Phenolic compounds, carotenoids and antioxidant activity of three tropical fruits

Original Article

Phenolic compounds, carotenoids and antioxidant activity of three tropical fruits

Christian Mertz a, Anne-Laure Gancel a, Ziya Gunata b, Pascaline Alter a, Claudie Dhuique-Mayer a,Fabrice Vaillant a, Ana Mercedes Perez c, Jenny Ruales d, Pierre Brat a,*a Centre de Cooperation Internationale en Recherche Agronomique pour le Developpement (CIRAD), Departement PERSYST, UMR Qualisud, TA B-95/16,

F-34398 Montpellier Cedex 5, Franceb UMR Qualisud, Universite Montpellier 2 place E. Bataillon, 34095 Montpellier Cedex 5, Francec Centro Nacional de Ciencia y Tecnologıa de Alimentos (CITA), Universidad de Costa Rica, San Jose, Costa Ricad Escuela Politecnica Nacional (EPN), Quito, Ecuador

Journal of Food Composition and Analysis 22 (2009) 381–387

A R T I C L E I N F O

Article history:

Received 31 October 2007

Received in revised form 19 May 2008

Accepted 2 June 2008

Keywords:

Solanum quitoense

Solanum betaceum

Rubus glaucus

Rubus adenotrichus

Phenolic compounds

Carotenoids

HPLC-MS

ORAC

Food composition

Food analysis

A B S T R A C T

Major compounds (i.e. phenolic compounds and carotenoids) were analysed in the extracts of the edible

part of three tropical fruits: the Andean blackberry, the naranjilla and the tree tomato. Ellagitannins and

anthocyanins were predominant in blackberries and phenolic composition can be used to differentiate

the two species studied. Similar phenolic composition occurred in red and yellow tree tomato except for

anthocyanins which were absent in the yellow tree tomato. Phenolic acids were detected in the naranjilla

pulp. Carotenoids were analysed in the fruits. The composition in carotenoids was similar in the two

varieties of tree tomato and their vitamin A activity was calculated. Carotenol fatty acid esters were

predominant. b-Cryptoxanthin esters and b-carotene were the major carotenoids. The carotenoid

content was high compared to literature data, providing an important high vitamin A activity. In

blackberries and naranjilla, this class of compounds was found only at trace level. Finally, ORAC values

were estimated in different solvent extracts and results were compared with published data in common

fruits.

� 2009 Published by Elsevier Inc.

Contents lists available at ScienceDirect

Journal of Food Composition and Analysis

journal homepage: www.e lsev ier .com/ locate / j fca

1. Introduction

The interest of consumers for novel foods is increasingconsiderably. The new use of food ingredients and new types offood in the European food supply is of economic interest for tropicalcountries. Nevertheless, the access for bio-diverse products to theEuropean market needs many requirements, notably the knowledgeof their composition and nutritional value. As part of the PAVUCproject (Producing Added Value from Under-utilized tropical fruitCrops), three tropical fruits (Andean blackberry, tree tomato andnaranjilla), native from Latin America and consumed as fruit juice ordesserts, were selected. Besides their organoleptic quality, their realnutritional potential remained undetermined.

Polyphenols, together with carotenoids, are markers of thenutritional quality of foods. Polyphenols are known for theirantioxidant activity as radical scavengers and possible beneficialroles in human health, such as reducing the risk of cancer,cardiovascular disease, and other pathologies (Bravo, 1998;Hollman et al., 1996; Sellappan et al., 2002).

* Corresponding author. Tel.: +33 4 6761 65 03; fax: +33 4 6761 44 33.

E-mail address: [email protected] (P. Brat).

0889-1575/$ – see front matter � 2009 Published by Elsevier Inc.

doi:10.1016/j.jfca.2008.06.008

Blackberries are currently promoted as being a rich source ofpolyphenols, with ellagitannins and anthocyanins being the majorones (Hakkinen et al., 1999; Siriwoharn and Wrolstad, 2004).Anthocyanins were detected in the tree tomato (Bobbio et al.,1983; Wrolstad and Heatherbell, 1974) and no data was found inliterature about the phenolic composition of naranjilla.

Carotenoids have been reported in tree tomato (Rodriguez-Amaya et al., 1983). Except for the well-known provitamin Aactivity of some carotenoids, they could be involved in protectiveeffects against degenerative or cardiovascular diseases (Olson,1999) and are known for having antioxidant capacity (Olson,1996). To exploit the nutritional potential of these fruits, theircomposition in polyphenols and carotenoids together with theirantioxidant capacity were studied in this work.

2. Materials and methods

2.1. Plant materials

Two varieties (red and yellow) of tree tomato (Solanum

betaceum Cav.) and naranjilla (Solanum quitoense Lam.) wereharvested in Ecuador. The Andean blackberries were native from

C. Mertz et al. / Journal of Food Composition and Analysis 22 (2009) 381–387382

Costa-Rica (Rubus adenotrichus Schlech.) and Ecuador (Rubus

glaucus Benth.). After freezing, fruits were transported in refri-gerated containers to our laboratory and stored at�20 8C for one tofive months until analysis.

2.2. Sample preparation

All analysed fruits were at fully ripe stage. Blackberries (10 kg)were ground for 3 min in a Waring blender (Vorwerk, Montpellier,France). Thirty red tree tomato and 40 yellow tree tomato fruitswere separately peeled, ground with an ultra turax in dark roomand seeds were removed by sieving. For naranjilla, 30 fruits weresieved to afford the edible part of the fruit. The moisture contentsof Andean blackberries, tree tomato and naranjilla were82.5 � 0.5%, 87.5 � 0.3% and 91.1 � 0.2%, respectively. For each fruit,portions of the resulting pulp were immediately freeze dried forpolyphenol analysis. The remaining pulp was kept at �20 8C untilanalysis (less than one month). It was used for carotenoid analysisand determination of the antioxidant activity.

