25Acta Chim. Slov. 2007, 54, 25–32 Prosen et al.: Analysis of Free and Bound Aroma Compounds in Grape Berries ... 1. Introduction Wine aroma is one of the key properties determining its quality. It depends on several factors, e.g. first aroma of grapes, reactions during extraction of juice and macera- tion (hydrolysis, oxidation), biochemical reactions during fermentation, chemical and enzymatic reactions during the aging of wine. The most important of these factors is the grape aroma, which again depends on the grape va- riety and quality, soil and climate properties and also the winegrowing practice.1 Volatile compounds constituting the aroma of fresh grapes fall into several chemical families: terpenes, nori- soprenoids, alcohols and polyols, aldehydes, organic acids, esters, methoxypyrazines, sulphur compounds.1–7 The greatest contribution to a pleasant olfactory percep- tion of aroma comes from the group of terpenes and nori- soprenoids, while sulphur compounds and methoxypyra- zines have a more unpleasant smell, but if carefully balan- ced add a more distinctive character to some grape varie- ties.1,8–9 However, the sensory perception of grape aroma tells nothing of the actual concentrations of the involved compounds, as their olfactory thresholds can differ for a factor of 106 or more.1,8–9 Low threshold values, meaning high contribution to the aroma in spite of low concentra- tion, are typical for some monoterpene alcohols, met- hoxypyrazines,1 sulphur compounds, norisoprenoids and some esters.9 The analytical method of choice for volatile aroma compounds is gas chromatography (GC), preferably with mass spectrometric (MS) detection,3–4,7–17 although flame ionisation detection (FID) is also used.5,9,11,18–21 It is some- times combined with olfactometry8 or aroma extract dilu- tion analysis (AEDA)9 in order to better evaluate the con- tribution of each compound to the overall aroma. Diffe- rent extraction techniques are employed before the actual analysis with the aim of separating the compounds of inte- rest from the matrix and to pre-concentrate them. While purge-and-trap,14 liquid-liquid extraction (LLE)8–9,14, 16–17,21–22 and solid-phase extraction (SPE)5,7,9–10,18–19 are Scientific paper Analysis of Free and Bound Aroma Compounds in Grape Berries Using Headspace Solid-Phase Microextraction with GC-MS and a Preliminary Study of Solid-Phase Extraction with LC-MS Helena Prosen,a* Lucija Jane{,a Matija Strli~,a Denis Rusjan,b Drago Ko~ara a University of Ljubljana, Faculty of Chemistry and Chemical Technology, A{ker~eva 5 SI-1000 Ljubljana, Slovenia. Tel.: +38612419176, Fax: +38612419220, E-mail: helena.prosen@fkkt.uni-lj.si. b University of Ljubljana, Biotechnical Faculty, Jamnikarjeva 101, SI-1000 Ljubljana, Slovenia Received 20-11-2006 Paper based on a presentation at the 12th International Symposium on Separation Sciences, Lipica, Slovenia, September 27–29, 2006. Abstract An extraction procedure for the aroma compounds from musts and wines has been developed, using solid-phase mi- croextraction (DVB/CAR/PDMS fibre) from the headspace of heated samples (50 °C). Analysis was performed with gas chromatography – mass spectrometry. The method was applicable to the analysis of different aroma compounds (aliphatic, aromatic aldehydes, terpenes) in a broad concentration range (1–5000 µg L–1). A stir-bar sorptive extraction procedure was also tested, but was not effective enough due to the lack of a suitable desorption device. Free aroma com- pounds in must samples of different grape varieties were analysed, and their release was evaluated after enzymatic or acidic hydrolysis. Different hydrolytic approaches were tested and the most successful was the enzymatic hydrolysis (with two different enzymes) and acidic hydrolysis at pH 3. Acidic hydrolysis at pH 1 resulted in substantial decompo- sition and re-arrangement reactions of terpenes. Non-hydrolysed terpene glycosides were extracted from the musts us- ing solid-phase extraction and the extract was analysed with liquid chromatography – mass spectrometry (electrospray interface). Some compounds were tentatively identified as terpene glycosides. Keywords: food aroma, grape, wine, extraction, chromatography, mass