ZHANG Jun , ZHANG Jing-yi SHAN You-xia, GUO Can HE Lian ZHANG Lin-yan LING Wei LIANG Yan ZHONG Ba-lian

1 National Engineering Research Center of Navel Orange, Gannan Normal University, Ganzhou 341000, P.R.China

2 South China Botanical Garden, Chinese Academy of Sciences, Guangzhou 510650, P.R.China

Abstract The present study investigates the chemical composition and antioxidant capacity of juice from the Gannan navel orange, which is harvested at one- to two-week intervals during the ripening period.The total soluble solid (TSS), total polyphenol content (TPC), total flavonoid content (TFC), sucrose and hesperidin contents gradually increase with the ripening of the fruit, followed by slight declines at the late maturity stage.Contrary to these observations, the contents of titratable acid (TA), vitamin C (Vc), and limonin trend downward throughout the ripening period.However, the contents of fructose, glucose, and narirutin fluctuate throughout the harvest time.Three in vitro antioxidant assays consistently indicate that the harvest time exerts no significant influence (P>0.01) on the antioxidant capacity.Furthermore, principal component analysis (PCA) and Pearson’s correlation test are performed to provide an overview of the complete dataset.This study provides valuable information for evaluating the fruit quality and determining when to harvest the fruit in order to meet the preferences of consumers.Meanwhile, our observations suggest that the fruits subjected to juice processing should be harvested at the late maturity stage to alleviate the “delayed bitterness” problem without compromising the antioxidant capacity and the flavonoid content in the juice.

Keywords: navel orange, antioxidant, harvest time, chemical composition

1.Introduction

Gannan navel orange (CitrussinensisOsbeck ‘Newhall’) is widely cultivated in Ganzhou (25.83°N, 114.93°E) of Jiangxi Province, China.Currently, Ganzhou is the largest navel orange production area in China, with an annual output of more than 1 million tons (GGC 2020).Due to its beautiful appearance, delicious taste, and high nutritional value, Gannan navel orange is designated as a national geographic indication fruit and enjoys great popularity among the consumers in China.

The orange is one of the most widely consumed fruits worldwide for its unique flavor and nutrition.Organic acids and sugars are among the primary components of oranges, and their concentrations largely affect the characteristic taste and organoleptic quality of orange juice.Sucrose, fructose and glucose are the major sugar components in orange juice, and their concentrations are usually determined by high-performance liquid chromatography (HPLC) or gas chromatography (GC) analysis (Kelebeket al.2009).Phenolic compounds, especially flavonoids (hesperidin and narirutin), are present in large amounts in orange juice, and have shown a wide range of health beneficial properties (Sdiriet al.2012; Wanget al.2016).In recent years, tremendous attention has been paid to the antioxidant potential of orange juice (Giuffrèet al.2017; Hunlunet al.2019; Katariyaet al.2020).The phenolic compounds present in orange juice were reported to be strongly correlated with the antioxidant capacity determined by the 1,1-diphenyl-2-picrylhydrazyl free radical (DPPH) and ferric ion reducing antioxidant power (FRAP) assays (Floegelet al.2011; Kumaret al.2019).However, some studies assumed that the total antioxidant capacity of orange juice was mainly attributed to Vc, giving rise to some differing opinions regarding which compounds might be the major contributors (Gardneret al.2000; Zuluetaet al.2007).

The maturity level is an important factor responsible for the quality of the fruits.The Valencia oranges (C.sinensis) harvested at the late maturity stage had a higher sugar/acid ratio (Baiet al.2009).The contents of total soluble solid (TSS), titratable acid (TA), and Vc of three grapefruit cultivars varied with different harvest times (Muhtaseb 2007).Therefore, it might be interesting to explore how the maturity level influences orange fruit quality, which might provide valuable information regarding when the orange fruit should be harvested to meet the preferences of consumers.

Eating the fresh fruit is the major consumption method of Gannan navel orange at present.One of the major difficulties in commercial juice production from Gannan navel orange is the “delayed bitterness” problem, which is mainly caused by the gradual formation of the bitter compound limonin from the non-bitter precursor limonate A-ring lactone under acidic conditions (Kimball and Norman 1990; Raithoreet al.2016).Our previous study revealed that the limonin content in the juice decreased with the increment of postharvest storage time of the fruit (Zhanget al.2018).It has been reported that the limonin contents in both Hamlin (C.sinensis) and Valencia oranges decreased with the ripening of the fruit (Baiet al.2016).To the best of our knowledge, there are no reports in terms of whether the harvest time influences the limonin content in Gannan navel orange.

Gannan navel orange is usually harvested from late October to late December, covering a span of around two months.A literature survey indicated that only limited research regarding the fruit quality of Gannan navel orange has ever been conducted, mainly focusing on the physicochemical properties such as TSS and TA (Haoet al.2011).The contents and types of functional compositions in the Gannan navel orange, as well as the antioxidant capacity of its juice, remain largely unknown.Therefore, the present study comprehensively investigated the fruit quality of Gannan navel orange during its ripening period and explored how the harvest time influenced the physicochemical properties (TSS, TA, sucrose, fructose and glucose), bioactive components (total polyphenol content (TPC), total flavonoid content (TFC), Vc, hesperidin, narirutin and limonin), and antioxidant capacity (DPPH, 2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS) and FRAP) of the Gannan navel orange juice.

