Food Chemistry 164 (2014) 324–331

Contents lists available at ScienceDirect

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

Effects of condensed tannins on anthocyanins and colour of authentic pomegranate (Punica granatum L.) juices Meltem Türkyılmaz a,⇑, Mehmet Özkan b a b

Institute of Food Safety, Ankara University, Diskapi Campus, Diskapi, 06110 Ankara, Turkey Department of Food Engineering, Faculty of Engineering, Ankara University, Diskapi, Ankara 06110, Turkey

a r t i c l e

i n f o

Article history: Received 8 February 2014 Received in revised form 19 April 2014 Accepted 7 May 2014 Available online 20 May 2014 Keywords: Pomegranate juice Condensed tannin Anthocyanin Polymeric colour

a b s t r a c t This study was conducted to determine the effects of condensed tannins (CT) on anthocyanins (ACNs) and colour of pomegranate juice (PJ) samples obtained from nine registered varieties in Turkey. CT-catechins (CTCs) reactive to vanillin and phloroglucinol adducts of CT contents were determined. CTC and ACN contents of PJs highly depended on variety (p < 0.01), and ranged from 31 to 155 mg/L juice and from 47 to 405 mg/L juice, respectively. As catechin–phloroglucinol content increased, ACN content also increased (r = 0.866). Strong logarithmic correlation between the ratio of ACN contents to catechin–phloroglucinol contents and polymeric colour (PC) values of the samples was found (r = 0.822). When PC value of PJs was P8% or ratio of ACN contents to catechin–phloroglucinol contents of PJs was 62.82, ACN contents of the samples determined by spectrophotometric method were higher than those determined by HPLC. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Colour has a great influence on the consumer acceptance of foods. Therefore, colour measurements are of great interest for both food scientists and food industry. Anthocyanins (ACNs) are responsible for the attractive red colour of pomegranate juices (PJs) as well as many other red-coloured fruit juices (Li, Pan, Cui, & Duan, 2010). ACN contents of fruit juices are commonly determined by the spectrophotometric (pH-differential method) and high-performance liquid chromatography (HPLC) methods (Lee, Rennaker, & Wrolstad, 2008). Since the pH-differential method is a simple and economical method to determine total monomeric ACN contents, it is especially preferred by laboratories which do not have the capability for expensive and time-consuming HPLC analysis (Lee et al., 2008). However, the presence of polymeric pigments in the sample may interfere with the measured values for monomeric ACNs, when the pH-differential method is used for quantification (Xu & Howard, 2012). In HPLC analysis, polymeric pigments are retained in the HPLC column; therefore, they do not interfere with monomeric ACN measurements. On the contrary, since polymeric pigments cannot be separated prior to spectrophotometric measurements, they may cause higher absorbance values (Lee, Durst, & Wrolstad, 2002).

⇑ Corresponding author. Tel.: +90 (312) 596 1087; fax: +90 (312) 317 8711. E-mail address: [email protected] (M. Türkyılmaz). http://dx.doi.org/10.1016/j.foodchem.2014.05.048 0308-8146/Ó 2014 Elsevier Ltd. All rights reserved.

Polymeric pigments are formed by the reaction of monomeric ACNs with condensed tannins (CTs) or flavan-3-ols, such as catechin or epicatechin. Multiple phenolic hydroxyl groups of CTs lead to the formation of complexes with ACNs (Xu et al., 2012). The formation of complexes of CTs with ACNs contributes to colour stability (Singleton & Trousdale, 1992). ACN–flavanol reaction occurs via two major ways: indirect or direct condensation (Drinkine, Glories, & Saucier, 2005). Indirect condensation is known as acetaldehyde mediated condensation. In this reaction, an ethyl bridge is formed between the flavonoids (Drinkine et al., 2005), when ethanol is oxidised to form acetaldehyde in beverages as well as during wine production. Direct polymerisation is a slower condensation reaction than indirect condensation and oxygen is not required for this reaction. Direct condensation reactions produce tannin–ACN (T–A+) and ACN–tannin (A+–T) adducts, resulting in polymeric pigments (Drinkine, Lopes, Kennedy, Teissedre, & Saucier, 2007). In the formation of A+–T, the ACN is in the flavylium form (A+) and plays a role as an electrophile (Salas, Fulcrand, Meuder, & Cheynier, 2003). The interaction between flavanol, which has nucleophilic properties, and the flavylium cation leads to formation of the colourless flavene (A–T), which can either be oxidised to the red flavylium (A+–T) (Liao, Cai, & Haslam, 1992) or proceeds to a colourless cyclic condensation product with an A-type bond [A–(4–8, 2-O-7)–T] (Salas et al., 2003). In the formation of T–A+, CT (T–T) plays a role as an electrophile, whilst the ACN, in its hydrated hemiketal form (AOH), plays a role as a nucleophile (Salas et al., 2003). This

