Research article Received: 2 May 2014

Revised: 25 June 2014

Accepted: 11 July 2014

Published online in Wiley Online Library

(wileyonlinelibrary.com) DOI 10.1002/jms.3441

Profiling of grape monoterpene glycosides (aroma precursors) by ultra-high performanceliquid chromatography-high resolution mass spectrometry (UHPLC/QTOF) Riccardo Flamini,* Mirko De Rosso, Annarita Panighel, Antonio Dalla Vedova, Fabiola De Marchi and Luigi Bavaresco A ‘suspect screening analysis’ method for grape metabolomics by ultra-high performance-liquid chromatography (UHPLC) and highresolution quadrupole-time of flight (QTOF) mass spectrometry was recently developed. This method was applied to study grape monoterpene glycosides, the main grape aroma precursors. Since standard compounds were not available, they were tentatively identified by overlapping various analytical approaches, in agreement with the indications recommended in mass spectrometry (MS)-based metabolomics. Accurate mass and isotopic pattern, MS/MS fragmentation, correlation between fragments observed and putative structures and between liquid chromatography coupled with mass spectrometry (LC/MS) and gas chromatography/ mass spectrometry signals were studied. Seventeen monoterpene glycosides were identified without performing the hydrolytic artifacts commonly used to study these compounds which may affect sample profile. This is the first time that a detailed study of these aroma precursors has been carried out by direct LC/MS analysis. Copyright © 2014 John Wiley & Sons, Ltd. Keywords: grape; monoterpene glycosides; aroma precursors; time-of-flight mass spectrometry; metabolomics

Introduction

1214

Many grape varieties typically contain monoterpenols, sometimes in large quantities. These compounds occur mainly in the berry skin, but also in the pulp, as both free and glycoside forms.[1–3] In general, free monoterpenes are floral fragrances which are immediately perceptible in grape and wine, due to their low sensory thresholds.[4] The main grape monoterpenols are linalool (3,7-dimethyl-1,6octadien-3-ol), nerol and geraniol [(Z) and (E)-3,7-dimethyl-2,6octadien-1-ol, respectively], α-terpineol (α,α,4-trimethyl-3-cyclohexene-1-methanol), citronellol (3,7-dimethyl-6-octen-1-ol), cis and trans furanlinalool oxides (5-ethenyltetrahydro-α,α,5-trimethyl-2furanmethanol) and cis and trans pyranlinalool oxides (6-ethenyltetrahydro-2,2,6-trimethyl-2H-pyran-3-ol). Monoterpenediols such as cis- and trans-8-hydroxydihydrolinalool [(E) and (Z)-2,6-dimethyl2,7-octadiene-1,6-diol], 7-hydroxynerol [(Z)-3,6-dimethyl-2-octene1,7-diol], 7-hydroxygeraniol [(E)-3,6-dimethyl-2-octene-1,7-diol], Ho-diendiol I (3,7-dimethyl-1,5-octadiene-3,7-diol) and p-menthenediol I (p-menth-1-ene-7,8-diol) may also occur in substantial quantities.[5] In general, in glycoside derivatives, the monoterpene is linked to arabinosyl-glucose, apiosyl-glucose, rhamnosyl-glucose and glucose.[6,7] The substantial presence of these compounds confers intense floral notes to some grape varieties. According to monoterpene content, they are classified as aromatic or semi-aromatic grapes and are included in the Moscati and Malvasie groups or defined as Traminer-type and Riesling-type varieties.[2,8–11] It is known that the release of terpene aglycones from their glycosides is very important to improve organoleptic characteristic

J. Mass Spectrom. 2014, 49, 1214–1222

of wine (i.e. aroma); the knowledge of the profile of these aroma precursors can be determinant for choosing the suitable oenological practices which favor their acid-catalyzed and/or enzymatic hydrolysis during winemaking and aging of wine.[12] For example, at the pH of must and wine, linalool is transformed by acid hydrolysis in α-terpineol and cis and trans 1,8-terpin [4-(2-hydroxypropan2-yl)-1-methylcyclohexan-1-ol], geraniol and nerol in linalool and α-terpineol.[13,14] Also, some non-floral diols such as Ho-diendiol I, Ho-diendiol II (3,7-dimethyl-1,7-octadiene-3,6-diol), endiol (3,7-dimethyl-1-octene-3,7-diol) and 3,7-dimethyl-1-octene-3,6,7-triol contribute to organoleptic characteristics of wines because they are the precursors of other compounds such as nerol oxide [4methyl-2-(2-methylprop-1-enyl)-3,6-dihydro-2H-pyran], rose oxide [tetrahydro-4-methyl-2-(2-methylpropenyl)-2H-pyran] and anhydrofuran and anhydropyran derivatives.[15,16] Enantiospecific reduction of geraniol and nerol by yeasts promotes the formation of (R)-(+)-citronellol.[13,17] Oenological glycosidic enzymes are used in wine-making to liberate the aromatic aglycones.[7,14,18,19] For this aim and for analytical purposes, the knowledge of the nature of the sugar residues linked to the aroma compound is essential, in order to select the most suitable enzyme for hydrolysis of the precursor.

* Correspondence to: Riccardo Flamini, Consiglio per la Ricerca e la Sperimentazione in Agricoltura, Centro di Ricerca per la Viticoltura (CRA-VIT), Viale XXVIII aprile 26, 31015 Conegliano (TV), Italy. E-mail: [email protected] Consiglio per la Ricerca e la Sperimentazione in Agricoltura, Centro di Ricerca per la Viticoltura (CRA-VIT), Viale XXVIII aprile 26, 31015, Conegliano, TV, Italy

Copyright © 2014 John Wiley & Sons, Ltd.

UHPLC/QTOF profiling of grape monoterpene glycosides

J. Mass Spectrom. 2014, 49, 1214–1222

structure was determined by using the software ‘Molecular Structure Correlator’ for defining molecular structure. Moreover, GC/MS analyses of aglycones produced by enzymatic hydrolysis of grape extracts were also carried out for further confirming the compounds identification by crossing with the LC/QTOF data. Finally, the identified monoterpene glycosides were added to GrapeMetabolomics database.