2.Materials and methods

2.1.Chemicals

Vitamin C, 2,6-dichloroindophenol sodium salt, gallic acid, quercetin, 2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS, 98%), 1,1-diphenyl-2-picrylhydrazyl (DPPH free radical, 95%), 6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid (trolox, 97%), 2,4,6-tris(2-pyridyl)-s-triazine (TPTZ), Folin-Ciocalteu reagent, D-sucrose, D-glucose, and D-fructose were purchased from Macklin Biochemical (Shanghai, China).Ammonium persulphate ((NH4)2S2O8), sodium carbonate (Na2CO3), iron(III) chloride hexahydrate (FeCl3∙6H2O), acetic acid (CH3COOH) and sodium acetate (CH3COONa), were of analytical grade and purchased from Sigma-Aldrich (Shanghai, China).Hesperidin and narirutin were obtained commercially from Shanghai Yuanye Bio-Technology Co., Ltd., (Shanghai, China).N, N-dimethylformamide (DMF) was of analytical grade and obtained from Damao Chemical Reagent Factory (Tianjin, China).Methanol and acetonitrile were of HPLC grade and obtained from Anaqua Chemicals Supply (Houston, USA).Deionized water was obtained from a Milli-Q Gradient A10 system (Millipore, Billerica, USA).

2.2.Fruit materials

Newhall navel oranges were harvested from an orchard at the National Engineering Research Center of Navel Orange located in Ganzhou of Jiangxi Province, China.Five healthy Newhall navel orange trees (8 years old) grafted on trifoliate orange (Poncirustrifoliata(L.) Raf.) and growing under the same climate and farming conditions were selected for sample collection.The orange trees were planted on a red-yellow loam soil (pH 5.39±0.3) and spaced at 4 m and 3 m between and along the rows, respectively.During the experiment, seedcake was applied at 50.0 kg/tree in March and the compound fertilizer was applied at 10.0 kg/tree in July.A drip irrigation system was used to irrigate the plants in relation to the environmental temperature and based on the demand for water by the trees.The orchard is located in the area of subtropical monsoon climate with an annual average air temperature of 20°C and an average annual rainfall of about 1 320 mm.Seven batches of fruits were collected at one- to two-week intervals from Oct.25 to Dec.26, 2018.Each batch of fruit contained 50 fruits (10 from each tree) of similar size (each fruit weight 250.0±15.0 mg) and color determined by using a colorimeter (Konica Minolta CM-5, Osaka, Japan), which were collected from different parts of the crown.All fruits were sent to the laboratory and juiced immediately by using a domestic semi-automatic juicer (Model 4161, De’Longhi Braun Household GmbH, Budapest, Hungary).Fresh juice was subjected to centrifugation (Eppendorf centrifuge 5430R, 1 000×g, at 4°C) for 5 min and the supernatant was transferred into sterile plastic Falcon tubes (50 mL) and stored in the refrigerator at –80°C for less than one week before analysis.The juice samples were thawed quickly under running tap water and directly subjected to TSS, TA, Vc, TPC, TFC and antioxidant assays, while for the other experiments, additional steps were needed to prepare the samples based on the analytical methods.

2.3.TSS analysis

TSS of juice was determined by using the handheld refractometer Atago N1 (Atago Co., Ltd., Tokyo, Japan) and expressed as ºBrix at 20°C.Briefly, the instrument was subjected to zero-set using deionized water.After removing the water, one drop of juice was added and the measurement was recorded as TSS (ºBrix).

2.4.TA and vitamin C analysis

TA was determined by titrating 20 mL of juice with a 0.1 mol L–1sodium hydroxide aqueous solution up to pH 8.1 according to a previous method and expressed as grams of citric acid equivalent per 100 g of juice (Giuffrè 2019).Vitamin C was determined by the titration method, using 2,6-dichlorophenolindophenol dye according to previous literature (Stincoet al.2012).The result was expressed as mg ascorbic acid per 100 g of juice.

2.5.HPLC analysis of sucrose, glucose and fructose

The sucrose, glucose and fructose in the orange juice were detected by HPLC according to a previous method with some modifications (Kelebeket al.2009).The juice was diluted ten-fold with 80% methanol/water, followed by centrifugation (Eppendorf centrifuge 5430R, 2 000×g, at 4°C) for 10 min.The supernatant was then filtrated through a 0.22 μm size hydrophilic polytetrafluoroethylene syringe filter to yield the juice sample for HPLC analysis.An Agilent 1260 system (Agilent Technology, CA, USA) consisting of an RI detector and an Agilent ZORBAX NH2(250 mm length, 4.6 mm i.d., 5 µm particle size) was employed for sugar analysis.The mobile phase was an isocratic solution (70% acetonitrile aqueous) at a flow rate of 1.0 mL min–1.The injection volume was fixed at 10 μL, and the temperature for both column and detector was kept at 40°C.D-sucrose (tR=8.84 min), D-glucose (tR=7.26 min), and D-fructose (tR=6.61 min) were identified by comparisons of their retention times with those of the corresponding standards and the concentration of each sugar was quantified by the external calibration method.The calibration curves of sucrose (y=191 728x+ 1 106.3,R2=0.9996), glucose (y=139 960x+887.85,R2=0.9999) and fructose (y=182 405x+6 141.9,R2=0.9994) were constructed by plotting the peak areasversusthe concentrations (from 0.1 to 5.0 g L–1) of the corresponding standard compound.

