Molecular Background of the Lychee Aroma of Vitis vinifera L. ‘Muscaris’

Muscaris is a modern white grape variety with good fungal resistance and a pleasant aroma, the molecular background of which was unknown. A comparative aroma extract dilution analysis applied to Muscaris grapes and grapes of the father variety Muskateller revealed little differences and resulted in 39 and 35 odorants, respectively. Sixteen odorants exceeded their odor threshold concentrations. Odor reconstitution and omission experiments showed that the distinct lychee note in the aroma of the Muscaris grapes was generated by the combination of (2S,4R)-rose oxide and geraniol. This finding will guide further molecular research on the transfer of the lychee note into wine and may also be helpful for the targeted breeding of new grape varieties.


■ INTRODUCTION
Muscaris is a relatively new Vitis vinifera variety with an intense aroma and good resistance to downy mildew, powdery mildew, and botrytis.Muscaris is early ripening.The berries remain green, even at high must weights.It is particularly suitable for producing dessert wine and dry wine. 1 The cultivation area of Muscaris increased in recent years, reaching 117 ha in Germany in 2022. 2 Muscaris was bred at the State Institute of Viticulture Freiburg, Germany, in 1987 from the mother variety Solaris and the father variety Gelber Muskateller, also known as Muscat àpetits grains blancs or Muscat Blanc, 3 with the idea to combine the disease tolerance and environmental adaptability of Solaris with the intense aroma of Gelber Muskateller.Solaris is a white grape variety bred in 1975, whose fungal resistance originates from the wild Asian species Vitis amurensis. 3,4ctually, it was shown that Muscaris has a similar resistance to downy mildew, powdery mildew, and botrytis as Solaris. 1,5elber Muskateller is one of the most widely planted white Muscat varieties.It is of Greek origin and has a long cultivation history in Germany.Its fungal resistance is low; 6 however, it is highly appreciated for its floral and fruity notes, which are intense in the grapes 6,7 as well as in the wine. 8,9Muskateller grapes are often used as a blending partner for other white grape varieties to boost the aroma of wine. 6everal studies have been conducted on Muscaris and addressed fungal resistance, 5,10,11 viticultural characteristics, 12,13 mechanical grape properties, 13 and grape composition including phenolic compounds, 14 minerals, and ascorbic acid. 15−20 However, there is currently no information in the literature on odor-active compounds in Muscaris grapes.Some phenolic compounds including 4-ethenylphenol, 4-ethenyl-2-methoxyphenol, 4hydroxy-3-methoxybenzaldehyde, 4-hydroxy-3,5-dimethoxy-benzaldehyde, and 4-hydroxy-3-ethoxybenzaldehyde were identified as Muscaris grape volatiles, but their odor contribution was not assessed. 14We found that the aroma of Muscaris grapes is characterized by a distinct lychee note, the molecular background of which piqued our interest.−24 A comparative gas chromatography−olfactometry (GC−O) analysis applied to Gewurztraminer wine, fresh lychees, and canned lychees detected 12 odorants common to all three samples. 22Among them, cis-rose oxide, β-damascenone, ethyl 2-methylpropanoate, linalool, and geraniol consistently showed concentrations beyond their odor threshold concentrations (OTCs).cis-Rose oxide, linalool, and geraniol had also been reported in odor-active amounts, that is in concentrations exceeding their OTCs, in Muskateller grapes and wines. 8,9,25,26iven the gaps in the literature detailed above, our aims were to characterize the major odor-active compounds in Muscaris grapes, compare them with the odor-active compounds in the grapes of its father variety, Muskateller, and elucidate the molecular background of the lychee note in the aroma of the Muscaris grapes.This knowledge is a prerequisite for further research, e.g., on odorant transfer from Muscaris grapes into wine and for targeted breeding.Thus, our study included (i) the screening for odor-active compounds in Muscaris and Muskateller grapes by application of a comparative aroma extract dilution analysis (cAEDA), (ii) the quantitation of potent odorants in grapes of both varieties was concentrated to a final volume of 1.0 mL using a Vigreux column (50 × 1 cm) and a Bemelmans microdistillation device. 35he grape volatile isolate was stepwise diluted 1:2 with dichloromethane to obtain dilutions of 1:2, 1:4, 1:8, 1:16, 1:32, 1:64, 1:128, 1:256, 1:512, 1:1024, 1:2048 1:4096, 1:8192, 1:16,384, 1:32,768, and 1:65,536.The undiluted sample, as well as each diluted sample, was subjected to GC−O analysis 36 using the GC−O/FID instrument detailed in the Supporting Information with the FFAP column.Each odorant was assigned a flavor dilution (FD) factor, representing the dilution factor of the highest diluted sample in which the odorant was detected at the sniffing port during GC−O.
