Research Article

Received: 28 June 2013, Revised: 29 September 2013, Accepted: 3 October 2013 Published online in Wiley Online Library: 21 November 2013

(wileyonlinelibrary.com) DOI 10.1002/pca.2483

Chemometric Discrimination of DifferentTomato Cultivars Based on Their VolatileFingerprint in Relation to Lycopene and TotalPhenolics ContentSonia A. Socaci,a Carmen Socaciu,a* Crina Mureşan,a Anca Fărcaş,aMaria Tofană,a Simona Vicaşb and Adela Pinteac

ABSTRACT:Introduction – The characteristic flavour of tomato is given by a complex mixture of sugars, acids, amino acids, minerals andvolatile metabolites. Of these, volatile compounds are considered to greatly influence the flavour of tomato fruits. The vola-tile aroma compounds and phytochemical content of tomatoes are dependent on genotype, environmental conditions andcultural practices, and can thus be used for cultivar discrimination.Objective – To assess the possibility of using the volatile profile of tomato to fingerprint and discriminate different tomatocultivars based on an ‘in-tube extraction’ technique coupled to gas chromatography, combined with mass spectrometry(ITEX/GC–MS) and a chemometric approach.Results – Using the ITEX/GC–MS technique, 61 volatiles were analysed and separated from tomato cultivars, with 58 being identified.The main volatiles identified in all tomato cultivars were: hexanal, trans-2-hexenal, 1-hexanol, 3-pentanone, 3-methylbutanol, 2-methylbutanol, 3-methylbutanal and 6-methyl-5-hepten-2-one. The lycopene content and total phenolic compound content of thetomato cultivars varied between 36.78 and 73.18mg/kg fresh weight (fw) and from 119.4 to 253.7mg of gallic acid equivalents(GAE) per kilogram fresh weight, respectively. Volatile fingerprint and phytochemical composition led to a good differentiation be-tween tomato cultivars, with the first two principal components explaining 89% of the variance in the data.Conclusion – The tomato cultivars studied were easily discriminated based on their characteristic volatile profile that wasobtained using the reliable ITEX/GC–MS technique. Principal component analysis revealed, in addition to volatile compounds,the important role played by the total phenolic content in tomato cultivar discrimination, which is highly correlated with phe-notypic and biochemical differences between tomato cultivars. Copyright © 2013 John Wiley & Sons, Ltd.

Supporting information can be found in the online version of this article.

Keywords: Anti-oxidant capacity; GC–MS; in-tube extraction; PCA; lycopene; phenolic compounds; tomatoes; volatiles

Introduction The characteristic flavour of tomato is given by a complex mix-

* Correspondence to: Carmen Socaciu, University of Agricultural Sciencesand Veterinary Medicine, Faculty of Food Science and Technology, 3–5Manastur St., 400372 Cluj-Napoca; Romania.E-mail: carmen.socaciu@usamvcluj.ro

a University of Agricultural Sciences and Veterinary Medicine, Faculty of FoodScience and Technology, 3-5 Manastur St., 400372 Cluj-Napoca, Romania

b University of Oradea, Faculty of Environmental Protection, 26 Gen. MagheruSt., 410048 Oradea, Romania

c University of Agricultural Sciences and Veterinary Medicine, Faculty ofVeterinary Medicine, 3-5 Manastur St., 400372 Cluj-Napoca, Romania

16

Tomato (Solanum lycopersicum) is a herbaceous plant of theSolanaceae family. It is cultivated all over the world, from thetropics to within a few degrees of the Arctic Circle, being alsothe most popular garden crop (Foolad, 2007). Its fruit is the sec-ond most consumed vegetable worldwide (http://faostat.fao.org). The contribution of tomatoes to a diverse and balanced dietis related to the biological active compounds they contain. Freshand processed tomatoes are an excellent source of vitamins,minerals and secondary phytochemicals, such as carotenoids,anthocyanins, flavonoids and other phenolic compounds (Luthriaet al., 2006; Odriozola-Serrano et al., 2008). For example, a mediumfresh tomato provides 47% of the recommended daily allowance(RDA) of vitamin C and 22% of the RDA of vitamin A (Foolad,2007). In addition, tomato is the main source of dietary lycopene.Lycopene is the predominant red carotenoid in tomatoes, knownto have a high antioxidant activity compared with other caroten-oids, such as astaxanthin and beta-carotene. Epidemiologicalstudies have demonstrated a positive correlation between a dietrich in lycopene and a reduced prostate cancer incidence, as wellas other health benefits (Giovannucci et al., 1995).

Phytochem. Anal. 2014, 25, 161–169 Copyright © 2013 John

ture of sugars, acids, amino acids, minerals and volatile compounds.The quantity and quality of phytochemicals detected in tomatofruits is known to be dependent on genotype characteristics as wellas on environmental factors and cultural conditions (Luthria et al.,2006). Thus, after genotype, the maturity stage at harvest is thesecond most important factor that influences the flavour andquality of tomatoes, considering that biosynthesis of aroma vola-tiles increases during maturation (Farneti et al., 2012).

Wiley & Sons, Ltd.

1

S. A. Socaci et al.

162

Based on their biosynthesis pathway, tomato volatiles havebeen classified into six groups: lipid-derived, carotenoid-related,amino-acid-derived, carbohydrate-derived, as well as being relatedto terpenoids and lignins (Buttery and Ling, 1993). The most abun-dant volatiles in tomato fruits are derived from lipids through theoxylipin pathway (Buttery and Ling, 1993). Tomato volatiles havebeen the subject of numerous studies, over 400 volatile com-pounds being identified, although only a limited number of theseare essentially contributing to tomato flavour (Tandon et al., 2000).Also worth mentioning are acetaldehyde, acetone, methanol,ethanol, l-penten-3-one, hexanal, cis-3-hexenal, 2-methylbutanol,3-methylbutanol, trans-2-hexenal, trans-2-heptenal, 6-methyl-5-hepten-2-one, cis-3-hexenol, geranylacetone, 2-isobutylthiazoleand β-ionone (Tandon et al., 2000; Krumbein et al., 2004).