Materials and methods Standards and reagents Heptyl glucoside and 1-heptanol were purchased from SigmaAldrich and Fluka (Milan, Italy), respectively. Methanol Super Purity Solvent (SpSTM), formic acid SpSTM, dichloromethane SpSTM and acetonitrile Ultra Purity Solvent (UpSTM) were purchased from Romil (Cambridge, GB). Citric acid RPE-ACS grade was purchased from Carlo Erba Analyticals (Milan Italy); Na2HPO4 dihydrated analytical grade (99.5%) was purchased from Riedel-de Haën (Germany). Sample preparation About 100 berries of Vitis vinifera grape varieties White Muscat, Riesling Italico and Glera were harvested in 2011 and 2012 at full technological ripeness (corresponding to the maximum sugar content measured during the grape ripeness by refractometer) from the CRA-VIT grapevine Germoplasm Collection (Susegana, Veneto, Italy). Berries were picked randomly from five different plants and immediately frozen at 20 °C. Fifty grape berries were weighed, and the pulps were separated from the skins and seeds. The skins were extracted with 35 ml of methanol for 4 h; then, the solution was homogenized and centrifuged at 4000 rpm/min for 10 min. To remove methanol, the solution was concentrated to 10 ml at 40 °C by rotary evaporator (model Laborota 4000, Heidolph, Germany), and the residue was adjusted to 100 ml by water. To remove polyphenols and tannins, the solution was treated with 1 g Polyclar AT under stirring for 20 min, and after centrifugation at 4000 g for 10 min the limpid supernatant was recovered. Pulps were added of 50 mg sodium metabisulphite, the solution was homogenized then centrifuged at 4000 rpm/min for 10 min. The volume was adjusted to 100 ml by water, and the solution was clarified with 40 mg of pectolytic enzyme without secondary glycosidase activity, Pectazina DC (Dal Cin, Concorezzo, MB, Italy), at room temperature for 4 h, then centrifuged. To perform LC/MS analysis, 10 ml of skin extract was combined with an equal volume of limpid pulp juice, and 200 μl of heptyl glucoside methanolic solution (40 mg/l) was added to the resulting solution as per internal standard. The solution was passed through a 1-g Sep-Pak C18 cartridge (Waters) previously activated by three consecutive passages of 5-ml dichloromethane, 5-ml methanol and 5-ml water, respectively. After cartridge washing with 5 ml of water, followed by 5 ml of dichloromethane, the glycoside compounds fraction was recovered with 5-ml methanol. The solution was filtered on GHP Acrodisc 0.22-μm filter (Waters) and collected in a vial for LC/MS analysis. For each variety, two samples were analyzed. The content of monoterpene glycosides was calculated as mg of heptyl glucoside per kg of grape. Sample preparation for GC/MS analysis was performed using the remaining 90 ml of skin extract combined with an equal volume of pulp juice and using the methods previously reported.[20] Briefly, 200 μl of 1-heptanol ethanolic solution (180 mg/l) was added as

Copyright © 2014 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/jms

1215

The traditional method of studying monoterpene glycosides and, generally, wine aroma precursors is by hydrolysis of extracts through enzymes with strong secondary glycosidase activity and identification of the aglycones by GC/MS.[5,20,21] Although this method cannot examine the aroma precursors in their integral form, it can identify the aglycones in the sample, but no information is available on the nature of the sugars linked to the monoterpenes. This approach also requires a hydrolytic artifact which may affect the sample profile. In general, liquid chromatography coupled with mass spectrometry (LC/MS) is suitable for studying non-volatile and labile compounds with high selectivity and sensitivity. Several ESI-MS and LC/MS methods have been used to study the metabolomics of natural extracts by both ‘untargeted’ and ‘targeted’ approaches.[22–27] In general, the former approach provides high sensitivity, good resolution and high-throughput capacity; instead, ‘targeted’ methods focus more on quantitative study of specific compounds.[28–30] In a recent study, twelve glycosidically bound terpenes were revealed in Moscato Giallo grape juice fractions; namely, (E) and (Z)-furanosyllinalooloxide-7-O-[α-D-apiofuranosyl-(1 → 6)-1-β-D-glucopyranoside], (E)-furanosyl-linalooloxide-7-O-[1-β-D-glucopyranoside], (Z)-8-hyd roxylinalool-8-O-[1-β-D-glucopyranoside], 1,2-dihydroxylinalool-1O-[1-β-D-glucopyranoside], linalool-3-O-[α-L-arabinofuranosyl-(1 → 6)1-β-D-glucopyranoside], linalool-3-O-[α-L-apiofuranosyl-(1 → 6)-1-βD-glucopyranoside], linalool-3-O-[α-L-rhamnopyranosyl-(1 → 6)-1-βD-glucopyranoside], nerol-1-O-α-D-apiofuranosyl-(1 → 6)-1-β-D-glu copyranoside, geraniol-1-O-[α-L-arabinofuranosyl-(1 → 6)-1-β-Dglucopyranoside], geraniol-1-O-[α-L-rhamnopyranosyl-(1 → 6)-1-βD-glucopyranoside] and a geranic acid disaccharide derivative were identified by using LC/MS and NMR methods. In particular, MS/MS had provided the identification of glycoside type, moiety sequence and terpene aglycone in the molecule, while NMR analysis of more concentrated and purified fractions allowed the structural elucidation of the molecules; anyway, the study did not propose a method for qualitative or quantitative analysis of these compounds.[31] The literature contains only another paper describing an LC/MS method for analysis of grape monoterpene glycosides, in which linalool and geraniol in glucoside, arabinosyl-glucoside and rhamnosyl-glucoside forms were determined.[32] A method for grape metabolomics has recently been developed.[33] It is based on a mid-way approach between ‘targeted’ and ‘untargeted’ analysis, called suspect screening analysis.[34] Identification of metabolites uses a new electronic database of putative grape and wine compounds – GrapeMetabolomics – constructed with molecular information from the literature and other electronic databases, such as their molecular formulas, isotopic patterns and structure. This database currently contains more one thousand molecular formulas. Two analyses – in positive and negative ionization mode – can yield the molecular formulas and putative identification of 320–450 compounds in grape extracts (depending on variety), providing a detailed profile of their chemical composition.[33] In the present work, this method was applied to study monoterpene glycosides profile in three aromatic and semi-aromatic grape varieties: White Muscat, Riesling Italico and Glera. Analyses were performed by ultra-high performance-liquid chromatography (UHPLC) and high-resolution quadrupole-time of flight (QTOF) mass spectrometry. Compounds were identified according to accurate mass measurements and isotope patterns, and identification was confirmed by high-resolution multiple mass spectrometry (MS/MS). Correlation between fragments observed and putative

