Characterization and Comparison of Unfermented and Fermented Seed-Watermelon Juice

Beijing Academy of Agriculture and Forestry Sciences, Beijing Vegetable Research Center, Beijing 100097, China Beijing Advanced Innovation Center for Food Nutrition and Human Health, Beijing Technology and Business University (BTBU), Beijing 100048, China Beijing Key Laboratory of Fruits and Vegetable Storage and Processing, Beijing 100097, China Key Laboratory of Biology and Genetic Improvement of Horticultural Crops (North China) of Ministry of Agriculture, Beijing 100097, China Key Laboratory of Vegetable Postharvest Processing of Ministry of Agriculture, Beijing 100097, China


Introduction
Watermelon juice is welcomed for its aroma and health for most people.Consumption of watermelon juice protects the chemical-induced hepatotoxicity in rats [1], increases plasma concentrations of β-carotene in humans [2], and increases antiproliferative activity on both human breast and liver cancer cell lines [3].Seed watermelon (Citrullus lanatus sp.vulgaris var.megalaspermus) is the same species with watermelon in Cucurbitaceae but cultivated for seed.Byproducts (pulp and rind) of seed watermelon constituting about 90% of fruit weight are directly thrown in the elds [4].e byproducts are rich in polyphenol, sugar, and some other biological-related compounds and present a strong radical scavenging capacity [5,6].However, a further application of the byproducts has not been reported.
Fermentation is a traditional way to preserve high sugar and high moisture products.Grape wine is one of the successful examples.Fermentation of wine reduces price uctuations and enhances the economic bene t of the fruits [7].Moreover, fermentation raises the radical scavenging capacity of red grape wine [8] and pomegranate wine as well [9] and maintains the typical aroma of grape or rice in their corresponding fermented products [10,11].Consequently, for the first time, the byproducts of seed watermelons were used to prepare the fermented juice.Our previous results proved that the wine active dry yeast (RV100, E80001478, Angel Yeast Co., Ltd., Yichang, China) maintains the typical watermelon aroma of the fermented juice [12].However, the radical scavenging capacity and aroma composition of the unfermented juice and fermented juice were neglected.Hence, both unfermented juice and fermented juice were prepared and their characteristics were compared.

Preparation of Unfermented and Fermented Juice.
Mature seed watermelons (Citrullus lanatus sp.vulgaris var.megalaspermus) were harvested in September 2015 in Beijing Sijiqing Farm.e fruits were round with regular green stripes and an average weight of 2.3 kg. e pulp was light yellow and full of black seeds.e thickness of the rind was 8∼12 mm.After stored at 4 °C for 24 h, the cold fruits were washed in 100 ppm ice sodium hypochlorite solution and blanched in 50 ppm ice sodium hypochlorite solution in a clean workshop.e washed fruits were drained in a big case for 30 min.e drained fruits were then cut into blocks (enabling access to the juicer), and the blocks were dropped into a juicer (Philips HR1876, Philips, Dongguan, China).
e juicer removed the seeds while collecting the juice.e juice was nominated as the unfermented juice.e soluble solid content and pH of the juice were adjusted to 20 °Brix by adding sucrose and pH 5.5 by adding 0.5 M/L sodium bicarbonate solution, respectively.
e resulting juice was inoculated with the activated yeast mixture at 0.2% (v/v).e activated yeast mixture was prepared by blending the dry yeast (Wine active dry yeast RV100, E80001478, Angel Yeast Co., Ltd,) of 500 g and 10 L of the sucrose solution (5%, w/v) at 30 °C for about 1 h.e inoculated juice was covered by gauze and stored at 25 °C for 15 d and then filtered via 4-layer gauze.e resulting juice was stored at 25 °C for 90 d for further analysis, which was nominated as the fermented juice.