The protocols most commonly used for the extraction of volatilecompounds from tomatoes or other vegetable matrices are thosebased on headspace sampling (Krumbein et al., 2004; Tikunovet al., 2005; Odriozola-Serrano et al., 2008; Rios et al., 2008;Lo Feudo et al., 2011) or solvent extraction (Xu et al., 2005). ‘In-tubeextraction’ (ITEX) is a relatively new technique (commerciallyavailable since 2006, by CTC Analytics AG, Zwingen, Switzerland)similar to purge-and-trap systems. The ITEX technique has beenapplied successfully for the extraction of volatile compounds fromdifferent matrices (Laaks et al., 2010; Socaci et al., 2013). Themetabolites extracted from the headspace above the plantmaterial using the ITEX technique are further separated by gaschromatography (GC) and fragmented to charged molecularfragments (ions) that are detected by mass spectrometry (MS).The unique fingerprint obtained by MS analysis is then used formetabolite recognition and identification. Thus, the ITEX techniqueefficiently combines the extraction, with selective concentration ofanalytes and rapid transfer of target compounds, to the GC–MSsystem (Tikunov et al., 2005; http://www.itex-headspace.com/pdf/ITEX.pdf). The data obtained from the ITEX/GC–MS analysis(volatile composition) are usually subject to different chemometricapproaches that allow sample classification based on their specificproperties (similarities and regularities in the data, i.e. patternrecognition; Gad et al., 2013).

The aim of the present work was to assess the possibility ofusing the volatile profile of tomato to fingerprint and discriminatedifferent tomato cultivars. In addition, the lycopene content andtotal phenolic compounds content were also determined for all to-mato cultivars included in the study. Discrimination between theselected tomato cultivars was achieved by chemometry, usingprincipal component analysis (PCA) and cluster analysis (CA) ofthe data obtained.

Experimental

Samples

Ten tomato (Solanum lycopersicum) cultivars were included in the study,as follows:

(1) five tomato fruits samples were collected at full maturity (red stage)from cultivars cultivated in conventional soil culture in the greenhousesof Vegetable growingDepartment of University of Agricultural Sciencesand Veterinary Medicine (UASVM) from Cluj-Napoca – cv. Tolstoi (TOL),cv. Balet (BAL), cv. Marisa (MAR), cv. Tamaris (TAM), cv. Cronos (CRO);

(2) five tomato fruit sampleswere imported cultivars, purchased from a localsupermarket (Cluj-Napoca, Romania) at commercial maturity – cv. SanMazzo from Holland (SM), cv. Chica from Holland (CHI), cv. Kumado fromBelgium (KUM), cv. cherry from Spain (CHE-ES), cv. cherry from Turkey(CHE-TK).

Copyright © 2013 Johnwileyonlinelibrary.com/journal/pca

All tomato fruits samples were collected/purchased in the sameperiod (May, 2012). Approximately 200 g of each fresh tomato samplewere ground and homogenised in a high-speed blender. Aliquots ofthe obtained puree were placed in 20mL vials (with no headspace)and kept at �20°C prior to extraction.

Extraction of volatile compounds

The extraction of volatile compounds was performed using the ITEXtechnique described in our previous work (Socaci et al., 2013). Five gramsof tomato puree were placed in a 20 mL headspace vial. The sealed vialwas incubated at 60°C for 20 min, with continuous agitation. After incuba-tion, using the headspace syringe, the volatile compounds from thegaseous phase from the vial were adsorbed repeatedly (30 strokes) into aporous polymer fibre microtrap (ITEX-2TRAPTXTA, Tenax TA 80/100 mesh).The thermal desorption of volatiles was made directly into the GC–MSinjector, after which the hot trap (250°C) was cleaned with N2. The above-mentioned procedures were performed automatically by a CombiPALAOC-5000 autosampler. All samples were analysed in triplicate.

GC–MS analysis

TheGC–MS analyseswere carried out on aGCMSQP-2010 (Shimadzu ScientificInstruments, Kyoto, Japan) model gas chromatograph and mass spectrometerequipped with a CombiPAL AOC-5000 autosampler. The volatiles were sepa-rated on a Zebron ZB-5ms capillary column of 50 m×0.32mm i.d and0.25μm film thickness. The carrier gas was helium, 1 mL/min, split ratio 5:1,injector temperature 250°C. The temperature programme used for thecolumn oven was: 35°C (hold for 10 min) to 50°C at 3°C/min to 150°C at6°C/min, to 200°C at 10°C/min and hold for 5 min. The ion-source temper-ature and interface temperature were set at 250°C and the MS mode wasmeans Electron Ionisation (EI). The mass range scanned was 35–350μm.

The identification of separated compounds was based on comparisonof the sample’s mass spectra with the mass spectra libraries, NIST27 andNIST147. All peaks found in at least two of the three total ion chromato-grams (TIC) were taken into account when calculating the total area ofpeaks (100%) and the relative areas of the volatile compounds.

Lycopene content

The total lycopene content of each tomato cultivar was determinedusing a rapid spectrometric technique, as proposed by Davis et al.(2003) for tomato and tomato products. The lycopene extraction wasachieved using reduced volumes of organic solvents. Briefly, approxi-mately 0.6 g of tomato puree were weighed from each cultivar andadded to a centrifuge tube that contained 5 mL of 0.05% (w/v) butylatedhydroxytoluene (BHT) in acetone, 5 mL of 95% ethanol and 10 mL of hex-ane. The samples were centrifuged at 180 rpm for 15 min on ice andthen 3 mL of deionised water were added to each vial and mixed. Thesamples were left at room temperature to allow phase separation. Usinga Shimadzu UV-1700 PharmaSpec spectrophotometer, the absorbanceof the upper hexane layer was measured at 503 nm using hexane asblank. The total lycopene content was expressed as milligrams lycopeneper kilogram tomato (Fish et al., 2002).

Total phenol compounds

The content of total phenolics was determined following a modified Folin–Ciocalteu method (Singleton et al., 1999). The extraction of phenoliccompounds was performed usingmethanol:water (80:20, v/v) as extractionsolvent. The samples were centrifuged at 6000 rpm, for 20 min at 4°C(Hettich centrifuge, model Micro 22R; Tuttlingen, Germany) and thenfiltered through Whatman No. 1 filter paper. An aliquot of 0.1mL of extractwas mixed with 6 mL of water and 0.5mL of Folin–Ciocalteu reagent. After4 min, 1.5mL Na2CO3 solution (7.5%) was added and the samples werebrought to a final volume of 10 mL with water. All samples were allowedto stand for 2 h at room temperature, before measuring the absorbanceat 725 nm. A calibration curve was performed using different

Phytochem. Anal. 2014, 25, 161–169Wiley & Sons, Ltd.

