Food Chemistry 122 (2010) 1327–1332

Contents lists available at ScienceDirect

Food Chemistry

journal homepage: www.elsevier .com/locate / foodchem

Analytical Methods

Analysis of free and glycosidically bound compounds of ash gourd (Benincasahispida): Identification of key odorants

Jyoti Sharma, Suchandra Chatterjee, Vivekanand Kumar, Prasad S. Variyar *, Arun SharmaFood Technology Division, Bhabha Atomic Research Centre, Mumbai 400085, India

a r t i c l e i n f o a b s t r a c t

Article history:Received 9 March 2009Received in revised form 27 October 2009Accepted 24 March 2010

Keywords:AcetoinAroma glycosidesAsh gourdVolatile essential oilGC–olfactometry

0308-8146/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.foodchem.2010.03.099

* Corresponding author. Tel.: +91 22 25592531; faxE-mail address: prasadpsv@rediffmail.com (P.S. Va

Free and glycosidically bound aroma compounds of ash gourd were investigated. Gas chromatogra-phy–mass spectrometry (GC/MS) analysis of the free aroma compounds isolated by both steam distil-lation and high vacuum distillation revealed the presence of acetoin as the major constituent in boththe oils. Acetoin was also identified as the major constituent existing as glycosidic conjugate. GC–olfactometry (GC–O) of the volatile oil based on detection frequency (olfactometry global analysis)was further used to characterise odour active compounds of the vegetable. Based on odour activityvalue (OAV) measurements the potent odorant among these were identified as acetoin (9522.5), oct-anal (8571) and nonanal (24,000). The region corresponding to the characteristic ash gourd aromawhen collected by cryotrapping from a packed GC column and subsequently analysed by GC/MSwas found to be mainly comprising of acetoin suggesting the role of this compound as the key odouractive compound of this vegetable.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Ash gourd (Benincasa hispida), also known as winter melon,white gourd or wax gourd is a vegetable widely consumed in In-dian households. Besides its use as a vegetable it is extensivelyused in traditional oriental medicine for treatment of various ail-ments such as gastrointestinal problems, respiratory diseases,heart diseases, vermifuge, diabetes mellitus and urinary diseases(Aslokar, Kakkar, & Chakre, 1992; Sivarajan & Balachandran,1994). While the intact fresh green vegetable is nearly odourless,a characteristic fatty green and buttery odour is perceived eitherduring cutting or cooking. Despite its wide use in culinary prepara-tions, aroma active constituents of this vegetable have not beenextensively investigated. To the best of our knowledge, only onereport exists on the nature of volatile aroma compounds in thisvegetable (Chung, Shu, Chang, & Chiang, 1987). The major volatilecompounds reported by these researchers include (E)-2-hexenal,n-hexanal and n-hexyl formate. However, the role of the individualcompounds in contributing to the characteristic odour of the veg-etable was not investigated.

GC–O techniques have in recent years proved to be a powerfultool in correlating chemical information with sensory impression.Based on these techniques odour active compounds in variousfoods have been investigated and reported (Acree, Barnard, &Cunningham, 1984; Jordan, Tandon, Shaw, & Goodner, 2001; Ull-

ll rights reserved.

: +91 22 25505150/222198.riyar).

rich & Grosch, 1993). Among the GC–O techniques, the dilutionsniffing methods (Aroma Extract Dilution Analysis, AEDA and Com-bined Hedonic Aroma Response Measurement, CHARM) based onodour detection thresholds have the drawback of reproducibilitydue to a very high variability within and between panelists (Pollienet al., 1997). The odour specific magnitude estimation (OSME)method based on odour intensity, although more precise thanthe other methods, requires at least four sensitive assessors whoare repeatable and trained in odour intensity. A more recent meth-od based on frequency of detection (olfactometry global analysis)rather than perceived intensity aided in identifying odour activecomponents in a given food stuff. Despite the use of untrained pa-nel and one dilution level, repeatability of the method was demon-strated to be satisfactory and is almost twice as fast as the othermethods (Guen, Prost, & Demaimay, 2000). So far, this techniquehas not been applied to characterise the aroma profile of ash gourd.

