Comparison on chemical compositions and antioxidant capacities of the green, oolong, and red tea from blueberry leaves

Abstract Blueberry leaves, by‐products of the blueberry industry, could be explored as source of functional foods, such as teas. Three different types of tea, including nonfermented green tea, semifermented oolong tea, and fully fermented red tea from blueberry leaves, were investigated on their chemical compositions and antioxidant capacities here. The contents of individual amino acids in three types varied, while the total amounts retained constant. A total of 167 volatiles were detected with alcohols, alkenes, and aldehydes as the dominant. More volatiles produced in the fermented teas. The total phenolic/flavonoid contents were highest in the green tea and decreased significantly in the oolong and red teas, correlating inversely with the fermentation degree. The highest levels of representative phenolics, that is, phenolic acids and flavonol glycosides, contributed to the strongest antioxidant capacity in the green tea. These indicated that blueberry leaves provided promising and prospective potential to develop new teas beneficial for health.

Manufacturers had made blueberry leaves available in the form as "herbal tea" (Gallaher, Gallaher, Marshall, & Marshall, 2006); however, they were actually the simple water infusions of the leaves without any conventional tea processing steps. The main types of tea are classified into green, white, yellow, oolong, red, and pu-erh teas based on different manufacturing processes (Yi et al., 2015).
Among them, the most well-known and commonly found in the market are unfermented green tea, semifermented oolong tea, and fully fermented red tea. Lacking the relevant study, the exploration of different types of tea from blueberry leaves is promising, and the chemical composition's changes during the manufacturing process of teas are worth well elucidating.
In the present study, three different types of tea samples, including minimally, partially, and maximally oxidized blueberry leaf teas, were attempted to produce, respectively, according to the making process of the green, oolong, and red tea from C. sinensis. These three types of teas were thus named as green tea (GT), oolong tea (OT), and red tea (RT) from blueberry leaves hereafter. The changes of main chemical compositions and antioxidant capacities during the processing of different blueberry leaf teas were evaluated and compared. Various profiles, such as amino acids, volatile aroma components, and phenolic compounds/flavonoids/proanthocyanidin, were further investigated on the development of tea quality. Meanwhile, the antioxidant capacities in vitro of blueberry leaf teas were comprehensively assessed through four common antioxidant activity methods. This study would contribute to current knowledge on the comprehensive utilization of the waste blueberry leaves as potential valuable resource, and also the systematically development of novel tea varieties.

| Plant materials and tea sample preparation
Fresh leaves of Garden Blue rabbiteye blueberry (Vaccinium ashei) were collected from the orchards of Nanjing Botanical Garden Memorial Sun Yat-Sen, Jiangsu, China, in August 2017. The shoots with one bud and several leaves were handpicked from the top of the newly grown branches and then brought back to the laboratory immediately for tea preparation. Three different types of tea samples were manufactured, including nonfermented GT, semifermented OT, and fully fermented RT. The manufacturing process of the GT samples was on the basis of corresponding traditional tea processing method of C. sinensis. The blueberry leaves were fixed by microwave at 500 W for 60 s with leaf quantity of 2 g/cm 2 after withering at 30°C and a relative humidity (RH) of 60% for 3 hr. Then, the leaves were tossed and rolled by hand gently for 30 min before they were dried at 70°C through hot air to a final moisture content of less than 4%. For the OT samples, the blueberry leaves were turned over at 25°C and a RH of 80% for 5 min using a rotary shaker for five times with 40-min intervals after withering. Afterward, the fermented samples were fixed by microwave, rolled at room temperature, and dried at 70°C to produce the OT. Similarly, the blueberry leaf RT samples underwent the process of withering, rolling, fermentation, and drying. Fermentation was performed at 28-30°C and a RH of 80% for 15-20 hr, and then terminated by hot air drying to a moisture content of less than 4%. The dried tea samples were ground to fine powder by a pulverizer and sifted through a 40-mesh sieve. The powder was stored at −20°C in a refrigerator for subsequent analysis.

| Preparation of extracts
The extracts were prepared according to the reported method with some modifications (Yang et al., 2009). The dried powder of different tea samples was extracted by 85% methanol solution containing 0.5% formic acid for three times. The extracts were stored at 4°C prior to further assay.

| Gas chromatography-mass spectrometer (GC-MS)
Separation and identification of the volatiles were conducted on an Agilent 7890A-5975C GC-MS system (Agilent Technologies) equipped with a DB-WAX capillary column (0.25 mm × 0.25 μm × 50 m). Helium was used as the carrier gas, with a rate of 1.5 ml/min. The oven temperature was programmed from 50°C (retained for 1 min) to 100°C (retained for 5 min) at a rate of 5°C/min, ramped up to 140°C (retained for 10 min) at a rate of 4°C/min, raised to 180°C (retained for 10 min) at a rate of 4°C/min, and finally increased to 250°C (retained for 5 min) at a rate of 4°C/min. Samples were injected into the GC injection port held at 250°C. MS was operated in full scan mode (mass range, m/z 50-550) with ionization voltage at 70 eV, and ion source temperature at 230°C. A library search was carried out using the in-house database of NIST11.L.