2.3. Chemicals

All solvents were of HPLC grade, purchased from Carlo Erba (Valde Reuil, France) except of methyl tert-butyl ether (MTBE) (Sigma–Aldrich, Steinheim, Germany). Folin-Ciocalteu reagent, sodiumchloride, anhydrous sodium sulphate, magnesium hydroxidecarbonate, potassium chloride and formic acid were purchasedfrom Carlo Erba (Val de Reuil, France). Chlorogenic and caffeicacids, cyanidin 3-O-glucoside, kaempferol, b-carotene, b-crypto-xanthin, lutein, and zeaxanthin were from Extrasynthese (Genay,France). Ferulic acid, quercetin, ellagic acid 2,6-di-tert-butyl-4-methylphenol (BHT) were from Sigma (L’isle d’Abeau, France).Fluorescein and 6-hydroxy-2,5,7,8-tetramethyl-2-carboxilic acid(Trolox) were purchased from Sigma (Steinheim, Germany). 2-20-azobis (2-amidinopropane) dihydrochloride (AAPH) was pur-chased from Wako Chemicals (Richmond, USA).

2.4. Polyphenolic compounds extraction

Two grams of powder were extracted twice for 15 min with60 mL of 70% aqueous acetone containing 2% formic acid. Theextracts were combined, filtered and concentrated under vacuumto remove acetone (40 8C). The aqueous layer was concentrated toyield aqueous acetone extract. An aliquot was analysed by liquidchromatography with diode array detection and mass spectro-metry (LC-DAD/MS). Minor phenolic compounds in blackberrieswere extracted as previously described (Mertz et al., 2007).

2.5. Total phenolic content (TPC)

The total phenolic content was determined by the Folin-Ciocalteu method optimized by George et al. (2005). 50 mL ofaqueous acetone extract were used for the quantification. Resultswere expressed as mg of gallic acid equivalent (GAE) per 100 g ofdry matter (DM). Analysis was made in triplicate.

2.6. HPLC-DAD-MSn analysis of phenolic compounds

Samples were filtered through a 0.45 mm filter (Millipore). TheHPLC analysis was carried out on a Waters 2690 HPLC systemequipped with Waters 996 DAD (Waters Corp., Milford, MA) andEmpower Software (Waters). The separation was performed at30 8C using a 250 mm � 4.6 mm, 5 mm particle size, endcappedreversed-phase Lichrospher ODS-2 (Interchim, Montlucon,France). The injection volume was 10 mL and the detection was

carried out between 200 and 600 nm. MSn analysis was carried outusing a LCQ ion trap mass spectrometer fitted with an electrosprayinterface (Thermo Finnigan, San Jose, CA, USA). The LC/MSparameters and HPLC solvents used are described in our previouspublished study (Mertz et al., 2007). Identifications were achievedon the basis of the ion molecular mass, MSn and UV–visible spectra.Blackberry phenolics were analysed as previously described(Mertz et al., 2007). Naranjilla and tree tomato phenolics wereanalysed using the following gradient: from 5% to 35% B in 50 min,from 35% to 50% B in 5 min, from 50% to 80% B in 5 min, after whichthe column was washed and equilibrated to the initial conditions.The percentage of formic acid in both solvents was 2% foranthocyanins and reduced to 0.1% for other phenolic compounds.

2.7. Quantitative analysis of polyphenols

The HPLC analysis was carried on a Dionex liquid chromato-graph equipped with model P680 pumps, an ASI 100 autosamplerand a UVD 340U diode array detector coupled to a HP ChemStation(Dionex, France), according to the previous published study (Mertzet al., 2007). Gradient conditions were the same as used foridentification. Hydroxycinnamic acids in naranjilla and treetomato extracts were quantified at 280 nm using calibrationcurves established with chlorogenic acid standard (R2 = 0.997).Each analysis was made in triplicate.

2.8. Carotenoid extraction

Carotenoid extraction was adapted from that described byTaungbodhitham et al. (1998). 3 g of puree was stirred for 5 minwith 80 mg of MgCO3 and 15 mL of extraction solvent (ethanol/hexane, 4:3 (v/v), containing 0.1% of BHT as antioxidant). Theresidue was separated from the liquid phase by filtration with afilter funnel (porosity no. 2) and washed with 15 mL of ethanol and30 mL of hexane. Organic phases were transferred in a separatoryfunnel and successively washed with 50 mL of 10% sodiumchloride and 3 mL � 50 mL of distilled water. The aqueous layerwas removed. The hexanic phase was dried under anhydroussodium sulphate, filtered and evaporated to dryness at 40 8C in arotary evaporator. The residue was dissolved in 500 mL ofdichloromethane and 500 mL of MTBE/methanol (80/20, v/v).Samples were placed in amber vials before HPLC analysis.

The saponification was carried out in 10% methanolic KOH,according to the method described by Fanciullino et al. (2006).Analyses were carried out under red light to avoid carotenoiddegradation during extraction and saponification.