R. Flamini et al. internal standard, and the solution was passed through a 10-g Sep-Pak C18 cartridge (Waters) previously activated using 30-ml solvent aliquots as described above. After washing the cartridge with 50-ml water and 50-ml dichloromethane, the fraction of glycosides was recovered with 30-ml methanol. The solvent was evaporated until dry using a rotary evaporator, the residue was dissolved in 5 ml of citrate-phosphate buffer and 100 mg of AR 2000 enzyme (DSM Food Specialties B.V., Delft, The Netherlands) was added to the solution. Reaction was carried out overnight at 40 °C, then the solution was passed through a 1-g Sep-Pak C18 cartridge previously activated and the aglycones were recovered with 6 ml dichloromethane. Before analysis the solution was concentrated at 0.2 ml using a nitrogen flow. For each variety, two samples were analyzed. The contents of the compounds were calculated as mg of 1-heptanol per kg of grapes.

Santa Clara, CA, US) equipped with a fused silica HPInnowax polyethylene glycol (PEG) capillary column (30 m × 0.25 mm, 0.25 μm i.d.) (Agilent Technologies, Santa Clara, CA, US) coupled with HP 5975C mass spectrometer and 7693A automatic liquid sampler injector (Agilent Technologies, Santa Clara, CA, US). Oven temperature program: 40 °C isothermal for 1 min, increase 2 °C/min until 160 °C, 3 °C/min until 230 °C, 230 °C isothermal for 15 min. Other experimental conditions: injector temperature, 230 °C; carrier gas helium at constant flow, 1.2 ml/min; sample volume, 1 μl; splitless injection mode; transfer line temperature, 250 °C; quadrupole temperature, 150 °C; mass range, m/z 20–550. Compounds were identified by searching in NIST08 (National Institute of Standards and Technology) Mass Spectral Database and the CRA-VIT homemade database of grape volatiles, ESTRATTI.

UHPLC/QTOF mass spectrometry

Results and discussion

The system used was Agilent UHPLC 1290 Infinity equipped with Agilent 1290 Infinity Autosampler (G4226A) and coupled to Agilent 6540 accurate-mass Q-TOF mass spectrometer (nominal resolution 40.000) with Jet Stream Ionization source (Agilent Technologies, Santa Clara, CA). Before each sample, a blank was run. The data acquisition software was Agilent MassHunter version B.04.00 (B4033.2). Chromatographic separation was performed using a Zorbax reverse-phase column (RRHD SB-C18 3 × 150 mm, 1.8 μm) (Agilent Technologies, Santa Clara, CA), and mobile phase was composed of A) 0.1% v/v aqueous formic acid and B) acetonitrile containing 0.1% formic acid (v/v). Gradient elution program: 5% B isocratic for 8 min, from 5% to 45% B in 10 min, from 45% to 65% B in 5 min, from 65% to 90% in 4 min and 90% B isocratic for 10 min. Flow rate 0.4 ml/min. Sample injection, 10 μl. Column temperature, 35 °C. Two repeated analyses of each sample in both positive and negative ionization mode with full scan acquisition (range m/z 100–1700, acquisition rate 2 spectra/s) were performed. QTOF conditions: sheath gas nitrogen 10l/min at 400 °C; drying gas nitrogen flow 8 l/min at 350 °C; nebulizer pressure, 60 psig; nozzle voltage, 1 kV; capillary voltage, 3.5 kV. Negative mass calibration was performed with standard mix G1969-85000 (Supelco Inc.) and had residual error for the expected masses between ±0.2 ppm. Lock masses in negative-ion mode: TFA anion at m/z 112.9856 and HP-0921(+formate) at m/z 966.0007. MS/MS conditions: collision energies between 20 and 60 eV used to fragment the parent ions in the m/z 100–1700 range. Data processing was performed with Agilent MassHunter Qualitative Analysis software version B.05.00 (5.0.519.0). Confidence of compound molecular formula identification was based on accurate mass measurements and isotope pattern and expressed by ‘overall identification score’ computed as a weighted average of the isotopic pattern signals of compound, such as exact masses, relative abundances and m/z distances (spacing). The weights of parameter were: Wmass = 100, Wabundance = 60, Wspacing = 50, mass expected data variation 2.0 mDa + 5.6 ppm, mass isotope abundance 7.5%, mass isotope grouping peak spacing tolerance 0.0025 m/z + 7.0 ppm. Software MassHunter Molecular Structure Correlator (MSC) was used to calculate the correlations between the MS/MS fragment ions observed and the molecular structures proposed. GC/MS analysis

1216

Gas chromatography/mass spectrometry (GC/MS) analysis was performed using a 6850 gas chromatograph (Agilent Technologies,