| Composition analysis
The Folin-Ciocalteu colorimetric method was employed to determine the total phenolic content (TPC) according to the previous report (Huang, Zhang, Liu, & Li, 2012). TPC values were expressed as gallic acid equivalent (GAE), that is, mg GAE/g dry weight (DW).
The total flavonoid content (TFC) was determined by a colorimetric method. The TFC values were expressed as rutin equivalents (RTE), that is, mg RTE/g DW. The vanillin-hydrochloric acid method was conducted for the determination of the proanthocyanidin content (PAC) (Nakamura, Tsuji, & Tonogai, 2003). The PAC was expressed as catechin equivalents (CTE), that is, mg CTE/g DW.

| Assay of antioxidant activity in vitro
The scavenging activity for DPPH radical was estimated through spectrophotometric method (Xiao et al., 2014). The scavenging activity for ABTS radical cation (ABTSˑ + ) was performed according to the previous method (Arts, Dallinga, Voss, Haenen, & Bast, 2004).
The DPPH and ABTS results were expressed in terms of trolox equivalent antioxidant capacity (TEAC), that is, µmol TEAC/g DW.
The ferric reducing antioxidant power (FRAP) assay was conducted based on the method of Razak, Rashid, Jamaluddin, Sharifudin, and Long (2015). The FRAP results were expressed as Fe (II) equivalent antioxidant capacity (FEAC), that is, μmol Fe (II)/g DW. The oxygen radical absorbance capacity (ORAC) assay was carried out to assess the total antioxidant activity (Li, Huang, Wang, & Liu, 2013
The phenolic compounds were detected at 280, 320, and 360 nm, and identified by comparing both retention times (t R ) and ultraviolet (UV) spectra with authentic standards. Ten phenolics were further quantified, expressed as milligram of each compound per gram of the dried weight, that is, mg/g DW.

| HPLC-ESI-MS analysis
HPLC-ESI-MS analysis was carried out on an Agilent 1100 HPLC/ MS(SL) system equipped with a UV detector and a LC-MSD Trap VL ion-trap mass spectrometer (MS) via an electrospray ionization (ESI) interface (Agilent Technologies). HPLC separation conditions were the same as described above. For MS conditions, the ESI capillary voltage was 3.0 kV in negative ion (NI) mode with the capillary temperature at 350°C. A nebulizing gas of 1.5 L/min and a drying gas of 10 L/min were applied for ionization using nitrogen (N 2 ). ESI was performed with the scan range at m/z 150-1000.

| Statistical analysis
All determinations were conducted in triplicate, and the results were calculated as mean value ± standard deviation (SD). The fluorescence decay curve and figure data in ORAC assay were obtained using GraphPad Prism Version 5.02 (GraphPad Software, Inc., CA, USA).
One-way analysis of variance (ANOVA) with Tukey's post hoc test was used to determine statistical differences among 3 tea groups and the tested indices. Differences were considered significant with p < .05.

| Amino acid content in blueberry leaf teas
The amino acid contents in blueberry leaf GT, OT, and RT are listed in Table 1. The GT and RT showed the high total amounts of amino acid at 28.734 and 28.535 mg/g DW, respectively, followed by the OT at 27.608 mg/g. The teas from blueberry leaves here revealed an elevated total amounts of amino acid than existing commercial teas from leaves of the plant C. sinensis L, which was reported previously to range from 13.2 to 20.2 mg/g (Horanni & Engelhardt, 2013).
The contents of individual amino acids varied in the GT, OT, and RT samples. Notably, proline (Pro) was the most abundant amino acid among the three teas. Its content in the GT (6.137 mg/g) was higher than that of the OT and RT (5.889 and 5.548 mg/g, respectively), indicating an inverse correlation with the degree of fermentation. Except for Pro, aspartic acid (Asp), glutamic acid (Glu), serine (Ser), glycine (Gly), arginine (Arg), alanine (Ala), tyrosine (Tyr), valine (Val), phenylalanine (Phe), leucine (Leu), and isoleucine (Ile) were the major amino acids, accounting for approximately 70.64%-73.20% of the total amino acids and 1.96%-2.09% of dry weight. It has been reported that most of the above amino acids contribute to the sweet or bitter taste of tea (Kirimura, Shimizu, Kimizuka, Ninomiya, & Katsuya, 1969).
The content and composition of amino acids were recognized to be closely related to the sensory property and consequently the quality of tea (Alcazar et al., 2007). Some amino acids are the main contributors for the umami taste of tea, especially Glu (Kaneko, Kumazawa, Masuda, Henze, & Hofmann, 2006). The content of Glu, representing 1.253 mg/g dry weight, was found to be the lowest in the GT. The OT and RT contained higher Glu contents at 1.390 and 1.346 mg/g, respectively. This demonstrated that fermentation process could increase the intensity of umami taste. In addition, the Ser and Phe levels in the OT and RT were higher than that in the GT. On the contrary, higher amounts of histidine (His), Gly, Arg, and methionine (Met) were present in the nonfermented GT. A clear corresponding relationship between their contents and fermentation process could be observed. During the fermented process, protein breakdown, polymerization, or transformation could occur, which led to great changes on the content and composition of amino acids (Horanni & Engelhardt, 2013).