Besides free aroma, the potential glycosidically bound aromacompounds have been reported to occur in several food stuffs (Sar-ry & Gunata, 2004). Release of these compounds during storage orprocessing can generate modification or enhancement in aroma.Glycoconjugates thus play a crucial role in the overall food quality.No reports exist till date on glycosidically bound aroma com-pounds of ash gourd. This work thus aims at isolation and identifi-cation of free and bound aroma constituents of ash gourd. Anattempt was made to identify aroma active constituents based onfrequency response (olfactometry global analysis) as well as keyodorants based on OAV as well as GC–O in combination with GCand GC/MS.

1328 J. Sharma et al. / Food Chemistry 122 (2010) 1327–1332

2. Materials and methods

2.1. Materials

Commercial samples of ash gourd were procured from a localmarket. All the standard compounds were obtained from Sigma/Al-drich chemical company, USA. Solvents (analytical reagent grade)were redistilled before use.

2.2. Methods

2.2.1. Isolation of free aroma compounds2.2.1.1. Steam distillation. Ash gourd (200 gm) was cut into smallpieces and subjected to simultaneous distillation extraction (SDE)technique (Nickerson & Likens, 1966) for 2 h using peroxide freediethyl ether (AR Grade, S.D. Fine-Chem. Ltd., Mumbai, India) asextracting solvent. The solvent was then removed by passing aslow stream of nitrogen to obtain the volatile oil.

2.2.1.2. High vacuum distillation (HVD). High vacuum distillationwas carried out according to the procedure reported in literature(Bandyopadhyay, Gholap, & Sreenivasan, 1973; Gholap & Bandyo-padhyay, 1977). Cut ash gourd (200 g) was frozen in liquid nitro-gen before use. The sample was placed in a glass tube (5 cmi.d. � 25 cm length) and then connected to a distillation unit main-tained under vacuum (1 � 10�3 torr). The distillate collected in areceiving tube (3 cm i.d. � 20 cm length) was maintained at lowtemperatures under liquid nitrogen. The isolate (50 ml) was ex-tracted with diethyl ether (3 � 5 ml). Solvent was removed by aslow stream of nitrogen to obtain oil that was subjected to GC/MS analysis as described below.

2.2.2. Analysis of free aroma compounds2.2.2.1. GC–MS analysis. The aroma concentrates obtained from SDEand HVD were subjected to GC–MS analysis on a Shimadzu GC–MSinstrument (Shimadzu Corporation, Kyoto, Japan) equipped with aGC-17A gas chromatograph and provided with a DB-5 (J&W Scien-tific, California, USA) capillary column ((5%-phenyl)-meth-ylpolysiloxane, length, 30 m; i.d., 0.25 mm and film thickness,0.25 lm). The operating conditions were: column temperatureprogrammed from 60 to 200 �C at the rate of 4 �C/min, held at ini-tial temperature and at 200 �C for 5 min. and further to 280 �C atthe rate of 10 �C/min, held at final temperature for 20 min; Injectorand interface temperatures, 210 and 280 �C, respectively; carriergas helium (flow rate, 0.9 ml/min); ionisation voltage, 70 eV; elec-tron multiplier voltage, 1 kV. Peaks were tentatively identified bycomparing its mass fragmentation pattern with that of standardcompounds wherever available, from standard spectra availablein the spectral library (Wiley/NIST Libraries) of the instrument aswell as by comparing RI (retention index) values of the compoundswith literature data. Amounts of the identified compounds in theessential oil obtained from various extraction methods were esti-mated from a standard curve (R2 = 0.99) of concentration versuspeak area prepared using different concentrations of standard acet-oin (0.01–20 lg/ll) and expressed as lg/kg of ash gourd.

2.2.3. Isolation and characterisation of major volatile oil constituentPreparative TLC of all the SDE isolates was carried out on silica

gel G plates (20 cm � 20 cm � 0.5 mm thickness) using hexane:ethyl acetate (80:20) as developing solvent system. After develop-ment, the plates were dried in air at room temperature. The bandat Rf 0.28, having ash gourd aroma was scrapped and eluted withdiethyl ether. Solvent was then removed by passing a slow streamof nitrogen and the pure compound was identified as acetoin byGC/MS and other spectral analyses (UV, IR, NMR) as detailed below.