wileyonlinelibrary.com/journal/jms

Identification of glycosidically bound monoterpenes In negative ionization mode, monoterpene glycosides produced intense signals of both [M H] and [M + HCOO] ions. Figure 1 shows the LC/QTOF extracted chromatogram of [M H] ions recorded for White Muscat grape extract. Seventeen compounds were found with molecular formulas matching those of some monoterpene derivatives (Fig. 1). Initial identification was based on accurate mass measurements of the monoisotopic signal and isotopic pattern of the compounds. Molecular formulae corresponding to more abundant signals were identifiable with a high confidence score (>90%). For the lower signals, molecular formula identification was achieved with lower score. Likewise for these compounds, in any event, accurate mass measurement of the monoisotopic ion signal was achieved with low error (maximum 2 ppm), although the identification score was probably affected by the low signal intensity of the lower isotopic ions (Table 1). The main difficulty in identifying the single monoterpenes was due to the fact that most of them have the same molecular formula and they fall into two groups of isomeric compounds: linalool, nerol, geraniol and α-terpineol with molecular formula C10H18O, 8hydroxylinalool, 7-hydroxynerol, 7-hydroxygeraniol, p-menthenediol I, furan linalool oxides and pyran linalool oxides with molecular formula C10H18O2. With accurate mass measurement, it was therefore possible to distinguish three different groups of glycoside terpenols: arabinosyl (or apiosyl)-hexoside, rhamnosyl-hexoside and arabinosyl (or apiosyl)-hexoside diols. Due to the presence of isomeric compounds, and the lack of both available standards and LC/MS studies in literature, further attempts to identify single monoterpenes were made by performing MS/MS experiments. The main fragments observed for each compound are listed in Table 2. For 15 putative monoterpenols, the aglycone or aglycone-hexose fragments produced from the [M-H]- precursor ion were observed. Only two isomeric rhamnosyl-hexoside derivatives, with retention times of 17.51 and 17.61 min, respectively, and molecular formula C22H38O10, did not show these two diagnostic fragments. Anyway, some diagnostic sugar fragments were observed for all the precursor ions. Figure 2 shows the structures of the putative compounds identified. In general, the MS/MS experiment identified sugar moiety of all the putative compounds: the pentosyl-hexoside derivatives showed the fragments at m/z 161.0455 and 119.0350, the former

Copyright © 2014 John Wiley & Sons, Ltd.

J. Mass Spectrom. 2014, 49, 1214–1222

UHPLC/QTOF profiling of grape monoterpene glycosides

Figure 1. LC/QTOF extract ion chromatogram of putative monoterpene glycoside [M bers correspond to compounds tentatively assigned in Table 1.

H] signals of White Muscat grape extract (harvest 2011). Peak num-

Table 1. Putative monoterpene glycosides identified by UPLC/QTOF analysis. Confidence of molecular formula identification (Id. Score) was calculated on accurate mass measurements of [M H] ion and isotopic pattern (weighted average exact masses, relative abundances, m/z distances). Numbers correspond to peaks in Fig. 1 [M Peak

RT (min)

H]

Formula

Experimental mass

Theoretical mass

Error ppm

Id. score

C21H36O11 C21H36O11 C21H36O11 C21H36O11 C21H38O11 C21H36O11 C21H36O11 C21H36O11

463.2192 463.2187 463.2188 463.2188 465.2341 463.2193 463.2195 463.2190

463.2185 463.2185 463.2185 463.2185 465.2341 463.2185 463.2185 463.2185

1.5 0.4 0.6 0.6 0.0 1.7 2.2 1.1

82.4 83.1 82.6 81.7 98.2 81.8 94.6 92.1

C21H36O10 C21H36O10 C13H26O6 C21H36O10 C21H36O10 C21H36O10 C21H38O10

447.2233 447.2247 277.1660 447.2238 447.2243 447.2245 449.2397

447.2236 447.2236 277.1657 447.2236 447.2236 447.2236 449.2392

0.7 2.5 1.1 0.4 1.6 2.0 1.1

99.3 96.4 99.4 98.6 98.8 97.6 98.7

C22H38O10 C22H38O10 C22H38O10

461.2396 461.2393 461.2383

461.2392 461.2392 461.2392

0.9 0.2 2.0

82.8 82.3 85.3

Monoterpendiols pentosyl-hexoside 1 2 3 4 5 6 7 8

13.72 13.96 14.28 14.53 14.59 14.66 15.24 15.39

p-Menthenediol I Furan/pyran linalool oxides Furan/pyran linalool oxides Furan/pyran linalool oxides 7-Hydroxygeraniol/7-hydroxynerol Diendiol I cis/trans 8-hydroxylinalool Isomeric compound Monoterpenols pentosyl-hexoside

9 10 11 13 14 17 18

16.75 17.09 17.25 17.32 17.51 17.65 18.08

α-Terpineol Isomeric compound Heptyl glucoside (internal standard) Linalool Nerol Geraniol Citronellol Monoterpenols rhamnosyl-hexoside

17.29 17.51 17.61

Linalool/nerol/geraniol Isomeric compound Isomeric compound

J. Mass Spectrom. 2014, 49, 1214–1222

Copyright © 2014 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/jms

1217

12 15 16

R. Flamini et al. a

Table 2. Main product ions observed in MS/MS fragmentation of putative monoterpene glycosides (the peak numbers correspond to Fig. 1). Match percentage between MS/MS fragments observed and putative molecular structure calculated by Molecular Structure Correlator software (MSC%) Precursor Peak

[M

H]

m/z fragment

Error

Principal

ppm

Putative ion

Experimental

Theoretical

Formula

447.2233 MS/MS fragment MS/MS fragment 447.2247 MS/MS fragment MS/MS fragment 447.2238 MS/MS fragment MS/MS fragment 447.2243 MS/MS fragment MS/MS fragment 447.2245 MS/MS fragment MS/MS fragment 449.2397 MS/MS fragment MS/MS fragment

315.1809 161.0451 131.0347 315.1819 161.0458 119.0348 315.1808 161.0452 131.0349 315.1804 161.0456 131.0352 315.1817 161.0455 131.0351 317.1967 161.0457 131.0354

315.1813 161.0455 131.0350 315.1813 161.0455 119.0350 315.1813 161.0455 131.0350 315.1813 161.0455 131.0350 315.1813 161.0455 131.0350 317.1970 161.0455 131.0350

C16H27O6 C6H9O5 C5H7O4 C16H27O6 C6H9O5 C4H7O4 C16H27O6 C6H9O5 C5H7O4 C16H27O6 C6H9O5 C5H7O4 C16H27O6 C6H9O5 C5H7O4 C16H29O6 C6H9O5 C5H7O4

1.3 2.5 2.3 1.9 1.9 1.7 1.6 1.9 0.8 2.9 0.6 1.5 1.3 0.0 0.8 0.9 1.2 3.1

Aglycone-hexose Hexose Pentose Aglycone-hexose Hexose Pentose Aglycone-hexose Hexose Pentose Aglycone-hexose Hexose Pentose Aglycone-hexose Hexose Pentose Aglycone-hexose Hexose Pentose

