phenolic composition and total antioxidant capacity of

Research Article Received: 13 April 2018

Revised: 14 June 2018

Accepted article published: 17 July 2018

Published online in Wiley Online Library:

(wileyonlinelibrary.com) DOI 10.1002/jsfa.9267

Phenolic composition and total antioxidant capacity of South African frozen concentrated orange juice as affected by varietal, seasonal and regional differences Cindy Hunlun,a* Dalene de Beer,a,b Gunnar O Siggea and Jessy Van Wykc Abstract BACKGROUND: There is epidemiological evidence that fruits and vegetables promote general health due to their phenolic composition. The phenolic composition of three commercially important citrus varieties (‘Mandarin’, ‘Navel’ orange and ‘Valencia’ orange), used for frozen concentrated orange juice (FCOJ) production in South Africa, were evaluated based on variety, production season and geographical region (Western Cape (WC) and Eastern Cape (EC)). RESULTS: FCOJ from the WC had significantly (P < 0.05) lower titratable acidity (TA) and higher total soluble solids (TSS):TA ratio compared to FCOJ produced in the EC. The ‘Navel’ FCOJ, irrespective of season and region, had the highest (P < 0.05) phenolic content (sum of quantified compounds, TP). Regional effects were clear for the ‘Mandarin’ variety, the EC ‘Mandarin’ FCOJ had the highest TP and WC ‘Mandarin’ had the lowest. Seasonal differences were less evident. Variation that could be ascribed to regional differences were also found for individual phenolic compounds. CONCLUSION: Robust data regarding the phenolic profile of FCOJ produced in South Africa, suitable for inclusion in food composition databases, were collected. © 2018 Society of Chemical Industry Supporting information may be found in the online version of this article. Keywords: Citrus sinensis; Citrus reticulata; phenolic compounds; total antioxidant capacity; ORAC; DPPH; food analysis; food composition; frozen concentrated orange juice

ABBREVIATIONS EC ET0 FCOJ FHX HD NRGLC NART NEOP QRG tR RUT TAC TP VIC2 WC

Eastern Cape province evapotranspiration frozen concentrated orange juice ferulic acid-O-hexoside hesperidin naringenin-7-O-rutinose-4′ -O-glucoside narirutin neoponcirin quercetin-3-O-rutinose-7-O-glucoside retention time rutin total antioxidant capacity total of quantified polyphenolic compounds vicenin-2 Western Cape province

not-from-concentrate (NFC) orange juice started overtaking demand for FCOJ.1 This quickly became a global trend, affecting markets in Brazil and South Africa. Many possible reasons account for this trend, but mostly it is ascribed to the perception of differing product quality.2,3 Food for health is a recurring theme in modern society. There is a demand for nutritional food with pharmacological properties that is readily available. Secondary plant metabolites such as flavonoids are abundant in fruits and vegetables. In particular, flavanones from citrus have been studied for their use as chemopreventive

∗

Correspondence to: C Hunlun, Plant Bioactives Group, Post-Harvest and Agro-Processing Technologies, Agricultural Research Council (ARC), Infruitec Nietvoorbij, Private Bag X5026, Stellenbosch 7599, South Africa. E-mail: [email protected]

a Department of Food Science, Stellenbosch University, Matieland, South Africa

INTRODUCTION Frozen concentrated orange juice (FCOJ) consumption became popular after its introduction in the USA in the mid-1940s, peaking in 1997/98. However, between 2004 and 2016 J Sci Food Agric (2018)

b Plant Bioactives Group, Post-Harvest and Agro-Processing Technologies, Agricultural Research Council (ARC), Infruitec-Nietvoorbij, Stellenbosch, South Africa c Department of Food Science and Technology, Cape Peninsula University of Technology, Bellville, South Africa