2.6.TPC analysis

TPC was determined spectrophotometrically with Folin-Ciocalteu reagent according to our previous report (Guoet al.2020).Briefly, pre-diluted juice (20 μL, 20-fold dilution with deionized water), deionized water (60 μL), and pre-diluted Folin-Ciocalteu reagent (15 μL, 2-fold dilution with deionized water) were sequentially added into a 96-well plate and mixed well.After slightly shaking the plate on a spectrophotometer (Tecan Spark 10M, Männedorf, Switzerland) for 4 min at room temperature, 75 μL sodium carbonate aqueous solution (2%, w/v) was added, followed by slight shaking for another 5 min.After standing stationary for 15 min at room temperature, the absorbance was then measured at 750 nm by a spectrophotometer (Tecan Spark 10M, Männedorf, Switzerland).Deionized water instead of juice was used as a blank control.TPC was calculated from the linear calibration curve of gallic acid, which was constructed by plotting the absorbance valuesversusthe concentrations of gallic acid.The calibration curve (y=0.0056x–0.012,R2=0.9998) was linear within the range of 6.25–100 mg L–1.TPC was expressed as µmol gallic acid equivalents per mL juice (µmol GAE mL–1).

2.7.TFC analysis

TFC was estimated by the aluminum chloride method with minor modifications (Guoet al.2020).To a flask containing 500 μL juice, was sequentially added 500 μL sodium nitrite (5%, w/v), 500 μL aluminium chloride (10%, w/v) and 500 μL sodium hydroxide (1 mol L–1) aqueous solution at an interval of 5 min and the solution was mixed well upon each addition.After standing stationary for 15 min at room temperature, the reaction mixture was diluted with deionized water to a final volume of 10 mL.Deionized water instead of juice was used as a blank control.The absorbance of the reaction mixture was measured at 415 nm by a UV-vis photospectrometer (Model 2450, Shimadzu Co., Ltd., Kyoto, Japan).TFC was calculated from the linear calibration curve of quercetin, which was constructed by plotting the absorbance valuesversusthe amounts of quercetin and expressed as µmol of quercetin equivalents per mL juice (µmol QE mL–1).The calibration curve (y=0.0013x+0.0031,R2=0.9991) was linear within the range of 31.25–500 mg.

2.8.HPLC analysis of hesperidin and narirutin

Hesperidin and narirutin in the juice were analyzed by our previous method (Zhanget al.2020).Briefly, the juice was subjected to 10-fold dilution with 20% dimethylformamide (DMF)/water, followed by centrifugation (Eppendorf centrifuge 5430R, 2 000×g, at 4°C) for 10 min.The supernatant was then filtered through a 0.22-μm pore-size hydrophobic polytetrafluoroethylene syringe filter to yield the juice sample for HPLC analysis.An Agilent 1260 System (Agilent Technology, CA, USA), which consisted of a UV detector, a binary gradient pump, and a sunfire™ C18 reverse-phase column (150 mm length, 4.6 mm i.d., 3.5 μm particle size), was employed for the analysis of hesperidin and narirutin.The mobile phase consisted of acetonitrile (A) and deionized water (B) at a flow rate of 1.0 mL min–1.The gradient profile was as follows: 0– 40 min, 10–25% A; 40–45 min, 25–50% A; 45–60 min, 50–65% A; 60–65 min, 65–90% A; 65–75 min, 90% A; 75–80 min, 90–10% A.The chromatogram was monitored at 280 nm with an injection volume of 20 μL and column temperature at 30°C.Both hesperidin and narirutin were identified by comparisons of their retention times and UV spectra with those of the standard compounds.The concentration of each compound was quantified by the external calibration curves (y=17.585x+1.8909,R2=0.9999;y=20.448x+0.5933,R2=0.9996, for hesperidin and narirutin, respectively), which were constructed by plotting the peak areasversusthe concentrations (from 1.0 to 100.0 mg L–1) of the standard compounds.