Volatile-profile Chemometric Discrimination of Tomato Cultivars

concentrations of standard gallic acid solutions (r2 = 0.999) in order to quan-tify the content in phenolic compounds of each tomato sample. The resultswere expressed as milligrams of gallic acid equivalents (GAE) per kilogramof fresh weight (fw) material.

Statistical analysis

The classification of tomato cultivars was achieved by subjecting the chro-matographic data, together with the results obtained for lycopene contentand total phenolics content, to principal component analysis (PCA) withcross-validation (full model size and centre data). Also a cluster analysis(CA) based on the K-means algorithm was performed on the chromato-graphic fingerprint of tomato cultivars. The chromatographic data werenormalised before PCA and CA. In order that all variables included in theanalysis had an equal chance to influence the model, we usedstandardisation as the scaling technique. All the statistical analyses wereperformed using Unscrambler X Version 10.1 software (CAMO SoftwareAS, Oslo, Norway).

16

Results and discussion

Fingerprinting of volatile components by ITEX/GC–MS analysis

The volatile profile of each tomato cultivar was determined usingan ITEX/GC–MS technique. A total of 61 volatile compounds wereseparated, of which 58 were identified by comparing their massspectra with NIST27 and NIST147 libraries, as well as based onthe retention indices from www.flavornet.org, www.pherobase.com or from the literature (Lo Feudo et al., 2011) (Table 1). Acharacteristic GC–MS chromatogram for the volatile profile ofone indigenous tomato cultivar (CRO) and one imported tomatocultivar (CHI) is presented in Fig. 1A and B.

The main volatile compounds identified in all the 10 tomato cul-tivars investigated in the study were: hexanal (16), trans-2-hexenal(17), 1-hexanol (20), 3-pentanone (6), 3-methylbutanol (11), 2-methylbutanol (12), 3-methylbutanal (1) and 6-methyl-5-hepten-2-one (38). Other major compounds found at least in one tomatocultivar were: cis-3-hexenol (18), cis-2-hexenol (19), 1-nitro-pentane(24), 3-pentenone (5), sabinene (49) and terpinolene (41). Butteryand Ling (1993) stated that a combination of hexanal, cis-3-hexenal,cis-3-hexenol, 3-pentenone, 3-methylbutanal, trans-2-hexenal,6-methyl-5-hepten-2-one, methylsalycilate, 2-isobutylthiazole andβ-ionone at appropriate concentrations are responsible for thearoma of fresh ripe tomato. Xu and Barringer (2010) also showedthe importance of these volatiles by measuring their release fromdifferent tomato varieties, in mouth space and nose space, duringchewing, using selected ion-flow tube mass spectrometry(SIFT/MS). They found that ‘green aldehydes’ such as hexanal,methylbutanal and nonanal were retained at a higher percentagein themouth after swallowing than cis-3-hexenal, trans�2-hexenal,3-pentenone and isobutyl alcohol. The concentration of hexanal(16) for the tomato cultivars used in our study varied between7.63% and 65.26% (Table 1). The highest concentrations of hexanal(16) were found in TOL and CHE-TK tomato cultivars (65.26% and62.59%, respectively) as well as in TAM, CHE-ES and KUM tomatosamples (45.27–54.08%). Methylbutanal was also found in allanalysed tomato samples, its concentration ranging from 1.02%to 5.89%. Instead, nonanal was characteristic of the imported to-mato cultivars (0.26–0.65%), being detected in only one indigenoustomato cultivar (Tamaris) and in a smaller concentration (0.11%).

Analysing the volatile fingerprint of tomato cultivars, from qual-itative and quantitative points of view, several remarks can bemade. The volatile profiles of BAL, CRO and MAR tomato cultivars

Phytochem. Anal. 2014, 25, 161–169 Copyright © 2013 John

are very similar. The main compound determined for these culti-vars was 1-hexanol (20) (25.05–27.34%). A characteristic of thesetomato samples was their high content in alcohols such as cis-3-hexenol (18), 3-methylbutanol (11) and 2-methylbutanol (12).According to Baldwin et al. (1998), cis-3-hexenol, 2-methylbutanoland 3-methylbutanol are compounds that are responsible for thesweetness of tomato fruits and significantly contribute to a to-mato-like flavour. The other two indigenous tomato cultivars(TAM and TOL) belong to another group with similar volatile com-position. In their case, hexanal (16) is the major compound,accounting for over 54% of total volatile constituents. Moreover,TAM and TOL tomato cultivars contained the highest concentra-tion of trans-2-hexenal (17) of all analysed samples (21.71% and13.51% respectively). These two compounds are derived fromlipids and are considered to be important for the fresh tomatoaroma (Buttery and Ling, 1993; Tandon et al., 2000). They impartfresh, grassy, green or floral notes to tomato aroma. Nevertheless,data presented by Tandon et al. (2000) suggested that qualitativeand quantitative changes in the perception of volatile compoundsin different media may occur. Thus, hexanal and trans-2-hexenalmay also contribute to stale notes in tomato flavour. The generalcharacteristic for the tomato cultivars cultivated in UASMV green-houses is that they contain a larger amount of nitrogen-containingcompounds, especially 1-nitropentane (24) than the importedtomato cultivars we analysed. The other nitro compound, 2-isobuthylthiazole (50) was detected in four indigenous tomatosamples (BAL, CRO, MAR, TAM) and in two imported tomato sam-ples (CHE-TK and SM), but in lower amounts than 1-nitropentane.These compounds are characteristic of tomato fruits and maycontribute to cultivar discrimination (Farneti et al., 2012).In the case of imported tomato cultivars, the CHI sample had a