461.2396 MS/MS fragment MS/MS fragment 461.2393 MS/MS fragment 461.2383 MS/MS fragment

315.1804 145.0505 161.0453 145.0501 161.0456 145.0501 161.0451

315.1813 145.0506 161.0455 145.0506 161.0455 145.0506 161.0455

C16H27O6 C6H9O4 C6H9O5 C6H9O4 C6H9O5 C6H9O4 C6H9O5

2.9 0.7 1.2 3.4 0.6 3.4 2.5

Aglycone-hexose Rhamnose Hexose Rhamnose Hexose Rhamnose Hexose

463.2192 MS/MS fragment MS/MS fragment 463.2187 MS/MS fragment MS/MS fragment 463.2188 MS/MS fragment MS/MS fragment MS/MS fragment 463.2188 MS/MS fragment MS/MS fragment 463.2193 MS/MS fragment MS/MS fragment MS/MS fragment 463.2195 MS/MS fragment MS/MS fragment 463.2190 MS/MS fragment MS/MS fragment 465.2341 MS/MS fragment MS/MS fragment MS/MS fragment

169.1231 161.0449 131.0350 331.1765 161.0454 119.0350 331.1769 169.1233 161.0455 131.0349 331.1764 161.0450 131.0353 331.1761 169.1237 161.0458 149.0452 331.1760 161.0453 131.0353 331.1767 161.0451 131.0341 333.1909 161.0449 131.0348 113.0246

169.1234 161.0455 131.0350 331.1762 161.0455 119.0350 331.1762 169.1234 161.0455 131.0350 331.1762 161.0455 131.0350 331.1762 169.1234 161.0455 149.0455 331.1762 161.0455 131.0350 331.1762 161.0455 131.0350 333.1919 161.0455 131.0350 113.0244

C10H17O2 C6H9O5 C5H7O4 C16H27O7 C6H9O5 C4H7O4 C16H27O7 C10H17O2 C6H9O5 C5H7O4 C16H27O7 C6H9O5 C5H7O4 C16H27O7 C10H17O2 C6H9O5 C5H9O5 C16H27O7 C6H9O5 C5H7O4 C16H27O7 C6H9O5 C5H7O4 C16H29O7 C6H9O5 C5H7O4 C5H5O3

1.8 3.7 0.0 0.9 0.6 0.0 2.1 0.6 0.0 0.8 0.6 3.1 2.3 0.3 1.8 1.9 2.0 0.6 1.2 2.3 1.5 2.5 2.6 3.0 3.7 1.5 1.8

Aglycone Hexose Pentose Aglycone-hexose Hexose Pentose Aglycone-hexose Aglycone Hexose Pentose Aglycone-hexose Aglycone Pentose Aglycone-hexose Aglycone Hexose Pentose Aglycone-hexose Hexose Pentose Aglycone-hexose Hexose Pentose Aglycone-hexose Hexose Pentose Pentose

a

MSC %

Monoterpenols pentosyl-hexoside 9

α-Terpineol

10

Isomeric compound

13

Linalool

14

Nerol

17

Geraniol

18

Citronellol

94.0 98.5 97.7

99.2 97.9 98.0 97.5 98.5

97.8 97.1 95.8

Monoterpenols rhamnosyl-hexoside 12

Linalool/nerol/geraniol

15

Isomeric compound

16

Isomeric compound

99.6 96.2 98.1

Monoterpendiols pentosyl-hexoside 1

p-Menthenediol I

2

Furan/pyran linalool oxides

3

Furan/pyran linalool oxides

4

Furan/pyran linalool oxides

6

Diendiol I

7

Cis/trans hydroxylinalool

8

Isomeric compound

5

7-Hydroxygeraniol/7-hydroxynerol

1218 wileyonlinelibrary.com/journal/jms

Copyright © 2014 John Wiley & Sons, Ltd.

99.6 98.1 98.5 99.5 98.1 90.0

J. Mass Spectrom. 2014, 49, 1214–1222

UHPLC/QTOF profiling of grape monoterpene glycosides

Figure 2. Structure of monoterpenes glycosides tentatively identified in the samples: 1) α-terpineol pentosyl-hexoside; 2) linalool pentosyl-hexoside; 3) geraniol/nerol pentosyl-hexoside; 4) citronellol pentosyl-hexoside; 5) linalool rhamnosyl-hexoside; 6) geraniol/nerol rhamnosyl-hexoside; 7) furan linalool oxide pentosyl-hexoside; 8) pyran linalool oxide pentosyl-hexoside; 9) diendiol I pentosyl-hexoside; 10) 8-hydroxylinalool pentosyl-hexoside; 11) 7hydroxygeraniol/7-hydroxynerol pentosyl-hexoside; 12) p-menthenediol I pentosyl-hexoside.

J. Mass Spectrom. 2014, 49, 1214–1222

between GC/MS area signal was correlated with the sum of signal intensity of the corresponding glycosides recorded by LC/MS in the three grape samples (data in Table 3). In particular, linalool, geraniol and nerol aglycones were correlated with the sum of their pentosyl-hexoside and rhamnosyl-hexoside derivatives, and the sums of pyran and furan linalool oxide GC/MS signals with their LC/MS signals. Despite in the case of some compounds, the distribution of GC/MS and LC/MS data had scarce normal distribution

Table 3. Correlation between the GC/MS signal intensity of the aglycones liberated by enzymatic hydrolysis and the sum of their corresponding LC/MS [M-H] signals (r-Pearson coefficients) in the three samples studied Compound p-Menthenediol I Linalool oxides tot. 8-Hydroxylinalool tot. 7-Hydroxygeraniol/7-hydroxynerol Linalool tot. α-Terpineol Nerol tot. Geraniol tot. Diendiol I

Copyright © 2014 John Wiley & Sons, Ltd.