spectrometry, 26 Acta Chim. Slov. 2007, 54, 25–32 Prosen et al.: Analysis of Free and Bound Aroma Compounds in Grape Berries ... already well-established for this purpose, several publica- tions have recently appeared using novel extraction tech- niques, e.g. simultaneous distillation-extraction (SDE),15 solid-phase microextraction (SPME)4,9,11–13,21 and stir bar sorptive extraction (SBSE).3,15,21 Especially SPME and SBSE are promising as they are simple-to-use, solvent- free and efficient. Beside the free volatiles constituting the aroma, odoriferous compounds, especially monoterpenols, terpe- ne polyols and norisoprenoids are also present in grapes in the form of diglycosides, involving glucose, arabinose, rhamnose and apiose.1–2,6,23 These glycosides are non-vo- latile and odourless, but rather constitute the aroma poten- tial of a grape variety since they are present in much hig- her concentrations than the free flavour compounds.2 Only a small part of them is hydrolysed during the various stages of wine production, e.g. due to endogenous or yeast glycosidases1 or because of the acidic medium during wi- ne ageing.2 However, odoriferous compounds can be re- leased from their glycosides by adding exogenous glyco- side hydrolases to the must or wine,2,6,23–24 resulting in an enhanced bouquet and therefore superior wine quality. Presently, the established method for diglycoside analysis in grapes involves an extraction step, usually SPE,10,20,22,25 followed by enzymatic2,10,16,20,22,24,25 or aci- dic hydrolysis.2,10 After the hydrolysis, free aglycones are extracted (SPE) and analysed, mainly by GC-MS.10,25 Ot- her methods are rarely used, e.g. enzyme analysis of libe- rated glucose after hydrolysis,20 Fourier-transform infra- red spectrometry,22 droplet countercurrent chromato- graphy or HPLC with fast atom bombardment tandem mass spectrometry.26 Our aim in the present work was to develop a fast and simple method for the screening of the flavour com- pounds in grapes. We have also tried some hydrolytic ap- proaches for the release of bound flavour components from their glycosilated form in order to assess the aroma potential of different grape varieties. Three aromatic and three non-aromatic varieties were chosen to compare the differences in their aroma potential. The emphasis was put on locally grown varieties, for which the aroma potential has – to our knowledge – never been assessed before. We also report the preliminary results of LC-MS analysis of terpene disaccharides extracted from grape berries. 2. Experimental 2.1. Materials Standard compounds used in this study were mono- terpenes α-terpineol (99% purity), nerol (90% purity), geraniol (96% purity), linalool (97% purity), all from Fluka (Buchs, Switzerland); as well as aldehydes benzal- dehyde (puris., Riedel-de Haën, Seelze, Germany), hexa- nal (98% purity, PolyScience, Niles, IL, USA), and (E)- 2-hexen-1-al (97% purity, Fluka, Buchs, Switzerland). Solvents n-hexane, acetone, methanol and acetonitrile were of HPLC grade purity, obtained from Sigma-Al- drich (Steinheim, Germany). Other chemicals used were of p.a. grade quality from different producers. Pectinoly- tic enzymes for wine clarification with side glycosidase activity were Rohavin MX from AB Enzymes (Darm- stadt, Germany) and Lallzyme BETA from Lallemand (St. Simon, France). For solid-phase extraction, Supelclean C18 (1 g) ex- traction cartridges from Supelco (Bellefonte, PA, USA) were used. For stir bar sorptive extraction (SBSE) a stir bar coated with polydimethylsiloxane (PDMS) coating, dimensions 20 × 1.0 mm from Gerstel (Mülheim an der Ruhr, Germany) was used. For solid-phase microextrac- tion, manual holder was used and fibres with different coatings: polydimethylsiloxane (PDMS), 100 µm, polya- crylate (PA), 85 µm, polydimethylsiloxane-divinylbenze- ne (PDMS/DVB), 65 µm, Carbowax-divinylbenzene (CW/ DVB), 65 µm, divinylbenzene-Carboxen-polydi- methylsiloxane (DVB/CAR/PDMS), 50/30 µm, all from Supelco (Bellefonte, PA, USA). For gas chromatography, helium (>99.999%) from Messer (Gumpoldskirchen, Austria) was used. The analy- tical capillary column was VOCOL, dimensions 60 m x 0.25 mm (i.d.), film thickness 1.5 µm, from Supelco (Bel- lefonte, PA, USA). Different grape varieties were screened for their fla- vour compounds and their glycoside precursors: aromatic varieties “Aurora”, “Beograjska rana”, “Muscat blanc”, and non-aromatic varieties “Perlette”, “Pinot noir”, “Da- nijela”. The listed varieties were selected to be as different as possible: locally and globally known, red and white, more and less aromatic, wine and table grape varieties. 