very specific volatile profile, defined by the largest content insabinene (49) (30.84%), 2-methylbutanol (12) (12.28%) andterpinolene (41) (12.25%). Several other monoterpenes, such asα-pinene (26), α-phellandrene (44) or o-cymene (33) were charac-teristic of CHI tomato samples. These biogenic volatile compoundsare the result of some typical monoterpene or sesquiterpenesynthases action on substrates such as geranyl diphosphate orneryl diphosphate. However, in contrast to the fruits of many otherspecies, the tomato fruits are mostly devoid of such volatile ter-penes, probably due to the fact that extensive breedingprogrammes focused primarily on larger fruit yields and may havedecreased the amount of defensive terpenoids produced in thevegetative part of the plant (Falara et al., 2011). Chica was therichest tomato cultivar in monoterpenes content. It must be men-tioned that the other Dutch tomato cultivar, San Mazzo (SM), alsohad relatively large amounts of sabinene (11.1%), α-pinene(6.12%), β-pinene (1.34%) and terpinolene (3.47%) compared withthe other tomato cultivars we analysed.As expected, cherry tomato samples that originated from Spain

and Turkey (CHE-ES and CHE-TK) had a similar volatile composition,with some qualitative and quantitative differences between the vol-atile fingerprints of the samples (Fig. 2A and B). Although the majorvolatile compound for both CHE-ES andCHE-TK sampleswas hexanal(16) (45.7% respectively 62.59%), the othermajor flavour compoundswere different. In the case of the CHE-ES sample, 3-pentenone(13.48%) and trans-2-hexenal (10.35%) are the other main volatiles,whereas the CHE-TK tomato sample is richer in α-pinene, β-pinene,2-methylbutanol, 3-methylbutanol and methylsalicylate. Apart fromcherry-tomato samples (CHE-ES and CHE-TK), methylsalicylate wasdetected in only one other analysed tomato cultivar, namely Cronos(CRO). Tikunov et al. (2005) found that this phenolic compound is one

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/pca

3

Table

1.Meanrelativ

econcen

trations

(exp

ressed

as%

oftotalp

eakareas)an

dstan

dard

deviations

(SD)o

fvolatile

compo

unds

from

tomatovarie

tiesan

alysed

bythehe

adspace

ITEX

/GC–M

Stechniqu

e

No.

Com

poun

dna

me

Tomatocultivar

BAL

±SD

CRO

±SD

MAR

±SD

TAM

±SD

TOL

±SD

CHE-ES

±SD

CHE-TK

±SD

CHI

±SD

KUM

±SD

SM±SD

13-Methy

lbutan

al3.18

±1.75

2.17

±1.03

3.32

±0.86

3.12

±0.30

2.77

±0.04

1.13

±0.12

2.43

±0.07

1.55

±0.25

1.02

±0.20

5.89

±0.81

2Acetic

acid,

1-methy

lethylester

0.24

±0.27

1.01

±0.89

1.81

±1.57

0.19

±0.17

1.10

±0.28

––

––

––

––

––

32-Methy

lbutan

al0.30

±0.12

––

1.00

±0.33

0.24

±0.06

0.39

±0.12

0.21

±0.09

––

0.72

±0.14

0.80

±0.09

0.99

±0.15

4Not

iden

tified

0.28

±0.27

––

––

––

––

0.13

±0.13

––

––

––

––

51-Pe

nten

-3-one

––

––

––

––

––

13.49

±2.52

2.13

±0.74

2.76

±0.28

6.03

±0.23

6.08

±2.32

63-Pe

ntan

one

7.09

±0.72

9.03

±0.43

1.90

±1.32

––

––

5.66

±0.97

4.32

±0.31

4.14

±1.71

––

––

7Not

iden

tified

––

––

4.04

±0.87

––

––

––

––

––

––

––

8Pe

ntan

al–

––

––

–5.82

±0.35

4.35

±0.34

––

––

––

19.60

±1.32

5.14

±1.76

9Bu

tano

icacid,

methy

lester

––

––

––

––

––

0.29

±0.03

0.30

±0.08

––

––

0.30

±0.12

103-Methy

l-bu

tane

nitrile

0.54

±0.13

0.47

±0.19

1.05

±0.21

0.43

±0.25

0.31

±0.07

––

––

0.61

±0.20

––

1.34

±0.33

113-Methy

l-1-butan

ol8.89

±0.29

4.48

±0.28

17.13

±2.44

2.28

±0.19

0.51

±0.09

0.85

±0.13

2.94

±0.48

2.12

±2.30

––

1.91

±0.06

122-Methy

l-1-butan

ol3.83

±3.52

4.26

±0.29

7.57

±0.66

2.59

±0.18

1.49

±0.35

1.99

±0.53

3.10

±0.14

12.28

±1.62

2.66

±0.80

8.35

±1.10

133-Methy

l-pen

tana

l–

––

––

––

––

––

––

––

––

–0.08

±0.06

141-Pe

ntan

ol0.83

±0.63

0.93

±0.27

0.67

±0.28

0.92

±0.17

0.74

±0.30

1.19

±0.31

0.65

±0.16

––

3.41

±0.23

0.35

±0.32

15Bu

tano

icacid,2

-methy

l-,methy

lester

––

––

––

––

––

––

––

––

––

0.07

±0.01

16Hexan

al12

.89

±2.47

15.65

±2.12

7.63

±1.06

54.08

±1.13

65.26

±0.64

45.70

±2.55

62.59

±1.32

11.26

±3.52

45.27

±1.19

24.45

±1.97

172-Hexen

al,(E)

9.22

±1.58

15.19

±1.48

7.75

±2.35

21.71

±0.48

13.51

±3.21

10.35

±1.87

1.87

±0.23

1.27

±0.63

––

2.89

±0.30

183-Hexen

-1-ol,(Z)

12.27

±1.73

10.31

±1.01

6.39

±0.81

0.85

±0.08

0.12

±0.10

0.60

±0.29

––

––

––

––

192-Hexen

-1-ol,(Z)