Determination of Phenolic Content.
e phenolic content was measured via the Folin-Ciocalteu reagent method with gallic acid as a reference [13].In brief, the sample of 25 g was blended with 50 mL cold 80% acetone for 5 min via a magnetic blender.
e mixture was filtered through Whatman #1 paper.e paper was rinsed by 10 mL cold 80% acetone for two times.All permeate was collected and evaporated to <10% initial weight at 45 °C with a rotary evaporator (Buchi Vacuum Rotavapor R215, Buchi Labortechnik AG, Flawil, Switzerland).e extract was diluted to 25 mL with 70% methanol and stored at −40 °C until further measurement.e measurement mixture contained 50 μL of the methanol extract, 250 μL of Folin-Ciocalteu reagent, 0.75 mL of 20% sodium carbonate, and 3 mL of deionized water.After inoculation at 25 °C for 2 h, the absorbance at 765 nm was measured with the reaction mixture without sample as control via an ultraviolet spectrophotometer (UV-1800, Shimadzu Corporation, Kyoto, Japan).e result was determined from the linear calibration curve and was expressed as mg gallic acid equivalent/L.
e absorbance of the resulting mixture was measured on an ultraviolet spectrophotometer (UV-1800, Shimadzu Corporation, Kyoto, Japan) at 515 nm after incubation at 30 °C for 1 h with butylated hydroxytoluene (BHT) as the control.
e DPPH radical scavenging activity was calculated as a percentage of DPPH discoloration as BHT equivalent.

Hydroxyl Radical Scavenging Capacity.
e hydroxyl radical scavenging capacity was examined based on Fenton reaction, whereas 5,5-dimethyl-N-oxide pyrroline (DMPO) was used as the trapping agent [15].e reaction mixture contained 10 μL of 3 mM freshly prepared FeSO 4 , 80 μL of 0.75 mM PBS, 15 μL of 10 mM H 2 O 2 , 15 μL of 1 M DMPO, and 30 μL of sample or solvents for the blank.e reaction mixture was evenly shaken and incubated for 60 min at 37 °C.
e absorbance of the resulting mixture was measured at 510 nm via an ultraviolet spectrophotometer (UV-1800, Shimadzu Corporation, Kyoto, Japan).e scavenging capacity of hydroxyl radical was calculated as (1 − absorbance of sample/absorbance of control) × 100%.
e scavenging activity of the sample was further expressed as IC50.Specifically, the sample was diluted to 1%, 5%, 10%, 25%, 50%, 75%, and 100% in deionized water.e scavenging activity of the sample dilution was measured and exponentially fitted.e IC50 of the sample was calculated based on the exponentially fitted curves.

ORAC Assay.
e oxygen radical absorbance capacity (ORAC) of the sample was measured according to the recently reported protocol [16,17].e sample was measured in a hydrophilic system.Specifically, the sample (20 μL) was pipetted to the proper well and incubated for 10 min with fluorescein as the fluorescent probe.e final assay solution contained 0.067 μM of fluorescein, 60 mM of 2,2′-azobis-2methyl-propanimidamide, 300 μL of sample, or 7% β-cyclodextrin as a reagent blank.e fluorescence of an assay mixture was measured and recorded once every minute.e trolox equivalent (TE) was calculated using a standard curve prepared with trolox and used to compare the ORAC of various samples via expression as µM TE/g.e fermented juice without ethanol was measured to show the contribution of ethanol to the oxygen radical scavenging capacity.Specifically, the fermented juice was concentrated to about 10% of the original volume via nitrogen blowing at 25 °C.e resulting solution was diluted to its original volume by adding deionized water.e ORAC of the diluted solution was measured.

2
Journal of Food Quality

Determination, Quantification, and Classification of
Volatiles.An aliquot of 2 mL of the sample was put into a 20 mL glass vessel.e glass vessel was placed in a water bath at 50 °C for 30 min.e headspace was purged with a constant flow of nitrogen at 20 mL/min for 10 min.e volatile compounds were trapped in a glass capillary tube (3 mm internal diameter) containing 100 mg of Tenax ® adsorbent.e Tenax ® trap was then disconnected from the system and placed in a Chrompack TCT/PTI 400 injector.e volatiles were evaluated on an Agilent 6890 gas chromatograph coupled to an Agilent 5973I mass selective detector (Agilent Technologies, Palo Alto, CA).e volatiles were separated on a DB-WAX column (30 m × 0.25 mm i.d., 0.25 μm film thickness, Agilent Technologies).e injection was performed in splitless mode (0.7 mm splitless inlet liner, Supelco), and the injector temperature was 250 °C.
e purge valve was opened at 0.5 min at a 50 mL/min flow rate.Helium (99.999%) was used as a carrier gas with a constant starting flow rate at 1 mL/min.e oven temperature was programmed as follows: 45 °C for 3 min, 5 °C/min to 120 °C, and 10 °C/min to a final temperature of 200 °C with a final holding time of 5 min.e detector was fitted with an electron impact ionization source set at 200 °C.
e quadrupole temperature was set to 150 °C, and the transfer line temperature was kept at 250 °C.
e solvent delay was set to 3 min.Total ion chromatograms were collected scanning from m/z 35 to 500 at a rate of 3.06 scans/s.
Volatiles were identified via comparison of their mass spectra and retention times to those of authentic standards or via comparison of Kovats' retention indexes and mass spectrum to those reported in the NIST Mass Spectral Search Program (version 2.0a) with <80% as a cutoff to match compounds.
Quantification of the volatile compounds was conducted via calibration curves in the selected ion monitoring mode.
e 2-methyl-3-heptanone with a concentration of 45 mg/L in the analyzed samples was used as internal standard to monitor instrument response and retention time stability.Quantitative analysis was performed using software Agilent MassHunter (USA).
e volatiles were classified via their chemical structure and aroma style.Specifically, the chemical structure included alkane, alcohol, aldehyde, ketone, acid, and ester.e aroma style included flower, fruit, grass, fragrant, oil, and bouquet based on the manual Perfumes [18].