2.2. Instrumentation Gas chromatograph was HP 5890 Series with mass spectrometric detector (MSD) 6890 from Hewlett-Pac- kard (Palo Alto, CA, USA). For LC-MS/MS experiments, liquid chromatograph Perkin Elmer Series 200 from Perkin Elmer (Shelton, CT, USA) and 3200 QTRAP LC/MS/MS System equipped with ESI and APCI ion sources from Applied Biosy- stems/MDS Sciex (Foster City, CA, USA) were used. The analytical balance was Mettler Toledo MX5 (Mettler Toledo, Kuesnacht, Switzerland), We used a Visi- prep SPE Vacuum Manifold from Supelco (Bellefonte, PA, USA). 2.3. Preparation of Standard Solutions Stock standard solutions of compounds were prepa- red by dissolving the weighed solid standard in methanol to obtain the concentration of 0.8–1.2 g L–1. These solu- tions were kept in the refrigerator and were stable for se- veral months. They were further diluted with n-hexane or 27Acta Chim. Slov. 2007, 54, 25–32 Prosen et al.: Analysis of Free and Bound Aroma Compounds in Grape Berries ... methanol to obtain working solutions for injection in gas chromatograph or water for extraction optimisation. 2.4. Solid-Phase Extraction (SPE) Extraction cartridge was conditioned with 5 mL of methanol and 5 mL of deionised water. A filtered homo- genisate of grape berries (cca. 20 g) diluted to 100 mL with deionised water was passed under vacuum through the cartridge using the vacuum manifold. The cartridge was rinsed with 20 mL of deionised water. Free aroma compounds were eluted with 8 mL of n-hexane and the re- maining water was sorbed on solid sodium sulphate. The solvent was transferred into a conical test tube and evapo- rated to 2 mL under the stream of nitrogen on a water bath (approx. 50 °C). Glycoside fraction was eluted with 10 mL of metha- nol. The solvent was transferred into a conical test tube and evaporated to dryness under the stream of nitrogen on a water bath (approx. 50 °C). The dry residue was recon- stituted in 1 mL of methanol for subsequent LC analysis. 2.5. Stir-Bar Sorptive Extraction (SBSE) A tentative procedure for SBSE without the thermal desorption unit was as follows: a sample solution (5 mL) was extracted for 30 min on a magnetic stirrer, stir bar was transferred to 2 mL of n-hexane and placed in an ultraso- nic bath for 10 min. The hexane extract (1 µL) was injec- ted into the gas chromatograph. 2.6. Solid-Phase Microextraction (SPME) Fibre with DVB/CAR/PDMS coating was used for the extraction of aroma compounds from grapes. Grape berries were homogenised (cca. 20 g diluted to 100 mL with deionised water) and 5 mL of homogenisate was measured into a 20-mL headspace vial, crimped and ther- mostated for 15 min at 50 °C (for analyses of terpenes and benzaldehyde). An SPME fibre was inserted in the heads- pace and the compounds were sampled for 35 min. The fi- bre was subsequently inserted into the injector port of a gas chromatograph and desorbed for 10 min. The proce- dure for aliphatic aldehydes was the same except that they were sampled at room temperature (25 °C) for 15 min. 2.7. GC-MS Conditions Volatile grape aroma compounds were analysed us- ing a gas chromatograph with mass spectrometric detec- tor. The temperature programme was: 50 °C (2 min) –10 °C min–1 –210 °C (40 min). Temperature of the injec- tor was 250 °C, temperature of the detector was 280 °C. Injection volume of n-hexane/methanol solutions or ex- tracts was 1 µL (splitless). SPME fibre was left in the in- jector for 10 min. In the mass spectrometer, electron impact (EI) ioni- sation was used and the chromatograms were recorded in the total ion current (TIC) mode. Compounds were identi- fied on the basis of their retention times (comparison with standards) and spectra using the searchable EI-MS spectra library (NIST02). The peak area for quantitation was mea- sured either in TIC chromatogram or in an extracted ion chromatogram in the case of coelution with other com- pounds. 