r-Pearson

p

0.999 0.997 1.000 0.982 0.999 0.944 1.000 0.999 1.000

0.03 0.05 0.00 0.12 0.02 0.21 0.02 0.02 0.00

wileyonlinelibrary.com/journal/jms

1219

characteristic of hexose and the latter of pentose; the rhamnosylhexoside derivatives showed the rhamnose fragment at m/z 145.0506 (Table 2). To further confirm the theoretical molecular structure proposed for each compound, the correlation between structure and MS/ MS fragments was calculated by MSC software. The program provided a percentage overall correlation score for each ion fragment, calculated on the energy required to break bonds, the mass accuracy of fragment ions and the overall percentage of fragment ion intensity which can plausibly be explained by substructures. Unfortunately, as the software did not recognize low-intensity signals and grouped together the signals of compounds with very similar fragmentation spectra, close eluting in the chromatograms (e.g. nerol and geraniol), this approach could not be used for all compounds. However, it confirmed the structure of four mono-terpenol pentosyl-hexosides (α-terpineol, linalool, nerol and citronellol), two diol pentosyl-hexosides (cis/trans hydroxylinalool pentosyl-hexoside and the isomeric compound at RT 15.39 min) and linalool rhamnosyl-glucoside. All the identified fragments had high ID scores (90.0–99.6%; Table 2) inferring high correlations with the structure proposed. Last, to confirm the correct identification of aglycones, the correlation between the GC/MS signal intensity of the compounds liberated by enzymatic hydrolysis and that of [M H] ions recorded in LC/MS analysis was studied by calculating r-Pearson coefficients. For the nine aglycones identified, the correlation

R. Flamini et al. Table 4. Peak area percentage normalized of the compounds identified in the grape samples. The percentage error for four repetitions (2011 and 2012 harvests, two samples per harvest), is reported. Bold: Total monoterpene glycosides contents calculated on formate adduct signals and expressed as mg heptyl glucoside/kg grape. n.f., not found Monoterpene glycoside

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

White Muscat

Glera

Riesling Italico

% mean

% mean

% mean

α-Terpineol pentosyl-hexoside Isomeric compound pentosyl-hexoside Linalool pentosyl-hexoside Nerol pentosyl-hexoside Geraniol pentosyl-hexoside Citronellol pentosyl-hexoside Linalool/nerol/geraniol rhamnosyl-hexoside Isomeric compound rhamnosyl-hexoside Isomeric compound rhamnosyl-hexoside p-Menthenediol I pentosyl-hexoside Furan/pyran linalool oxides pentosyl-hexoside tot. Diendiol I pentosyl-hexoside Cis/trans 8-hydroxylinalool pentosyl-hexoside Isomeric diol pentosyl-hexoside 7-Hydroxygeraniol/7-hydroxynerol pentosyl-hexoside

5.2 ± 0.8 5.1 ± 1.2 32.6 ± 4.1 100.0 ± 28.3 36.5 ± 9.5 1.1 ± 0.3 6.4 ± 0.1 1.4 ± 0.3 2.8 ± 0.8 1.3 ± 0.4 23.2 ± 2.9 45.3 ± 3.1 81.9 ± 1.0 n.f. 17.4 ± 4.2

6.5 ± 2.1 5.3 ± 0.1 5.1 ± 0.0 27.9 ± 9.6 23.8 ± 6.6 n.f. 2.6 ± 0.1 n.f. 2.9 ± 0.6 8.8 ± 1.7 100.0 ± 12.6 n.f. 14.7 ± 0.6 n.f. 42.3 ± 10.6

47.6 ± 13.4 n.f. 1.9 ± 0.4 11.7 ± 2.4 41.0 ± 0.6 n.f. n.f. n.f. n.f. 88.7 ± 28.9 100.0 ± 37.2 100.0 ± 77.9 20.5 ± 4.0 42.2 ± 9.6 28.5 ± 3.6

Total monoterpene glycosides (mg IS/kg grape)

82.30 ± 3.16

11.72 ± 0.65

9.10 ± 0.39

(in particular total 8-hydroxylinalool and diendiol I), and the higher p value of some compounds was probably linked to the low number of samples studied, the high correlations found gave a further reasonable confirmation of the putative assignments in Table 1. Differing to other studies, the presence of glucoside derivatives found in our samples was not significant.[30,32] At all events, this finding is generally in accordance with a lower presence of terpenol glucosides in grapes reported similarly by Ribéreau-Gayon et al.[7]

Samples profiling

1220

The profile and semi-quantitative data of monoterpene glycosides in the three grape varieties were determined. A semi-quantitative index was calculated by relating the signal intensity of each analyte with that of the internal standard and expressed as mg of heptyl glucoside per kg of grapes. Unfortunately, a significant difference of ionization yield in producing the [M H] ion between the internal standard and the analytes was observed, due probably to the fact that the standard is a saturated molecule whereas all the analytes are unsaturated compounds. By contrast, the formation of an intense formate adduct ([M + HCOO] ) was observed for both the internal standard and the analytes, in agreement with existing data on the negative ionization of mono- and di-saccharides.[35] For example, by using the [M + HCOO] ion signals, the analyte/IS area ratio decreased by around 3.3-fold for linalool pentosyl-hexoside, and fivefold for geraniol pentosyl-hexoside. Consequently, the semi-quantitative index was calculated by comparing the signal intensity of [M + HCOO] ion of the analytes with that of internal standard. In any event, resulting data were higher than those of aglycones determined by GC/MS and expressed as mg 1-heptanol per kg of grapes. This was explained by: (1) possible artifacts occurring in sample hydrolysis, (2) different ionization yield of heptyl glucoside with respect to analytes and (3) specific selectivity of the glycosidic enzyme used for some compounds, with consequent loss of the others.

wileyonlinelibrary.com/journal/jms

Table 4 lists the data percentage normalized of the compounds identified in each sample. White Muscat showed the highest occurrence of rhamnosyl-hexoside derivatives; nerol-pentosyl-hexoside and 8-hydroxylinalool were the higher compound found. The LC/ MS chromatogram of Riesling Italico showed a peak at retention time 15.39 min with molecular formula C21H35O11 (MW 463.2185), not found in the other two varieties, which was putatively assigned to a pentosyl-hexoside diol. Diendiol I pentosyl-hexoside and total pyran linalool oxides were the more abundant compound found in this sample, but the former showed high variability among the biological repetitions analyzed (100 ± 77.9%, Table 4). As expected for an aromatic grape variety, White Muscat had the highest contents of monoterpene glycosides expressed as mg heptyl glucoside/kg grape, particularly linalool, geraniol and nerol in pentosyl-hexoside and rhamnosyl-hexoside forms.[10,36] The profiles of Riesling Italico and Glera (the grape variety used to produce Prosecco wine) showed minor but significant contents of monoterpenols and confirmed the semi-aromatic character of these varieties.[37–39] As all samples had been grown in the same vineyard, they were presumably not affected by viticultural or environmental variables, and the differences found are essentially due to their varietal expression and – to a lesser extent – the two different harvests.[10]

Conclusions Suspect screening metabolomics identified 17 grape monoterpene derivatives without applying the hydrolytic artifacts which are commonly used to study these compounds but which may affect sample profiles. Various analytical methods (accurate mass and isotopic pattern measurements, MS/MS, calculation of correlations between fragments observed and putative structure, matching between LC/MS and GC/MS signals) were overlapped, and compounds were ‘putatively identified by a match to database mass

Copyright © 2014 John Wiley & Sons, Ltd.