www.soci.org

© 2018 Society of Chemical Industry

www.soci.org and therapeutic agents.4–9 However, the effect of food processing, used to extend shelf-life and availability of food, on the quantity and quality of flavonoids and specifically their biological activity is of concern. This is especially important to educated consumers and inform their food choices. However, misperceptions exist that frozen orange juice is inferior and possesses fewer health benefits compared to fresh juice. This can also explain consumers’ willingness to pay a premium for fresh orange juice. This has led to a decline in consumption of ‘processed’ products such as FCOJ. The functional properties of citrus fruit juice are due to its phenolic composition and vitamin C content.10,11 Various studies have been published evaluating the effect of heat treatment, industrial extraction methods and stability of major citrus flavonoids in citrus juice.11–14 Most studies have concluded that citrus flavonoids are relatively stable and are dependent on the raw material composition and processing system employed, and that processing even results in higher flavonoid levels. Thus variability in flavonoid composition is mainly attributable to the specific genotype and possibly geographical region. Currently, data describing the phenolic composition of various citrus varieties and products are limited, even in databases such as Phenol-Explorer and the US Department of Agriculture (USDA) database on flavonoids.15,16 Moreover, phenolic data will be useful for inclusion in food composition databases such as the South African Food Data System (SAFOODS) operated by the South African Medical Research Council (SAMRC), which is currently not available. This will enable wider access to this information in order to further educate and transform consumers’ perceptions regarding healthy food choices. The addition of phenolic composition data to such databases will also allow for determination of current and recommended intake levels. Information on the variation in phenolic composition of different South African citrus fruit due to varietal and seasonal effects has been published in a previous baseline study.17 Furthermore, the phenolic composition of various citrus-based products needs to be established. Therefore, the aim of this study was to characterize the phenolic composition and total antioxidant capacity of FCOJ produced in the WC and EC regions of South Africa. The phenolic composition of three varieties (‘Mandarin’, ‘Navel’ orange and ‘Valencia’ orange) was compared with a view to evaluate the effect of varietal, seasonal and regional differences, if any, and to obtain representative data for inclusion in food composition databases.

EXPERIMENTAL Chemicals High-performance liquid chromatography (HPLC)-grade water was obtained from deionized water that was purified using an Elix Millipore system (Merck Millipore, Darmstadt, Germany) and subsequently treated with a Milli-Q Reference A+ (Merck Millipore) water purification system. HPLC gradient-grade acetonitrile was purchased from Sigma-Aldrich (St Louis, MO, USA). All analytical-grade chemicals and reagents were obtained from either Merck Millipore or Sigma-Aldrich. Reference standards (purity ≥ 95%) included rutin from TransMIT (Gießen, Germany), narirutin from Extrasynthese (Genay, France), hesperidin and ferulic acid from Sigma-Aldrich, as well as vicenin-2 and neoponcirin from Phytolab (Vestenbergsgreuth, Germany). Sample preparation FCOJ samples, from three varieties, were collected from two citrus fruit processors over a 3-year period (2012: n = 53; 2013: n = 63;

wileyonlinelibrary.com/jsfa

C Hunlun et al.