3.78

±0.80

2.37

±0.63

1.65

±0.46

0.37

±0.15

––

––

––

––

––

––

201-Hexan

ol27

.34

±1.15

26.31

±1.98

25.05

±2.24

3.10

±0.87

2.42

0.44

0.73

±0.40

0.87

±0.11

3.31

±1.07

––

0.15

±0.06

211-Bu

tano

l,3-methy

l-,acetate

––

0.16

±0.03

0.39

±0.26

––

––

––

––

––

––

––

221-Bu

tano

l,2-methy

l-,acetate

0.12

±0.07

0.34

±0.04

0.17

±0.16

0.11

±0.02

––

––

––

––

––

––

23Non

ane

0.69

±0.17

0.62

±0.04

0.89

±0.29

0.35

±0.08

0.49

±0.23

0.70

±0.10

0.96

±0.22

1.53

±0.19

1.37

±0.16

1.01

±0.16

241-Nitro-pe

ntan

e3.80

±0.27

2.21

±0.11

7.46

±0.90

1.24

±0.08

2.12

±0.14

0.44

±0.03

0.59

±0.02

0.71

±0.07

1.79

±0.03

3.27

±0.44

25Hexan

oicacid,

methy

lester

0.12

±0.04

0.23

±0.04

0.10

±0.09

––

––

––

0.21

±0.04

––

––

––

26α-Pine

ne–

––

––

––

––

–0.44

±0.03

2.77

±1.21

1.68

±0.05

––

6.12

±0.77

27Cam

phen

e–

––

––

––

––

––

––

––

––

–0.05

±0.03

282-Hep

tena

l,(Z)-

––

––

––

0.33

±0.05

0.13

±0.06

0.50

±0.03

0.46

±0.04

––

0.57

±0.09

0.49

±0.25

29Be

nzalde

hyde

0.14

±0.05

––

0.14

±0.05

0.11

±0.06

0.30

±0.08

0.83

±0.09

0.82

±0.19

0.49

±0.16

0.49

±0.26

0.54

±0.03

302H

-Pyran

-2-

carboxalde

hyde

,5,6-dihy

dro-

––

––

––

0.06

±0.05

––

––

––

––

––

––

S. A. Socaci et al.

Phytochem. Anal. 2014, 25, 161–169Copyright © 2013 John Wiley & Sons, Ltd.wileyonlinelibrary.com/journal/pca

164

Table

1.(Con

tinued)

No.

Com

poun

dna

me

Tomatocultivar

BAL

±SD

CRO

±SD

MAR

±SD

TAM

±SD

TOL

±SD

CHE-ES

±SD

CHE-TK

±SD

CHI

±SD

KUM

±SD

SM±SD

31Dim

ethy

ltrisulph

ide

––

––

––

0.22

±0.04

––

0.43

±0.05

0.38

±0.03

0.87

±0.21

0.30

±0.07

1.45

±0.64

322,3-Dim

ethy

l-4-

hydroxy-2-

buteno

iclacton

e

––

––

––

––

––

0.08

±0.09

––

––

––

––

33o-Cym

ene

––

––

––

––

––

––

––

1.00

±0.02

––

0.33

±0.09

341-Hep

tano

l–

–0.08

±0.07

––

––

––

––

0.07

±0.03

––

––

––

35β-Pine

ne–

––

––

––

––

––

–0.36

±0.14

––

––

1.34

±0.26

361-Octen

-3-one

––

––

––

––

––

––

0.12

±0.04

––

––

––

37Ph

enol

––

––

0.19

±0.05

––

––

0.36

±0.36

0.51

±0.22

––

––

0.32

±0.05

386-methy

l-5-

hepten

-2-one

2.31

±0.53

1.60

±0.05

1.62

±0.44

0.45

±0.06

1.48

±0.77

4.55

±0.63

4.86

±1.81

3.44

±0.52

7.16

±0.60

3.42

±0.37

392-pe

ntyl-Furan

0.35

±0.21

0.59

±0.16

0.19

±0.03

0.45

±0.07

1.41

±0.16

1.42

±0.34

1.18

±0.59

0.37

±0.08

3.89

±0.56

2.16

±0.92

40Bicyclo[3.1.0]he

x-2-

ene,4,4,6,6-

tetram

ethy

l-

––

––

––

––

––

0.77

±0.12

––

––

––

––

41Terpinolen

e–

––

––

––

––

––

––

–12

.25

±0.24

––

3.47

±0.64

42Decan

e0.65

±0.14

0.56

±0.06

0.76

±0.34

0.32

±0.06

0.81

±0.27

0.98

±0.23

1.52

±0.50

2.05

±0.37

1.87

±0.38

1.29

±0.32

43Octan

al0.21

±0.20

0.48

±0.07

0.19

±0.17

0.09

±0.02

––

0.21

±0.03

0.21

±0.03

––

0.40

±0.08

0.19

±0.01

44α-Ph

elland

rene

––

––

––

––

––

––

––

1.30

±0.14

––

0.49

±0.10

45Acetic

acid,

hexyle

ster

0.14

±0.15

0.38

±0.05

0.19

±0.19

––

––

––

––

––

––

––

46Cyclohe

xene

,4-methy

l-3-

(1-m

ethy

lethyliden

e)-

––

––

––

––

––

––

––

0.58

±0.04

––

0.22

±0.05

47p-Cym

ene

––

––

––

––

––

0.11

±0.02

––

1.29

±0.10

––

0.75

±0.13

48Not

iden

tified

––

––

––

––

––

0.49

±0.05

––

––

––

––

49Sabine

n–

––

––

––

––

––

––

–30

.84

±1.48

0.86

±0.15

11.10

±2.58

502-Isob

utylthiazole

0.48

±0.11

0.10

±0.11

0.37

±0.09

0.14

±0.02

––

––

0.60

±0.09

––

––

0.62

±0.04

512-Ph

enylacetalde

hyde

––

––

––

––

––

––

––

––

––

0.32

±0.02

522-Octen

al,(E)

––

––

––

0.28

±0.08

0.26

±0.07

1.01

±0.30

0.78

±0.08

––

1.50

±0.16

1.09

±0.52

53Acetoph

enon

e0.19

±0.10

0.13

±0.05

0.14

±0.03

0.07

±0.02

––

0.96

±0.11

0.77

±0.13

0.42

±0.03

0.46

±0.33

0.56

±0.03

54Ocimen

e–

–0.13

±0.02

0.12

±0.08

––

––

1.47

±0.07

––

0.90

±0.09

––

0.27

±0.01

55Non

anal

––

––

––

0.11

±0.02

––

0.65

±0.02

0.49

±0.03

0.26

±0.03

0.55

±0.09

0.29

±0.02

56Be

nzoicacid

––

––

––

––

––

––

––

––

––

0.10

±0.09

57Methy

lsalicylate

––

0.13

±0.02

––

––

––

0.33

±0.35

0.66

±0.16

––

––

––

58Decan

al–

––

––

––

––

–0.09

±0.05

0.18

±0.06

––

––

0.07

±0.02

59β-Iono

ne–

––

––

––

––

–0.26

±0.06

––

––

––

––

60Geran

ylaceton

e0.14

±0.07

0.10

±0.01

0.11

±0.02

––

––

0.37

±0.13

0.34

±0.10

––

––

0.52

±0.15

611-Dod

ecan

ol–

––

––

––

––

–0.23

±0.07

––

––

––

0.15

±0.09

Volatile-profile Chemometric Discrimination of Tomato Cultivars

Phytochem. Anal. 2014, 25, 161–169 Copyright © 2013 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/pca