UV/Vis spectrum was taken on a UV spectrophotometer (Uni-cam Hekios UVA, Cambridge, UK) in methanol, kmax: 218 nm. IRspectrum was recorded with a JASCO FTIR 4100 spectrophotometer(Jasco Corporation, Tokyo, Japan). mmax (KBr): 3404, 2987, 2939,2884, 1714, 1644, 1452, 1380, 1187, 1116, 1024, 939 cm�1. TheNMR spectra were recorded with a Bruker AC-200 MHz FT NMRspectrometer (Bruker, Fallanden, Switzerland). The usual abbrevia-tions employed are: t = triplet, q = quartet, J = coupling constant (inHz), d = chemical shift in ppm. 1H NMR (CDCl3, 200 MHz): d ppm,1.37 (3H, d, J = 5.8) 2.22 (3H, s) 4.42 (1H, q). 13C NMR (CDCl3,200 MHz): d ppm, 18, 72, 107, 210. EIMS was recorded using DIprobe facility in the GC/MS equipment described above. m/z (%,rel. int.) 88(M+, 1.4), 73.05(33.41), 61(30.62), 45(100),44.05(21.33), 43(93.43), 42.05(8.86).

2.2.4. Isolation of aroma glycosidesAsh gourd (1 kg) was cut in small pieces and soaked in aqueous

methanol overnight. After filtration the remaining ash gourd wascrushed in omni mixer and extracted twice with aqueous metha-nol. These extracts were pooled and concentrated under vacuum(760 mm Hg) on a rotary evaporator (35 �C) to obtain an aqueoussolution. The aqueous solution so obtained was washed withdiethyl ether to remove traces of lipophilic substances includingfree aroma compounds and subsequently passed through anAmberlite XAD-2 column according to the procedure reported ear-lier (Gunata, Bayonove, Baumes, & Cordonnier, 1985). The boundaroma precursors were desorbed from the resin with methanolas eluting solvent, evaporated to dryness under vacuum and theresidue was made to 2% in methanol (w/v).

2.2.5. Separation and estimation of aroma glycosides2.2.5.1. TLC analysis. Analytical TLC was carried out on ammoniumsulphate impregnated silica gel G plate (10 cm � 25 cm � 0.25 mmthickness) using toluene:ethanol:formic acid (60:35:5, v/v/v) asthe developing solvent system. The separated spots were visual-ised by heating the plate in oven at 120 �C for 20 min. PreparativeTLC was carried out on silica gel plates (20 cm � 20 cm � 0.5 mmthickness) using the same solvent system as above. Five distinctlyresolved bands were scrapped and eluted with methanol. The elu-ate was evaporated to dryness to make 1% solution (w/v).

2.2.5.2. Estimation by densitometry. Density of the individual spots/bands was determined on a dual wavelength flying spot scanningdensitometer Shimadzu CS-9301PC (Shimadzu, Kyoto, Japan). Thedensity of the spots was determined in the reflectance mode at awavelength of 528 nm. Acetoin glucoside was used as externalstandard. Aliquots of suitably diluted sample were spotted on theplate in increasing concentration ranging from 0.05 to 6 mg/ml.The amount of different bands was estimated from a standardcurve (R2 = 0.99) of spot density versus concentration and ex-pressed as mg/kg of ash gourd.

2.2.6. Analysis of bound aroma compoundsThe XAD extract was subjected to acid hydrolysis (1 N HCl, 1 h,

80 �C). The hydrolysate was extracted with diethyl ether(3 � 20 ml) and made free of acid by repeated washing with dis-tilled water. The organic layer was dried over sodium sulphateand then concentrated for subsequent identification of the agly-cone moiety on GC/MS. The five bands isolated from preparativeTLC as described above were also subjected to acid hydrolysis.The free volatiles obtained after subsequent workup of each bandwere also analysed by GC/MS. The remaining aqueous solutionfrom the total XAD eluate as well as individual bands was neutra-lised with 1 N KOH, dried under vacuum and the residue dissolvedin methanol. This methanol solution was subjected to TLC in orderto identify the sugar residue. The sugar residue obtained from the

Table 2Major aroma glycosides in ash gourd as estimated by TLC-densitometry.

S. No. Compound Retentionfactor (Rf)

Concentration(mg/kg)

1. 2-Ethyl-1-hexanol glucoside + glucosidesof aliphatic long chain saturated alcohols

0.38 19.57 ± 1.254

2. Acetoin glucoside 0.44 24.00 ± 1.9453. 4-Hydroxy-2-butanone glucoside 0.52 0.067 ± 0.0054. 4-Penten-2-ol glucoside + 2-nonanol

glucoside0.59 0.24 ± 0.008

5. Nonanal glucosides + decanal glucoside 0.77 0.16 ± 0.014

Data are the mean of six replicates ± standard deviation.