M. Türkyılmaz, M. Özkan / Food Chemistry 164 (2014) 324–331

generates the colourless dimer (T-AOH) that dehydrates to the red flavylium form (T–A+). These reactions are expected to be pHdependent, as the proportion of cationic forms (A+ and T+) increases with increasing acidity (Salas et al., 2003). This study was conducted to determine the effects of CTs on ACNs in PJ samples produced from nine registered varieties in Turkey. Moreover, the comparison of commonly used methods, i.e., spectrophotometric and HPLC methods, for monomeric ACN determination was carried out. Furthermore, the correlations of CT contents with polymeric colour (PC) and colour density (CD) values of authentic PJs were also determined. Finally, the effects of individual ACNs in PJ samples on PC formation were investigated. 2. Materials and methods 2.1. Chemicals and reagents Cyanidin 3-O-glucoside and cyanidin 3,5-O-diglucoside standards were purchased from Sigma (St. Louis, MO); pelargonidin 3-O-glucoside and pelargonidin 3,5-O-diglucoside were purchased from Fluka (Seelze, Germany); and delphinidin-3-O-glucoside was purchased from Polyphenols Laboratories AS (Sandnes, Norway). Catechin and epicatechin standards were purchased from Extrasynthèse (Genay, France); rutin, chlorogenic acid, gallic acid and ferulic acid were purchased from Fluka (Seelze, Germany). All reagents used for liquid chromatography were HPLC grade and purchased from Merck (Darmstadt, Germany). All other reagents were analytical grade and obtained from Merck. In all analyses, ultra-pure water was used (Millipore Simplicity UV, Molsheim, France). 2.2. Samples Nine fully-ripe registered pomegranate (Punica granatum L.) varieties were harvested in November 2008. Eight out of 9 registered pomegranate varieties (Izmir 1255, 1264, 1473, 1477, 1478, 1479, 1508 and 1513) were obtained from Aegean Agricultural Research Institute (Menemen, Izmir). The last variety (Hicaznar), which is the most preferred variety by Turkish fruit juice industry due to its attractive red-violet colour and sour–sweet taste, was obtained from the Alata Horticultural Research Institute (Erdemli, Mersin). The total soluble contents (Brix), titratable acidity and pH values of the samples were published by Türkyılmaz (2013). The pomegranates were processed immediately. Before juice extraction, pomegranates were washed in cold tap water and drained. Damaged pomegranates were discarded. The top and bottom of the pomegranate husks were removed with a sharp stainless steel knife to prevent microbial contamination. Then, the outer skins of 10 kg of pomegranates from each variety were hand-peeled. The juicy sacs from fruit pericarp were separated by hand, placed in a muslin cloth and pressed with a laboratory hand press. The yield of the unprocessed PJs obtained from each variety was ca. 32%. 2.3. Monomeric ACN analysis The total monomeric ACN content was determined using the pH differential method described by Giusti and Wrolstad (2005). The pH of juice samples was brought to 1.0 with potassium chloride and 4.5 with sodium acetate buffers. The dilutions were then allowed to equilibrate for 15 min at room temperature (22 °C). Prior to absorbance measurements, the solutions were filtered through a 0.45-lm PVDF (polyvinylidene fluoride) filter (Sartorious AG, Goettingen, Germany) to remove the haze. The absorbance of equilibrated solutions at 512 nm (kmax) for ACN content and

325

700 nm for haze correction was measured on a UV–Vis double beam spectrophotometer (ThermoSpectronic Helios-a, Cambridge, England) with 1-cm path length disposable cuvettes (Brand Gmbh, Wertheim, Germany). All absorbance measurements were carried out at room temperature against deionised water as a blank. Pigment content was calculated as cyanidin-3-O-glucoside equivalents with a molecular weight of 449.2 and an extinction coefficient of 26 900 L/(cm mol). The difference in absorbance values at pH 1.0 and 4.5 was directly proportional to ACN concentration. All ACN measurements were replicated two times. 2.4. CD and PC content CD and percent PC contents were determined using the bisulphite bleaching method described by Giusti et al. (2005). Total CD is a measure of the colour strength of the solution, whilst PC is an indicator of polymerised pigments, including tannins, and brown compounds (Somers, 1971). Prior to absorbance measurements, the solutions were filtered through a 0.45-lm PVDF filter (Sartorious AG) to remove the haze. The absorbance of bisulphite-treated and non-treated solutions were measured at 420 nm for brown pigments, 530 nm (kmax) for monomeric ACNs and 700 nm for haze correction on the UV–Vis double-beam spectrophotometer (ThermoSpectronic Helios-a). Disposable cuvettes (Brand Gmbh) of 1cm path length were used. Absorbance measurements were carried out at room temperature and made against deionised water as a blank. CD and PC measurements were replicated two times. 2.5. HPLC separation of ACNs 2.5.1. ACN purification The ACNs were purified on a C-18 cartridge (Waters Corp., Milford, MA) using a vacuum manifold system (Waters). Prior to sample loading, the cartridge was activated with 5 mL ethyl acetate followed by 5 mL methanol (containing 0.01% HCl, v/v) and 2 mL aqueous 0.01% HCl (v/v). After loading 1 mL of PJ, the cartridge was washed with 2 mL aqueous 0.01% HCl to remove compounds not adsorbed by the column, such as sugars and organic acids (Skrede, Wrolstad, & Durst, 2000). The cartridge was then dried under a stream of nitrogen for 10 min. Elution of ACNs was carried out by rinsing the cartridge with 2 mL methanol (containing 0.01% HCl, v/v). The methanolic extract containing ACNs was then evaporated to dryness under a stream of nitrogen at 40 °C (Caliper TurboVap LV, Hopkinton, MA), and ACNs were dissolved in 2 mL aqueous 0.01% HCl. The resulting extract was filtered through a 0.22-lm PVDF filter (Sartorius AG) directly into an amber-coloured autosampler vial, and the filtered extract was immediately analysed by-high performance liquid chromatography (HPLC). 2.5.2. Instrumentation and chromatography Separation and quantification of ACNs were performed using HPLC (Agilent 1200 series, Waldbronn, Germany) with a binary pump, a photodiode array (PDA) detector, a thermostatted autosampler, a degasser and a thermostatted column compartment. Chromatographic data were recorded and processed using Agilent 1200 series ChemStation rev.B.02.01 software. ACNs were separated on a C18 (5 lm) column (250  4.6 mm) (Phenomenex, Inc., Torrance, CA) with a C18 (5 lm) guard column (4  3 mm, 5 lm) (Phenomenex). The eluents used were (A) 100% acetonitrile and (B) O-phosphoric acid, acetic acid, acetonitrile and water (1:10:5:84; v/v/v/v) with a flow rate of 1 mL min 1. Separation was performed with gradient elution using a modification of the elution profile described by Skrede et al. (2000). The linear gradient programme for the separation of pomegranate ACNs was as follows: from 0% to 12% A in 10 min, from 12% to 22%