2.9.HPLC analysis of limonin

The determination of limonin content was performed according to our previous method with some minor modifications (Zhanget al.2018).Juice (20 mL) was acidified by adding 1.0 mol L–1hydrochloric acid aqueous to pH 2.0, followed by heat treatment at 50°C for 10 min.After cooling down to room temperature, deionized water (15 mL) was added.The juice solution was then extracted with dichloromethane (40 mL×3).The combined dichloromethane extract was concentrated by rotary evaporator under reduced pressure at 50°C to give a residue, which was dissolved and diluted with methanol to a final volume of 5.0 mL.The methanol solution was then filtered through a 0.22-μm pore-size hydrophobic polytetrafluoroethylene syringe filter to yield the juice sample for HPLC analysis.An Agilent 1260 Infinity System (Agilent Technology, CA, USA), which consisted of a UV detector, a binary gradient pump, and a sunfire™ C18 reverse-phase column (150 mm length, 4.6 mm i.d., 3.5 μm particle size), was employed.The mobile phase consisted of acetonitrile (A) and deionized water (B) at a flow rate of 1.0 mL min–1.The gradient profile was as follows: 0–30 min, 10–70% A; 30–40 min, 70–90% A; 40–50 min, 90% A; 50–60 min, 10% A.The detector wavelength was 210 nm with an injection volume of 20 μL and column temperature at 30°C.The limonin was identified by comparisons of its retention time and UV spectrum with those of the standard compound.The limonin content was quantified by the external calibration method.The external calibration curve of limonin (y=30 271x+5 851,R2=0.9995) was linear within the range of 1.0 to 100.0 mg L–1, and constructed by plotting the peak areasversusthe concentrations of the standard compound.

2.10.DPPH assay

The DPPH assay was performed according to a previous method with minor modifications (Fioreet al.2005).DPPH was freshly prepared in methanol at a concentration of 0.1 mmol L–1.The juice was subjected to sequential two-fold dilution with deionized water to obtain various concentrations of juice samples.A juice sample (50 μL) was added to a 96-well plate, followed by the addition of DPPH solution (150 μL).Deionized water (50 μL) instead of a juice sample was added in the control well, and methanol (150 μL) in place of the DPPH solution was used as a blank.The plate was kept stationary and placed in the dark for 20 min at 37°C, then the absorption values were recorded at 517 nm by a microplate reader (Tecan Spark 10M, Männedorf, Switzerland).The inhibitory rate against the DPPH radical was calculated according to the formula: Inhibition (%)=[1–(Atreated–Ablank)/Acontrol]×100, where the Atreatedrepresents the average absorption of wells containing both DPPH and juice sample, Ablankis the average absorption of wells containing no DPPH, and Acontrolis the average absorption of wells containing only DPPH.A series of concentrations of trolox aqueous solution (6.25–100 mg L–1) was used for the construction of a calibration curve (y=0.0069x+0.2756,R2=0.9986).The results were expressed as µmol trolox equivalent antioxidant capacity per mL of juice (µmol TEAC mL–1) by a comparison with the calibration curve of trolox.

2.11.ABTS assay

ABTS was determined according to a previous method with some modifications (Arenaet al.2001).An aqueous solution of ABTS (7 mmol L–1) was incubated with ammonium persulphate aqueous solution (2.45 mmol L–1) at a volume ratio of 1:1 in the dark at room temperature for 12 h to generate the ABTS radical solution, which was diluted with deionized water to an absorbance of 0.70±0.02 at 734 nm.Various concentrations of two-fold diluted juice solutions (50 mL) were added into the wells of a 96-well plate, followed by the addition of ABTS radical solution (200 mL).Deionized water (50 μL) instead of a juice sample was added in the control well, and water (200 μL) in place of the ABTS radical solution was used as a blank.The plate was steadily shaken for 10 min, then the absorbance was measured at 734 nm by a microplate reader (Tecan Spark 10M, Männedorf, Switzerland).The scavenging ability against ABTS radical was calculated according to the formula: Inhibition (%)=[1–(Atreated–Ablank)/Acontrol]×100, where Atreatedrepresents the average absorption of wells containing both ABTS and a juice sample, Ablankis the average absorption of wells containing no ABTS, and Acontrolis the average absorption of wells containing only ABTS.The results were expressed as µmol trolox equivalent antioxidant capacity per mL of juice (µmol TEAC mL–1) by a comparison with the calibration curve (y=23.97x+0.2024,R2=0.9989) of trolox, which was constructed by plotting the inhibitory percentagesversusthe concentrations (1.56–25.0 mg L–1) of trolox.

2.12.FRAP assay

The FRAP assay was carried out according to the literature with a minor modification (Giuffrè 2019).A fresh FRAP solution was prepared by mixing TPTZ solution (10 mmol L–1) in 40 mmol L–1hydrochloric acid, acetate buffer (0.1 mmol L–1, pH 3.6), and ferric chloride (20 mmol L–1) solution in 40 mmol L–1hydrochloric acid, at a volume ratio of 1:10:1.Various concentrations of two-fold diluted juice solutions (50 µL) were added into a 96-well plate, followed by the addition of freshly prepared FRAP reagent (200 µL).The plate was then steadily shaken at room temperature for 10 min on a spectrophotometer (Tecan Spark 10M, Männedorf, Switzerland), followed by the measurement of absorbance at 593 nm by the same spectrophotometer.The results were compared to the trolox calibration curve (y=19.875x–0.0077,R2=0.9999), which was constructed by plotting the absorbance valuesversusthe concentrations (3.12–100.0 mg L–1) of trolox and expressed as µmol trolox equivalent antioxidant capacity per mL of juice (µmol TEAC mL–1).