J. Sharma et al. / Food Chemistry 122 (2010) 1327–1332 1329

total XAD hydrolyzate as well as from the individual bands wasfurther converted to its acetyl derivative using pyridine:aceticanhydride (1:1) (overnight, room temperature). The acetylatedderivative was concentrated by slow stream of nitrogen and thensubsequently analysed by GC/MS. Amounts of the volatile aromacompounds liberated from their glycosidic precursors were esti-mated from a standard curve (R2 = 0.99) prepared as above and ex-pressed as lg/kg of ash gourd.

2.2.7. GC–O analysisAroma isolates obtained from SDE and HVD were subjected to

GLC on a Shimadzu GC-14A instrument equipped with a TCDdetector, a Shimadzu C-R6A integrator and provided with dual col-umns (a packed Silar 10C column (3 mm o.d. � 3 m) and a DB-5capillary column). The operating conditions were: column temper-ature programmed from 50 to 150 �C at the rate of 4 �C/min, heldat initial temperature and also at 150 �C for 5 min with the TCDcurrent maintained at 75 mA. Carrier gas (He) flow rate for Silar10C packed (Gaschrome Q, 60/80 mesh) column was howevermaintained at 40 ml/min. The compounds eluting out of both thecolumn were sniffed at the exit port of the TCD (TCD vent).

Different aroma notes such as oily, cooked, cut vegetable andgreen were perceived at sniffing port. Peaks corresponding to dif-ferent notes were identified by a sensory panel made up of ninejudges who were familiar with ash gourd aroma. Region corre-sponding to ash gourd aroma (Rt 7.5–9.5 min) eluting out of thepacked column was cryotrapped (liquid nitrogen, �196 �C) inwidebore glass capillary tube having intermittent constrictionthroughout the length according to the procedure reported earlier(Gholap & Bandyopadhyay, 1977). The capillary tube was jacketedwith a test tube (2.5 cm i.d. � 20 cm) that was immersed in liquidnitrogen (�196 �C). The exit end of the detector and the collectingcapillary tube was connected via short Teflon tubing. After the col-lection was over the condensate was dissolved in 0.1 ml of solvent.The sample thus collected after repeated injections (5 times) corre-sponding to the characteristic ash gourd aroma was directly in-jected into the GC/MS instrument under the conditions describedin Section 2.2.2.1.

2.2.8. Olfactometry global analysisAnalysis was carried according to the conditions recommended

by Guen et al. (2000). A panel of nine judges trained in odour rec-ognition and experienced in GC–O was selected. Panelists assigned

Table 1Quantitative distribution (lg/kg of ash gourd) of volatile oil components identified in diff

No.e Compound RIa Concentration (lg/kg of ash

SDE HV

1. Ethyl acetate 806 0.029 ± 0.026 0.02. Acetoin 812 7.618 ± 1.18 7.53. 2-Hexenal 854 0.006 ± 0.0014. 1-Hexanol 868 0.098 ± 0.045. 1-Heptanol 965 0.012 ± 0.016. 2-Octanone 985 0.011 ± 0.0097. Octanal 994 0.006 ± 0.0038. 2-Ethyl-1-hexanol 1021 0.068 ± 0.0089. Nonanal 1093 0.024 ± 0.01610. Nonanol 1170 0.005 ± 0.00311. a-Terpineol 1185 0.003 ± 0.00112. Decanal 1193 0.003 ± 0.00113. Nerolidol 1550 0.014 ± 0.009

Data are the mean of six replicates ± standard deviation.a Retention index on DB-5 column.b Odour description as perceived by panelists during olfactometry global analysis.c Detection frequency (of nine panelists).d Odour activity value.e Numbers refer to peak numbers depicted in Fig. 1.

odour properties to each compound detected. Detection of odourby fewer than four panelists was considered as noise (VanRuth &Roozen, 1994). Both SDE volatiles as well as hydrolysed totalbound fraction were analysed. Sniffing was divided into two partsof 13 min. Each person participated in the sniffing of both parts butduring two distinct sessions to avoid tiredness. A final aromagramof detection frequency versus RI was obtained after summing up allthe nine individual aromagrams.