and 2014: n = 41). The FCOJ types included ‘Mandarin’ (containing ‘Satsuma’ and ‘Clementine’ varieties, which were processed together), n = 62; ‘Navel’ orange, n = 63; and ‘Valencia’ orange, n = 32. The fruit processors were situated in different regions of South Africa: one in Citrusdal in the WC Province (n = 117) and the other in Kirkwood in the EC Province (n = 40). The processor situated in the WC did not process ‘Valencia’ orange in 2012 due to unfavourable fruit prices that season. The FCOJ samples (250 mL, total soluble solids (TSS) ∼65 ∘ Brix) were retention samples taken from a production batch (typically 16 500 L) by the quality control department of each company. Samples were distinguished on the basis of processing or batch date and variety. All samples were kept frozen at −18 ∘ C and defrosted and diluted with HPLC-grade water to standardize the TSS content to ∼11.5 ∘ Brix prior to analysis. This is between the international minimum ∘ Brix level (11.2–11.8) for reconstituted fruit juices and reconstituted purée as indicated for Citrus sinensis.18 Prior to total antioxidant capacity (TAC) determinations the samples were centrifuged at 10 000 × g (BOECO, Hamburg, Germany) for 5 min and the supernatant was diluted with HPLC-grade water if necessary. For HPLC analysis all samples were filtered using 0.45 μm Acrodisc syringe filters with GHP membranes (Separations, Pall Life Sciences, Port Washington, NY, USA). Determination of citrus juice characteristics Titratable acidity (TA) was determined using an automatic titrator (Metrohm Titrino 702 autotitrator, Johannesburg, South Africa) and results were expressed as g kg−1 citric acid. TSS (expressed as ∘ Brix) was measured using a hand-held refractometer (Atago model PAL 1, Tokyo, Japan). Additionally, a TSS:TA ratio was calculated and the pH recorded for each sample. Determination of phenolic composition using high-performance liquid chromatography coupled with diode array detection (HPLC-DAD) The individual phenolic compounds in FCOJ were quantified using the method previously described.17 In short, an Agilent 1200 series HPLC instrument (Agilent Technologies, Santa Clara, CA, USA) equipped with an in-line degasser, quaternary pump, autosampler, column thermostat and diode array detector was used. Chemstation software controlled the system and was used for the data acquisition and analysis. A Gemini-NX C18 (3 μm particle size, 110 Å pore size,150 × 4.6 mm ID; Phenomenex, Santa Clara, CA, USA) column was used and fitted with a guard column (4.0 × 3.0 mm ID, Phenomenex). Separation was performed at 31 ∘ C at a flow rate of 1.0 mL min−1 using 0.2% acetic acid (A) and acetonitrile (B) with the following gradient: 0–3 min, 5% B; 3–40 min, 5–27% B; 40–41 min, 27–100% B; 41–43 min, 100% B; 43–45 min, 100–5% B; 45–55 min, 5% B. The flavone-O-glycosides rutin (RUT) and quercetin-3-O-rutinose-7-O-glucoside (QRG) were quantified at 255 nm, flavanone-O-glycosides narirutin (NART), naringenin-7-O-rutinose-4′ -O-glucoside (NRGLC), hesperidin (HD) and neoponcirin (NEOP) were quantified at 288 nm, ferulic acid-O-hexoside (FHX, a hydroxycinnamic acid) was quantified at 320 nm and lastly vicenin-2 (VIC2, a flavone-C-glucoside) was quantified at 350 nm. Total antioxidant capacity 2,2′ -Diphenyl-1-picrylhydrazyl (DPPH) radical scavenging and oxygen radical absorbance capacity (ORAC) assays were used to determine the TAC of juice samples expressed as μmol Trolox

© 2018 Society of Chemical Industry

J Sci Food Agric (2018)

Phenolic characterization of citrus juice

www.soci.org

equivalents (TE) L−1 . A BioTek Synergy HT microplate reader with Gen5 software for data acquisition (Winooski, VT, USA) was used for absorbance and fluorescence readings. The DPPH and ORAC assays were performed as previously described.19,20 The ORAC protocol was slightly modified by adding 300 μL deionized water to the outside wells of each plate as a thermal barrier. Statistical analysis Data were expressed as mean ± standard deviation (SD). All analyses were performed in duplicate except for DPPH and ORAC, in triplicate. Univariate analysis of variance (ANOVA) and least significant difference Student’s t-test were used as a post hoc test. SAS statistical software (SAS® , Version 9.2, SAS Institute Inc., Cary, NC, USA) was used for data processing. The Shapiro–Wilk test was used to test for normality and to identify outliers. In addition, multivariate statistical analysis was performed on the data using XLSTAT software (Version 7.5.2, Addinsoft, New York, USA). Principal component analysis (PCA) was used to evaluate relationships between the characteristics, FCOJ type, seasons and region. Differences were considered statistically significant at 95% confidence level (P-values < 0.05).