Peak areas of the analytes and the respective internal standards were collected from the extracted ion chromatograms using the quantifier ions detailed in Table S2.The concentration of each odorant in the grapes was calculated from the area counts of the analyte peak, the area counts of the internal standard peak, the amount of grapes used in the workup, and the amount of internal standard added by employing a calibration line equation.To obtain the calibration line equation, solutions of the analyte and the respective internal standard were mixed in different concentration ratios and analyzed under the same conditions, followed by linear regression.The calibration line equations are available in Table S2.Individual concentrations and standard deviations can be found in Tables S3−S5.
Odor Threshold Concentrations.Orthonasal OTCs of (3E)hex-3-enal (3), heptanal (5), and (2S,4R)-rose oxide (10) were determined in pure water according to the American Society for Testing and Materials (ASTM) standard practice for determination of odor and taste thresholds by a forced-choice ascending concentration series method of limits. 37Assessors (6 males and 12−14 females, aged 21−60 years) were recruited from the trained panel of the Leibniz-LSB@TUM.The tests were carried out in separate booths of a room exclusively dedicated to sensory evaluations.The room temperature was 22 ± 2 °C.Further details including the GC−O approach for the purity testing of the odorants prior to their use in the OTC determinations are available in the literature. 38dor Reconstitution Models.The basis of the odor reconstitution models was aqueous solutions of tartaric acid (3.57g/L for the Muscaris grape models and 4.23 g/L for the Muskateller grape models).The tartaric acid concentrations in the grapes were previously determined using an enzymatic test kit (R-Biopharm; Darmstadt, Germany).An individual ethanolic stock solution was prepared for each odorant for which an ≥1 OAV had been determined.The absence of odor-active impurities in the reference odorants was checked by GC−O. 38Aliquots of the stock solutions were combined and diluted with the appropriate tartaric acid solution to obtain final odorant concentrations in the models that represented the concentrations previously determined in the grapes.Final ethanol concentrations were kept below 200 μL/kg.Using aqueous potassium hydroxide (2 mol/L), the pH of the models was adjusted to 3.8, the value previously determined in both the Muscaris and the Muskateller grapes.The complete odor reconstitution models contained 16 odorants; incomplete models were additionally prepared from which either (2S,4R)-rose oxide (10), geraniol (26), or both were omitted.
Quantitative Olfactory Profiles.Samples (10 g), either freshly crushed frozen-thawed grapes or odor reconstitution models, were placed in cylindrical polytetrafluoroethylene vessels (5.7 cm height, 3.5 cm i.d., 50 mL nominal volume) with lids (Bohlender; Grunsfeld, Germany) and presented to 14 assessors (5 males and 9 females, aged 24−60 years) recruited from the trained panel of the Leibniz-LSB@ TUM.The tests were carried out in the room described before.