2.9. HPLC-MSn analysis of carotenoids and related fatty acid esters

The HPLC apparatus was a Surveyor plus model equipped of anautosampler, a PDA detector and LC pumps (Thermo ElectronCorporation, San Jose, CA, USA). Carotenoids and related fatty acidesters were analysed according to the previously publishedmethod of Dhuique-Mayer et al. (2005). Carotenoids wereseparated along a C30 column (250 mm � 4.6 mm, 5 mm particlesize), YMC (EUROP, GmbH). The mobile phases were water/20 mMammonium acetate as eluent A, methanol/20 mM ammoniumacetate as eluent B and MTBE as eluent C. Flow rate was fixed at1 mL/min and the column temperature was set at 25 8C. A gradientprogram was performed: 0–2 min, 40% A/60% B, isocratic elution;2–5 min, 20% A/80% B; 5–10 min, 4% A/81% B/15% C; 10–60 min, 4%A/11% B/85% C; 60–71 min, 100% B; 71–72 min, back to the initialconditions for reequilibration. The injection volume was 10 mL andthe detection was carried out between 250 and 600 nm. Afterpassing through the flow cell of the diode array detector the

C. Mertz et al. / Journal of Food Composition and Analysis 22 (2009) 381–387 383

column eluate was split and 0.5 mL was directed to the ion trapmass. Experiments were performed in positive ion mode. Scanrange was 100–2000, scan rate: 1 scan/s. The desolvationtemperature was set at 250 8C.

2.10. Quantification of carotenoids and related fatty acid esters

Carotenoids and related fatty acid esters were quantified byHPLC using an Agilent 1100 System (Massy, France). The columnand gradient conditions were the same as used in mass spectro-metry analysis. The injection volume was 20 mL. Absorbance wasfollowed at 290, 350, 400, 450 and 470 nm using an Agilent 1100photodiode array detector. Chromatographic data and UV–visiblespectra were collected, stored and integrated using an AgilentChemstation plus software. Because reference standards forcarotenoid esters are not commercially available, the total amountof carotenoid esters was calculated as b-carotene equivalents inmicrograms per 1 g of puree. For this purpose, only the areascorresponding to peaks that disappear after alkaline treatmentwere summed. Quantification of b-carotene, b-cryptoxanthinzeaxanthin and lutein before and after saponification was achievedusing calibration curves established at 450 nm with authenticstandards, with five concentrations. Concentrations of standardsolutions were calculated by spectrophotometric measurementdissolving standard with the appropriate solvent and using a molarextinction coefficient (emol) (Britton et al., 1995). Purity ofstandards was verified by HPLC and photodiode array detection.Correlation coefficients ranged from 0.996 to 0.999. Limits ofdetection (LOD) and quantification (LOQ) were calculated for b-carotene, b-cryptoxanthin and lutein by preparing serial dilutionsof these compounds in mobile phase. Calibration curves and thenLOD and LOQ were determined with LOD = 3 � S/a andLOQ = 10 � S/a (where S is the standard deviation of the blanksignal and a is the slope of the calibration curve).

The concentration of each carotenoid was expressed asmicrograms per 1 g of fresh weight (FW). Concentrations aregiven as the mean of data of three assays. Recoveries weredetermined by adding internal standard (b-apo-80-carotenal)before extraction for unsaponified extracts and after saponificationfor saponified extracts.

The vitamin A activity was calculated using the followingformula: Vitamin A = (mg/b-carotene)/6 + (mg/b-cryptoxanthin)/12 and was expressed as retinol equivalents (RE) per kg of freshweight.

2.11. Antioxidant activity

2.11.1. Preparation of extracts

Two grams of edible part of the fruits were weighted inEppendorf tubs and centrifuged 5 min at 14,000 rpm. The super-natants were then recovered and constituted the crude extract(CE). Others extracts were obtained as follows: 3 g of edible part ofthe fruits were extracted during 30 min with 7 mL of acetone andthe mixtures were filtrated (Whatman, England) and washed with10 mL of acetone/water (7:3, v/v). The filtrates were evaporatedunder vacuum at 40 8C and then diluted with water in a 10 mL flaskand constituted the acetone extracts (AE). 5 mL of AEs werewashed with 2 mL � 5 mL hexane. Thus, we obtained the washedacetone extract (WAE). 3 mL of WAEs were loaded onto a 3 mLXAD-7 column to remove the non-phenolic compounds. Thecolumn was then washed with 2 mL � 5 mL of water. Phenoliccompounds were desorbed with 2 mL � 5 mL of methanol/water(8:2, v/v) and 2 mL � 5 mL of pure acetone. The eluates wereevaporated under vacuum at 40 8C and then diluted with water in a10 mL flask and constituted the XAD-7 Extracts (XAD-7E). The

extraction of carotenoids was achieved as described above toafford the Hexane extracts (HEs). The residues obtained afterevaporation of the hexanic phase were dissolved in pure acetone ina 10 mL flask to constitute the HEs.

2.11.2. ORAC assay

ORAC assays were performed according to Huang et al. (2002).We used with a microplate spectrofluorimeter TECAN Infinite 200(TECAN Austria GmbH) in 96-well polypropylene plates. Theexcitation and emission wavelengths were 485 � 9 nm and520 � 20 nm, respectively. Solutions were all prepared with75 mM Phosphate buffer (pH 7.4) except for the measurement ofthe ORAC value of HEs where the buffer was replaced by acetone inthe Trolox solutions and sample solutions. Each well was filled with160 mL of a 78.75 nM fluorescein (FL) solution (63 nM final in thewell) and 20 mL of the buffer or acetone (blank), the standard 0–40 mM Trolox solution (0–4 mM final), or the sample (crude, acetone,washed acetone, and hexane extracts with appropriate dilution). Theplate was incubated at 37 8C during 15 min before 20 mL of a 178 mMAAPH solution (17.8 mM final) were added. Fluorescein and troloxsolution were daily made using a 787.5 mM and a 500 mM stocksolution, respectively, and stored in the dark at 4 8C. AAPH was madedaily and discarded within 8 h if unused. After the AAPH addition, thefluorescence decay is measured every minute during 60 min.