2.8. LC-MS Conditions Extracts of glycosides in methanol were analysed using a liquid chromatograph coupled to mass spectrome- ter through electrospray ion source. HPLC column was Hypersil ODS (Agilent Technologies), dimensions 250 mm x 4 mm, 5 µm particles. Mobile phase was com- posed of acetonitrile (phase A) and 0.5% acetic acid in de- ionised water (phase B). Gradient elution was applied: 0–5 min 0% A, 5–60 min from 0% to 90% A and hold for 10 min. Mobile phase flow was 0.8 mL min–1, injection volume was 10 µL. Electrospray ion source (ESI) voltage was 5500 V and the temperature was 400 °C. 2.9. Hydrolysis of Terpene Disaccharides Two protocols for the hydrolysis of terpene disaccha- rides were adopted: with pectinolytic enzymes for wine cla- rification with side glycosidase activity; acid hydrolysis. Enzyme hydrolysis: an enzyme preparation (50 mg) was added to the homogenisate of grape berries (20 g) di- luted to 100 mL with deionised water, pH adjusted to 4.5. The flask was sealed and placed in a water bath (40 °C) for 24 h. Acid hydrolysis: a homogenisate of grape berries (20 g) diluted to 100 mL with deionised water was prepa- red and the pH was adjusted to either 3 or 1 with an addi- tion of H2SO4 solution. The flask was sealed and placed in a thermostated oven (100 °C) for 30 min. 3. Results and Discussion 3.1. Optimisation of SPME Conditions Several different fibres were initially tested to estab- lish their efficiency for extracting the selected terpenes and benzaldehyde from the headspace of homogenised grapes: polydimethylsiloxane (PDMS), polyacrylate (PA), polydimethylsiloxane-divinylbenzene (PDMS/DVB), Carbowax-divinylbenzene (CW/DVB), divinylbenzene- Carboxen-polydimethylsiloxane (DVB/CAR/PDMS), as shown in Figure 1. The latter (mixed) phase was found to extract the highest amount of all analytes except geraniol (see Figure 1), for which the PDMS fibre was slightly mo- re efficient. 28 Acta Chim. Slov. 2007, 54, 25–32 Prosen et al.: Analysis of Free and Bound Aroma Compounds in Grape Berries ... These findings are in good agreement with the re- sults obtained by Tat et al.,11 who found the DVB/CAR /PDMS phase to be the most suitable for slightly less volatile and more polar compounds of wine aroma (tR > 15 min), while for those with tR < 15 min, the CAR/ PDMS fibre (which we did not test) was more suitable. However, the compounds of our interest are of the first type. In another study of SPME conditions,12 CAR/PDMS fibre was also found to be the most suitable for the extrac- tion of acids, aldehydes and esters in wine aroma, while DVB/ CAR/PDMS was not tested. DVB/CAR/PDMS fibre was therefore chosen for further experiments. Next, the sample temperature during headspace sampling with SPME was optimised. The sam- ples were placed in headspace vials, sealed and left in a water bath for 15 min to allow for equilibration at the cho- sen temperature. The septum was pierced with a SPME needle and the analytes were extracted from the headspa- ce. We tested sampling at temperatures in the range 25– 50 °C. At 50 oC, the amount of analytes on the fibre was still increasing except for linalool, but we decided not to increase the temperature any further for safety reasons. The last step of the optimisation was to choose the opti- mal extraction time. The equilibrium amount of com- pounds on the fibre (at 50 °C) was reached at 25–35 min, so we chose 35 min as the optimal extraction time. However, for SPME of the two aliphatic aldehydes, the chosen temperature (50 °C) was found to be too high: because of the higher volatility of these analytes and their relatively high concentration in the grape aroma, the headspace concentration was very high and the fibre pha- se was saturated already at lower concentration, leading to a sub-optimal linear range of the method for these com- pounds. The optimal conditions for SPME of hexanal and (E)-2-hexenal were found to be the extraction at room temperature (25 °C) for 15 min. At the optimal sampling conditions, a part of the method evaluation was performed using the synthetic aqu- eous solutions of the analytes. The results are shown in Table 1. It is evident that the method shows good repeata- bility and excellent linearity over a wide linear range, ma- king it suitable for determination of concentrations of aro- ma compounds both in aromatic and non-aromatic grape varieties. 