■ RESULTS AND DISCUSSION
Odorant Screening.GC−O in combination with cAEDA applied to the volatile isolates obtained from fresh Muscaris and Muskateller grapes, both of which showed a distinct lychee Odorants showing an FD factor of ≥1 in either of the two samples; odorants were identified by comparing the retention indices on two columns of different polarity (DB-FFAP, DB-5), the mass spectra obtained by GC−MS as well as the odor quality as perceived at the sniffing port during GC−O to data obtained from authentic reference compounds analyzed under equal conditions.b Odor as perceived at the sniffing port during GC− O. c Retention index; calculated from the retention time of the odorant and the retention times of adjacent n-alkanes by linear interpolation.d Flavor dilution factor; dilution factor of the highest diluted grape volatile isolate in which the odorant was detected during GC−O by any of three assessors.e An unequivocal mass spectrum of the compound could not be obtained; identification was based on the remaining criteria detailed in footnote a and by spiking experiments using GC−O/FID.f Odor-active enantiomers as identified by analysis of the volatile isolates using GC−O/ FID in combination with a chiral column.g The compounds were not separated on the column used for AEDA; the FD factor refers to the mixture.note, resulted in 39 odor-active compounds, 35 of which were present in both samples.FD factors ranged from 1 to 32,768 for Muscaris grapes and from 1 to 8192 for Muskateller grapes (Table 1).Structure assignments were achieved with the following approach: the RIs obtained for Muscaris and Muskateller grape odorants with two GC columns of different polarity (DB-FFAP and DB-5) in combination with the odor descriptions were compared to published data, foremost those compiled in the Leibniz-LSB@TUM Odorant Database. 39The resulting structure proposals were confirmed by GC−O and GC−MS analyses of the grape volatile isolates in parallel with authentic reference compounds.In the case of coelution problems, the comprehensive two-dimensional GC × GC−MS instrument was employed for the mass spectral analyses.For enantiospecific structure assignments, the approach was repeated using two chiral GC columns with differently substituted β-cyclodextrin phases.
As a result, the structures of 38 odorants were successfully determined.Only compound 33, despite all efforts, remained unknown.Among the odorants identified in Muscaris grapes, vanilla-like smelling 4-hydroxy-3-methoxybenzaldehyde (vanillin; 39) was the only compound which had previously been reported in grapes of this variety. 14he compound with by far the highest FD factor among the Muscaris grape odorants was floral, rose-like smelling geraniol (26; FD factor 32,768).With 8192, geraniol also showed the highest FD factor among the Muskateller grape odorants.Nineteen odorants were assigned FD factors ≥32 in at least one of the two samples.Twelve of them showed comparable FD factors in both grape varieties.This compound group included green, grassy smelling (3Z)-hex-3-enal ( 4 and 8).In contrast, higher FD factors in the Muskateller grapes were found for cooked potato-like smelling 3-(methylsulfanyl)propanal (13; FD factors 512 and 2048), citrusy, floral smelling isomers (3R)-and (3S)-linalool (18; FD factors 64 and 512), and green, grassy smelling hexanal (2; FD factors 16 and 64).In summary, however, the odorant spectrum in the Muscaris grapes showed a huge similarity to the odorant spectrum in the Muskateller grapes, thus reflecting the close genetic relationship between the two varieties.This was particularly the case for some monoterpenes considered iconic for Muskateller such as cis-rose oxide, linalool, and geraniol.A detailed review on their biosynthesis in grapes has been published by Schwab and Wust. 40dorant Quantitation and OAV Calculation.Twenty odorants for which high FD factors in at least one of the two grape samples had been determined during the odorant screening (cf.Table 1) were selected.Quantitation was accomplished by GC−MS using stable isotopically substituted odorants as internal standards.The grape samples subjected to quantitation were from the same batch as those used for odorant screening.Given the time required to collect the screening data and assign the structures, it was impossible to perform the quantitations with fresh material.Accordingly, a majority of odorants were quantitated in frozen-thawed grapes.Exceptions were the lipoxygenase products 41 (3Z)-hex-3-enal (4) and (3E)-hex-3-enal (3), which were quantitated in the fresh grapes in parallel to the screening experiments.This was done because it had been reported that the concentration of (3Z)-hex-3-enal can substantially differ between fresh and stored as well as frozen-thawed plant materials. 28,42nantiospecific concentrations of important chiral odorants were calculated from the sum of enantiomers as determined via stable isotopically substituted internal standards in the quantitation assays and the enantiomeric distribution as determined by GC with chiral columns.Chromatograms showing the separation of the enantiomers are available in Figure S1.The enantiomeric distributions of cis-rose oxide, linalool, and β-citronellol are listed in Table 2.In both linalool and β-citronellol, the (S)-isomer clearly predominated with percentages of 91−98%.With 94 and 99.8%, similar data were reported for linalool in two other Muscat grape varieties. 43To the best of our knowledge, the enantiomeric distribution of βcitronellol in Muscat grapes has not been examined so far.In cis-rose oxide, the major enantiomer was the odor-active (2S,4R)-isomer with 71% in the Muscaris grapes and 75% in the Muskateller grapes.In other six Muscat grape varieties previously investigated, the (2S,4R)-isomer showed even higher percentages of 88−97%. 32In total, the data obtained for Muscaris and Muskateller grapes were highly similar, again illustrating their close genetic relation.