326

M. Türkyılmaz, M. Özkan / Food Chemistry 164 (2014) 324–331

A in 10 min, and holding at 22% A for 5 min. The sample injection volume was 50 lL, the column temperature was 25 °C, and the detector was set at 520 nm. Identification of ACNs was carried out by comparing retention times and absorption spectra of unknown peaks with external reference standards. Quantification of ACNs was carried out using calibration curves of the following external reference standards: cyanidin 3-O-glucoside (r2 = 0.9995), cyanidin 3,5-O-diglucoside (r2 = 0.9834), delphinidin 3-O-glucoside (r2 = 0.9749), pelargonidin 3-O-glucoside (r2 = 0.9944) and pelargonidin 3,5-O-diglucoside (r2 = 0.8980). The calibration curves for each ACN standard contained 7 data points. Quantification of total ACNs by HPLC was calculated based on cyanidin 3-O-glucoside. Five out of 6 ACNs in PJ samples were identified by comparing retention times and absorption spectra of unknown peaks with external reference standards. Because the commercial standard for delphinidin-3,5-diglucoside was not available at the time of analysis, mass spectrometric detection was carried out for delphinidin-3,5-diglucoside on the same HPLC system used for the identification of ACNs in PJ. The mass/charge (m/z) ratio of 627 for delphinidin-3,5-diglucoside was used (Brito et al., 2007). The diode-array detector (DAD) was interfaced with a mass spectrometer (Agilent Series 1200 HPLC system) with an electrospray ionisation (ESI) source operating in positive ionisation mode. Nitrogen was used at a flow rate of 12 L min 1 and a pressure of 35 psi both as a drying and a nebulising gas. The nebuliser temperature was set at 250 °C, and a potential of 2000 V was used on the capillary. The mobile phase consisted of 100% acetonitrile (eluent A) and 1% formic acid (eluent B). The other chromatographic conditions were the same as described above for HPLC separation of major ACNs in PJ. 2.6. HPLC separation of phenolics except for tannins 2.6.1. Purification The phenolics were purified on a C-18 cartridge (Waters) using a vacuum manifold system (Waters). Prior to sample loading, the cartridge was activated with 10 mL methanol and 2.5 mL water (adjusted to pH 7.0 with 0.1 N NaOH) (Revilla, Bourzeix, & Alonso, 1991). After loading 1 mL of PJ (adjusted to pH 7.0 with 0.1 N NaOH), the cartridge was washed with 10 mL water (adjusted to pH 7.0 with 0.1 N NaOH) to elute phenolics except for tannins (Revilla et al., 1991). The resulting extract was filtered through a 0.22-lm PVDF filter (Sartorius AG) directly into an amber-coloured autosampler vial, and the filtered extract was immediately injected into the same HPLC used for ACN analysis. 2.6.2. Instrumentation and chromatography The phenolics were separated on a C18 (5 lm) column (250  4.6 mm) (Phenomenex) with a C18 (5 lm) guard column (4  3 mm, 5 lm) (Phenomenex). The eluents used were (A) 100% acetonitrile and (B) 1% formic acid with a flow rate of 0.7 mL min 1. Separation was performed with gradient elution using a modification of the elution profile described by Lee et al. (2002). The linear gradient program for the separation of pomegranate phenolics was as follows: from 2.5% to 25% A in 45 min, from 25% to 50% A in 5 min, from 50% to 100% A in 5 min, holding at 100% A for 13 min and from 100% to 2.5% A in 17 min. The sample injection volume was 30 lL, the column temperature was 20 °C, and the detector was set at 280 nm and 360 nm. Identification of phenolics was carried out by comparing retention times and absorption spectra of unknown peaks with external reference standards. Quantification of phenolics was carried out using calibration curves of the following external reference standards: gallic acid (r2 = 0.999), chlorogenic acid (r2 = 0.999), catechin (r2 = 0.999), epicatechin (r2 = 0.985), ferulic acid (r2 = 0.999) and rutin (r2 = 0.990).

The calibration curves for each phenolic standard contained 7 data points. 2.7. Condensed tannin-catechin (CTC) contents CTC reactive to vanillin were analysed according to Tanner and Brunner (1979). The analyses were carried out in three test tubes. In the first test tube, 4.0 mL of vanillin solution were added to 2 mL of diluted juice samples (1:100 ratio, juice:deionised water). In the second test tube, 4.0 mL of H2SO4 solution (70%, v:v) were added to 2 mL of diluted juice samples. In the third test tube, 4.0 mL of vanillin solution were added to 2 mL deionised water as a blank. After 15 min, the absorptions of the mixtures in all tubes were measured at 500 nm. However, after adding the vanillin and H2SO4 solutions, the mixtures in the test tubes were cooled in an ice-water bath so that the temperature of the tube contents could not exceed 35 °C. Quantification of CTC was carried out using calibration curves of catechin as an external standard. The calibration curves contained 6 data points. For all points of standard concentrations, the net absorbance value was calculated by subtracting the absorbance values for the mixtures in the second and third tubes from the absorbance value for the mixture in first tube. The net absorbance values were plotted against the standard concentrations, and a linear regression equation was obtained. Similarly, the net absorbance values were also determined for all samples, and CTC contents were calculated using the linear equation. The content of total CTC was expressed as mg (+)-catechin/L PJ. All samples were analysed in two replications. 2.8. Phloroglucinolysis of CTs for the determination of the nature and proportion of constitutive units of CTs To determine the monomeric composition and mean degree of polymerisation (mDP) of CTs, depolymerisation of CTs was carried out in the presence of excessive phloroglucinol according to phloroglucinolysis procedure described by Kennedy and Jones (2001). For this purpose, CTs were purified (Gil et al., 2012) and then treated with phloroglucinol. 2.8.1. CTs purification The CTs were purified on a C-18 cartridge (Waters) using a vacuum manifold system (Waters). Prior to sample loading, the cartridge was activated with 10 mL methanol and 15 mL deionised water. After loading 1 mL of PJ, the cartridge was washed with 15 mL water and elution of CTs was carried out by rinsing the cartridge with 12 mL methanol (Gil et al., 2012). The methanolic extract containing CTs was immediately evaporated to dryness under a stream of nitrogen at 40 °C (Caliper TurboVap LV), and CTs were dissolved in 2 mL methanol. 2.8.2. Phloroglucinolysis of CTs The resulting extract (100 lL) was reacted with 100 lL phloroglucinol solution containing 50 g phloroglucinol/L and 10 g ascorbic acid/L in methanol (containing 0.1 N HCl). The reaction was carried out at 50 °C for 20 min and quenched by adding 1000 lL sodium acetate solution (40 mM). This mixture was then analysed by the same HPLC used for ACN analysis. The monomeric composition and phloroglucinol adducts of CTs were separated on a C18 (5 lm) column (250  4.6 mm) (Phenomenex) with a C18 (5 lm) guard column (4  3 mm, 5 lm) (Phenomenex). The eluent used consisted of water, methanol and phosphoric acid (85:15:0.1; v/v/v). Separation was performed with gradient elution using a modification of the elution profile described by Nishitani and Sagesaka (2004). The linear gradient program for the separation of the monomeric composition and phloroglucinol adducts of CTs