Sensory Evaluation.
A panel of 12 judges was composed of seven males and five females between the ages of 22 and 55 years.Each judge was an experienced wine taster, and eight of them were employees of our institution.e panel had already been trained to assess wines using the quantitative descriptive analysis technique for previous projects.However, the panel attended an additional four-hour training session to ensure reliability and accuracy of the data.e panel was trained in both definition and recognition of appearance, flavor, sweetness, acidity, hotness, viscosity, and aroma and the use of nine-point scales.
e aim of the evaluation and definitions of rating items were clearly explained.e sensory evaluation was performed between 10 a.m. and 12 a.m. with three rounds over three weeks.In each round, three cups of unfermented juice (30 mL) and three cups (30 mL) of fermented juice were served in randomized order for all judges in ISO wine tasting glasses in tasting booths at 25 °C.Remarkably, the unfermented juice was prepared just prior to each round and served at 18 °C.Each bottle of the fermented juice was opened just prior to each round and served at 18 °C.e judges rated the intensity of appearance, flavor, sweetness, acidity, hotness, viscosity, and aroma using a nine-point scale.For the appearance, flavor, hotness, sweetness, and viscosity attributes, these were light, light-medium, medium, medium-full, and full in a ninepoint scale.For all other attributes, the word anchors describing intensity were low, low-medium, medium, medium high, and high in a nine-point scale.e judges rated the items with a cross in the scale.is protocol was repeated for three times over three weeks.e result was the average of the three rounds.

Statistical Analysis.
e data were expressed as average ± standard deviation of at least three repetitions.Analysis of variance was used to compare mean differences of the results.If differences among means were detected, multiple comparisons were performed using Duncan's multiple range test.Directional difference tests were employed to enable direct comparison of the results of the sensory evaluations.All analyses were conducted in SPSS (Windows Version 19).

Characterization of Unfermented and Fermented Seed-Watermelon Juice.
e characterization of both unfermented and fermented juice of seed watermelon is listed in Table 1.Soluble solid content and sugar-to-acid ratio are generally taken as a basic characteristic of juice quality, which show good relationships with consumer acceptability [19].USDA defines the sugar-to-acid ratio between 12 and 20 as the standard of level A of orange juice [20], which means a higher consumer acceptability of the juice.e soluble solid content and sugar-to-acid ratio of the fermented juice were (5.20 ± 0.1) °Brix and (12.4 ± 0.5), respectively.e sugar-to-acid ratio of the fermented juice was laid in the range of level A of orange juice.However, the soluble solid content and sugar-to-acid ratio of the unfermented juice were (2.3 ± 0.1) °Brix and (23.9 ± 0.6), respectively.Consequently, the unfermented juice was plain and unbalanced.Compared to the unfermented juice, the fermented juice was more balance with a light sweetness.
e phenolic content of the fermented juice was significantly higher than that of the unfermented juice.Similar results were also reported for cherry wine [21] and grape wine [22], where the phenolic content is raised after fermentation.
is phenomenon could result from the degradation of hydrolytic tannin and condensed tannin during fermentation [23].