2.13.Statistical analysis

At least three independent experiments were performed for each analysis and data were expressed as mean±standard deviation (SD).The analysis of one way variance (ANOVA) was used to compare the means, and determine significant differences (P<0.01).Statistical analysis of the data was done using the SPSS Software (version 13, SPSS, Chicago, USA).PCA and Pearson’s correlation analysis were performed in R language (R Core Team 2017).

3.Results and discussion

3.1.Effects of harvest time on TSS and TA

The TSS is an important factor reflecting the maturity level of fruit, which is normally reported as grams of glucose equivalent per 100 g juice and recorded as °Brix (Giuffrè 2019).In the present study, the TSS of Gannan navel orange showed an incremental tendency from Oct.25 (11.41 °Brix) to Dec.19 (13.15 °Brix) (Table 1).However, there was a slight decline of TSS on Dec.26 (12.79 °Brix) when compared to that on Dec.19 (13.15 °Brix), indicating that the TSS might have reached a maximal level around Dec.19.Contrary to the TSS, the TA showed a declining trend during the whole ripening period from Oct.25 (0.69 g 100 g–1) to Dec.26 (0.48 g 100 g–1).A significant difference (P<0.01) was found in the TA contents among fruits harvested at the early ripening period, from 0.69 (Oct.25) to 0.61 (Nov.8), and then declining to 0.56 g 100 g–1(Nov.22).At the late ripening period (from Dec.5 to Dec.26), TA only decreased slightly, from 0.49 (Dec.5) to 0.48 g 100 g–1(Dec.26), showing no significant difference (P>0.01) between them.The TSS/TA ratio is an important parameter related to the quality characteristics of citrus fruits (Giuffrè 2019).Due to the increment trend of TSSversusthe declining tendency of TA during the harvesting period, the TSS/TA ratio of Gannan navel orange gradually increased from Oct.25 (16.53) to Dec.26 (26.64) (Table 1).Similar to our observations, the TSS and TA of Gannan navel orange harvested on Dec.20 were reported as 13.5 °Brix and 0.75 g 100 g–1, respectively (Zenget al.2012).It has been reported that the TSS of Valencia oranges (C.sinensisL.Osbeck ‘Valencia’) produced in Florida remained almost constant (10.3 to 11.0 °Brix) regardless of the harvesting time (Baiet al.2009), which is different from the trend in the present study, and could be mainly ascribed to the differences in both citrus varieties and climatic conditions.

Table 1 Physicochemical properties of juice from the Gannan navel orange harvested at different time1)

3.2.Effects of harvest time on the contents of fructose, glucose, and sucrose

Fructose, glucose, and sucrose are the primary sugar components in orange juice.As shown in Table 1, Gannan navel orange contains a high amount of sucrose, ranging from 53.39 to 67.46 g L–1, followed by glucose (25.73–29.49 g L–1) and fructose (23.14–25.29 g L–1) with a content ratio of about 1:1.The sucrose content displayed an increasing tendency along with the ripening of the fruit, from 53.39 g L–1on Oct.25 to 67.46 g L–1on Dec.19, and then it slightly decreased to 66.42 g L–1on Dec.26, achieving an increasing percentage of about 26% across the whole maturity period.The sucrose content exhibited a significant difference (P<0.01) between the first two batches of fruits harvested at the early maturity stage on Oct.25 (53.39 g L–1) and Nov.8 (58.78 g L–1), while the growth rate gradually decreased for fruits harvested in the late ripening period, showing no significant differences between the adjacent batches (P>0.01).There is no obvious increasing or declining tendency for either the fructose or glucose contents during the whole harvesting period, which fluctuated between 23.14 and 25.29 g L–1, and between 25.79 and 29.49 g L–1, respectively.The ratio of sucrose, glucose, and fructose contents was about 2:1:1 in the present study, being consistent with those reported previously (Lee and Coates 2000; Kelebeket al.2009).The Kozan orange (C.sinensisL.Osbeck ‘Kozan’) contained sucrose (59.34 g L–1), glucose (32.30 g L–1) and fructose (28.55 g L–1) as its main sugar compositions (Kelebeket al.2009), which was in agreement with the present result.The content of sucrose in Valencia orange (C.sinensisL.Osbeck ‘Valencia’) increased from 49.0 to 56.0 g L–1over the ripening period from February to May, while both fructose and glucose remained stable at around 20 g L–1(Baiet al.2016).This tendency was consistent with our present observations.