The final values were calculated by using a regression equationbetween the trolox concentration and the net area under the FLdecay curve. The area under the curve (AUC) and the net AUC arecalculated as follow:

AUC ¼ 0:5þ f 1

f 0

þ f 2

f 0

þ f 3

f 0

þ . . .þ f 59

f 0

þ 0:5f 60

f 0

� �

where f0 is the initial fluorescence reading at 0 min and fi is thefluorescence reading at time i.

net AUC ¼ AUC sample � AUCblank

The relative trolox equivalent ORAC value is calculated as rela-tive ORAC value = [(AUCsample � AUCblank)/(AUCtrolox � AUCblank)](molarity of trolox/molarity of the sample)

The ORAC values were expressed as micromole Troloxequivalents per gram of fresh weight (mM TE/g FW).

2.11.3. Statistical analysis

All extractions were made in quadruplicates and four dilutionswere made for each extract. Each value was the mean of the ORACvalues from the four extractions that were the mean of the ORACvalues obtained for the four dilutions. The standard deviationswere also calculated for the four extractions. Analyses of variancewere performed with the XLSTAT 2007.6 software (Addinsoft).Differences at P < 0.05 were considered significant.

3. Results and discussion

3.1. Analysis of phenolic compounds

The identification of phenolic compounds in blackberries wasreported in our previous study (Mertz et al., 2007). Briefly, majorcompounds were anthocyanins (A1, A2, A3) and ellagitannins (E1,E2) and other phenolic compounds were minor (data not shown).These identifications were supported by literature data (Mullenet al., 2002; Mullen et al., 2003; Tanaka et al., 1993).

UV–visible spectra, LC–MS, and the subsequent fragmentationof the predominant ions in MS–MS were used to analyse aqueousacetone extracts of tree tomato and naranjilla. Whenever possible,chromatographic retention and literature data were used tosupport the identification of the compounds. Chromatographic

Fig. 1. 15–50 min segment of LC-DAD chromatograms at 510 nm of red tamarillo

extract (gradient 2% 5–35% en 50 min ). A4: delphinidin glucosyl rutinoside; A5:

delphindin rutinoside; A6: cyaniding rutinoside; A7: pelargonidin rutinoside.

C. Mertz et al. / Journal of Food Composition and Analysis 22 (2009) 381–387384

profiles of red and yellow tree tomato are very similar (data notshown) except of the presence of anthocyanins in the red variety.Fig. 1 shows the LC-DAD chromatogram at 510 nm of the red treetomato extract. Four compounds were detected with UV–visibledata characteristic of anthocyanins (Table 1). LC–MS data in thepositive ion mode showed molecular ions at m/z 773 (peak A4), 611(peak A5), 595 (peak A6), and 579 (major peak A7). MS2 spectrum ofA4 had fragment ions at m/z 627 (M-146, loss of a rhamnosyl unit),465 and 303 (loss of two hexosyl units). Peaks A5, A6 and A7 hadmolecular ions at m/z 611, 595 and 579, respectively, with thesame fragmentation scheme (minor ion at M-146, major ion at M-308). Thus, A4, A5, A6 and A7 were expected as being delphinidinglucosyl rutinoside, delphinidin, cyanidin, and pelargonidinrutinoside, respectively. Major anthocyanins A5, A6 and A7 havebeen previously identified in tree tomato fruit from New Zealand(Wrolstad and Heatherbell, 1974).

The analysis of tree tomato and naranjilla extracts revealed thepresence of two compounds with UV spectra matching withhydroxycinnamic acids (Table 1). LC/MS experiments in thenegative mode afforded ions at m/z 515 and 353 which fragmentedto produce ion at m/z 191. According to MS and literature data(Clifford et al., 2007; Lai et al., 2007) these compounds were

Table 1Identification of phenolic compounds by using their spectral characteristics in LC-

DAD, positive or negative ions in LC–MS and MSn, and previous identification dataa.

Tentative identification LC-DAD data (nm) LC–MS data (m/z)b

MS MS2/MS3

Tree tomato

Delphinidin glucosyl

rutinoside (A4)

280, 526 773 (+) 627, 303

Delphinidin rutinoside (A5) 280, 526 611 (+) 465, 303

Cyanidin rutinoside (A6) 280, 516 595 (+) 449, 287

Pelargonidin rutinoside (A7) 280, 502 579 (+) 433, 271

Dicaffeoylquinic acid 292, 318 515 (�) 353, 191

Caffeoylquinic acid 290sh, 326 353 (�) 191

Caffeoyl glucose 296sh, 328 341 (�) 179

Feruloyl glucose 298sh, 329 355 (�) 193

Naranjilla

Dicaffeoylquinic acid 292, 318 515 (�) 353, 191

Caffeoylquinic acid 290sh, 326 353 (�) 191

a Abbreviations: sh, maximum of the shoulder in the spectrum; (+): positive

mode; (�): negative mode.b In MS–MS, the most abundant parent ion in LC–MS is fragmented.

tentatively identified as dicaffeoylquinic (DCQA) and caffeoylqui-nic acids, respectively. Two other compounds with similar UVspectra, present in the tree tomato extracts only, had m/z at 341and 355 in the negative mode. They are identified as caffeoyl andferuloyl hexose, respectively, according to literature data (Cliffordet al., 2007; Maatta-Riihinen et al., 2003; Maatta-Riihinen et al.,2004).

3.2. Total phenolic content

Blackberries had an average concentration of TPC of 4250 mg/100 g DM and 6300 mg/100 g DM in R. adenotrichus and R. glaucus,respectively. This TPC is significantly higher than those found fornaranjilla (650) and tree tomato fruits (308–570). Blackberries aretraditionally a rich source of polyphenols. It is difficult to comparethe TPC content found in this study with literature data on berriesbecause most of the authors did not substract interferingcompounds (e.g. reducing sugars or ascorbic acid) leading thusto an overestimation of the TPC (George et al., 2005). Nevertheless,values obtained are higher than those mentioned in literature(Heinonen et al., 1998; Sellappan et al., 2002; Wada and Ou, 2002)suggesting that these blackberries have a high antioxidantpotential. No data about total phenolic content in the literaturewere found for tree tomato and naranjilla.