Radical Scavenging Capacities of Unfermented and Fermented Seed-Watermelon Juice.
e DPPH radical is usually defined as chemically relevant radicals.Consequently, the DPPH radical has been used to assess the ability of phenolic compounds to transfer labile H atoms to the radical, which is linearly related to the phenolic content [24].Fermentation significantly raises the DPPH radical scavenging capacity of red grape wine [8] and raises the hydroxyl radical scavenging capacity of pomegranate wine as well [9].
e radical scavenging capacities of unfermented and fermented juice were explored (Table 1).
e DPPH scavenging capacity of the fermented juice was significantly higher than that of the unfermented juice.Remarkably, the DPPH radical scavenging capacity of the fermented juice was higher than that of coconut wine (25.6%) [25] but was lower than that of red grape wine (60%∼70%) [8].Furthermore, the DPPH radical scavenging capacities of the unfermented and fermented juice were further converted to (0.513 ± 0.011) and (0.592 ± 0.014) mg BHT equivalent/mL, respectively.is result indicated that the intake of 100 mL unfermented or fermented juice was equivalent to the intake of BHT of 51.3 or 59.2 mg, respectively.
Hydroxyl and oxygen radicals belong to reactive oxygen species, which are the main reasons leading to an oxidative stress and thereby chronic diseases such as heart disease, cancer, diabetes, and Alzheimer's disease [26][27][28].Hence, both hydroxyl and oxygen radicals are usually defined as the biologically relevant radicals.e IC50 of hydroxyl radical scavenging capacities of the fermented juice was 14.6 ± 0.56%, which was significantly lower than that of the unfermented juice of 28.1 ± 1.02%.Hence, the fermented juice was more effective to scavenge hydroxyl radicals than the unfermented juice.
e ORAC values of the unfermented juice, fermented juice, and fermented juice without ethanol were 56, 60, and 59 µM E/g, respectively.
e fermented juice with or without ethanol presented a stronger oxygen radical scavenging capacity than the unfermented juice.e ORAC value of the fermented juice was about 7.14% higher than that of the unfermented juice.e results were in good agreement with the higher total phenol content of the fermented juice.Moreover, the ORAC value of the fermented juice was higher than that of the fermented juice without ethanol due to the contribution of ethanol [29].
e ethanol in the fermented juice contributed the ORAC value of 1.6%.Furthermore, the ORAC value of our products is comparable to that of other wines and fruits (Table 2).
e ORAC value of the fermented juice was higher than that of coconut wine (27.9 µM TE/g) [25].e ORAC values of both unfermented and fermented juice were higher than that of muskmelon, carrot, and tomato but lower than that of blueberry, apple, avocado, and broccoli [16].
In summary, the radical scavenging capacities of the fermented juice were stronger than that of the unfermented juice for both chemical and biologically relevant radicals.

Volatiles of Unfermented and Fermented
Seed-Watermelon Juice.Volatiles of both unfermented and fermented juice were analyzed via GC-MS analysis.e total ion chromatography of the unfermented and fermented juice is shown in Figure 1. e peak presented at 4.2 min in the total ion chromatography profile of the fermented juice was assigned to ethanol, which constituted 6.3% of the total area.e ethanol content of 0.5%∼0.75%raises the fruity character of the aroma, while a higher content does not [30].Hence, the contribution of ethanol to aroma of the fermented juice was ignored in the following analysis.
A total of 29 volatiles were assigned in the fermented juice with ethyl decanoate, phenylethanol, ethyl laurate, ethyl octanoate, hexanol, and ethyl acetate as main components.Remarkably, the total content of 1-nonanol, 3,6nonadien-1-ol, and nonanal was reduced from 51.68 mg/L in  ese volatiles possibly resulted from the physiological metabolism via the lipoxygenase pathway in the beginning stage or the subsequent fermentation [35].Consequently, the content of the typical watermelon aroma was reduced from 92.08 mg/L in the unfermented juice to 26.41 mg/L in the fermented juice.Furthermore, a total of 14 esters were detected in the fermented juice with ethyl octanoate, ethyl acetate, and ethyl laurate as main components.
e volatiles of both unfermented juice and fermented juice are further classified based on their chemical structure and aroma style in Figure 2. Ester and alkane were main components of the fermented juice, while alcohol was the main component of the unfermented juice.Ester usually shows typical bouquet of grape wine, which is mainly formed during the fermentation [9,22,36].Moreover, the acid and ester of the fermented juice were higher than that of the unfermented juice, while alcohol, aldehyde and ketone, and alkane of the fermented juice were lower than that of the unfermented juice. is phenomenon resulted from the formation of the acids and esters during the fermentation.Remarkably, isoamylol and ethyl octanoate were detected in the fermented juice, which induced the aroma of calvados and brandy.Hence, the fermented juice presented both typical watermelon and wine bouquet aroma.e aroma of these volatiles was further classified as flower, fruit, grass, fragrant, oil, and bouquet based on their characteristic smells [18].e bouquet and fragrant aromas of the fermented juice were stronger than that of the unfermented juice, while the other aroma was weaker than that of the unfermented juice.Combining the distribution of the chemical structure and aroma style of the volatiles, the bouquet of the fermented juice possibly originated from esters, while the fruits and flower mainly originated from alcohol, aldehyde and ketone, and alkane.