165

A

B

Figure 1. Chromatograms (TIC, total ion chromatogram) of headspaceITEX/GC–MS analysis of volatiles from tomato cultivars (A) Cronos and(B) Chica. The numbering of the peaks refers to Table 1.

A

B

Figure 2. Chromatograms (TIC, total ion chromatogram) of headspaceITEX/GC–MS analysis of volatiles from cherry tomato cultivars from (A)Spain and (B) Turkey. The numbering of the peaks refers to Table 1.

S. A. Socaci et al.

166

of the most variable volatiles in tomato flavour and can act as amarker compound, being responsible for volatile profile differencesbetween cultivars. β-Ionone (59), a β-carotene-derived flavourcompound, was found only in CHE-ES tomato cultivar. This ketoneimparts a fruity aroma to tomato flavour, being associated with a‘sweet’ or ‘floral’ note (Tandon et al., 2000).

Kumato tomato cultivar has a distinctive volatile profile. These to-matoes contained the highest quantity of pentanal (19.60%) as wellas a considerable amount of 3-pentenone (6.03%) and 6-methyl-5-hepten-2-one (7.16%). The importance of these compounds infresh tomato aroma is well known (Buttery and Ling, 1993).Whereas pentanal imparts a green, grassy odour, 3-pentenoneand 6-methyl-5-hepten-2-one contribute to the fresh and floralnotes, respectively (Tandon et al., 2000; Farneti et al., 2012).

Total lycopene content

Total lycopene content varied from 36.78 to 73.18mg/kg fw,depending on the tomato cultivar (Table 2). The highest concen-trations of lycopene were observed for San Mazzo and Chica culti-vars (73.18mg/kg fw and 73.14mg/kg fw, respectively), bothimported from Holland. For the other imported tomato cultivars,the concentration of lycopene ranged from 36.78 to 65.1mg/kgfw, whereas for the indigenous tomato cultivars it was between45.35 and 69.68mg/kg fw. From the 10 cultivars analysed, the low-est lycopene content was found for cherry cultivars (36.78mg/kgfw from Spain and 42.15mg/kg fw from Turkey). Other studieshave shown that the lycopene content of tomatoes is influenced

Copyright © 2013 Johnwileyonlinelibrary.com/journal/pca

not only by their genetic background but also by extrinsic factors,such as: growing medium, post-harvest treatment, stage of fruitmaturity, light intensity, day/night temperatures, irrigation systemor fertilisation (Kuti and Konuru, 2005; Helyes et al., 2012).Nevertheless, the results obtained in the present work for lycopenecontent are similar to those reported in the literature by otherauthors: from 20 to 193 mg of lycopene per kilogram of fresh to-mato (Odriozola-Serrano et al., 2008; Georgé et al., 2011; Helyeset al., 2012). The carotenoid degradation pathway is consideredto be one of the most important routes for aroma compoundformation in tomato fruits. Many recent studies focused on thepathways of tomato volatile biosynthesis, including biosynthesisof apocarotenoid derivatives. The main volatiles of tomatopredicted to derive from carotenoids by enzymatic cleavage are6-methyl-5-hepten-2-one, geranylacetone, β-ionone, pseudoiononeand citral (Tieman et al., 2006). Consistent with it being a lycopene-derived flavour, 6-methyl-5-hepten-2-one was found in all tomatosamples analysed, with higher amounts present in imported tomatocultivars. However, we could not establish a linear correlation be-tween the amount of lycopene and the 6-methyl-5-hepten-2-oneconcentration. Another volatile apocarotenoid found in our tomatosamples was geranylacetone (60), present at similar concentra-tions in cherry cultivars as well as in SM, BAL, CRO and MAR sam-ples. Although SM, BAL, CRO and MAR cultivars have a highlycopene content and cherry cultivars have the lowest lycopeneconcentration, the concentration of geranylacetone (60) waslower in the cultivars cultivated in UASMV greenhouses. Tiemanet al. (2006) suggested that geranylacetone must be derived from

Phytochem. Anal. 2014, 25, 161–169Wiley & Sons, Ltd.

Table 2. Lycopene content and total phenolic compoundscontent of the tomato cultivars analysed

Tomato cultivar Lycopene (mg/kg fw) TPC (mg GAE/kg fw)

BAL 69.67 ± 0.13 135.02 ± 2.85CRO 65.21 ± 3.83 195.67 ± 4.10MAR 52.47 ± 3.51 119.41 ± 0.95TAM 45.35 ± 0.44 189.61 ± 0.95TOL 62.01 ± 0.18 150.79 ± 0.95CHE-ES 36.78 ± 1.80 247.60 ± 3.95CHE-TK 42.15 ± 0.33 247.65 ± 1.98CHI 73.14 ± 1.24 253.69 ± 1.90KUM 65.10 ± 3.38 165.90 ± 1.03SM 73.18 ± 3.18 173.51 ± 1.98

Values are expressed as mean± SD.

Volatile-profile Chemometric Discrimination of Tomato Cultivars

a carotenoid, before lycopene in the synthetic pathway, whichdoes not accumulate to substantial levels in ripe tomato fruits.Another volatile apocarotenoid identified in only one of thetomato sample (CHE-ES) we analysed was β-ionone (59), producedby oxidative cleavage of β-carotene (Simkin et al., 2004). Consideringthat the β-carotene content is the determinant factor in β-iononeproduction, it can be predicted that CHE-ES has the larger contentin this carotenoid. Even though during tomato fruit ripening,the content in carotenoids (mainly lycopene) increases,this does not necessarily correlate with an increase ingeranylacetone or β-ionone. Simkin et al. (2004) suggested thatone reason for the low accumulation rates of these volatileapocarotenoids, compared with the significant increase ofcarotenoids during ripening, may be the lower availability ofthe substrate to a non-plastid targeted enzyme.