2.2.9. Determination of odour activity values (OAV)OAV values were determined by dividing concentration of a

compound to its odour threshold. The odour threshold data wasobtained from reported literature data.

2.2.10. Statistical analysisAll data on the distribution of free and bound aroma com-

pounds (Tables 1–3) are a mean of three independent analyses,each analysed in duplicate. Data are thus expressed as mean ± S.D.(n = 6).

3. Results and discussion

3.1. Free aroma compounds

SDE, although a widely used technique for the isolation of vol-atile oils, has the disadvantages that some of the heat labile con-stituents may undergo degradation, or the low boiling volatilesmay be lost at the temperatures employed (Bandyopadhyay &Variyar, 1988). A characteristic cooked odour is also perceived in

erent isolates.

gourd) Odour descriptionb DFc OAVd

D

14 ± 0.006 Fruity 2 5.894 ± 1.15 Buttery/cut ash gourd type 8 9522.5

Green fruity/vegetable like 6 352.9Green flowery 3 39.2Earthy/herbaceous 7 4000Fruity 5 220Fatty green 8 8571Floral 2 ndGreen/fatty/soapy 8 24,000Fatty/green 5 100Fruity/anise 1 8.57Green/nutty 6 1500Dry wood/hay 1 nd

Table 3Quantitative distribution (lg/kg ash gourd) of free volatile components obtained fromglycosidic precursors.

S. No. RI Compound Concentration(lg/kg ash gourd)

1. 784 1,3-Butane diol 0.084 ± 0.022. 812 Acetoin 826.56 ± 3.4563. 820 4-Hydroxy-2-butanone tr4. 959 2-Pentanol 0.084 ± 0.0255. 961 Benzaldehyde tr6. 1021 2-Ethyl-1-hexanol 1.68 ± 0.1367. 1072 4-Penten-1-ol tr8. 1093 Nonanal tr9. 1150 3-Methyl butanol 0.168 ± 0.03610. 1193 Decanal tr10. 1368 Trans-2-undecen-1-ol 0.170 ± 0.03111. 1391 3-Hydroxy-4-methoxy benzaldehyde 1.01 ± 0.58112. 1407 Dodecanal 0.505 ± 0.062413. 1473 Dodecanol 0.162 ± 0.00814. 1879 Hexadecanol tr15. 2035 Octadeca-9,12-dien-1-ol 0.165 ± 0.01216. 2082 Cis-9-Octadecen-1-ol 0.926 ± 0.512

Data are the mean of six replicates ± standard deviation.

1330 J. Sharma et al. / Food Chemistry 122 (2010) 1327–1332

many cases. Milder techniques such as distillation under reducedpressure including high vacuum distillation (HVD) have thereforebeen attempted to obtain isolates that represent truer aroma (Gho-lap & Bandyopadhyay, 1977) The aroma volatiles were thus ex-tracted using both SDE and HVD in order to ascertain generationof artefacts as gas chromatographic (GC) patterns are largely influ-enced by extraction procedures (VanRuth, Roozen, Cozijnsen, &Posthumus, 1995).

Table 1 provides a quantitative distribution of the constituentsidentified in the essential oils isolated by both the methods. Irre-spective of the distillation techniques used, acetoin was found tobe the major constituent accounting for 96–99% of the total con-stituents identified. Ethyl acetate was the only other constituentdetected in both the isolates. Unlike SDE, however, the GLC profileof HVD isolates was characterised by a negligible content of ali-phatic C6-C10 alcohols and aldehydes. The C5, C6, and C8 alcoholsand carbonyl compounds responsible for green odour of severalvegetables are known to be derived from linoleic and linolenicacids (De Lumen, Stone, Kazeniac, & Forsythe, 1978; Hatanka,1996). Aldehydes are also known to be derived from thermal Strec-ker oxidative degradation of amino acids and fatty acids (Tressel &Rewicki, 1999). The aliphatic C6–C10 aldehyde and alcohols iden-tified in significant amounts in the essential oil of SDE volatilescould thus possibly arise from thermal degradation of these pre-cursors during SDE. Wu et al. in their study on volatile compoundsof ash gourd identified E-2-hexenal (green fruity, pungent vegeta-ble like odour), n-hexanal (fatty green grassy) and n-hexyl formate(ethereal fruity and refreshing green) as the major volatile constit-uents of this vegetable. n-Hexanal and n-hexyl formate were how-ever detected only in trace amounts in the present study. Othervolatile constituents such as 1-octen-3-ol, undecen-1-ol, 2-hexylfuran, deca-2, 4-dienal, isoamyl alcohol and 2-heptenal earlier re-ported by Wu et al. (1987) was also not detected by us in thisstudy. It is interesting to note that despite a high content of acetoindetected in the essential oil in our study this compound was notearlier reported even in trace amounts as a constituent of the ashgourd volatiles.