Sensory Evaluation of Unfermented and Fermented Seed-Watermelon Juice.
e sensory score of the unfermented and fermented juice was compared in Figure 3. e sensory score rated the intensity of appearance, flavor, sweetness, acidity, hotness, viscosity, and aroma in a nine-point category scale.e "1" in the nine-point category scale represented light or low, while the "9" represented full or high.
Appearance includes properties such as color, reflectivity, transparency, self illumination, and texture [37].e unfermented juice was cloudy with a light yellow, while the fermented juice was transparent with an attractive golden yellow.Moreover, the fermented juice was clearly reflective.
Hence, the appearance score of the fermented juice was significantly higher than that of the unfermented juice, thus being more attractive for consumers.e flavor of the fermented juice was statistically similar to that of the unfermented juice.e sweetness, acidity, and hotness were the main factors influencing the flavor.Specifically, both the sweetness and acidity of the fermented juice were significantly stronger than that of the unfermented juice.ese experiences were in accordance with the results of the soluble solid content and titratable acid content, respectively.Very low level of hotness was present in the fermented juice, which was present in most alcoholic drinks.erefore, the fermented juice would easily be accepted by most consumers.
Ten judges recognized strong watermelon aroma when the unfermented juice was approaching the booth, while 11 judges recognized the watermelon aroma of the fermented juice during the evaluation.Consequently, the aroma intensity of the fermented juice was significantly lower than that of the unfermented juice. is phenomenon was consistent with the content variation of the typical watermelon aroma content.

Conclusion
e DPPH, hydroxyl, and oxygen radical scavenging capacity of the fermented juice were stronger compared to those of the unfermented juice.e ORAC value of the fermented juice was 60 µM TE/g, which was 7.14% higher than that of the unfermented juice.GC-MS analysis showed a total of 26 and 29 volatiles the in unfermented and fermented juice, respectively.e content of the Journal of Food Quality erefore, the fermented seedwatermelon juice showed a strong radical scavenging capacity and presented the typical watermelon aroma with a grape-wine bouquet.
Data Availability e data used to support the ndings of this study are available from the corresponding author upon request.Natural Science Foundation (6172013), Beijing Academy of Agricultural and Forestry Sciences Youth Funds (QNJJ201802), and Beijing Advanced Innovation Center for Food Nutrition and Human Health Open Funds (20181053).

Figure 1 :
Figure 1: Total ionic chromatogram profiles of the unfermented juice (a) and fermented juice (b).

Figure 2 :
Figure 2: Volatile structure (a) and aroma style (b) distribution of the unfermented and fermented juice.

Figure 3 :
Figure 3: Sensory score of the unfermented and fermented juice.◆ is the maximum of the score; • is the minimum of the score; is the mean of the score.e histogram is the distribution of the score except the maximum and the minimum.* and * * represent the signi cant di erence between the mean of the unfermented juice and fermented juice at levels 0.05 and 0.01, respectively.

Table 1 :
Characterization of the unfermented and fermented juice.A significant difference with data in the same line (p < 0.05). *

Table 2 :
ORAC values of some fruits.

Table 3 :
Volatiles of the unfermented and fermented juice.-ol,nonanal, trans-2-nonenal, and trans,cis-2,6-nonadienal, was reduced from 92.08 mg/L in the unfermented juice to 26.41 mg/L in the fermented juice.Furthermore, isoamylol, ethyl octanoate, ethyl acetate, and ethyl laurate contributed the typical grape-wine bouquet to the fermented juice.esensory evaluation further veri ed the results of the characterization.e fermented juice presented an attractive appearance and watermelon aroma.

Table 3 :
Continued., not detected or not found; * a signi cant di erence with data in the same line. -