3.3. Content of phenolic classes

3.3.1. Blackberries

In agreement with previous papers, ellagitannins and antho-cyanins are the major phenolic classes that characterize Rubus

species (Kahkonen et al., 2001; Maatta-Riihinen et al., 2004). Theamounts found in R. glaucus and R. adenotrichus ranged from 1000to 3000 mg/100 g DM (data not shown).

3.3.2. Tree tomato

Pelargonidin rutinoside (115 mg/100 g DM) and delphinidinrutinoside (33 mg/100 g DM) were the two major anthocyanins inthe red variety. As expected, this phenolic class having a typicalpurple colour was absent in the yellow tree tomato. Allhydroxycinnamic acids identified were quantitatively higher inthe red tree tomato than in the yellow one (Table 2). These resultsare in agreement with the TPC, which is higher in the red treetomato.

3.3.3. Naranjilla

Only mono and dicaffeoyl quinic acids were quantified. The firstone was predominant (106 mg/100 g DM) while DCQA content waslower (5.4 mg/100 g DM). No data was found in the literature aboutthe phenolic composition of this fruit.

3.4. Analysis of carotenoids and related fatty acid esters

3.4.1. Qualitative analysis

The HPLC chromatograms of unsaponified extracts of treetomato fruit (red and yellow) indicated that carotenol fatty acidesters eluting after b-carotene were predominant (data notshown). LC/MS analysis in the positive mode revealed that estersof b-cryptoxanthin were the major compounds, with b-cryptox-anthin esterified with myristic and palmitic acids. These resultswere established on the basis of LC/MS and literature data(Breithaupt and Bamedi, 2001; Wingerath et al., 1996). Wheneverpossible, co-chromatography with authentic standards supportedthe identifications. b-Carotene was the major free carotenoidpresent in these unsaponified extracts. After saponification, themajor xanthophylls found in the tree tomato extracts were

Fig. 2. Fluorescent decay curves of fluorescein in the presence of Trolox at different

concentrations and crude extract (CE) of red tree tomato.

Table 2Contentsa of phenolic compounds in tree tomato.

Phenolic compounds Red tree tomato Yellow tree tomato

Anthocyanins

Delphinidin glucosyl rutinoside (A4) 5.3 � 0.2c b

Delphinidin rutinoside (A5) 32.7 � 0.4 b

Cyanidin rutinoside (A6) 12.1 � 0.2 b

Pelargonidin rutinoside (A7) 115.0 � 1.3 b

Hydroxycinnamic acids

Dicaffeoylquinic acid 21.0 � 0.3 17.1 � 0.2

Caffeoylquinic acid 54.8 � 0.4 32.8 � 0.2

Caffeoyl glucose 9.7 � 0.1 3.7 � 0.01

Feruloyl glucose 9.8 � 0.1 6.3 � 0.05

a The contents are expressed in milligrams standard equivalents per 100 g of dry

matter and are the average of three assays.b Not detected.c Standard deviation.

C. Mertz et al. / Journal of Food Composition and Analysis 22 (2009) 381–387 385

b-cryptoxanthin, lutein and zeaxanthin, together with two otherunidentified compounds at m/z 601 and 585, in the positive mode.In the unsaponified extracts, carotenol mono- and bis-fatty acidesters were detected. Thus, according to literature data, UVcharacteristics and MS analyses, lutein dimyristate, b-cryptox-anthin myristate and palmitate, zeaxanthin dimyristate, andzeaxanthin dipalmitate were tentatively identified (Breithauptet al., 2002; Khachik et al., 1988; Wingerath et al., 1996). Nostructural elucidation was performed to distinguish between theregioisomers. The chromatographic profiles of saponified andunsaponified extracts were identical in both red and yellow treetomato. The only difference was the slightly higher amount ofcarotenol esters (and free carotenoids after saponification) in thered variety. In blackberries and naranjilla, b-carotene was themajor carotenoid.

3.4.2. Quantification

Quantitative analysis of tree tomato extracts was performedbefore and after saponification. b-Apo-80-carotenal was used asinternal standard for the quantitative determination of somecarotenoids and carotenol fatty acid esters. In unsaponifiedextracts, carotenoid esters were major and represent 78% of thetotal carotenoid content (Table 3). b-Carotene was the major freecarotenoid (4.6 and 5.1 mg/g FW in the yellow and red tree tomato,respectively). The recovery of the internal standard withoutsaponification was about 93%.

After saponification, b-cryptoxanthin, lutein and zeaxanthinwere quantified using calibration curves established with theavailable corresponding standards. The results are presented in

Table 3Contentsa of carotenoids in tree tomato.

Carotenoid Yellow tree tomato Red tree tomato

Before saponification

Zeaxanthinb 0.1 � 0.02d 0.3 � 0.06

b-Cryptoxanthin 1.1 � 0.1 1.5 � 0.08

b-Carotene 4.6 � 0.3 5.1 � 0.3

Estersc 22.6 � 1.0 25.7 � 1.0

After saponification

Lutein 0.98 � 0.05 1.25 � 0.05

Zeaxanthinb 0.59 � 0.02 1.7 � 0.06

b-Cryptoxanthin 13.5 � 0.1 15.8 � 0.1

a Contents are expressed as mg standard equivalents per gram of fresh weight.

Results are the mean of three independent determinations. The limits of detection

(LOD) and the limits of quantification (LOQ) were 0.07 mg and 0.23 mg for b-

cryptoxanthin, 0.004 and 0.013 mg for b-carotene, 0.0051 and 0.017 mg for lutein.b Contents are expressed as lutein equivalents.c Contents are expressed as b-carotene equivalents.d Standard deviation.