The major compound present in the volatile essential oil andcorresponding to acetoin was isolated by preparative TLC. Subse-quent spectral analysis confirmed its identity as 3-hydroxy-2-butanone (acetoin). Thus acetoin was confirmed as the major con-stituent of free aroma compound of Indian ash gourd. Acetoin isknown to possess an intense creamy, fatty, buttery odour and is

perceptible at concentrations above 0.2 ppm (Chyau, Ko, Chang,& Mau, 2003). Presence of this compound in high amounts sug-gested its possible role as contributor of ash gourd aroma. Besidesthe characteristic buttery note, the vegetable is known to possessintense green herbaceous vegetable odour. Other prominent con-stituents identified in this study such as hexanol, heptanol(earthy/herbaceous odour) and nonanal (green fatty and soapyodour) may thus play an important role in contributing to the over-all odour of this vegetable.

3.2. Glycosidically bound aroma compounds

Glycosidically bound aroma compounds or potential aromacompounds are known to play a major role in contributing to aro-ma quality. It was therefore of interest to elucidate the nature ofthese glycoconjugates in ash gourd. The bound fraction obtainedfrom the XAD column was subjected to TLC analysis. The isolate re-solved into five distinct bands at Rf values of 0.38, 0.44, 0.52, 0.59and 0.77. Hydrolysis of each of the bands isolated from a prepara-tive TLC followed by GC/MS analysis of ether extract of the hydro-lyzate aided in identifying the aglycone moiety. Glucose was theonly sugar residue detected in each of the isolated band.

Table 2 provides quantitative distribution of the major aromaglycosides identified in ash gourd. Acetoin was the major constitu-ent existing as glucoconjugate. This compound was present inamount far higher than in the free form. The second major bandat Rf 0.38 was made up of a mixture of several constituents. 2-ethyl-1-hexanol accounted for 20% of the total aglycones, while ali-phatic long chain saturated alcohols (hexadecanol, 1-dodecanoland 1-decanol) were the other prominent constituents identifiedin this band by GC/MS. The alcohols identified, however, have nocharacteristic odours and hence have no significant contributionto ash gourd aroma. Other glycosides identified (Table 2) in thisstudy were present in minor amounts.

Table 3 lists the nature of aglycones liberated from the totalbound fraction of ash gourd. Acetoin accounted for 98% of thebound aroma compounds. Besides the aglycones existing as theirglycosides and listed in Table 2, several other constituents werealso identified. A low content of their glycoconjugates could possi-bly account for these compounds being not detected on TLC. Theoccurrence of acetoin as a major glycoconjugate further confirmsthis compound as an important aroma constituent of ash gourd.Among the other bound aroma constituents identified in thisstudy, 2-pentanol (minty), 2-ethyl-1-hexanol (floral), 3-methylbutanol (ripe onion) and vanillin (vanilla like) with low thresholdsmay also contribute to the overall odour of the vegetable duringprocessing including cooking.

3.3. Identification of odour active compounds

The presence of acetoin in considerable amount both in free andin bound form suggests its possible role as an aroma active com-pound of the vegetable. This work was thus focused on the identi-fication of the most potent odorants of ash gourd. We employedthe GC–O method based on detection frequency (Olfactometry Glo-bal Analysis) of odours as the method was rapid and required notrained panelist. Eight odour active compounds were sniffed byat least more than four assessors when SDE volatiles were injectedon a DB-5 column.

Table 1 presents odour descriptors perceived by the panel forthe aroma constituents in order of elution in this column. A varietyof odour qualities such as cut ash gourd, green, herbaceous, fattygreen were perceived. The nine judges constituting the sensory pa-nel agreed that the odour of the extract was typical buttery greenvegetable like. Two regions at Rt 7.5–9.5 min and 17–18 min corre-sponding to the cut vegetable and fatty green, respectively closely

Fig. 2. GC/MS chromatogram of cryotrapped fraction (Rt 7.5–9.5 min) isolated fromSilar 10C packed column corresponding to ash gourd aroma.