327

M. Türkyılmaz, M. Özkan / Food Chemistry 164 (2014) 324–331

was as follows: 0.3 mL/min for 12 min, from 0.3 to 0.45 mL/min in 8 min, and from 0.45 to 1.0 mL/min in 25 min. The sample injection volume was 50 lL, the column temperature was 40 °C, and the detector was set at 280 and 210 nm for the monomer (+)-catechin and phloroglucinol adduct of (+)-catechin, respectively. Identification of these compounds was carried out by comparing retention times. Quantification of these compounds was carried out using calibration curves of external (+)-catechin standard (r2 = 0.9998). The apparent mDP was calculated by diving the sum of all subunits [(+)-catechin and phloroglucinol adduct of (+)-catechin, in moles] to the sum of all (+)-catechin (in moles). 2.9. Statistical analyses Results of the ACN and CT contents were analysed using the Minitab statistical software, version 14 (Minitab Inc., State College, PA). Statistical differences amongst means were determined by Duncan’s multiple range test at a significance level of 1%. 3. Results and discussion 3.1. Phenolic composition except for tannins of authentic PJs Since hydrolysable tannins are the main phenolic group in PJs, hydrolysable tannin compositions of PJs have been investigated in many studies (Gil, Tomas-Barberan, Hess-Pierce, Holcroft, & Kader, 2000; Li, Percival, Bonard, & Gu, 2011). However, other than tannins, there are only few studies about phenolic composition of

PJs. In addition, the separation of phenolic peaks in the HPLC chromatograms of these studies was not very successful. In the present study, the phenolic compositions of PJs (except for tannins) from different varieties were identified. The HPLC chromatogram and phenolic contents (except for tannins) of PJs from different varieties are given in Fig. 1A and B, respectively. In all PJ samples, chlorogenic acid (8.6–71.4 mg/L), catechin (32.7–47.6 mg/L), ferulic acid (11.1–30.5 mg/L), epicatechin (2–31 mg/L), rutin (6.3–22.1 mg/L) and gallic acid (7.0–18.6 mg/L) were identified. Results showed that there were significant differences between phenolic contents of the PJ samples (p < 0.01). The lowest phenolic content was determined in Izmir 1255 variety. Amongst PJ samples, Izmir 1477 and 1478 varieties had the highest phenolic contents. Similar to the differences in phenolic contents, the predominant phenolic of PJ samples also showed differences amongst varieties. Whilst the predominant phenolic of Izmir 1477, 1478 and Hicaznar varieties was chlorogenic acid, that of PJs from the other varieties was identified as catechin. Similarly, Poyrazog˘lu, Gökmen, and Artık (2002) determined different major phenolics in PJs obtained from various varieties [chlorogenic acid (in 01-N-07 and 33-N-12 varieties), catechin (in 07-N-08, 31-N-01 and 33-N-20 varieties), gallic acid (in 33-N-23 and 33-N-24 varieties) and caffeic acid (in 33-N-15 variety)]. 3.2. CTC contents and CT compositions of authentic PJs The contents of CTC in PJs highly depended on variety and ranged from 31 to 155 mg/L juice (p < 0.01) (Fig. 2B). Whilst the

(A)

Individual phenolic contents (mg/L)

150 Gallic acid Epicatechin

120

Catechin Ferulic acid

Chlorogenic acid Rutin

(B)

90

60

30

0 1255

1264

1473

1477

1478

1499

1508

1513

Hicaz

Fig. 1. HPLC chromatograma of phenolic profile (A) and phenolic contents (B) of PJs obtained from different varieties (n = 2). aThe chromatogram shows phenolic profile of Hicaznar variety.

328

M. Türkyılmaz, M. Özkan / Food Chemistry 164 (2014) 324–331

(A) Catechin

C-P

4,5

(D)

CTC

Catechin-phloroglucinol

(B)

Catechin

3,6

250 Ratio ACN content to Catechin-phloroglucinol content

200

150

100

50

0 1255

1264

1473

1477

1478

1499

1508

1513

2,7

1,8

0,9

0,0

Hicaz

1255

1264

1473

1477

1478

1499

1508

1513

45

500

(C)

ACN content by spectrophotometer

Polimerik renk oranı (%)

Hicaz

(E)

Color density

ACN content by HPLC

4,5

36 Polymeric color (%)

ACN content (mg/L)

400

300

200

100

3,6 27 2,7 18 1,8 9

0 1255

1264

1473

1477

1478

1499

1508

1513

Hicaz

5,4

Color density

Content of CTC and phloroglucinol adduct (mg/L)

