LAVIGNE-CRUÈGE V., CUTZACH I. and DUBOURDIEU D., Interprétation chimique du vieillissement aromatique défectueux des vins blancs. Incidence des modalités d élevage. In Œnologie 99, 6 ème Symposium International d Œnologie. Edition Tec & Doc, Paris, MACLEOD G. and COPPOCK B.M., Comparison of the chemical composition of boiled and roasted aromas of heated beef. J. Agric. Food Chem., 25, MAGA J.A., and SIZER C.E., Pyrazines in foods. A Review. J. Agric. Food Chem. 21, MAGA J.A., Sensory and stability properties of added methoxypyrazines to model and authentic wines. In Flavors and off-flavors, Proceedings of the 6th international flavor conference. Charalambous, G. Ed., Elsevier Science Publishers B.V., Amsterdam, Netherlands., MAGA J.A., Pyrazine update. Food Reviews International, 4, MAGALETTA R.L. and HO C.T., Effect of roasting time and temperature on the generation of nonvolatile (polyhydroxyalkyl)pyrazine compounds in peanuts, as determined by high performance liquid chromatography. J. Agric. Food Chem. 44, MARAIS J. and SWART E., Sensory impact of 2-methoxy-3-isobutylpyrazine and 4- mercapto-4-methylpentan-2-one added to a neutral Sauvignon blanc wine. S. Afric. J. Enol. Vitic., 20, MCIVER R.C. and REINECCIUS G.A., Synthesis of 2-methoxy-3-alkylpyrazines by Pseudomonas perolens. In Biogeneration of aromas, PARLIMENT T.H. and CROTEAU R. Eds., Washington D.C. 19

20 MERMET G., CROS E. and GEORGES G., Etude préliminaire de l optimisation des paramètres de torréfaction du cacao. Consommation des précurseurs d arôme, développement des pyrazines, qualité organoleptique. Café Cacao Thé, XXXVI, MILLER III A., SCANLAN R. A., LEE J. S., LIBBEY L. M. and MORGAN M. E., Appl. Microbiol., 25, MOREL G., Le problème de la transformation tumorale chez les végétaux. Physiol. Vég., 8, MORGAN M.E., The chemistry of some microbially induced flavor defects in milk and dairy foods. Biotechnol. and Bioengin., 18, MORRISON J.C. AND NOBLE A.C., The effects of leaf and cluster shading on the composition of Cabernet Sauvignon grapes and on fruit and wine sensory properties. Am. J. Enol. Vitic., 41, MURASHIGE T. and SKOOG F., A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol. Plant., 15, MURRAY K.E., SHIPTON J. and WHITEFIELD F.B., methoxypyrazines and the flavour of green peas (Pisum sativum). Chem. and Ind., MURRAY K. E., WHITFIELD F. B., The occurrence of 3-alkyl-2-methoxypyrazines in raw vegetables. J. Sci. Food Agric., 26, NOBLE A.C., Sensory and instrumental evaluation of the aroma. In Analysis of foods and beverages, Charalambous G. Ed., Academic press : New York, NOBLE A.C., ELLIOT-FISK D.L. and ALLEN M.S., In Fruit flavors : biogenesis characterization and authentication, ROUSEFF R.L. and LEAHY M.M., Eds., ACS Symposium series 596, American Chemical Society : Washington, D.C., NURSTEN H.E. and SHEEN M.R., Volatile flavor components of cooked potato. J. Sci. Fd. Agric., 25, OLLAT N., GENY L. and SOYER J.P., Les boutures fructifères de vigne : validation d un modèle d étude de la physiologie de la vigne. II. Principales caractéristiques de l appareil végétatif. J. Int. Sci. Vigne Vin, 32, 1-9. OLLAT N. and GAUDILLÈRE J.P., The effect of limiting leaf area during stage I of berry growth on development and composition of berries of Vitis vinifera L. cv. Cabernet Sauvignon. Am. J. Enol. Vitic., 49, PIERI PH., OLLAT N. and TANDONNET J.P., Growth of vines and maturation of berries as influenced by the soil water balance. In Œnologie 95, 5 ème Symposium International d Œnologie. Edition Tec & Doc, Paris, RAPP A. HASTRICH H. and ENGEL L., Gas chromatographic investigations on the aroma constituents of grape berries. I. Concentration and separation by capillary glass columns. Vitis, 15, RIBEREAU-GAYON P. and STONESTREET E., Le dosage des anthocyanes dans le vin rouge. Bull. Soc. Chim., 9, RIBEREAU-GAYON P., Le dosage des composés phénoliques totaux dans les vins rouges. Chimie Analytique, 52, RIBEREAU-GAYON P., DUBOURDIEU D., DONECHE B. and LONVAUD A., In Traité d œnologie, Tome 1 : Microbiologie du vin, vinifications, Dunod, Paris, ROJAS-LARA B.A. and MORRISON J.C., Differential effects of shadin fruit or foliaige on the development and composition of grape berries. Vitis, 28, ROUJOU DE BOUBÉE D. and DUBOURDIEU D., Incidence des conditions de maturation et des pratiques viticoles sur la maturation des raisins de Cabernet-Sauvignon et de Merlot à Bordeaux. In Œnologie 99, 6 ème Symposium International d Œnologie. Edition Tec & Doc, Paris,