J. Sharma et al. / Food Chemistry 122 (2010) 1327–1332 1331

resembled the odour of the vegetable. GC–O analysis of the hydro-lyzate from the total bound fraction, on the other hand, resulted inonly one region corresponding to cut vegetable (Rt 7.5–9.5 min).This could possibly be due to the predominance of acetoin in thebound fraction with a negligible content of other constituentsincluding nonanal.

The odour detection frequency chromatogram of SDE volatileconstituents analysed by GC–O is shown in Fig. 1. The DF datashowed highest frequencies for peaks two (buttery/cut ash gourdtype), seven (fatty green), nine (green/fatty/soapy) while some-what lower frequencies were found for three (green fruity/vegeta-ble like), four (green flowery), five (earthy/herbaceous), six (fruity),10 (fatty/green) and 12 (green/nutty). The identities of the eightcompounds listed were established based on their retention index,odour quality, and mass spectral data with those of standard com-pounds. Among the odour active regions with highest DF, acetoinwith a characteristic creamy, fatty and buttery aroma, octanal witha fatty green odour and nonanal with a green, fatty and soapyodour has not been reported earlier as an aroma compound of thisvegetable. The successive elution of the latter two compounds inthe column could possibly account for only two regions (cut vege-table and fatty green) being perceived by the sensory panel ascharacteristic odour of the vegetable.

Acetoin is widely reported to occur in apples, asparagus, blackcurrants, banana, lychee, guava, blackberry, wheat, broccoli, brus-sel sprouts, cantaloupe, butter and yogurt (Chyau et al., 2003;Acree et al., 1984; Jordan, Margaria, Shaw, & Goodner, 2003; Ott,Fay, & Chaintreau, 1997). It has also been proposed to be a majorcharacteristic compound of lichee aroma (Chyau et al., 2003). Alke-nals and alkanals are known to contribute fatty-oily, slightly rancidodours and are reported to be present in several vegetables (Guenet al., 2000). Thus acetoin, octanal and nonanal could be proposedas the most potent odorants of ash gourd. The latter two com-pounds namely octanal and nonanal possess very low thresholdsand thus were well perceived during global analysis. Despite a highodour threshold, acetoin was also well perceived by the sensorypanel. This could possibly be due to the high concentration of thiscompound in the vegetable. A high OAV value of this compound(Table 1), even higher than octanal clearly suggests that its contri-bution to the overall odour of the vegetable might be significant. In

Fig. 1. Odour detection frequency chromatogram of ash gourd SDE volatiles (referto Table 1 for peak numbers).

order to further ascertain its role as the odour active constituent ofash gourd the region corresponding to buttery/cut ash gourdodour, as sensed by the panel was collected from the TCD ventand further subjected to GC/MS analysis. Acetoin was the only con-stituent detected in the isolate (Fig. 2), further confirming the roleof this compound as the key aroma active compound of thevegetable.

4. Conclusion

This work demonstrates the role of acetoin, octanal and nonanalas major aroma contributing compounds of ash gourd. The charac-teristic odour of the vegetable thus arises from a combination ofbuttery note of acetoin with green fatty notes of octanal andnonanal.

References

Acree, T. E., Barnard, J., & Cunningham, D. G. (1984). A procedure for the sensoryanalysis of gas chromatographic effluents. Food Chemistry, 14, 273–286.

Aslokar, L. V., Kakkar, K. K., & Chakre, O. J. (1992). Glossary of Indian medicinal plantswith active principles. Part I (1st ed.). New Delhi: CSIR. p. 119.

Bandyopadhyay, C., & Variyar, P. S. (1988). On the volatiles of pepper plant. PafaiJournal, 1, 25–28.

Bandyopadhyay, C., Gholap, A. S., & Sreenivasan, A. (1973). Studies on aroma of ripemango pulp: Isolation and concentration of aroma (var. Alphonso). IndianJournal of Technology, 11, 275–277.

Chung, M. W., Shu, E. L., Chang, Y. H., & Chiang, W. (1987). Volatile compounds ofthe wax gourd (Benincasa hispida, Cogn) and a wax gourd beverage. Journal ofFood Science, 52, 132–134.