RESULTS AND DISCUSSION Principal component analysis (PCA) Information on climatic growing conditions pertaining to the sampling sites (Citrusdal and Kirkwood) from the two regions (WC and EC) used during this study was obtained from the Agricultural Research Council of South Africa. In brief, the monthly average maximum and minimum temperatures of the two regions were compared from 2011 to 2014 since the climatic conditions in 2011 could have affected the quality of fruit that was harvested in 2012, and so on (Fig. S1, Supporting Information). The biplots compiled from the PCA are provided in Fig. 1. ANOVA was used to evaluate and compare the individual phenolic characteristics of the FCOJ variety, season (year) and region (Table 1). The PCA biplot (Fig. 1a) depicts the distribution of the three varieties sampled in 2012 from the WC and EC sampling sites based on phenolic composition and explains 71% of the inherent variability in the data. The ‘Mandarin’ variety shows clear clusters corresponding to the two regions. ‘Mandarin’ from the EC sampling site (B_man) was associated with higher total quantified phenolic compounds (TP) and TACORAC levels, while ‘Mandarin’ from the WC sampling site (A_man) had lower TP and TACORAC levels. The ‘Navel’ orange samples from both sampling sites (A_nav and B_nav) was closely associated in terms of phenolic composition and overlapping with the ‘Valencia’ orange (B_val) from the EC sampling site, although some B_val samples were associated with higher levels of FHX. The PCA biplot (Fig. 1b) depicts the distribution of the three varieties sampled in 2013 from the WC and EC sampling sites based on phenolic composition and explains 65% of the inherent variability in the data. The distribution of the different varieties is widespread for the 2013 season and differs from the 2012 season. However, some varietal and regional differences in terms of associations may be noted. Most of the ‘Mandarin’ samples from the WC (A_man) were positively associated with higher RUT, QRG and TACORAC . The ‘Mandarin’ samples from the EC (B_man) were scattered with no distinct associations. Table 2 shows the large variation in the results for the ‘Mandarin’ varieties. ‘Valencia’ orange varieties from both regions were associated with higher FHX and TACDPPH . The ‘Navel’ orange varieties formed a cluster and were associated with higher NRGLC, HD, NEOP and TP. J Sci Food Agric (2018)