3.2. Optimisation of SBSE Conditions Due to the lack of suitable thermal desorption unit (TDU) for the sorptive stir-bar thermal desorption, we conducted a series of improvised experiments in SB sorp- tive extraction, the results of which hint at the method’s real potential. The sorption time was 1 h, after which the stir-bar was transferred into 2 mL of n-hexane and placed in an ul- trasonic bath to desorb the analytes into the organic sol- vent. Although different variations to this procedure were tried, the best extraction recoveries obtained were 18–40% for the analysed terpenes, 17% for hexanal and an inexplicably high value of 208% for benzaldehyde (possible carry-over effect). The extraction recoveries for the same terpenes using TDU are reported to be 72–78%, but only 6% for benzaldehyde and 7% for hexanal.15 In another study, no extraction recoveries for wine aroma compounds using SBSE are reported, although the aut- hors conclude it is more efficient then SPME.21 3.3. Determination of Free Aroma Compounds in Grapes In the preliminary experiments done on synthetic solutions, we established the optimal conditions for HS- SPME of aroma compounds. Grape berries (cca. 20 g) Figure 1. Comparison of different stationary phases on the SPME fibre for the extraction of grape aroma compounds. Table 1: Parameters of the HS–SPME–GC–MS method for the quantified compounds. compound lin. range/ µ L–1 r2 tR / min RSD (tR) / % RSD (area) / % α-terpineol 1.3–630 0.9962 24.0 0.07 2.3 nerol 1.6–790 0.9993 24.3 0.08 2.8 geraniol 2.7–530 0.9993 24.9 0.07 2.9 linalool 1.2–600 0.9940 20.6 0.07 3.1 benzaldehyde 1.0–970 0.9957 19.2 0.09 6.3 hexanal 19.4–1940 0.9987 14.1 0.05 7.1 (E)–2–hexen–1–al 5.1–5080 0.9980 15.7 0.04 6.9 29Acta Chim. Slov. 2007, 54, 25–32 Prosen et al.: Analysis of Free and Bound Aroma Compounds in Grape Berries ... were homogenised with water to add up to 100 mL of the final macerate. An aliquot (5 mL) was sealed in a heads- pace vial and extracted under the same conditions as synthetic solutions used for the evaluation of the method. In Figure 2, track a, a HS-SPME-GC-MS chromato- gram of a “Muscat blanc” grape variety is shown. The hig- hest peaks in this chromatogram belong to the aliphatic al- dehydes hexanal, (E)-2-hexen-1-al, nonanal and decanal, which were the prevailing compounds also in all other analysed grape varieties. In the subsequent experiments we separately analysed the pulp and the skins of grape berries. The results revealed that aliphatic aldehydes were present mainly in the skin: 55–65% more than in the pulp. A small amount of terpenes (mainly linalool and ge- raniol) was present only in the most aromatic “Muscat blanc” and “Aurora” varieties. A part of the quantitative results for some compounds in three grape varieties is shown in Table 2, column 3 (“before hydrolysis”). 3.4. Different Hydrolytic Approaches for the Release of Bound Aroma Compounds in Grapes The glycosidically bound terpenes can be released by enzymatic treatment (glycosidase)2,10,16,20,22,24,25 or by acidic hydrolysis.2,10 We optimised and tested both. Enzymatic hydrolysis of terpene diglycosides is usually a two-step process, involving two different enzy- mes, although one-step hydrolysis has also been descri- bed.2,24 The glycosidases can be of plant or microbial ori- gin, but are not readily available. A frequently used practi- cal approach is to apply pectinase enzymatic preparations with substantial glycosidase side-activity.2,10,16,25 We te- sted two such preparations: Rohavin MX from AB Enzy- mes and Lallzyme BETA from Lallemand. The later has, according to the producer, a strong glucosidase activity, and may be used to enhance the intensity of wine aroma. Figure 2. Comparison of HS-SPME-GC- MS chromatograms of the same grape va- riety (Muscat blanc) before (a) and after (b) the enzymatic treatment (Lallzyme BETA). Peak assignments: 1 – hexanal, 2 – (E)-2-he- xen-1-al, 3 – linalool, 4 – nonanal, 5 – deca- nal, 6 – α-terpineol, 7 – nerol, 8 – geraniol. Table 2: Concentration (in µg g–1 of grape berries; RSD 2–7%) of some aroma compounds in different grape varieties before and after hydrolytic treatment. ND – not detected. grape variety compound before Rohavin MX Lallzyme acid acid hydrolysis hydrolysis BETA hydrolysis, hydrolysis, hydrolysis pH 3 pH 1 Muscat blanc linalool 0.15 4.40 5.27 2.79 0.02 geraniol 0.11 0.33 1.00 0.31 ND benzaldehyde 0.04 ND ND ND ND hexanal 34.4 15.9 20.9 23.2 4.31 Danijela linalool ND ND ND ND ND geraniol ND ND ND ND ND benzaldehyde 0.03 0.06 0.03 0.03 0.04 hexanal 6.87 1.08 11.6 4.45 1.68 Pinot noir linalool ND ND ND ND ND geraniol ND 0.01 0.03 ND ND benzaldehyde 0.02 0.05 0.02 0.02 0.01 hexanal 43.9 8.39 19.0 13.2 1.72 30 Acta Chim. Slov. 2007, 54, 25–32 Prosen et al.: Analysis of Free and Bound Aroma Compounds in Grape Berries ... We first conducted the hydrolysis with an amount of enzy- mes specified by the producers (5 mg/100 mL); however, no change in the aroma profile was observed after such a treatment. Glycosidases in general are inhibited by pH be- low 424 and presence of high levels of phenolic com- pounds,25 while glucosidase is strongly inhibited by glu- cose.24 Most of the research on the enzymatic hydrolysis of the terpene disaccharides was focused on the more-or- less glucose-free wine as sample,24 therefore no clear gui- delines are available about the amount of enzymes that should be used in the treatment of non-fermented musts. In the available literature on musts, enzymes were applied mostly on extracted glycosides,10,22,25 although one study reports good results on a heat-treated must sample, using approximately the above quantity of the enzyme.16 Having in mind these difficulties, we used a ten-fold recommended amount of enzyme, incubating the macera- te for 24 h at 40 °C. This approach was successful in re- leasing a substantial part of terpenes, as can be seen from Figure 2, track b, for the “Muscat blanc” grape variety. Li- nalool content was substantially increased, while at least three new terpenes appeared in the GC-MS chromato- gram: nerol, geraniol and α-terpineol. A part of the quan- titative results is shown in Table 2. Results obtained with Lallzyme BETA were slightly better than those for Roha- vin MX, which is not marketed for the purpose of aroma- enhancement. As is also obvious from Table 2 and Figure 2, the enzymatic treatment had practically no effect on the benzaldehyde content, while the content of the aliphatic aldehydes was significantly decreased. The later fact is al- so favourable, as these compounds contribute to the un- pleasant, “herbaceous” smell of wines.17 Another observation was that several new peaks ap- peared in the GC-MS chromatogram after hydrolysis. The compounds yielding these peaks were preliminary identi- fied by the aid of the library of mass spectra, an example for seldom identified benzothiazole16 is shown in Fig. 3. Table 3 lists some compounds that were most often obser- ved only after the enzymatic treatment. Some of them are reported to appear in the aroma profile of musts and wi- Table 3: Some compounds identified in the grape aroma only after hydrolytic treatment (identification from mass spectra). Compound Hydrolytic treatment Grape variety benzothiazole EH–L, EH–R Beograjska rana, Muscat blanc, Pinot noir, Perlette, Danijela furfural EH–L Perlette AH3 Aurora thymine EH–L Perlette 2–phenylethyl acetate EH–R Perlette linalool oxide, pyranic form EH–L, EH–R, AH1 Muscat blanc AH1 Aurora 2,2–dimethyl–5–(1–methyl–1– AH1 Muscat blanc, Aurora, Danijela propenyl)– tetrahydrofuran 3,6–dihydro–4–methyl–2–(2–methyl–1– AH1, AH3 Muscat blanc propenyl)–pyran eucalyptol / 1,8–cineol AH1 Aurora, Muscat blanc Abbreviations: EH–R = enzymatic hydrolysis with Rohavin MX, EH–L = enzymatic hydrolysis with Lallzyme BETA, AH3 = acid hydrolysis at pH 3, AH1 = acid hydrolysis at pH 1. Figure 3. Mass spectrum of the compound present in the aroma of several grape varie- ties (see Table 3) after enzy- matic hydrolytic treatment with either Rohavin MX or Lallzyme BETA (identified from its mass spectrum as benzothiazol). 