Figure 3. Principal component analysis bi-plot of volatile compoundsfrom the 10 tomato cultivars analysed. Two components togetherexplained 85% of the data variation.

16

Total phenol compounds

The concentration of total phenolics in the 10 tomato cultivars var-ied between 119.4 (MAR) and 253.7mg GAE/kg fw (CHI) (Table 2).Similar values for phenol derivatives found in CHI sample were alsofound in CHE-ES and CHE-TK samples (247.6mg GAE/kg fw and247.7mg GAE/kg fw, respectively). For the other samples, the totalconcentration of phenolics was under 200mgGAE/kg fw. The phe-nolics concentrations we determined are in agreement with datareported by others for different tomato cultivars (160–560mgGAE/kg fw) (Odriozola-Serrano et al., 2008; Georgé et al., 2011;Helyes et al., 2012). In addition to their antioxidant properties,the phenolic derivatives are also precursors of some volatile com-pounds in tomato fruits, through the phenylpropanoid pathway(Tikunov et al., 2010). One of the phenolic derivatives released asvolatile in tomato fruit is methylsalicylate (57), detected in onlythree of all analysed tomato samples (CHE-ES, CHE-TK and CRO).These tomato cultivars had a high phenolic content, with thehighest concentration for methylsalicylate (0.66%) being deter-mined in the CHE-TK sample. Another phenolic-derived volatile is2-phenylacetaldehyde (51), which was detected only in the SMsample in a concentration of 0.32%. In a complex metabolomicstudy, Tikunov et al. (2010) showed that the ability of tomato fruitsto release phenylpropanoid volatiles (such as methylsalicylate) isinfluenced by tissue disruption and it is due to cleavage of thecorresponding hexose-pentoside. Also, the detection of low quan-tities of phenylpropanoids volatiles in tomato fruits may be

Phytochem. Anal. 2014, 25, 161–169 Copyright © 2013 John

explained by the hypothesis that these volatiles are in differentconjugated forms (glycoside species) as a storage reserve offlavour volatiles (Tikunov et al., 2010).

Principal component and cluster analysis

The data matrix obtained from the concentrations of the volatilecompounds we detected in tomato samples was subjected initiallyto cluster analysis. The tomato cultivars were classified in 10 clus-ters, each cultivar being distributed into a different cluster. WhenPCA was applied on the chromatographic data matrix, the samecorrelation was observed between volatile composition and to-mato cultivar (total of 91.4% information in the first three compo-nents). Thus, the first two principal components explained 85% ofthe variance of data, showing a good discrimination between to-mato samples. Even if the two cherry tomato samples (CHE-ESand CHE-TK) were well differentiated, other tomato cultivars (Baletand Cronos) were not differentiated clearly enough (Fig. 3). Thus,as variables for PCA, total lycopene content and total phenolicswere also introduced in addition to the chromatographic profile.As shown in Fig. 4A, this allowed a better discrimination betweentomato cultivars, the first two principal components explaining89% of the variance of data (total of 95.4% information in the firstthree components). Using lycopene content and total phenolicstogether with volatile composition in PCA, led to a good differen-tiation between Balet and Cronos tomato cultivars.Principal component analysis is a useful tool to establish interre-

lationships between different variables, allowing us to detect andinterpret the sample patterns, their similarities and differences(Lo Feudo et al., 2011). The loading-plot of the first two principalcomponents (data not shown) revealed the importance of severalvariables (total phenolic content and lycopene content, as well ashexanal (16), 1-hexanol (20) and sabinene (49) concentrations) fortomato cultivar discrimination. In order to visualise more clearlythe importance of each variable, the correlation loadings plot(Fig. 4B) was computed. The compounds from the inner ellipseindicate 50% of the explained variance, whereas those foundin the outer ellipse (e.g. hexanal, 2-methylbutanol, 1-nitropentane,o-cymene, terpinolene, sabinene, phenolics) indicate 100% of theexplained variance. Regarding the importance of 1-nitropentane(24) as a discrimination marker, our findings corroborate withthose of Farneti et al. (2012), who also mentioned that this

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/pca

7

Figure 4. Principal component analysis bi-plot of (A) volatile com-pounds, lycopene and total phenolic content and (B) correlation loadingbi-plot for the tomato cultivars analysed. Two components togetherexplained 89% of the data variation. The numbering of the compoundsrefers to Table 1.

S. A. Socaci et al.

168

compound significantly contributes to the differences observedamong volatile profiles of tomato cultivars. The crucial role thatphenolicsmight play in determining the phenotypic and biochem-ical differentiation among tomato cultivars has also been revealedby other authors (Tikunov et al., 2010).

According to the results obtained from the PCA, the tomatocultivars taken in our study can be discriminated easily basedon their characteristic volatile profile obtained using the ITEX/GC–MS technique. As stated previously, ITEX/GC–MS hasproven to be a suitable, rapid and eco-friendly (no solvent usedfor extraction of volatiles) technique for fingerprinting thevolatile composition of different vegetable matrices, therebyallowing us to obtain valuable information for cultivardifferentiation.

Acknowledgements

This work was supported financially by a Research Grant of theUniversity of Agricultural Sciences and Veterinary MedicineCluj-Napoca, and by a grant from the Romanian NationalAuthority for Scientific Research, CNCS – UEFISCDI, project num-ber PN-II-ID-PCE-2011-3-0721. The authors also thank Mr BeatSchilling for technical support.

Copyright © 2013 Johnwileyonlinelibrary.com/journal/pca

ReferencesBaldwin EA, Scott JW, Einstein MA, Malundo TMM, Carr BT, Shewfelt RL,

Tandon KS. 1998. Relationship between sensory and instrumentalanalysis for tomato flavor. J Amer Soc Hort Sci 123: 906–915.

Buttery RG, Ling LC. 1993. Volatile components of tomato fruit and plantparts: relationship and biogenesis. In Bioactive Volatile Compoundsfrom Plants. American Chemical Society: Washington, DC; 22–33.