The PCA biplot (Fig. 1c) depicts the distribution of the three varieties sampled in 2014 from the WC and EC sampling sites based on chemical and phenolic composition and explains 67% of the inherent variability in the data. The ‘Mandarin’ variety was scattered across the plot and did not correlate with specific phenolic attributes. Although most of the B_man samples were positively associated with higher NRGLC, RUT, NART, HD, NEOP and TP others were negatively associated with these parameters, with lower concentrations for these compounds. Once more, the majority of the ‘Valencia’ samples from both regions were associated with higher FHX and TACDPPH . Similarly, the ‘Navel’ orange samples, irrespective of region, were mostly associated with higher VIC2. Differences between ‘Valencia’ orange and ‘Navel’ orange based on the distribution of phenolic compounds were unclear, with overlap between the two varieties. Although ‘Valencia’ orange from the WC region, irrespective of season, was positively associated with higher FHX and TACDPPH values, ‘Navel’ orange was positively associated with higher VIC2, NEOP, HD, NRGLC, TP and TACORAC . Effect of variety, season and region on chemical and phenolic composition of FCOJ The main effects variety and region were the most significant (P < 0.05), with season (year) having a significant (P < 0.05) effect on some parameters such as TSS:TA, pH, HD and NEOP (Table 1). In addition, there were various significant (P < 0.05) interactions between region × variety and region × year. The results obtained indicated that the effect of variety was not significant (P ≥ 0.05) for TSS:TA, QRG or TACORAC . Moreover, the effect of region was not significant (P ≥ 0.05) for pH, RUT or TACDPPH . Finally, the effect of season (year) was not significant (P ≥ 0.05) for most parameters, except for the two major flavanones (HD and NART). In terms of chemical composition, TSS:TA and pH differed significantly (P < 0.05) between the different seasons. The significant (P < 0.05) effects for individual parameters will be discussed further. Varietal differences (FCOJ type) There were no significant differences (P ≥ 0.05) for TSS since all FCOJ samples were prepared by reconstitution to a standardized single-strength TSS level (∼11.5 ∘ Brix). However, significant differences (P < 0.05) were found between the TA of the three varieties (Table 2). As expected, the ‘Valencia’ orange variety was characterized with the highest TA of 12.2 g kg−1 citric acid, followed by ‘Mandarin’, which was not expected since ‘Mandarin’ is described as having low acid levels.19 The ‘Navel’ orange variety was found to have the lowest acidity level (9.3 g kg−1 citric acid), which can be ascribed to the fact that, in general, this variety is the earliest-maturing variety of all orange sector varieties. The lower TA could be due to the accumulation of carbohydrates and water in the fruit, which results in a decrease in the tartness or characteristic acidity because of a dilution effect as the fruit matures.19,20 In addition, when citrus fruit (especially ‘Navel’ oranges) are grown in warmer climates (where higher temperatures are observed), the respiration rate will be higher and the citric acid reserves will decrease as a result.21,22 This makes ‘Navel’ oranges less suited to warm environments. The growing conditions in terms of average temperature experienced by ‘Navel’ oranges grown in South Africa may have contributed to the lower TA. The pH was also found to differ significantly (P < 0.05) between ‘Navel’ orange and the other varieties. ‘Navel’ orange fruit were found to have a mean pH of 3.6 compared to ‘Mandarin’ and ‘Valencia’ orange, having a pH of 3.5 and 3.4, respectively. The higher pH of ‘Navel’ orange can be ascribed to the lower TA levels.

© 2018 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

(a)

5

6 RUT

5

B_man

A_man

3

1 0 -1 -2 -3

2

NART

F2 (24.54 %)

A_man A_man A_man A_man A_man A_man A_man A_man A_man A_man

ORAC B_man TP A_man A_man A_man A_man A_man A_man NRGLC A_man B_man A_nav A_nav A_man A_nav A_man A_nav A_nav A_nav A_nav A_nav B_nav B_nav NEOP B_val A_nav B_val A_nav A_nav A_nav A_nav A_nav A_nav B_val B_nav A_nav B_val A_nav pH

DPPH

VIC2 -4

-3

-2

-1

0

1

DPPH

-1

2

3

-5 4

5

6

7

B_man NEOP

8

TP NRGLC B_man

A_nav A_nav A_nav A_nav A_nav A_nav A_man A_man A_nav A_nav A_nav B_val A_nav A_nav A_navB_val A_nav A_nav A_val A_nav A_nav B_nav A_nav A_nav A_val B_nav A_val A_nav B_nav A_val A_val A_nav A_val B_val A_man B_val A_val A_valB_valA_man A_val B_val A_val A_man A_val A_val A_val

NART

A_man A_man A_man A_man -5

-4

-3

-2

QRG A_man

-1

0

1

B_man

ORAC

B_man

-4

-5 -5

0

-3

FHX

-6

1

FHX

-2

HD

-4

HD

pH

3

QRG

2

VIC2

4

B_man

4

F2 (20.34 %)

Biplot (axes F1 and F2: 64.95 %)

(b)

Biplot (axes F1 and F2: 70.78 %)

-6

C Hunlun et al.