To gain more information on the odor activity of the individual odorants, OAVs were calculated as the ratios of the odorant concentrations in the grapes and the OTCs in water.The results (Table 3) revealed an OAV of ≥1 in both grape varieties for 16 of the 20 quantitated odorants.The highest OAV was calculated for geraniol (26) in both Muscaris (OAV 1100) and Muskateller (OAV 400) grapes.Whereas in Muscaris grapes no previous data was available in the literature, a huge data set was found for geraniol in Muskateller grapes.The majority of the geraniol concentrations in Muskateller grapes ranged between ∼70 45 and ∼300 μg/kg, 46 corresponding to OAVs of ∼60 and ∼300.
Comparing the OAVs of individual odorants provided in Table 3 between Muscaris and Muskateller grapes resulted in little differences for a majority of compounds; that is, the OAVs did not differ by more than a factor of 2. Important exceptions were geraniol (26) and (2E,6Z)-nona-2,6-dienal (19), which not only featured the highest two OAVs in the Muscaris grapes but additionally showed OAVs ∼3 times lower in the Muskateller grapes.Another exception was (3S)-linalool (18b) with an OAV in the Muskateller grapes almost 5 times higher than that in the Muscaris grapes.However, when the OAVs of both citrusy, floral smelling linalool enantiomers (cf.Table 2) were summed, the combined values differed only by a factor of ∼2 between the Muskateller and the Muscaris grapes.
Odor Reconstitution and Omission Experiments.To verify the analytical data and identify the odorants responsible for the characteristic lychee note, odor reconstitution and omission experiments were performed.First, an aqueous odor reconstitution model was prepared for each of the two grape varieties containing all 16 odorants with OAVs ≥1 in the previously determined concentrations (cf.Table 3).Orthonasal evaluation of the two models by a trained sensory panel resulted in quite similar olfactory profiles for both grape  S3 and S4.d Orthonasal odor threshold concentration in water.e Odor activity value; calculated as the ratio of concentration to odor threshold concentration.f Data taken from the Leibniz-LSB@TUM Odorant Database. 39g Concentrations of individual enantiomers were calculated from the concentration of the sum of enantiomers as obtained in the quantitation assays and the enantiomeric distribution depicted in Table 2. h Data obtained in the current study. 37i Data from literature. 44arieties.The Muscaris model as well as the Muskateller model showed a balanced combination of green-grassy, floral, lychee, and green apple-like notes (Figure 1).Moreover, the distinct lychee note perceived in the fresh Muscaris grapes was well reflected in the aroma of the Muscaris reconstitution model.A direct comparison of the two reconstitution models with the fresh materials, however, was not possible due to the short shelf life of the grapes.Using fresh grapes of the following harvest as a reference was not considered appropriate as it is known that the vintage can have a huge influence on the grape aroma. 49s an alternative to evidence of the correctness of the analytical data, the frozen-thawed Muscaris and Muskateller grapes were used as references for reconstitution models fully based on odorant concentrations in the frozen-thawed materials.This included (3E)-hex-3-enal (3) and (3Z)-hex-3enal (4), which therefore were additionally quantitated in the frozen-thawed grapes (Table S5).The resulting quantitative olfactory profiles (Figure S2) revealed good agreements between the models and the grapes, thus providing evidence that all major aroma-contributing compounds in Muscaris and Muskateller grapes were correctly identified and quantitated.