Table 3. The recovery of the internal standard was more than 85%and b-cryptoxanthin content was the higher with values rangingfrom 14 to 16 mg/g FW. Lutein and zeaxanthin contents were lower(1.25 and 1.7 mg/g FW, respectively, in the red tree tomato). Thesevalues are in agreement with previous studies (Rodriguez-Amayaet al., 1983). The carotenoid content of the tree tomatoes is higherthan that of most tropical fruits, making them nutritionallyinteresting (Breithaupt and Bamedi, 2001; Kimura et al., 1991;Rodriguez-Amaya, 1999). Vitamin A values calculated were about2000 RE/kg FW. In blackberries and naranjilla, carotenoids werefound only at low level (data not shown).

3.5. Antioxidant activity assays

Many assays exist to evaluate the antioxidant activity but theORAC method remains the most utilized. It enables to determinethe antioxidant activity in both hydrophilic and lipophilic media. Inthe ORAC assay, the peroxyl radical reacts with the fluorescein toform non-fluorescent product (Prior et al., 2005). Fig. 2 shows thefluorescent decay curves of fluorescein in the presence of Trolox atdifferent concentrations and in the crude diluted extract (CE) of thered tree tomato. Out of concern for clarity, the curves of othersextract were not drawn on the Fig. 2 but appeared similar to thecurve of the red tree tomato extract (CE).

Table 4 shows the ORAC value (expressed as mmol Troloxequivalents per gram of fresh weight) of different extracts obtainedfrom the edible parts of the fruits. The crude extract correspondedto the centrifuged edible part of fruits. The acetone (AE), washedacetone (WAE) and XAD-7 (XAD-7E) extracts were more or lesspurified phenolic fractions. The hexane extract corresponded to thefraction of carotenoids (Fig. 3).

The ORAC values of the crude extracts ranged from 6.5 to 18.7for the yellow tree tomato and the Andean blackberry, respec-tively. These values are rather high when compared to those offruit juice usually consumed in Europe such as apple (ORACvalue = 1.9 mM TE/g of fresh fruit), tomato (1.6), red grape (4.0),orange (6.8) (Wang et al., 1996). The ORAC value of the Andeanblackberry (18.7) is in total agreement with the ORAC values (20.3–24.6) from other cultivars of blackberries obtained by Wang andLin (2000). For the CE extracts, none extraction was performed.Free hydrosoluble compounds, such as organic acids, vitamin C,some phenolic compounds, hydrosoluble, carotenoids, proteins,minerals, etc, were present. The antioxidant activities of these CEresulted from the interaction of all these compounds havingantioxidant or pro-oxidant activities.

The ORAC values of the AEs were higher compared to those of CE(from 8.1 to 42.6). 70% acetone enables the extraction of phenoliccompounds bounded to insoluble cell wall material, therefore not

Table 4ORAC values expressed as mM TE/g FW of different extracts of naranjilla, tree tomatoes and Andean blackberry.

Fruit CEa AEb WAEc XAD-7Ed HEe

Yellow tree tomato 6.5f � 0.3g Ah 8.1 � 0.4 B 7.3 � 0.3 C 4.5 � 0.4 D 0.14 � 0.02

Red tree tomato 10.0 � 0.3 A 14.8 � 0.5 B 14.7 � 0.7 B 12.1 � 0.1 C 0.18 � 0.06

Naranjilla 11.6 � 0.4 A 16.4 � 1.3 B 15.0 � 1.2 B 11.8 � 1.6 A 0.15 � 0.02

Andean blackberry 18.7 � 0.9 A 42.6 � 1.0 B 43.6 � 1.4 B 38.6 � 2.2 C 0.12 � 0.01

a Crude extract (CE).b Acetone extract (AE).c Washed acetone extract (WAE).d XAD-7 extract (XAD-7E).e Hexane extract (HE).f Mean of four repetitions.g Standard deviation.h Values with the same letter do not present significant difference (P = 0.05).

Fig. 3. Preparation of crude, acetone, washed acetone, XAD-7 and hexane extracts for ORAC assays.

C. Mertz et al. / Journal of Food Composition and Analysis 22 (2009) 381–387386

present in the crude centrifuged extract. It is particularly true forthe Andean blackberry extract. The acetone 70% solvent improvedconsiderably the extraction of anthocyanins. Thus, for the Andeanblackberry, the antioxidant activity of AE was more than twicehigher than the antioxidant activity of CE. Concerning the otherfruits, the ORAC values increased from 25% to 48% between CE andAE for the yellow and red tree tomato, respectively. Although theantioxidant activity was mainly due to the phenolic compounds, itmay be also due to the more hydrophilic carotenoids (i.e.

xanthophylls), organic acids, vitamin C and reducing sugars, alsoextracted with this solvent mixture.

The further step consisted in removing carotenoids by twoconsecutive liquid–liquid extractions with hexane (WAE). Nosignificant difference appeared for the naranjilla, red tree tomatoand Andean blackberry. However, there was a significantdifference between the AE and WAE for the yellow tree tomato,even if values were closed from each other. The combination ofmore hydrophilic carotenoids with other antioxidant compoundsdo not seem having neither a synergic nor an antagonist effect. TheORAC value of the WAE of Andean blackberry (43.6) is inagreement with the values obtained by Moyer et al. (2002). Theseauthors reported ORAC values between 41.6 and 78.8 mM TE/g FWfor different species of blackberry.