3.3.Effects of harvest time on TPC, TFC and Vc

TPC showed a gradually increasing trend over the harvest period from Oct.25 (4.39 µmol GAE mL–1) to Dec.19 (5.12 µmol GAE mL–1) (Table 2).This trend was also applicable to TFC, which increased from 1.90 µmol QE mL–1on Oct.25 to 2.30 µmol QE mL–1on Dec.19, for a percentage increase of about 21%.However, both the TPC and TFC declined to 4.79 µmol GAE mL–1and 2.16 µmol QE mL–1on Dec.26, respectively, which might be ascribed to the low metabolic ability of senescent fruit cells for producing phenolic and flavonoid compounds (Romani 1978).Contrary to the change tendencies of both TFC and TPC, the content of Vc exhibited a declining trend during the whole ripening period, from around 61.10–62.54 mg 100 g–1for the early harvest fruits (Table 2) to around 50.75–52.85 mg 100 g–1for the late harvest fruits, showing a significant difference (P<0.01) between them.In good agreement with the present observations, the Vc of Valencia (C.sinensisL.Osbeck ‘Valencia’) and Bergamot fruits (C.bergamiaRisso) decreased throughout the harvest season (Baiet al.2009; Giuffrè 2019).The Vc content was 58.4 mg 100 g–1for Washington navel orange (C.sinensisL.Osbeck ‘Washington’) (Kimball and Norman 1990), and 55.37 and 55.11 mg 100 g–1for Tarocco (C.sinensisL.Osbeck ‘Tarocco’) and Moro (C.sinensisL.Osbeck ‘Moro’) sweet orange varieties, respectively (Rapisardaet al.2008), which are similar to the Vc content reported in this study.It has been reported that the TPC was 8.81 µmol GAE mL–1for Hamlin (C.sinensisOsbeck ‘Hamlin’) (Xuet al.2008), 5.66 µmol GAE mL–1for Salustiana orange (C.sinensisOsbeck ‘Salustiana’) (Roussos 2011), and 1.43 µmol GAE mL–1for a commercially available orange in the Belgian market (Keverset al.2007).The variations of TPC between our present study and the previous reports might be attributed to the differences in factors such as the genotypes, cultivation conditions, postharvest processes, and maturity stages of the fruits.The TFC in Fantastico (C.bergamia‘Fantastico’) increased with fruit ripening, from 361 mg L–1in October to 678 mg L–1in March of the next year (Giuffrè 2019).The concentration of flavonoids in Persian lime (C.latifoliaTanaka) increased during the whole fruit growth period, whereas the phenolic compounds increased at first, followed by a decline during the rest of the maturation phase (Ledesma-Escobaret al.2018).The tendency of change of TFC/TPC in these reports was almost consistent with the present observations.However, it has been reported that in the flesh of Chachiensis (C.reticulataBlanco ‘Chachiensis’), the vitamin C content increased, while the TPC decreased by about 15% throughout the ripening period (Wanget al.2016), which was exactly the opposite of what we found in the present study.

Table 2 Bioactive components of juice from the Gannan navel orange harvested at different time1)

3.4.Effects of harvest time on hesperidin and narirutin

As shown in Table 2, the hesperidin content increased continuously from Oct.25 (478.21 mg L–1) to Dec.19 (839.62 mg L–1) with the ripening of fruit, then decreased slightly to 827.63 mg L–1on Dec.26.There was a significant difference in the hesperidin content between the fruits harvested at the early and the late stages (P<0.01), with an increasing percentage of about 73% during the whole harvest period.The content of narirutin, however, fluctuated between 249.58 (Dec.19) and 294.91 (Nov.8) mg L–1throughout the maturity period, exhibiting no obvious tendency of increase or decline.The contents of hesperidin and narirutin were 192.2 and 43.0 mg L–1, respectively, in blood orange juice (C.sinensisL.Osbeck ‘Moro’) (Giuffrèet al.2017), 171.17 and 39.91 mg L–1, respectively, in Kozan (C.sinensisL.Osbeck ‘Kozan’) (Kelebeket al.2009), 489.64 and 102.77 mg L–1, respectively, in Hamlin (C.sinensisL.Osbeck ‘Hamlin’) (Xuet al.2008), and about 470–761 and 185–222 mg L–1, respectively, in the pigmented orange juice (C.sinensisL.Osbeck) (Leuzziet al.2000).The difference between our present results and those reported previously might be attributable mainly to the differentCitrusspecies and sample preparation methods used.In general, the hesperidin content was determined by the HPLC method, whereas the sample preparation method before HPLC analysis may have highly influenced the results (Zhanget al.2020).In contrast with our present observations, the hesperidin content in the Chachiensis (C.reticulataBlanco ‘Chachiensis’) flesh increased with the ripening of fruit, at 896.47 µg g–1DW for immature fruit, 928.57 for semi-mature and 1 076.54 µg g–1DW for mature fruit, while the narirutin content was almost the same throughout the ripening period (Wanget al.2016).The content of narirutin in Valencia orange (C.sinensisL.Osbeck ‘Valencia’) remained constant in March and April, but decreased from April to May (Baiet al.2009).Naringin was a main flavonoid in bergamot (C.bergamiaRisso), and its highest content was obtained in the fruits harvested at the last sampling date on March 2017/2018 (Giuffrè 2019), showing a tendency of change similar to that of the hesperidin in the present study.