| Volatile compositions of blueberry leaf teas
The volatile compositions were found to vary significantly in the three different tea types, both qualitatively and quantitatively. A total of 167 volatile compounds were structurally identified in the three tea samples from blueberry leaves, 98 of which with integration area percentage > 0.2% and match quality > 40 are listed in Table 2 The volatile compounds identified mainly belonged to eight chemical categories, including alcohols (20), aldehydes (18), esters (5), acids (7), ketones (10), alkenes (18), heterocyclic compounds (10), and linear TA B L E 1 Content of amino acids (mg/g on dry weight basis) in the different tea samples from blueberry leaves were dominant (Figure 1). In addition, the proportions of these eight chemical categories changed in different tea samples. There tended to contain the most amount of alcohols (31.0145%) and alkenes (28.8934%) in the GT. Obviously, (±)-3,7-dimethyl-1,6-octadien-3-ol (11.1936%) and (+)-dipentene (9.5143%) were the most abundant compounds in the GT, which might be the major contributors for the green tea flavor. The major components in the RT were 2-nonanone (6.3884%), n-hexane (5.7039%), and decamethylcyclopentasiloxane (5.5769%). The aldehyde substances retained substantially in the three different tea samples after fermentation, with the proportion ranging from 15.6045% to 16.5481%.
Alpha-terpineol (6.2882%-7.9837%) was also observed as a main volatile with relatively consistent contents in all the tea samples. The proportions of alcohols and alkenes were observed to decrease in the OT and RT in accordance with the degree of fermentation process. On the contrary, the proportion of ketones, esters, heterocyclic compounds, and hydrocarbons increased in the OT and RT. This revealed that these four chemical categories were generated and released gradually during fermentation, which were accountable for the fermented tea flavor.
Different volatiles contributed to different flavors. For example, beta-ionone, linalool, and benzeneacetaldehyde were contributors for sweet floral aroma and honey-like flavor, respectively (Ho, Zheng, & Li, 2015). The contents of these compounds showed an increasing trend from the GT to the fermented OT and RT, which demonstrated that the fermentation promote the formation of delightful flavor. On the contrary, the nonanal, imparting inferior aroma, reduced significantly in the fermented tea samples (Xu et al., 2018).
There was a great difference on the volatile compositions of the teas from blueberry leaves in comparison with that of the commercial teas. Indole, alpha-farnesene, and (E)-nerolidol were reported as the characteristic volatiles and major contributors for the aroma of teas made from the leaves of C. sinensis L (Gui et al., 2015). However, all these components were found at trace level in the blueberry leaf teas in this study due to species differences.

| TPC, TFC, and PAC in blueberry leaf teas
As shown in Figure 2,  were identified as one of the major components in the leaves of rabbiteye blueberry (Matsuo et al., 2010). In this study, all the three teas had significantly (p < .001) larger amounts of proanthocyanidin than the fresh blueberry leaves (27.64 ± 5.41 mg CTE/g DW) (our unpublished data).
According to previous reports, the proanthocyanidins with mean polymerization degree of 3.1 from the leaves of rabbiteye blueberry could inhibit the proliferation of human T-cell lymphotropic virus type 1-associated cell lines (Nagahama et al., 2014), and the ones with polymerization degree of 8-9 have obvious potential to suppress the expression of hepatitis C virus RNA (Takeshita et al., 2009). Similarly, large amounts of procyanidin dimers and trimers were also detected in the leaf samples of other berry fruits, including lingonberry (Vaccinium vitis-idaea), hawthorn (Crataegus spp.), and saskatoon (Amelanchier alnifolia) (Tian et al., 2018).