Chyau, C. C., Ko, P. T., Chang, C. H., & Mau, J. L. (2003). Free and glycosidically boundaroma compounds in lychee (Litchi chinensis Sonn.). Food Chemistry, 80,387–392.

De Lumen, B. O., Stone, E. J., Kazeniac, S. J., & Forsythe, R. H. (1978). Formation ofvolatile compounds in green beans from linoleic and linolenic acids. Journal ofFood Science, 43, 698–708.

Gholap, A. S., & Bandyopadhyay, C. (1977). Characterisation of green aroma of rawmango (Magnifera indica L.). Journal of Science of Food and Agriculture, 28,885–888.

Guen, S. L., Prost, C., & Demaimay, M. (2000). Critical comparison of threeolfactometric methods for the identification of the most potent odorants incooked mussels (Mytilus edulis). Journal of Agricultural and Food Chemistry, 48,1307–1314.

Gunata, Y. Z., Bayonove, C. L., Bumes, R. L., & Cordonnier, R. E. (1985). The aroma ofgrapes I. Extraction and determination of free and glycosidically boundfractions of some grape aroma components. Journal of Chromatography, 331,83–90.

Hatanka, A. (1996). The fresh green odor emitted by plants – Review. Food ReviewsInternational, 12, 303–350.

Jordan, M. J., Margaria, C. A., Shaw, P. E., & Goodner, K. L. (2003). Volatilecomponents and aroma active compounds in aqueous essence and fresh pink

1332 J. Sharma et al. / Food Chemistry 122 (2010) 1327–1332

guava fruit puree(Psidium guajava L.) by GC-MS and multidimensional GC/GC-O.Journal of Agricultural and Food Chemistry, 57, 1421–1426.

Jordan, M. J., Tandon, K., Shaw, P. E., & Goodner, K. L. (2001). Aromatic profile ofaqueous banana fruit by gas chromatography–mass spectrometry (GC–MS) andgas chromatography–olfactometry (GC–O). Journal of Agricultural and FoodChemistry, 49, 4813–4817.

Nickerson, G. B., & Likens, S. T. (1966). Gas chromatography evidence for theoccurrence of hop oil components in beer. Journal of Chromatography, 21, 1–5.

Ott, A., Fay, L. B., & Chaintreau, A. (1997). Determination and origin of the aromaimpact compounds of yogurt flavor. Journal of Agricultural and Food Chemistry,45, 850–858.

Pollien, P., Ott, A., Montigon, F., Baumgartner, M., Mouz-Box, R., & Chaintreau, A.(1997). Hyphenated headspace–gas chromatography–sniffing technique:Screening of impact odorants and quantitative aromagram comparisons.Journal of Agricultural and Food Chemistry, 45, 2630–2637.

Sarry, J. E., & Gunata, Z. (2004). Plant and microbial glycoside hydrolases: Volatilerelease from glycosidic aroma precursors. Food Chemistry, 87, 509–521.

Sivarajan, V. V., & Balachandran, I. (1994). Ayurvedic drugs and their plant sources(1st ed.). New Delhi: Oxford and IBH Publishing.

Tressel, R., & Rewicki, D. (1999). Heat generated flavours and precursors. In R.Teranishi, E. L. Wick, & I. Horstein (Eds.), Flavour chemistry: 30 years of progress(pp. 305–325). New York: Kluwer Academic/Plenum Publishers.

Ullrich, F., & Grosch, W. (1993). Detection of potent odorants in foods by aromaextract dilution analysis. Trends in Food Science and Technology, 4, 68–73.

VanRuth, S. M., & Roozen, J. P. (1994). Gas chromatography/sniffing port analysisand sensory evaluation of commercially dried bell peppers (Capsicum annuum)after rehydration. Food Chemistry, 51, 165–170.

VanRuth, S. M., Roozen, J. P., Cozijnsen, J., & Posthumus, M. A. (1995). Volatilecompounds of rehydrated French beans, bell peppers and leeks. Part II. Gaschromatography/sniffing port analysis and sensory evaluation. Food Chemistry,54, 1–7.

Wu, C. M., Liou, S. E., Chang, Y. H., & Chaiang, W. (1987). Volatile compounds of thewax gourd (Beincasa hispida, Cogn) and a wax gourd beverage. Journal of FoodScience, 52, 132–134.