3.5.Effects of harvest time on limonin content

Limonin is believed to be the primary bitterness compound in navel orange juice, which is formed gradually from a non-bitter precursor during juice processing (Raithoreet al.2016).In this study, the non-bitter precursor in navel orange juice was completely transferred to limonin in the presence of acid and heat according to our previous report (Zhanget al.2018).As shown in Table 2, the limonin content was the highest (13.98 mg L–1) in the fruits harvested on Oct.25, and it exhibited a declining tendency with the ripening time, down to 3.35 mg L–1on Dec.26, which represents about a 76% reduction over the whole maturity period.The sensitivity threshold of limonin content in the juice is suggested to be 6.0 mg L–1, so a concentration less than that is considered to be acceptable for the consumers (Raithoreet al.2016).The limonin content in Gannan navel orange juice was below 6.0 mg L–1on Nov.29 (5.21 mg L–1), therefore the fruits harvested after this time should be suitable for juice production.Similar to the present results, the limonin concentration in the early Washington navel orange (C.sinensisL.Osbeck ‘Washington’) juice was about 17.2 mg L–1(Kimball and Norman 1990).It has been reported that the limonin content in Valencia orange (C.sinensisL.Osbeck ‘Valencia’) declined from February (0.90 mg g–1) to May (0.53 mg g–1) in 2007, and from February (5.6 mg g–1) to May (2.4 mg g–1) in 2012 (Baiet al.2016).The limonin content in the Thai tangerine (C.reticulataBlanco) decreased with the ripening of the fruit, from 4.0 to 1.0 mg L–1(Jungsakulrujirek and Noomhorm 1998).The average limonin concentration in Hamlin oranges (C.sinensisL.Osbeck ‘Hamlin’) fell from a high of 6.2 mg L–1in September to 1.8 mg L–1in January of the next year during the maturity season based on TLC (thin-layer chromatography) analysis (Albachet al.1981).These reports were in agreement with our present results, showing the limonin content in juice decreased with the increase of the fruit maturity level.It is worth pointing out that the values of limonin content were different between the present study and some of the previous reports, which might be attributed to the differences in the citrus species and the sample preparation methods employed.

3.6.Effects of harvest time on antioxidant capacity

To comprehensively estimate the antioxidant capacity of orange juice, the three platforms of DPPH, ABTS, and FRAP were used in the present study.The antioxidant capacity was expressed as µmol trolox equivalents per mL orange juice (µmol TEAC mL–1) for each of these three antioxidant platforms.As shown in Table 3, the FRAP values did not tend to increase or decrease with the ripening of the fruit, varying between 3.60 (Dec.5) and 3.88 µmol TEAC mL–1(Dec.19) throughout the harvesting period.Consistent with the FRAP, the ABTS value displayed fluctuations between 3.43 (Nov.8) and 4.00 µmol TEAC mL–1(Dec.19) with no significant difference between them (P>0.01).The antioxidant potential against the DPPH radical was not significantly different among the fruits harvested at different maturity stages (P>0.01), ranging from 5.28 (Oct.25) to 6.01 µmol TEAC mL–1(Dec.5), similar to those observed in both the FRAP and ATBS assays.Therefore, all three of the antioxidant platforms consistently indicated that the harvest time of the orange exerted no significant influence (P>0.01) on the antioxidant capacity of Gannan navel orange juice.Vitamin C, phenolic compounds and flavonoids have been reported to be the most important antioxidants in orange juice (Sdiriet al.2012).In the present study, the content of Vc declined, while the TPC/TFC increased during the fruit ripening period, imposing two opposite effects on the antioxidant capacity of juice, which might be responsible for the fluctuations of the values of FRAP, ABTS, and DPPH.The potency of bergamot juice against the DPPH radical was highly correlated with the content of its flavonoids, while the Vc was found to be responsible for the antioxidant capacity determined by the FRAP assay (Giuffrè 2019).It was reported that the ABTS values of the ready-to-drink orange juices available in Brazil were 0.58–3.15 µmol TEAC mL–1(Stellaet al.2011), and those of the freshly squeezed and processed blood orange juices were 2.38–5.18 µmol TEAC mL–1(Arenaet al.2001), which are similar to the values in the present report.The DPPH and ABTS radical scavenging capacity of seven commercial red orange juices were 1.26–3.08 and 2.28–3.53 µmol TEAC mL–1, respectively (Fioreet al.2005), slightly less than those reported in the present study, which might be due to the deterioration of antioxidants during the processing and storage period of the commercial orange juice (Klimczaket al.2007).

Table 3 Antioxidant capacity of juice from the Gannan navel orange harvested at different time1)