ORAC values of the XAD-7E ranging from 4.5 to 38.6 were alllower than the ORAC values of WAEs. The losses of antioxidant

activity, after purification on XAD-7, represented 11–38% of theWAEs. The higher the WAEs values, the lower were the losses ofantioxidant activity. We can note that the fruits were ranked in thesame order regarding their total phenolic content (expressed as mgGAE per 100 g of fresh weight) but no linear correlation can beestablished between them (data not shown). The purification ofWAEs on XAD-7 enabled to remove all the hydrophilic compoundsthat were not phenolic compounds, such as vitamin C, organicacids, minerals, etc. Thus, for all the studied fruits, phenoliccompounds were the main contributors of the hydrophilicantioxidant power.

Finally the ORAC values of the HEs that represented the extractof carotenoids were very low compared to the ORAC values of theXAD-7E. They ranged from 0.12 to 0.18 mM TE/g FW for the Andeanblackberry and the red tree tomato, respectively. These results aresimilar to those obtained by Wu et al. (2004), with ORAC valuesbeing 0.11 for the honeydew, 0.24 for the tomato and the kiwi, and0.35 for the grapefruit.

4. Conclusions

The analysis of polyphenols and carotenoids was achieved in theAndean blackberry, the tree tomato and the naranjilla. Ellagitanninsand anthocyanins were major phenolic compounds in blackberry. Toour knowledge, anthocyanins and hydroxycinnamic acids were

C. Mertz et al. / Journal of Food Composition and Analysis 22 (2009) 381–387 387

identified and quantified in tree tomato and naranjilla for the firsttime. Carotenol fatty acid esters were abundant in tree tomato withb-cryptoxanthin esters being the major one. b-Carotene was themajor hydrocarbon carotenoid. The carotenoid composition wassimilar in both varieties of tree tomato. The three fruits studied havea higher antioxidant potential than that of most of fruits, makingthem nutritionally interesting.

Acknowledgements

Tropical fruits were supplied by the Centro Nacional deInvestigacion en Tecnologia de Alimentos (CITA, Costa-Rica) andthe Escuela Politecnica Nacional (EPN, Ecuador). We express ourgratitude to Emmanuelle Meudec (INRA, UMR 1083, plateformePolyphenols) for assistance with LC–MS analysis; Gilles Morel(CIRAD) and Helene Fulcrand (Institut National de la RechercheAgronomique) for help and advices in this study. This work wasfinancially supported by the European Union (PAVUC project, INCOno. 015279).

References

Bobbio, F.O., Bobbio, P.A., Rodriguez-Amaya, D.B., 1983. Anthocyanins of the Bra-zilian Fruit Cyphomandra betacea. Food Chemistry 12, 189–195.

Bravo, L., 1998. Polyphenols: Chemistry, dietary sources, metabolism, and nutri-tional significance. Nutrition Reviews 56, 317–333.

Breithaupt, D.E., Bamedi, A., 2001. Carotenoids esters in vegetables and fruits: ascreening with emphasis on b-cryptoxanthin esters. Journal of Agricultural andFood Chemistry 49, 2064–2070.

Breithaupt, D.E., Wirt, U., Bamedi, A., 2002. Differentiation between lutein monoe-sters regioisomers and detection of lutein diesters from Marigold flowers(Tagetes erecta L.) and several fruits by liquid chromatography-mass spectro-metry. Journal of Agricultural and Food Chemistry 50, 66–70.

Britton, G., Liaanen-Jensen, S., Pfander, H., 1995. Carotenoids. Vol. 1B: Spectroscopy.Birkhauser, Basel, Switzerland.

Clifford, M.N., Wu, W., Kirkpatrick, J., Kuhnert, N., 2007. Profiling the chlorogenicacids and other caffeic acid derivatives of herbal Chrysanthemum by LC-MSn.Journal of Agricultural and Food Chemistry 55, 929–936.

Dhuique-Mayer, C., Caris-Veyrat, C., Ollitraut, P., Curk, F., Amiot, M.-J., 2005. Varietaland interspecific influence on micronutrient contents in citrus from the med-iterranean area. Journal of Agricultural and Food Chemistry 53, 2140–2145.

Fanciullino, A.L., Dhuique-Mayer, C., Luro, F., Casanova, J., Morillon, R., Ollitrault, P.,2006. Carotenoid diversity in cultivated citrus is highly influenced by geneticfactors. Journal of Agricultural and Food Chemistry 54, 4397–4406.

George, S., Brat, P., Alter, P., Amiot, M.J., 2005. Rapid determination of polyphenolsand vitamin C in plant-derived products. Journal of Agricultural and FoodChemistry 53, 1370–1373.

Hakkinen, S.H., Karenlampi, S.O., Heinonen, M., Mykkanen, H.M., Torronen, A.R.,1999. Content of the flavonols quercetin, myricetin, and kaempferol in 25 edibleberries. Journal of Agricultural and Food Chemistry 47, 2274–2279.

Heinonen, I.M., Meyer, A.S., Frankel, E.N., 1998. Antioxidant activity of berryphenolics on human low-density lipoprotein and liposome oxidation. Journalof Agricultural and Food Chemistry 46, 4107–4112.

Hollman, P.C.H., Hertog, M.G.L., Katan, M.B., 1996. Analysis and health effects offlavonoids. Food Chemistry 57, 43–46.

Huang, D., Ou, B., Hampsch-Woodill, M., Flanagan, J.A., Prior, R.L., 2002. High-throughput assay of oxygen radical absorbance capacity (ORAC) using a multi-channel liquid handling system coupled with a microplate fluorescence readerin 96-well format. Journal of Agricultural and Food Chemistry 50, 3122–3128.

Kahkonen, M.P., Hopia, A.I., Heinonen, M., 2001. Berry phenolics and their anti-oxidant activity. Journal of Agricultural and Food Chemistry 49, 4076–4082.

Khachik, F., Beecher, G.R., Lusby, W.R., 1988. Separation and identification ofcarotenoids and carotenol fatty acid esters in some squash products by liquidchromatography. 2. Isolation and characterisation of carotenoids and relatedesters. Journal of Agricultural and Food Chemistry 36, 938–946.