The primary aim of the omission experiments was to elucidate the molecular basis of the lychee note.The tests were based on the reconstitution models depicted in Figure 1.In the first experiment, (2S,4R)-rose oxide (10) was omitted from the models as it was the only compound among the odorants identified in Muscaris and Muskateller grapes with a specific lychee-like odor (cf.Table 1).The incomplete odor reconstitution models were then compared to the complete odor reconstitution models in quantitative olfactory profile analyses.Surprisingly, the omission of (2S,4R)-rose oxide did not result in a substantially reduced intensity rating for the lychee note, in either the Muscaris or the Muskateller model (Figure 2).Using paired t-tests, p-values of 0.8 (Muscaris) and 0.1 (Muskateller) were calculated, suggesting that there was no significant difference in the lychee note.In contrast, when geraniol (26) was additionally omitted, the intensity rating for the lychee note dropped considerably: in the Muscaris model by 0.6 units and in the Muskateller model by 0.7 units.The ttests resulted in p-values of 0.03 (Muscaris) and 0.01 (Muskateller), suggesting a significant difference between the complete and incomplete reconstitution models.Interestingly, when only geraniol was omitted, the effect on the lychee note was small, similar to the effect observed when only (2S,4R)rose oxide was omitted.p-Values were 1 in the Muscaris models and 0.2 in the Muskateller models.Thus, it became obvious that the lychee note in the Muscaris grapes, but also in the Muskateller grapes, was generated by the combination of (2S,4R)-rose oxide and geraniol.Whereas (2S,4R)-rose oxide showed a specific lychee odor but rather moderate OAVs of 19 and 26 in Muscaris and Muskateller grapes, respectively, geraniol showed high OAVs of 1100 and 400 in combination with a floral and rosy odor not distinctly lychee-like.To the best of our knowledge, the combinatorial effect of (2S,4R)-rose oxide and geraniol on the lychee note in the aroma of Muscaris and Muskateller grapes has not yet been described in the literature.Further studies are necessary to better understand  the interaction of (2S,4R)-rose oxide and geraniol during olfactory perception.
In summary, this work bridged some of the gaps in the literature.It revealed, for the first time, the major odorants in Muscaris grapes and demonstrated that the combination of (2S,4R)-rose oxide and geraniol was responsible for the characteristic lychee note in the aroma.These results form the basis for further studies, including research on the transfer of Muscaris grape odorants into wine and their role in the hedonic value of wine.Furthermore, the knowledge of the molecular background of the lychee note may be useful in the targeted breeding of new grape varieties with distinct aroma properties.

Figure 1 .
Figure 1.Quantitative olfactory profiles of the odor reconstitution models of Muscaris (A) and Muskateller (B) grapes.Assessors rated the intensity of each descriptor on a scale from 0 to 3 with 0.5 increments and 0 = not detectable, 1 = weak, 2 = moderate, and 3 = strong.

Figure 2 .
Figure 2. Quantitative olfactory profiles of the odor reconstitution models of Muscaris (A) and Muskateller (B) grapes from which (2S,4R)-rose oxide, geraniol, or both were omitted.The complete models depicted in Figure 1 are additionally included for comparison.Assessors rated the intensity of each descriptor on a scale from 0 to 3 with 0.5 increments and 0 = not detectable, 1 = weak, 2 = moderate, and 3 = strong.

Table 1 .
Odorants in the Volatile Isolates Obtained from Muscaris and Muskateller Grapes

Table 2 .
Enantiomeric Distribution of Important Chiral Odorants in Muscaris and Muskateller Grapes a Mean of triplicates.

Table 3 .
Concentrations and OAVs of Important Odorants in Muscaris and Muskateller Grapes Numbering according to Table 1.b Odorants in order of decreasing OAVs in Muscaris grapes.c Mean of duplicates or triplicates; individual values and standard deviations are available in Tables a