300

0

0,9

1255

1264

1473

1477

1478

1499

1508

1513

Hicaz

0

Fig. 2. Variations in phenolics and colour of PJs (n = 2). (A) HPLC chromatograma of CT cleavage products in the presence of phloroglucinol (C–P: catechin–phloroglucinol). (B) CTC and phloroglucinol adduct contents. (C) ACN contents determined by spectrophotometry and HPLC. (D) Ratio of ACN content to CT content. (E) PC and CD values. aThe chromatogram shows CT cleavage products of Hicaznar variety.

highest CTC content was found in Izmir 1264 variety, the lowest was found in Izmir 1508. The CTC content of Hicaznar variety (122 mg/L) was 75% higher than Izmir 1508 variety, but 21% lower than Izmir 1264 variety. Different from our results, higher CTC content (320 mg/L) was published for different pomegranate varieties grown in Iran (Mousavinejad, Emam-Diomeh, Rezaei, & Khodaparast, 2009). As mentioned before, since polymeric pigments are formed by the reaction of monomeric ACNs with CTs and/or catechins, the effect of CTC contents on colour of the juice samples was investigated in the present study. However, to determine the effect of only CTs on colour, CTs in PJ samples were acid-catalysed in the presence of phloroglucinol (Fig. 2A). Significant correlations were found between CTC contents determined by using vanillin assay with catechin–phloroglucinol (r = 0.874) and catechin (r = 0.856) contents. In PJ samples, (+)-catechin (112–148 mg/L, Fig. 2B) and phloroglucinol product of (+)-catechin (66.8–116 mg catechin equivalent/ L, Fig. 2B) were identified (Fig. 2A), whilst no epicatechin and phloroglucinol product of epicatechin were identified. In the literature, no study investigating phloroglucinolysis of CTs in PJs was

found. Therefore, these results could not be compared with the phloroglucinol products of CTs in other pomegranate varieties. However, in Brazilian Vitis vinifera red wines, catechin, epicatechin, gallocatechin, epigallocatechin, epicatechin gallate and their phloroglucinol products were determined and the amounts of total terminal (catechin, epicatechin, gallocatechin, epigallocatechin, epicatechin gallate) and extension units (phloroglucinol products of catechin, epicatechin, gallocatechin, epigallocatechin, epicatechin gallate) of Brazilian V. vinifera red wines were reported as 48.1–94.6 mg/L and 215.9–568.3 mg/L, respectively (Gris et al., 2011). These results showed that PJ samples had lower extension units but higher terminal units than the red wines. CTs are divided into three groups: dimers and trimers, oligomers and polymers. The mean degree of polymerisation (mDP) shows the polymerisation degree of CTs. Whilst the mDP values of PJs from eight different varieties ranged from 1.0 to 3.2, the mDP value of Izmir 1499 was 4.9. These results showed that, except for Izmir 1499, the other PJ samples had dimer and trimer groups since CTs were evaluated as dimer and trimer when the mDP value was below 4. On the contrary, Izmir 1499 variety contained oligomer CTs (mDP > 4).

329

M. Türkyılmaz, M. Özkan / Food Chemistry 164 (2014) 324–331

3.3. ACN content and profiles of authentic PJs

were found higher than 10% (Fig. 2E). Higher PC values (>10%) indicate prolonged storage of the fruits or vegetables. However, in the present study, the pomegranates prior to juice processing were stored at 4 °C for only 2 days after harvest. Thus, these results showed that, prior to juice processing, pomegranates stored even at a suitable storage temperature for a very short time could contain PC higher than 10%. Similar high PC values for pomegranate juices were also reported by Turfan (2008). The high PC values in unclarified PJ samples may be due to their CT contents. Turfan (2008) found high PC in PJ samples stored even at 18 °C. PC formation in these samples is attributable to the polymerisation of ACNs with CTs rather than ACN degradation. To prove this point, the correlation between the ratio of total monomeric ACN contents to catechin–phloroglucinol contents and PC values of the samples was investigated, and strong linear correlation was found between two variables (r = 0.822). The strong correlation revealed that there was a very important effect of the ratio of total monomeric ACN contents to CT contents on the degree of polymerisation. The effect of individual ACNs on polymerisation was also investigated. For this purpose, correlations between the ratios of individual ACN content to catechin–phloroglucinol contents and PC values were determined. The results from these correlations showed that cyanidin-3-glucoside amongst ACNs in PJ samples had the most significant adverse effect on the degree of polymerisation (r = 0.883). As the content of cyanidin-3-glucoside increased, the degree of polymerisation reduced. In comparing the number of sugars attached to anthocyanidins, ACN monoglucosides in the PJ samples had more significant adverse effect on the degree of polymerisation than ACN diglucosides. For example, although good correlation (r = 0.779) was found between the ratio of cyanidin-3,5-diglucoside contents to catechin–phloroglucinol contents and PC values, stronger correlation (r = 0.883)

Total monomeric ACN contents from HPLC of the same nine PJ samples were previously reported by Türkyılmaz (2013). In this study, the same ACN values were again used to show the effects of CTC and catechin–phloroglucinol contents on colour of the PJ samples. Total monomeric ACN contents from HPLC of the PJ samples ranged from 28 (Izmir 1508) to 447 mg ACN/L juice (Izmir 1513) (Fig. 2C). Similar variations (56–301 mg/kg) in total monomeric ACN contents were reported for different pomegranate varieties grown in Iran (Tehranifar, Zarei, Esfandiyari, & Nemati, 2010). Our results and the results from Iranian varieties clearly showed that the variety had very significant effect on total monomeric ACN contents of PJs (p < 0.01). The HPLC chromatogram of monomeric ACNs in Hicaznar variety is given in Fig. 3A as an example. Six different ACNs in this variety were detected and identified by HPLC, with the elution order as delphinidin-3,5-diglucoside, cyanidin-3,5-diglucoside, delphinidin-3-glucoside, pelargonidin-3,5-diglucoside, cyanidin3-glucoside and pelargonidin-3-glucoside. However, as shown in Fig. 3B, there were significant differences between ACN profiles and contents of the registered pomegranate varieties (p < 0.01). 3.4. Effects of CTC and catechin–phloroglucinol contents on colour of authentic PJs The PC values of PJ samples are presented in Fig. 2E. Significant differences (p < 0.01) were found amongst the PC values (5–34%) of the samples. The highest PC value was determined in Izmir 1255 _ _ (34%), followed by Izmir 1473 (26%) and Izmir 1508 (21%) varieties. Although PC values of freshly squeezed fruit or vegetable juices usually are less than 10% (Giusti et al., 2005), PC values of all varieties, except for Izmir 1513 (5%) and Hicaznar (8%) varieties,