A_manA_man A_man A_man A_man A_man

A_man RUT

2

3

4

5

6

F1 (40.40 %)

F1 (50.44 %)

Biplot (axes F1 and F2: 67.02 %)

(c)

VIC2 5

HD FHX

4

F2 (23.60 %)

3 NEOP

2

DPPH

A_valB_nav A_val A_val B_val B_val B_val B_nav B_val B_val A_nav B_val A_nav B_man A_nav A_nav B_nav A_nav A_nav B_man B_nav A_nav B_nav B_nav A_nav A_navA_man A_nav pH A_man A_man

1 0 -1

NART QRG

A_man A_man

-3 A_man A_man A_man

-4 -5

ORAC

B_man

-2

-5

-4

-3

-2

TP NRGLC

A_man A_man

B_man

RUT -1

0

1

2

3

4

5

6

7

8

F1 (43.42 %) Figure 1. PCA biplot of associations between variety and region based on chemical and phenolic composition for (a) 2012, (b) 2013 and (c) 2014 processing seasons. (QRG, quercetin-3-O-rutinose-7-O-glucoside; FHX, ferulic acid-O-hexoside; VIC2, vicenin-2; NRGLC, naringenin-7-O-rutinose-4′ -O-glucoside; NART, narirutin; HD, hesperidin; NEOP, neoponcirin; TP, sum of quantified compounds; DPPH and ORAC, total antioxidant capacity; nav, ‘Navel’ orange; man, ‘Mandarin’; val, ‘Valencia’ orange; A, Western Cape sampling site; B, Eastern Cape sampling site).

Seven phenolic compounds were identified and quantified for the three varieties (Table 2). The flavone-O-glycosides included the 3-O-rutinosides QRG and RUT. Flavanone-O-glycosides included the 7-O-rutinosides NRGLC, NART, HD and NEOP. One hydroxycinnamic acid, namely FHX, and lastly a flavone-C-glucoside, namely VIC2, were quantified. No significant difference (P ≥ 0.05) was found in the QRG of the three varieties. There was, however, a significant difference (P < 0.05) in the RUT levels, with ‘Mandarin’ having the highest concentration (7.13 mg L−1 ). The RUT concentration of ‘Navel’ orange and ‘Valencia’ orange did not differ significantly (P ≥ 0.05). These two varieties are closely associated since they are subspecies of C. sinensis, explaining their similar phenolic composition. A significant difference (P < 0.05) were observed for all the quantified flavanone-O-glycosides and ‘Navel’ oranges had the highest concentration of NRGLC (18.58 mg L−1 ),

wileyonlinelibrary.com/jsfa

HD (40.66 mg L−1 ) and NEOP (6.46 mg L−1 ). The NART and NEOP concentration of ‘Navel’ oranges (45.70 and 6.46 mg L−1 , respectively) were similar to that of ‘Mandarin’ (54.14 and 5.77 mg L−1 , respectively) (P ≥ 0.05), while the HD concentration was similar for ‘Navel’ and ‘Valencia’ oranges. In addition, the NRGLC of ‘Mandarin’ and ‘Valencia’ orange was found not to differ (P ≥ 0.05). FHX concentration varied significantly (P < 0.05) between the varieties, with ‘Valencia’ orange having the highest concentration (16.35 mg L−1 ), followed by ‘Navel’ orange (9.10 mg L−1 ) and ‘Mandarin’ (6.97 mg L−1 ). ‘Mandarin’ was found to contain significantly (P < 0.05) lower VIC2 levels (19.06 mg L−1 ) compared to ‘Navel’ and ‘Valencia’ oranges (34.28 and 31.28 mg L−1 , respectively). Ultimately, the ‘Navel’ orange variety was characterized by the highest TP of 161.94 mg L−1 , which was significantly (P < 0.05) higher than that of ‘Valencia’ (141.58 mg L−1 ) and ‘Mandarin’

© 2018 Society of Chemical Industry

J Sci Food Agric (2018)

Phenolic characterization of citrus juice

www.soci.org

Table 1. P-values (P < 0.05 in bold) for ANOVA of individual chemical parameters, phenolic composition and total antioxidant capacity for different citrus varieties, growing regions and processing seasons Source

Region (R)

Variety (V)

R×V

Year (Y)

R×Y

V ×Y

TA TSS:TA pH QRG FHX VIC2 NRGLC RUT NART HD NEOP TP TACDPPH TACORAC

0.0002 0.0018 0.7194 0.0407 0.0094