J. Mass Spectrom. 2014, 49, 1214–1222

UHPLC/QTOF profiling of grape monoterpene glycosides spectrometry tags’, matching the recommendations of Scalbert et al. (2009) for compound identification in MS-based metabolomics.[40] This is the first time that a detailed study of these aroma precursors has been carried out by direct LC/MS analysis. Due to the lack of standards commercially available and use of a saturated compound as IS, just a semi-quantitative study has been possible. The quantitative accuracy of method could be increased by using a glycoside unsaturated monoterpene as IS, and by improving the chromatographic resolution. Anyway, the semi-quantitative method proposed here is useful, especially for comparative studies of various samples. For example, it can be used for (1) characterizing grape varieties according to the profiles of secondary metabolites, (2) studying climate change and cultural variable effects (e.g. growing areas or cultural practices), (3) improving and refining enological techniques (e.g. by monitoring grape compound extraction in wine-making and studying the evolution of aroma precursors during wine aging) and (4) estimating the aromatic potential of various kinds of grapes.

References

J. Mass Spectrom. 2014, 49, 1214–1222

Copyright © 2014 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/jms

1221

[1] Y. Z. Gunata, C. L. Bayonove, R. L. Baumes, R. E. Cordonnier. The aroma of grapes. Localisation and evolution of free and bound fractions of some grape aroma components c.v. Muscat during first development and maturation. J. Sci. Food Agric. 1985, 36, 857. [2] B. Wilson, C. R. Strauss, P. J. Williams. The distribution of free and glycosidically-bound monoterpenes among skin, juice and pulp fractions of some white grape varieties. Am. J. Enol. Vitic. 1986, 37, 107. [3] E. Gomez, A. Martinez, J. Laencina. Localization of free and bound aromatic compounds among skin, juice and pulp fractions of some grape varieties. Vitis 1994, 33, 1. [4] A. Rapp, H. Hastrich, L. Engel, W. Knipser. In Flavor of food and beverages, G. Charalambous (Ed). Academic Press: New York, 1978, pp. 391–417. [5] R. Flamini, P. Traldi. In Applications of Mass Spectrometry in Grape & Wine, R. Flamini, P. Traldi (Eds). John Wiley & Sons Inc.: Hoboken, 2010, pp. 97–116. [6] P. J. Williams, C. R. Strauss, B. Wilson, R. A. Massy-Westropp. Use of C18 reversed-phase liquid chromatography for the isolation of monoterpene glycosides and nor-isoprenoid precursors from grape juice and wines. J. Chromatogr. A 1982, 235, 471. [7] P. Ribéreau-Gayon, Y. Glories, A. Maujean, D. Dubourdieu. Traité d’Oenologie 2. Chimie du vin – stabilisation et traitements. Dunod, Paris, 1998. [8] G. Versini, S. Inama, G. Sartori. A capillary golumn gaschromatographic research into the terpenesconstituents of "Riesling Renano" (Rhine Riesling) wine from Trentino AltoAdige: Their distribution within berries, their passage into must and their presence in the wine according to different wine-making procedures. Organoleptic considerations. Vini d’Italia 1981, 23, 189. [9] C. Bayonove, Z. Gunata, R. Cordonnier. Mise en evidence de l’intervention des enzymes dans le developpement de l’arome du jus de muscat avant fermentation: la production des terpenols. Bulletin de l’OIV 1984, 57, 741. [10] R. Flamini, A. Dalla Vedova, A. Calò. Study onmonoterpene contents of 23 accessions of Muscat grape: correlation between aroma profile and variety. Riv. Vitic. Enol. 2001, 2/3, 35. [11] D. Borsa, D. Carniel, A. Asproudi, L. Monticelli, M. Crespan, A. Costacurta. Characterization of Malvasia grapes by secondary metabolites. Rivista di Viticoltura e di Enologia 2005, 58, 167. [12] A. Rapp, H. Mandery, H. Niebergall. Neue Monoterpendiole in Traubenmost und Wein sowie in Kulturen von Botrytis cinerea. Vitis 1986, 23, 79. [13] R. Di Stefano, G. Maggiorotto, S. Gianotti. Transformation of nerol and geraniol induced by yeasts (Italian). Rivista di Viticoltura e di Enologia 1992, 45, 43. [14] S. Maicais, J. J. Mateo. Hydrolysis of terpenyl glycosides in grape juice and other fruit juices: a review. Appl. Microbiol. Biotechnol. 2005, 67, 322.