mAU

Cy-3,5-diglu

(A)

700 600

200 100 0

3

Individual ACN content (mg/L)

200

4

Dp-3,5-diglu Pg-3,5-diglu

5

6

Cy-3,5-diglu Cy-3-glu

Pg-3-glu

300

Dp-3-glu

400

Pg-3,5-diglu

Cy-3-glu

Dp-3,5-diglu

500

7

8

10

9

11

Dp-3-glu Pg-3-glu

min

(B)

150

100

50

0 1255 a

1264

1473

1477

1478

1499

1508 a

1513

Hicaz

Fig. 3. HPLC chromatogram of ACNs (A) and ACN contents (B) of PJs obtained from different varieties (n = 2). The chromatogram shows ACN profile of Hicaznar variety.

330

M. Türkyılmaz, M. Özkan / Food Chemistry 164 (2014) 324–331

was found for cyanidin-3-glucoside. Similar results were shown for delphinidin-3-glucoside (r = 0.632) and delphinidin-3,5-diglucoside (r = 0.523). Similar to PC values, the differences in CD values give also information about the polymerisation effect of CTs on ACNs in PJs. CD values of PJs varied from 0.65 to 3.78 (Fig. 2E). The highest CD values were determined in Izmir 1264 and 1513 varieties, whilst the lowest values were determined in Izmir 1473 and 1508 varieties. CD value of Hicaznar variety is 1.28. The reason why the CT and total monomeric ACN contents had significant effect on CD values of PJs is that CTs react with ACNs to create the effect known as copigmentation and condensation. The density of ACN colour is fortified by copigmentation and condensation (Rein, 2005). Therefore, CD values of the samples showed parallel changes with catechin–phloroglucinol (r = 0.919) and total monomeric ACN (r = 0.774) contents. As catechin–phloroglucinol and ACN contents increased, CD values also increased. Previous studies also showed that CD values increased linearly with the increase in concentration of copigments added (Rein, 2005), such as CTs. Moreover, Kovac, Alonso, and Revillo (1995) investigated the effect of CT contents on CD of red wines and they reported that high CT content led to an increase in CD values of red wines. Similarly, in the present study, strong correlation (r = 0.852) between the ratio of total monomeric ACN contents to catechin–phloroglucinol contents and CD values of the samples was found. Since the structure of ACNs, as well as of copigments, affects the magnitude of copigmentation, thus CD (Rein, 2005), the influence of individual ACNs on CD values were also investigated. Significant correlations (r = 0.644–0.982) between CD values and individual ACN contents were found. However, no correlation (r = 0.014– 0.489) between CD values and the ratios of the individual ACN content to CT content was observed. Amongst individual ACNs, delphinidin-3,5-diglucoside had the strongest effect on CD values in PJ samples (r = 0.982). Similarly, strong correlation (r = 0.941) between delphinidin-3,5-diglucoside contents and CD values was reported in grape juices containing rosemary extracts of 0.4% (Talcott, Brenes, Pires, & Pozo-Insfran, 2003). Compared with the effect of ACN monoglucosides contents (r = 0.644–0.782), ACN diglucoside contents (r = 0.783–0.928) in the PJs, whose pH values ranged from 2.7 to 3.2, had more significant effect on CD values (r = 0.928) than the ACN monoglucoside contents. This is probably because 3-monoglucosides have maximum effect on copigmentation at pH 3.5–4.2, whilst 3,5-diglucosides have maximum effect at pH 3.1 (Williams & Hrazdina, 1979). Brown pigments, monomeric ACNs as well as degraded ACN polymers contribute to CD. To determine the contribution of enzymatic browning resulting from the reaction of phenolics with polyphenoloxidase (PPO) enzyme, correlations between individual phenolic contents and CD values were taken as good pointers. Strong correlations were found between CD values and the contents of epicatechin (r = 0.952), gallic acid (r = 0.887), ferulic acid (r = 0.756) and catechin (r = 0.733), but weak correlation was found between CD value and rutin content (r = 0.462). The difference between the correlation strengths may be that flavonols such as rutin are either not direct substrates or poor substrates for PPO, whilst catechin, chlorogenic acid, gallic acid and ferulic acid are good substrates for PPO. These results clearly showed that, other than monomeric ACNs, browning and polymerisation had also a substantial effect on the high CD values of the PJ samples.