3.7.PCA and Pearson’s correlation analysis

PCA is a statistical method that reduces a large number of variables into a smaller number of uncorrelated variables and provides an overview of the complete dataset (Kawaiiet al.1999).Two principal components (PCs) were obtained from the orange juice data sets, accounting for 99.9% of the cumulative percentage of the total variation, with PC 1 and PC 2 accounting for 97.89 and 2.03% of the variance, respectively.The results of PCA are presented as the scores and loading plots (Fig.1), where the contribution of a variable to PC 1 and PC 2 is indicated by the direction and length of the vector.As shown in Fig.1, the oranges harvested on Oct.25, Nov.8, Nov.22, and Dec.19 were grouped very well, while those harvested on Nov.29, Dec.5, and Dec.26 overlapped each other partially.The proximity of the early harvesting fruits (Oct.25, Nov.8, and Nov.22) to the loading vectors involving Vc, limonin, and TA, revealed that the orange fruits harvested at the early ripening stage were characterized by high contents of Vc, limonin, and TA.The loading vectors for hesperidin, sucrose and TSS were close to the fruits harvested at the late maturity stage (Dec.19 and Dec.26), indicating these three were the main chemical components in the late harvest fruits.Pearson’s correlation analysis revealed that the hesperidin content was highly negatively correlated with the contents of limonin (coefficient=–0.96,P<0.01), TA (coefficient=–0.92,P<0.01), and Vc (coefficient=–0.78,P<0.01) in fruit, while positively correlated with TSS (coefficient=0.88,P<0.01), sucrose (coefficient=0.90,P<0.01), TPC (coefficient=0.73,P<0.01), and TFC (coefficient=0.74,P<0.01) (Fig.2).This pattern was in good agreement with the PCA results, where the variables of hesperidin, TSS, and sucrose had the same vector direction which was opposite to the vector direction of limonin, TA, and Vc.As shown in Fig.2, DPPH demonstrated low correlations with TPC (coefficient=0.21,P<0.01), TFC (coefficient=0.48,P<0.01), and Vc (coefficient=–0.43,P<0.01).Similar to DPPH, both ABTS and FRAP were poorly correlated with TPC (coefficient=0.22, 0.42, respectively,P<0.01), TFC (coefficient=–0.07, 0.26, respectively,P<0.01), and Vc (coefficient=–0.14, –0.15, respectively,P<0.01).These data indicated that the antioxidant capacity of Gannan navel orange juice was not determined by only one of the three factors including TPC, TFC, and Vc, which should indicate a synergistic effect among all these factors.Xuet al.(2008) suggested that the Vc, not the phenolic compounds, was the major contributor to the total antioxidant capacity of citrus juice.However, some studies reported that the phenolic compounds were primarily responsible for the total antioxidant capacity of citrus (Rapisardaet al.1999; Hunlunet al.2019).The discrepancy between the previous and the present observations may arise from the differences such as the citrus varieties, fruit maturity stages, sample preparation procedures and analytical methods used.The present study indicated that the fruits harvested at the early ripening time exhibited high contents of Vc and TA, which might be potentially beneficial for human health and desirable for some consumers who prefer oranges with a more acid taste.Meanwhile, the early harvested fruit was suggested to have an enhanced ability to withstand longdistance transportation and postharvest storage (Eckert and Eaks 1989).For the oranges harvested at the middle stage of maturation, they accumulated more sugars and exhibited a relatively high value of TSS/TA, thereby making them more popular as a fresh eating fruit among the consumers inclined to eat sweeter fruits.However, the oranges harvested at the late maturity stage were very suitable for juice processing due to the much lower amount of limonin as compared to the other maturity stages while still maintaining relatively high antioxidant capacity and flavonoid content.

Fig.1 The loading and score plots of the PCA describing data sets obtained from the Gannan navel orange juice.

Fig.2 Pearson’s correlation analysis among the data sets from the Gannan navel orange juice.DPPH, 1,1-diphenyl-2-picrylhydrazyl free radical; FRAP, ferric ion reducing antioxidant power; ABTS, 2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt; TSS, total soluble solid; TPC, total polyphenol content; TFC, total flavonoid content; Vc, vitamin C; TA, titratable acid.*, P<0.05; **, P<0.01; ***, P<0.001.

4.Conclusion

In this study, the effect of harvest time on the chemical composition and antioxidant capacity of Gannan navel orange juice was comprehensively investigated for the first time.Our results indicated that the fruits harvested during the early ripening time had high contents of TA, Vc, and limonin, and all of them trended downward throughout the maturity period of the fruit.The contents of TSS, sucrose, hesperidin, TPC and TFC, however, demonstrated increasing trends with the ripening of fruit, followed by a slight decline at the late maturity stage.Throughout the harvest period, the contents of fructose, glucose, and narirutin fluctuated without an apparent tendency of either increase or decrease.Furthermore, the harvest time exerted no obvious effect on the antioxidant capacity of the Gannan navel orange juice, evident from the consistent results of DPPH, ABTS, and FRAP assays showing that there were no statistical differences (P>0.01) in antioxidant capacity for the fruits harvested at different maturity stages.The present study provides valuable information for evaluating the fruit quality and determining when to harvest the fruit to meet the preferences of different consumers.Meanwhile, our observations suggest that the fruits subjected to juice processing should be harvested at the late maturity stage to alleviate the “delayed bitterness” problem without compromising the antioxidant capacity or the flavonoid content in the juice.

Acknowledgements

This research work was financially supported by the National Natural Science Foundation of China (31860091), the Natural Science Foundation of Jiangxi Province, China (20171BCB24011), the Open Foundation of National Engineering Research Center of Navel Orange, and the Research Foundation of Ganzhou, Jiangxi, China (2017179 and 201960).The authors also thank Dr.Kumaravel Kaliaperumal from Gannan Normal University for proofreading this manuscript.

Declaration of competing interest

The authors declare that they have no conflict of interest.