Kimura, M., Rodriguez-Amaya, D.B., Yokoyama, S.M., 1991. Cultivar differences andgeographic effects on the carotenoid composition and vitamin A value ofpapaya. Lebensmittel-Wissenschaft und Technologie 24, 415–418.

Lai, J.P., Lim, Y.H., Su, J., Shen, H.-M., Ong, C.N., 2007. Identification and character-ization of major flavonoids and caffeoylquinic acids in three Compositae plantsby LC/DAD-APCI/MS. Journal of Chromatography B 848, 215–225.

Maatta-Riihinen, K.R., Kamal-Eldin, A., Torronen, A.R., 2003. High performanceliquid chromatography (HPLC) analysis of phenolic compounds in berrieswith diode array and electrospray ionization mass spectrometric (MS)detection: Ribes species. Journal of Agricultural and Food Chemistry 51,6736–6744.

Maatta-Riihinen, K.R., Kamal-Eldin, A., Torronen, A.R., 2004. Identification andquantification of phenolic compounds in berries of Fragaria and Rubusspecies (family Rosaceae). Journal of Agricultural and Food Chemistry 52,6178–6187.

Mertz, C., Cheynier, V., Gunata, Z., Brat, P., 2007. Analysis of phenolic compounds intwo blackberry species (Rubus glaucus and Rubus adenotrichus) by high-perfor-mance liquid chromatography with diode array detection and electrospray iontrap mass spectrometry. Journal of Agricultural and Food Chemistry 55, 8616–8624.

Moyer, R.A., Hummer, K.E., Finn, C.E., Frei, B., Wrolstad, R.E., 2002. Anthocyanins,phenolics, and antioxidant capacity in diverse small fruits: Vaccinium, Rubus,and Ribes. Journal of Agricultural and Food Chemistry 50, 519–525.

Mullen, W., McGinn, J., Lean, M.E.J., MacLean, M.R., Gardner, P., Duthie, G.G., et al.,2002. Ellagitannins, flavonoids, and other phenolics in red raspberries and theircontribution to antioxidant capacity and vasorelaxation properties. Journal ofAgricultural and Food Chemistry 50, 5191–5196.

Mullen, W., Yokota, T., Lean, M.E.J., Crozier, A., 2003. Analysis of ellagitannins andconjugates of ellagic acid and quercetin in raspberry fruits by LC-MSn. Phyto-chemistry 64, 617–624.

Olson, J.A., 1996. Benefits and liabilities of vitamin A and carotenoids. Journal ofNutrition 126, 1208–1212.

Olson, J.A., 1999. Carotenoids and human health. Archivos Latinoamericanos deNutricion 49, 7S–11S.

Prior, R.L., Wu, X., Schaich, K., 2005. Standardized methods for the determination ofantioxidant capacity and phenolics in food and dietary supplements. Journal ofAgricultural and Food Chemistry 53, 4290–4302.

Rodriguez-Amaya, D.B., 1999. Latin American food sources of carotenoids. ArchivosLatinoamericanos de nutricion 49, 74S–84S.

Rodriguez-Amaya, D.B., Bobbio, P.A., Bobbio, F.O., 1983. Carotenoid compositionand vitamin A value of the brasilian fruit Cyphomandra betacea. Food chemistry12, 61–65.

Sellappan, S., Akoh, C.C., Krewer, G., 2002. Phenolic compounds and antioxidantcapacity of Georgia-grown blueberries and blackberries. Journal of Agriculturaland Food Chemistry 50, 2432–2438.

Siriwoharn, T., Wrolstad, R.E., 2004. Characterisation of phenolics in Marion andEvergreen blackberries. Journal of Food Science 69, 233–240.

Tanaka, T., Tachibana, H., Nonaka, G., Nishioka, I., Hsu, F.-L., Kohda, H., et al., 1993.Tannins and related compounds. CXXII. New dimeric, trimeric and tetramericellagitannins, lambertianins A-D, from Rubus lambertianus Seringe. Chemicaland Pharmaceutical Bulletin 41, 1214–1220.

Taungbodhitham, A.K., Jones, G.P., Walhlqvist, M.L., Briggs, D.R., 1998. Evaluation ofmethod for the analysis of carotenoids in fruits and vegetables. Food Chemistry63, 577–584.

Wada, L., Ou, B., 2002. Antioxidant activity and phenolic content of Oregon cane-berries. Journal of Agricultural and Food Chemistry 50, 3495–3500.

Wang, S.Y., Lin, H.-S., 2000. Antioxidant activity in fruits and leaves of blackberry,raspberry, and strawberry varies with cultivar and developmental stage. Jour-nal of Agricultural and Food Chemistry 48, 140–146.

Wang, H., Cao, G., Prior, R.L., 1996. Total antioxidant capacity of fruits. Journal ofAgricultural and Food Chemistry 44, 701–705.

Wingerath, T., Stahl, W., Kirsch, D., Kaufmann, R., Sies, H., 1996. Fruit juice carotenolfatty acid esters and carotenoids as identified by matrix-assisted laser deso-rption ionization (MALDI) mass spectrometry. Journal of Agricultural and FoodChemistry 44, 2006–2013.

Wrolstad, R.E., Heatherbell, D.A., 1974. Identification of Anthocyanins and Distribu-tion of Flavonoids in Tamarillo Fruit (Cyphomandra betacea (Cav.) Sendt.).Journal of the Science of Food and Agriculture 25, 1221–1228.

Wu, X., Gu, L., Holden, J., Haytowitz, D.B., Gebhardt, S.E., Beecher, G., et al., 2004.Development of a database for total antioxidant capacity in foods: A prelimin-ary study. Journal of Food Composition and Analysis 17, 407–422.