(pH-differential method) and HPLC methods (Fig. 2C). Since HPLC is an invaluable tool for quantifying individual ACNs, thus total ACNs in a sample (Lee et al., 2008), total monomeric ACN contents obtained from spectrophotometry were compared with those obtained from HPLC. Although strong correlation (r = 0.943) was found between total monomeric ACN contents obtained from spectrophotometry and HPLC, spectrophotometric values were higher (1.1–1.8 times) than the values obtained from HPLC, except for Izmir 1513 variety. Lee et al. (2002) also found similar results for blueberry juices, whose total ACN contents determined by spectrophotometrically were 1.5–1.9 times higher than those determined by HPLC. This difference may be attributable to the various amounts of polymeric pigments present in juice samples as well as the different solvent systems and wavelengths used in HPLC (k = 520 nm) and spectrophotometer (k = 512 nm) (Lee et al., 2002). Polymeric pigments, which result from the formation of a complex between ACNs and CTs, may lead to higher absorbance values, since polymeric pigments cannot be separated prior to spectrophotometric measurements (Lee et al., 2002). For example, Izmir 1264 variety (116 mg/L, Fig. 2B) contained 1.7 times higher catechin– phloroglucinol than Hicaznar variety (69 mg/L, Fig. 2B). Similarly, the difference between ACN contents analysed by both methods in Izmir 1264 variety was also 1.5 times higher than those of Hicaznar variety. These findings clearly indicated that the catechin– phloroglucinol content of PJ is closely related with ACN content determined by spectrophotometry. As the catechin–phloroglucinol content of PJ samples increased, ACN contents obtained from spectrophotometry also increased (r = 0.866). In a study conducted in our laboratory, similar correlation (r = 0.766) between CTC and monomeric ACN contents obtained from spectrophotometry was also determined for black carrot juice samples (Özkan, Türkyılmaz, Dereli, & Yemisß, 2009). Thus, the CT and CTC contents, and in turn PC ratio, were a very important contributor to the monomeric ACN content of PJ by spectrophotometry. The results of the present study also revealed that when PC value of the PJ samples was P8% or the ratio of ACN contents to catechin–phloroglucinol contents was 62.82, ACN contents analysed by spectrophotometry were higher than those by HPLC (Fig. 2C–E). For unclarified PJ (Turfan, 2008) and pasteurised blueberry juice and concentrates (Lee, Yusof, Hamid, & Baharin, 2006) whose PC ratios were higher than 8%, similar results were also reported. PC values for unclarified PJ and pasteurised blueberry juice were 29% and 50%, respectively, and total monomeric ACN contents obtained from spectrophotometry 1.2 and 1.9 times, respectively, higher than those obtained from HPLC. However, there is only one study found in the literature, showing the differences between ACN contents obtained from both analytical methods, for mulberry juice samples whose PC ratio (ranging from 2% to 5%) was lower than 8% (Özkan, Erbay, Türkyılmaz, Tag˘ı, & Küçüköner, 2012). In this study, ACN contents of mulberry juice measured by spectrophotometry were found to be 1.2–1.3 times lower than total monomeric ACN contents measured by HPLC (Özkan et al., 2012). Similarly, Izmir 1513 variety had the lowest PC value (5%) and, in contrast with the other varieties, the total ACN contents obtained from both methods were very close. Total ACN content of Izmir 1513 variety obtained from spectrophotometry was only 1.1 times lower than that from HPLC.

4. Conclusion 3.5. The effects of catechin–phloroglucinol contents and PC on monomeric ACN contents The total monomeric ACN contents in the registered pomegranate varieties were determined by spectrophotometric

Depending on variety, there were significant differences between CTC, catechin–phloroglucinol and total monomeric ACN contents of the PJ samples (p < 0.01). As CTC, catechin–phloroglucinol and ACN contents increased, PC and CD also increased.

M. Türkyılmaz, M. Özkan / Food Chemistry 164 (2014) 324–331

Moreover, catechin–phloroglucinol contents made a significant contribution to the ACN contents obtained from spectrophotometry. When PC value of PJs was P8%, ACN content analysed by spectrophotometry was higher than that analysed by HPLC. If the PJ sample from the investigated variety in the present study has PC which is lower than 8%, monomeric ACN content can be measured by spectrophotometry. In that case, HPLC is not required for the determination of total monomeric ACN contents in PJs. However, if the PJ has PC which is higher than 8%, HPLC should be used for the accurate measurement of total monomeric ACNs in place of spectrophotometry. Acknowledgements The authors thank the Alata Horticultural Research Institute (Erdemli, Mersin) and Aegean Agricultural Research Institute (Menemen, Izmir) for providing the pomegranates. References Brito, E. S. D., Araujo, M. C. P. D., Alves, R. E., Carkeet, C. C., Clevidence, B. A., & Novatny, J. A. (2007). Anthocyanins present in selected tropical fruits: Acerola, Jambolao, Jussara and Guajiru. Journal of Agricultural and Food Chemistry, 55, 9389–9394. Drinkine, J., Glories, Y., & Saucier, C. (2005). (+)-Catechin–aldehyde condensations: Competition between acetaldehyde and glycowylic acid. Journal of Agricultural and Food Chemistry, 53, 7552–7558. Drinkine, J., Lopes, P., Kennedy, J. A., Teissedre, P. L., & Saucier, C. (2007). Ethylidenebridged flavan-3-ols in red wine and correlation with wine age. Journal of Agricultural and Food Chemistry, 55, 6292–6299. Gil, M. I., Tomas-Barberan, F. A., Hess-Pierce, B., Holcroft, D. M., & Kader, A. A. (2000). Antioxidant activity of pomegranate juice and its relationship with phenolic composition and processing. Journal of Agricultural and Food Chemistry, 48, 4581–4589. Gil, M., Kontoudakis, N., Gonza´lez, E., Esteruelas, M., Fort, F., Canals, J. M., et al. (2012). Influence of grape maturity and maceration length on color, polyphenolic composition, and polysaccharide content of Cabernet Sauvignon and Tempranillo wines. Journal of Agricultural and Food Chemistry, 60, 7988–8001. Giusti, M. M., & Wrolstad, R. E. (2005). Unit F1.2: characterization and measurement of anthocyanins by UV–visible spectroscopy. In R. E. Wrolstad, T. E. Acree, E. A. Decker, M. H. Penner, D. S. Reid, S. J. Schwartz, C. F. Shoemaker, D. M. Smith, & P. Sporns (Eds.), Handbook of food analytical chemistry (pp. 19–31). New York: John Wiley & Sons. Gris, E. F., Mattivi, F., Ferreira, E. A., Vrhovsek, U., Pedrosa, R. C., & Bordignon-Luiz, M. T. (2011). Proanthocyanidin profile and antioxidant capacity of Brazilian Vitis vinifera red wines. Food Chemistry, 126, 213–220. Kennedy, J. A., & Jones, G. P. (2001). Analysis of proanthocyanidin cleavage products following acid-catalysis in the presence of excess phloroglucinol. Journal of Agricultural and Food Chemistry, 49, 1740–1746. Kovac, V., Alonso, E., & Revillo, E. (1995). The effect of supplementary quantities of seeds during fermentation on the phenolic composition of wines. American Journal of Enology and Viticulture, 46, 363–367. Lee, J., Durst, R. W., & Wrolstad, R. E. (2002). Impact of juice processing on blueberry anthocyanins and polyphenolics: Comparison of two pretreatments. Journal of Food Science, 67, 1660–1667. Lee, W. C., Yusof, S., Hamıd, N. S. A., & Baharın, B. S. (2006). Optimizing conditions for enzymatic clarification of banana juice using response surface methodology (RSM). Journal of Food Engineering, 73, 55–63.