31Acta Chim. Slov. 2007, 54, 25–32 Prosen et al.: Analysis of Free and Bound Aroma Compounds in Grape Berries ... nes, e.g. 2-phenylethyl acetate,3,9,12,17 linalool oxide7,13,16 and furfural.16–17 These compounds contribute to sweet and fruity aroma of wines.17 Acidic hydrolysis was also tested as the means to re- lease the bound terpenes. This procedure is thought to si- mulate the processes taking place during the storage and ageing of wines.2,10 The hydrolysis was conducted at two different pH values: pH 3, resembling pH conditions in wine,10 and pH 1, chosen as at these more drastic condi- tions rearrangement reactions of terpenes are probably in- duced.2 A part of the results is shown in Table 2. Acidic hydrolysis at pH 3 obviously results in a release of terpe- nes from glycosides, although to a lesser extent than enzy- matic hydrolysis. Benzaldehyde concentration is not sig- nificantly affected, while the concentration of aliphatic al- dehydes is decreased. After hydrolysis at pH 1, the con- tent of monitored terpenes is even lower than before hydrolysis, which is in agreement with the literature data.2 Aliphatic aldehydes are also significantly decomposed at these conditions. However, it is interesting to observe the release (or formation) of several compounds not noticed after enzymatic hydrolysis (examples in Table 3), e.g. eu- calyptol (or 1,8-cineol), a typical product of too harsh hydrolytic conditions.2 3.4. LC-MS Analysis of Terpene Glycosides We conducted some LC-MS analyses (electrospray interface) of non-hydrolysed terpene glycosides extracted from the grape berries. The results of these preliminary experiments were quite promising. An extracted LC-MS chromatogram (m/z 310-314) of the grape extract is shown in Figure 4. The ions at m/z 310-314 interval were chosen as possible fragments of the disaccharide part of terpene glycosides. At least two compounds eluting at tR 25,0 min and tR 30,9 min could be identified as terpene glycosides based on their mass spectra (example for the compound eluting at tR 30,9 min shown as an insertion in Figure 4). Typical ions featuring in these spectra are at m/z 156, 174, which could be attributed to the aglycone (ter- pene) part of the terpene glycosides.26 Ions at m/z 115 and 133 could be fragments of the monosaccharide moieties.26 However, in spite of these interesting results, the content of terpene glycosides in the extract was obviously quite low and their peaks were not clearly visible in the chroma- togram. Therefore, the extraction procedure seems to be less then satisfactory and has to be further optimised. 4. Conclusions We developed an effective and rather simple extrac- tion procedure for the aroma compounds from musts and wines, using solid-phase microextraction from the heads- pace of heated samples. Combined with GC-MS analysis, the overall method proved to be applicable to different aroma compounds (aliphatic, aromatic aldehydes, terpe- nes) in a broad concentration range (1–5000 µg L–1). The extraction method was also compared with a stir-bar sorp- tive extraction procedure for the same samples, however, SBSE was not very effective due to the lack of a suitable desorption device. The method was applied to the analysis of free aro- ma compounds in must samples of different grape varie- ties and to follow their release after enzymatic or acidic hydrolysis of bound aroma compounds. Different hydro- lytic approaches were tried and the most successful in re- leasing terpene compounds was the enzymatic hydrolysis (with two different enzymes) and acidic hydrolysis at pH 3. Figure 4. LC-MS chromatogram (extracted m/z: 310-314) of a glycoside extract from the grape variety Muscat blanc. Insertion: mass spectrum of peak at 30.9 min. 32 Acta Chim. Slov. 2007, 54, 25–32 Prosen et al.: Analysis of Free and Bound Aroma Compounds in Grape Berries ... Acidic hydrolysis at pH 1 resulted in substantial decom- position and re-arrangement reactions of terpenes. Non-hydrolysed terpene glycosides were extracted from the musts using solid-phase extraction and the ex- tract was analysed with LC-MS (ESI interface). The preli- minary results are quite promising, but the extraction pro- cedure has to be further optimised. 