Davis AR, Fish WW, Perkins-Veazie P. 2003. A rapid spectrophotometricmethod for analyzing lycopene content in tomato and tomatoproducts. Postharvest Biol Technol 28: 425–430.

Falara V, Akhtar TA, Nguyen TTH, Spyropoulou EA, Bleeker PM,Schauvinhold I, Matsuba Y, Bonini ME, Schilmiller AL, Last RL,Schuurink RC, Pichersky E. 2011. The tomato terpene synthase genefamily. Plant Physiol 157: 770–778.

Farneti B, Cristescu SM, Costa G, Harren FJM, Woltering EJ. 2012. Rapidtomato volatile profiling by using proton-transfer reaction mass spec-trometry (PTR–MS). J Food Sci 77: C551–C559.

Fish WW, Perkins-Veazie P, Collins JK. 2002. A quantitative assay for lyco-pene that utilizes reduced volumes of organic solvents. J Food CompAnal 15: 309–317.

Foolad MR. 2007. Genome mapping and molecular breeding of tomato.Int J Plant Genomics 2007: 1–52.

Gad HA, El-Ahmady SH, Abou-Shoer MI, Al-Azizi MM. 2013. Application ofchemometrics in authentication of herbal medicines: a review.Phytochem Anal 24: 1–24.

Georgé S, Tourniaire F, Gautier H, Goupy P, Rock E, Caris-Veyrat C. 2011.Changes in the contents of carotenoids, phenolic compounds and vi-tamin C during technical processing and lyophilisation of red and yel-low tomatoes. Food Chem 124: 1603–1611.

Giovannucci E, Ascherio A, Rimm EB, Stampfer MJ, Colditz GA, Willett WC.1995. Intake of carotenoids and retinol in relation to risk of prostatecancer. J Natl Cancer Inst 87: 1767–1776.

Helyes L, Lugasi A, Pék Z. 2012. Effect of irrigation on processing tomatoyield and antioxidant components. Turk J Agric For 36: 702–709.

Krumbein A, Peters P, Brückner B. 2004. Flavour compounds and a quan-titative descriptive analysis of tomatoes (Lycopersicon esculentumMill.) of different cultivars in short-term storage. Postharvest BiolTechnol 32: 15–28.

Kuti OJ, Konuru HB. 2005. Effect of genotype and cultivation environ-ment on lycopene content in red-ripe tomatoes. J Sci Food Agric 85:2021–2026.

Laaks J, Jochmann MA, Schilling B, Schmid TC. 2010. In-tube extrac-tion of volatile organic compounds from aqueous samples: aneconomical alternative to purge and trap enrichment. Anal Chem82: 7641–7648.

Lo Feudo G, Macchione B, Naccarato A, Sindona G, Tagarelli A. 2011. Thevolatile fraction profiling of fresh tomatoes and triple concentrate to-mato pastes as parameter for the determination of geographical ori-gin. Food Res Int 44: 781–788.

Luthria DL, Mukhopadhyay S, Krizek DT. 2006. Content of total phenolics andphenolic acids in tomato (Lycopersicon esculentumMill.) fruits as influencedby cultivar and solar UV radiation. J Food Comp Anal 19: 771–777.

Odriozola-Serrano I, Soliva-Fortuny R, Martin-Belloso O. 2008. Effect ofminimal processing on bioactive compounds and color attributes offresh-cut tomatoes. LWT 41: 217–226.

Rios JJ, Fernández-García E, Mínguez-Mosquera I, Pérez-Gálvez A. 2008.Description of volatile compounds generated by the degradation ofcarotenoids in paprika, tomato and marigold oleoresins. Food Chem106: 1145–1153.

Simkin AJ, Schwartz SH, Auldridge M, Taylor MG, Klee HJ. 2004. The to-mato carotenoid cleavage dioxygenase 1 genes contribute to the for-mation of the flavor volatiles β-ionone, pseudoionone, andgeranylacetone. Plant J 40: 882–892.

Singleton VL, Orthofer RM, Lamuela-Raventos RM.1999. Analysis of totalphenols and other oxidation substrates and antioxidants by meansof Folin–Ciocalteu reagent. Meth Enzymol 299: 152–178.

Socaci SA, Socaciu C, Tofană M, Raţi IV, Pintea A. 2013. In-tube-extractionand GC-MS analysis of volatile components from wild and cultivatedsea buckthorn (Hippophae rhamnoides ssp. Carpatica) berry varietiesand juice. Phytochem Anal 24: 319–328.

Tandon KS, Baldwin EA, Shewfelt RL. 2000. Aroma perception of individ-ual volatile compounds in fresh tomatoes (Lycopersicon esculentum,Mill.) as affected by the medium of evaluation. Postharvest BiolTechnol 20: 261–268.

Phytochem. Anal. 2014, 25, 161–169Wiley & Sons, Ltd.

Volatile-profile Chemometric Discrimination of Tomato Cultivars

Tieman DM, Zeigler M, Schmelz EA, Taylor MG, Bliss P, Kirst M, Klee HJ.2006. Identification of loci affecting flavour volatile emissions in to-mato fruits. J Exp Bot 57: 887–896.

Tikunov Y, Lommen A, Ric de Vos CH, Verhoeven HA, Bino RJ, Hall RD,Bovy AG. 2005. A novel approach for nontargeted data analysis formetabolomics. Large-scale profiling of tomato fruit volatiles. PlantPhysiol 139: 1125–1137.

Tikunov Y, Ric de Vos CH, González Paramás AM, Hall RD, Bovy AG. 2010. Arole for differential glycoconjugation in the emission of phenylpropanoidvolatiles from tomato fruit discovered using a metabolic data fusionapproach1[W][OA]. Plant Physiol 152: 55–70.

Phytochem. Anal. 2014, 25, 161–169 Copyright © 2013 John

Xu Y, Barringer S. 2010. Comparison of volatile release in tomatillo and differ-ent varieties of tomato during chewing. J Food Sci 75: C352–C358.

Xu N, Gruber M, Westcott N, Soroka J, Parkin I, Hegedus D. 2005. A Methodfor the Solvent Extraction of Low-boiling-point Plant Volatiles.Phytochem Anal 16: 239–245.

Supporting Information

Supporting information can be found in the online version ofthis article.

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/pca

169