331

Lee, J., Rennaker, C., & Wrolstad, R. E. (2008). Correlation of two anthocyanin quantification methods: HPLC and spectrophotometric methods. Food Chemistry, 110, 782–786. Li, Z., Pan, Q., Cui, X., & Duan, C. (2010). Optimization on anthocyanins extraction from wine grape skins using orthogonal test design. Food Science and Biotechnology, 19, 1047–1053. Li, Z., Percival, S. S., Bonard, S., & Gu, L. (2011). Fabrication of nanoparticles using partially purified pomegranate ellagitannins and gelatin and their apoptotic effects. Molecular Nutrition and Food Research, 55, 1096–1103. Liao, H., Cai, Y., & Haslam, E. (1992). Polyphenol interactions. Anthocyanins: Copigmentation and color changes in red wines. Journal of the Science of Food and Agriculture, 59, 299–305. Mousavinejad, G., Emam-Diomeh, Z., Rezaei, K., & Khodaparast, M. H. H. (2009). Identification and quantification of phenolic compounds and their effects on antioxidant activity in pomegranate juices of eight Iranian cultivars. Food Chemistry, 115, 1274–1278. Nishitani, E., & Sagesaka, Y. M. (2004). Simultaneous determination of catechins, caffeine and other phenolic compounds in tea using new HPLC method. Journal of Food Composition and Analysis, 17, 675–685. Özkan, M., Türkyılmaz, M., Dereli, U., & Yemisß, O. (2009). Changes in phenolics and anthocyanins of black carrot juice concentrate during production and storage and their relation with antioxidant activity. Ankara: Ankara University Research Foundation. Project Number: 07B4343002, (pp. 106). Özkan, M., Erbay, B., Türkyılmaz, M., Tag˘ı, Sß., & Küçüköner, E. (2012). Thermal and storage stability of mulberry anthocyanins and, some chemical and antimicrobial properties of mulberries. Ankara: The Scientific and Technological Research Council of Turkey. Project Number: 111R004, (pp. 88). Poyrazog˘lu, E., Gökmen, V., & Artik, N. (2002). Organic acids and phenolic compounds in pomegranates (Punica granatum L.) grown in Turkey. Journal of Food Composition and Analysis, 15, 567–575. Rein, M. J. (2005). Copigmentation reactions and color stability of berry anthocyanins (dissertation). EKT series 1331. Helsinki, Finland: University of Helsinki. Revilla, E., Bourzeix, M., & Alonso, E. (1991). Analysis of catechins and proanthocyanidins in grape seeds by HPLC with photodiode array detection. Chromatographia, 31, 465–468. Salas, E., Fulcrand, H., Meuder, E., & Cheynier, V. (2003). Reactions of anthocyanins and tannins in model solutions. Journal of Agricultural Food Chemistry, 51, 7951–7961. Singleton, V. L., & Trousdale, E. K. (1992). Anthocyanin–tannin interactions explaining differences in polymeric phenols between white and red wines. American Journal of Enology and Viticulture, 43, 63–70. Skrede, G., Wrolstad, R. E., & Durst, R. W. (2000). Changes in anthocyanins and polyphenolics during juice processing of highbush blueberries (Vaccinium corymbosum L.). Journal of Food Science, 65, 357–364. Somers, T. C. (1971). The polymeric nature of wine pigments. Phytochemistry, 10, 2175–2186. Talcott, S. T., Brenes, C. H., Pires, D. M., & Pozo-Insfran, D. (2003). Phytochemical stability and color retention of copigmented and processed muscadine grape juice. Journal of Agriculture and Food Chemistry, 51, 957–963. Tanner, H., & Brunner, H. R. (1979). Getraenke Analytik. Verlag Heller Chemie-und Verwaltungsgeselschaft mbH. D-7170. Germany: Schwaebisch Hall. Tehranifar, A., Zarei, M., Esfandiyari, B., & Nemati, Z. (2010). Physicochemical properties and antioxidant activities of pomegranate fruit (Punica granatum) of different cultivars grown in Iran. Horticulture Environment Biotechnology, 51, 573–579. Turfan, Ö. (2008). Changes in anthocyanins of pomegranate juice concentrate during production and storage, Master thesis. Ankara: Ankara University (pp.146). Türkyılmaz, M. (2013). Anthocyanin and organic acid profiles of pomegranate (Punica granatum L.) juices from registered varieties in Turkey. International Journal of Food Science and Technology, 48, 2086–2095. Williams, M., & Hrazdina, G. (1979). Anthocyanins as food colorants: Effect of pH on the formation of anthocyanin–rutin complexes. Journal of Food Science, 44, 66–68. Xu, Z., & Howard, L. R. (2012). Analysis of antioxidant-rich phytochemicals. Cichester: John Wiley & Sons.