| Antioxidant activity of blueberry leaf teas
The results of DPPH, ABTS, FRAP, and ORAC assays are shown in  F I G U R E 1 Volatile categories of different tea samples from blueberry leaves F I G U R E 2 Total phenolic content (TPC), total flavonoid content (TFC), and proanthocyanidin content (PAC) of the different tea samples from blueberry leaves Among these in vitro antioxidant models, the DPPH and FRAP assays were normally used to evaluate antioxidants in lipophilic and hydrophilic system, respectively, while the ABTS assay used in both (Lee, Seo, Lim, & Cho, 2011). The ORAC assay is a reliable technique that combines the inhibition percentage of several varieties of reactive oxygen species of biologically relevant source with time (Prior & Cao, 1999). Therefore, ORAC is largely utilized to assess the total antioxidant capacity. It could also be accounted for the results that ORAC values were much higher than the DPPH/ABTS/FRAP values.
The results in vitro indicated that the blueberry leaf GT, OT, and RT all could be potent sources of antioxidants. Their antioxidant activities decreased in the following order: GT > OT>RT, which were highly correlated with their fermentation degrees. Previous researchers had reported that there was a direct and positive relationship between the antioxidant capacity and the total phenolic content (Lin et al., 2016). Hence, phenolic compounds in the blueberry leaf teas could serve as a major contributor to their antioxidant capacity, and this was in agreement with the result of TPC measurements.
The same as TPC, the antioxidant activities of different tea samples were also dependent on their corresponding fermentation degree.
Fermentation process led to significantly decreased antioxidant activities. Previous studies had also reported that unfermented tea possessed higher antioxidant activity than fermented tea (Satoh, Tohyama, & Nishimura, 2005).

| Phenolic constituents of blueberry leaf teas
Eleven phenolic compounds were differentiated and identified in blueberry leaf tea samples, ten of which were quantified ( The typical chromatograms of the GT, OT, and RT showed visually distinct differences (Figure 4). The highest contents of most phenolic compounds were detected in the GT, for example, 114.206 ± 6.751 mg 3-caffeoylquinic acid/g DW and 8.357 ± 0.885 mg 4,5-dicaffeoylquinic acid/g DW, respectively. Similar to the total phenolic content measurements, a significant decrease (p < .001) was observed in the contents of the identified phenolic compounds with relevance to the fermentation degree of the tea samples. During the fermentation process for making oolong tea and red tea, phenolic compounds were transformed to various oxidation products in varying degrees by the endogenous enzymes, mainly polyphenol oxidase and peroxidase from fermentative microorganisms (Tan, Engelhardt, Lin, Kaiser, & Maiwald, 2017). Accompanying the oxidation and conversion of phenolics accomplished, the phenolic compounds measured consequently achieved a considerable decrease in the concentrations.
Similar results were found in unfermented and fermented Sri Lankan tea (Jayasekera et al., 2014). This was also in consistence with the total phenolic content measurements.
Since the phenolics were considered as the most significant contributor to the total antioxidant capacity of tea (Ehlenfeldt & Prior, 2001), the high content of the major phenolics in the GT, specifically chlorogenic acid, could be responsible for its strong free radical scavenging activities and total antioxidant capacity.
On the contrary, fermented tea, that is, the OT and RT, with lower levels in both total and individual content of phenolic compounds, showed relatively decreased antioxidant capacities, which was in agreement with previous studies (Satoh et al., 2005). In addition, phenolic compounds were also particularly important to the color, taste (especially astringency), and overall quality of teas (Shih et al., 2017). Different content of phenolic compounds could be accounted for various sensory characteristics of teas. For example, the high level of the flavonol glucosides in the GT was reported to induce a silky, mouth-drying, and mouth-coating sensation (Wu, Xu, Héritier, & Andlauer, 2012).
Similar to the volatile compositions and amino acid contents, the blueberry leaf teas also showed great differences on the phenolic constituents in comparison with the commercial teas made from the leaves of C. sinensis L, which mainly included catechins and their gallates in green tea, and dimeric theaflavins and polymeric thearubigins in fermented oolong and red teas (Menet, Sang, Yang, Ho, & Rosen, 2004). However, these components were not found in the blueberry leaf teas in this study. Instead, large amount of phenolic acids and flavonols were detected. Plenty of epidemiological studies had indicated that phenolics, such as chlorogenic acid, or flavonols like quercetin, were considered having health-promoting or chronic disease-preventing properties (Cai, Sun, Xing, Luo, & Corke, 2006). These functional phenolic compounds rich in blueberry leaf teas are water-soluble, to develop tea products with health benefits from blueberry leaves would be practical and significant.

| CON CLUS IONS
The nonfermented green tea, semifermented oolong tea, and fully fermented red tea were manufactured from the fresh blue-