5. Acknowledgements The donation of pectinolytic enzymes from AB Enzymes, Darmstadt, Germany (Mr Reinhold Urlaub) and from Jurana, Maribor, Slovenia (representatives for Lalle- mand, St. Simon, France) is gratefully acknowledged. This work has been supported by the Ministry of Higher Education, Science and Technology of the Republic Slo- venia through Grant P1-0153. 6. References 1. P. Ribereau-Gayon, Y. Glories, A. Maujean, D. Duboudieu, Handbook of Enology, Vol.2: The Chemistry of Wine and Sta- bilization and Treatments, J. Wiley & Sons, USA, 2000, pp. 187–206. 2. S. Maicas, J. J. Mateo, Appl. Microbiol. Biotechnol. 2005, 67, 322–335. 3. M. R. Salinas, A. Zalacain, F. Pardo, G. L. Alonso, J. Agric. Food Chem. 2004, 52, 4821–4827. 4. J. S. Camara, P. Herbert, J. C. Marques, M. A. Alves, Anal. Chim. Acta 2004, 513, 203–207. 5. S. C. Dieguez, L. C. Lois, E. F. Gomez, L. G. de la Pena, Le- bensm.-Wiss. u-Technol. 2003, 36, 585–590. 6. J. J. Mateo, M. Jimenez, J. Chromatogr. A 2000, 881, 557– 567. 7. J. M. Oliveira, I. M. Araujo, O. M. Pereira, J. S. Maia, A. J. Amaral, M. Odete Maia, Anal. Chim. Acta 2004, 513, 269– 275. 8. N. Guarrera, S. Campisi, C. Nicolosi Asmundo, Am. J. Enol. Vitic. 2005, 56, 394–399. 9. V. Ferreira, N. Ortin, A. Escudero, R. Lopez, J. Cacho, J. Agric. Food Chem. 2002, 50, 4048–4054. 10. M. Fernandez-Gonzalez, R. Di Stefano, Lebensm.-Wiss. u- Technol. 2004, 37, 467–473. 11. L. Tat, P. Comuzzo, I. Stolfo, F. Battistutta, Food Chem. 2005, 93, 361–369. 12. R. Castro, R. Natera, P. Benitez, C. G. Barroso, Anal. Chim. Acta 2004, 513, 141–150. 13. E. Sanchez-Palomo, M. C. Diaz-Maroto, M. S. Perez-Coel- lo, Talanta 2005, 66, 1152–1157. 14. M. E. O. Mamede, G. M. Pastore, Food Chem. 2006, 96, 586 –590. 15. D. J. Caven-Quantrill, A. J. Buglass, J. Chromatogr. A 2006, 1117, 121–131. 16. S. Rocha, P. Coutinho, A. Barros, M. A. Coimbra, I. Delga- dillo, A. Dias Cardoso, J. Agric. Food Chem. 2000, 48, 4802 –4807. 17. M. Franco, R. A. Peinado, M. Medina, J. Moreno, J. Agric. Food Chem. 2004, 52, 3905–3910. 18. E. Sanchez-Palomo, M. C. Diaz-Maroto, M. A. Gonzalez Vi- nas, A. Soriano-Perez, M. S. Perez-Coello, Food Control 2007, 18, 398–403. 19. I. Luki}, M. Banovi}, \. Per{uri}, S. Radeka, B. Sladonja, J. Chromatogr. A 2006, 1101, 238–244. 20. L. Moio, M. Ugliano, A. Gambuti, A. Genovese, P. Piombi- no, Am. J. Enol. Vitic. 2004, 55, 7–12. 21. D. Komes, D. Ulrich, T. Lovric, K. Schippel, Vitis 2005, 44, 187–193. 22. R. Schneider, F. Charrier, M. Moutounet, R. Baumes, Anal. Chim. Acta 2004, 513, 91–96. 23. O. Shoseyov, B. Bravdo, Enhancement of Aroma in Grapes and Wines: Biotechnological Approaches, In: Molecular Biology & Biotechnology of the Grapewine (Ed. Roubelakis- Angelakis, K.A.), Kluwer Academic Publishers 2001, pp. 225–239. 24. J.-E. Sarry, Z. Günata, Food Chem. 2004, 87, 509–521. 25. S. M. Bureau, R. L. Baumes, A. J. Razungles, J. Agric. Food Chem. 2000, 48, 1290–1297. 26. V. A. Marinos, M. E. Tate, P. J. Williams, J. Agric. Food Chem. 1994, 42, 2486–2492. Povzetek Razvili smo ekstrakcijski postopek za sestavine arome mo{ta in vina, kjer smo uporabili mikroekstrakcijo na trdno fazo (DVB/CAR/PDMS vlakno). Ekstrahirali smo iz plinske faze nad ogretim vzorcem (50 °C). Analiza je potekala s plin- sko kromatografijo z masno spektrometri~no detekcijo (GC-MS). Metodo smo uporabili za analizo razli~nih sestavin arome (alifatski in aromatski aldehidi, terpeni) v {irokem koncentracijskem obmo~ju (1–5000 µg L–1). Preizkusili smo tudi ekstrakcijo na me{alo s sorbentom (SBSE), vendar u~inkovitost postopka ni bila zadovoljiva, saj nismo imeli na vo- ljo primerne naprave za desorpcijo. V mo{tu razli~nih grozdnih sort smo analizirali proste sestavine arome, sledili pa smo tudi njihovemu spro{~anju med encimsko ali kislinsko hidrolizo. Preizkusili smo razli~ne postopke hidrolize. Naj- bolj u~inkovita je bila encimska hidroliza (dvoje razli~nih encimov) in kislinska hidroliza pri pH 3. Med kislinsko hidro- lizo pri pH 1 je pri{lo do znatnih razgradenj in premestitev terpenskih spojin. Nehidrolizirane terpenske glikozide smo ekstrahirali iz mo{ta s pomo~jo ekstrakcije na trdno fazo, ekstrakt pa smo analizirali s teko~insko kromatografijo in ma- sno spektrometrijo (LC-MS, ESI vmesnik). Nekatere spojine smo lahko identificirali kot terpenske glikozide.