[15] P. J. Williams, C. R. Strauss, B. Wilson. Hydroxylated linalool derivatives as precursors of volatile monoterpenes of Muscat grapes. J. Agric. Food Chem. 1980, 28, 766. [16] C. R. Strauss, B. Wilson, P. R. Gooley, P. J. Williams. In Biogenertion of aromas, T. H. Parliment, R. Croteau (Eds), ACS Symp. Series 317. ACS Pubblications: Washington DC, 1986, pp. 222–242. [17] P. Gramatica, P. Manitto, B. M. Ranzi, A. Dalbianco, M. Francavilla. Stereospecific reduction of geraniol to (R)-(+)-citronellol by Saccharomyces cerevisiae. Experientia 1982, 38, 775. [18] H. Michlmayr, S. Nauer, W. Brandes, C. Schümann, K. D. Kulbe, A. M. del Hierro, R. Eder. Release of wine monoterpenes from natural precursors by glycosidases from Oenococcus oeni. Food Chem. 2012, 135, 80. [19] P. González-Pombo, L. Fariña, F. Carrau, F. Batista-Viera, B. M. Brena. Aroma enhancement in wines using co-immobilized Aspergillus niger glycosidases. Food Chem. 2014, 143, 185. [20] J. J. Mateo, N. Gentilini, T. Huerta, M. Jimenez, R. Di Stefano. Fractionation of glycoside precursors of aroma in grapes and wine. J. Chromatogr. A 1997, 778, 219. [21] A. T. Toci, P. Crupi, G. Gambacorta, T. Dipalmo, D. Antonacci, A. Coletta. Free and bound aroma compounds characterization by GC-MS of Negroamaro wine as affected by soil management. J. Mass Spectrom. 2012, 47, 1104. [22] L. Mattoli, F. Cangi, A. Maidecchi, C. Ghiara, E. Ragazzi, M. Tubaro, L. Stella, F. Tisato, P. Traldi. Metabolomic fingerprinting of plant extracts. J. Mass Spectrom. 2006, 41, 1534. [23] R. C. De Vos, S. Moco, A. Lommen, J. J. Keurentjes, R. J. Bino, R. D. Hall. Untargeted large-scale plant metabolomics using liquid chromatography coupled to mass spectrometry. Nat. Protoc. 2007, 2, 778. [24] J. L. Wolfender, G. Marti, E. Ferreira Queiroz. Advances in techniques for profiling crude extracts and for the rapid identification of natural products: dereplication, quality control and metabolomics. Curr. Org. Chem. 2010, 14, 1808. [25] B. Houshyani, P. Kabouw, D. Muth, R. C. de Vos, R. J. Bino, H. J. Bouwmeester. Characterization of the natural variation in Arabidopsis thaliana metabolome by the analysis of metabolic distance. Metabolomics 2012, 8, 131. [26] J. J. van der Hooft, R. C. de Vos, L. Ridder, J. Vervoort, R. J. Bino. Structural elucidation of low abundant metabolites in complex sample matrices. Metabolomics 2013, 9, 1009. [27] K. Dettmer, P. A. Aronov, B. D. Hammock. Mass Spectrometry based metabolomics. Mass Spectrom. Rev. 2007, 26, 51. [28] P. Arapitsas, M. Scholz, U. Vrhovsek, S. Di Blasi, A. Biondi Bartolini, D. Masuero, D. Perenzoni, A. Rigo, F. Mattivi. A metabolomic approach to the study of wine micro-oxygenation. PLoS One 2012, 7(5), e37783. [29] A. Cuadros-Inostroza, P. Giavalisco, J. Hummel, A. Eckardt, L. Willmitzer, H. Peña-Cortés. Discrimination of wine attributes by metabolome analysis. Anal. Chem. 2010, 82, 3573. [30] L. Vaclavik, O. Lacina, J. Hajslova, J. Zweigenbaum. The use of high performance liquid chromatography-quadrupole time-of-flight mass spectrometry coupled to advanced data mining and chemometric tools for discrimination and classification of red wines according to their variety. Anal. Chim. Acta 2011, 685, 45. [31] E. Schievano, M. D’Ambrosio, I. Mazzaretto, R. Ferrarini, F. Magno, S. Mammi, G. Favaro. Identification of wine aroma precursors in Moscato Giallo grape juice: A nuclear magnetic resonance and liquid chromatography–mass spectrometry tandem study. Talanta 2013, 116, 841. [32] A. Nasi, P. Ferranti, S. Amato, L. Chianese. Identification of free and bound volatile compounds as typicalness and authenticity markers of non-aromatic grapes and wines through a combined use of mass spectrometric techniques. Food Chem. 2008, 110, 762. [33] R. Flamini, M. De Rosso, F. De Marchi, A. Dalla Vedova, A. Panighel, M. Gardiman, I. Maoz, L. Bavaresco. An innovative approach to grape metabolomics: stilbene profiling by suspect screening analysis. Metabolomics 2013, 9, 1243. [34] M. Krauss, H. Singer, J. Hollender. LC–high resolution MS in environmental analysis: from target screening to the identification of unknowns. Anal. Bioanal. Chem. 2010, 397, 943. [35] G. Verardo, I. Duse, A. Callea. Analysis of underivatized oligosaccharides by liquid chromatography/electrospray ionization tandem mass spectrometry with post-column addition of formic acid. Rapid Commun. Mass Spectrom. 2009, 23, 1607. [36] G. Mazza, P. Cascio, E. Barbieri. Composti volatili liberi e glicoconiugati presenti nelle foglie e negli acini della Vitis vinifera cv. Moscato bianco. Rivista di Viticoltura e di Enologia 2003, 56, 57.

R. Flamini et al. [37] A. Rapp, M. Guentert, W. Heimann. Characterization of the grape variety White Riesling. I. Investigation of the volatiles of foreign white wines with the name Riesling, but not actually produced from the variety White Riesling. Z. Lebensm. Unters. Forsch. 1985, 181, 357. [38] D. Tomasi, R. Flamini, A. Calò, M. Pol, A. Dalla Vedova, A. Costacurta. Studio dei precursori di aromi varietali contenuti nel mosto di uva Prosecco. L’Enologo 2000, 6, 97.

[39] R. Flamini, A. Dalla Vedova, R. Castagner, M. Salvador. Il profilo aromatico della grappa di Prosecco: gli aromi dalle vinacce al distillato. L’Enologo 2002, 12, 89. [40] A. Scalbert, L. Brennan, O. Fiehn, T. Hankemeier, B. S. Kristal, B. van Ommen, E. Pujos-Guillot, E. Verheij, D. Wishart, S. Wopereis. Massspectrometry-based metabolomics: limitations and recommendations for future progress with particular focus on nutrition research. Metabolomics 2009, 5, 435.

1222 wileyonlinelibrary.com/journal/jms

Copyright © 2014 John Wiley & Sons, Ltd.

J. Mass Spectrom. 2014, 49, 1214–1222

QTOF).

A 'suspect screening analysis' method for grape metabolomics by ultra-high performance-liquid chromatography (UHPLC) and high-resolution quadrupole-ti...
387KB Sizes 1 Downloads 11 Views