Antioxidant capacity of phenolics in Camellia nitidissima Chi flowers and their identification by HPLC Triple TOF MS/MS

Camellia nitidissima Chi (CNC) is a valuable medicinal and edible plant in China. In this study, CNC flowers were extracted with 95% ethanol, then partitioned into dichloromethane, ethyl acetate, n-butanol, and water fractions, with the antioxidant capacity of flavonoids and other phytochemicals in CNC flowers investigated for the first time. Results showed that the ethyl acetate fraction exhibited the strongest antioxidant capacity and highest total phenolic content (TPC) compared with the other fractions. Furthermore, in the ethyl acetate fraction, the 50% effective concentrations (EC50) of ABTS+ and DPPH radical scavenging activities were 64.24 ± 1.80 and 78.80 ± 0.34 μg/mL, respectively, and the ferric reducing antioxidant power (FRAP) was 801.49 ± 2.30 μM FeSO4 at 1,000 μg/mL. Pearson’s correlation coefficients and principal component analyses (PCA) for the TPC and antioxidant capacity of the five fractions indicated that the phenolic compounds were the major antioxidant constituents in the flowers. To exploit the antioxidants in CNC flowers, 21 phenolic compounds in the ethanolic extract fraction were identified by HPLC Triple TOF MS/MS, next, 12 flavonoids were isolated and elucidated, of which compounds 1–5 showed potent antioxidant capacity. In addition, the potential structure-activity relationship among these 12 flavonoids showed that (1) the o-catechol group in the B-ring was primarily responsible for the antioxidant capacity of flavonoids and (2) steric hindrance, produced by glycosides and other groups, could reduce the antioxidant capacity of the flavonoids.


Determination of total phenolic content (TPC)
Total phenolic content (TPC) was determined by the Folin-Ciocalteu method [11]. Briefly, 100 μL of the five fractions at suitable concentrations, 1.15 mL of deionized water, 0.25 mL of Folin-Ciocalteu's reagent, and 1 mL of 7.5% sodium carbonate solution were mixed. Samples were placed in darkness at room temperature for 1 h after vortexing. Absorbance was read at 760 nm using a spectrophotometer (Thermo Electron Corp., Waltham, MA, USA). Results were expressed as milligrams of gallic acid equivalent per gram of material (mg GAE/g material).

Measurements of antioxidant activity
ABTS radical cation scavenging activity assay. The ABTS radical cation scavenging activity of the samples were examined in accordance with an earlier study [13], with minor modifications. To generate ABTS + , ABTS (7 mM) and potassium persulfate (2.45 mM) were incubated in the dark at room temperature for 16 h. The freshly prepared ABTS + solution was diluted with ethanol to obtain an absorbance at 734 nm of 0.70 ± 0.02 before the test. Approximately, 100 μL of each sample at different concentrations was added to 400 μL of the ABTS + solution and adequately mixed. The concentrations of the fractions were 0, 25, 50, 100, 200, 400, 600, 800, and 1,000 μg/mL, and of the compounds were 0, 2.5, 5, 12.5, 25, 50, 100, 200, and 400 μg/mL. The reactive mixture was placed in the dark at room temperature for 6 min. Absorbance was then recorded at 734 nm. The ABTS + scavenging activity was calculated as follows, ABTS + scavenging activity (%) = [1 -A sample /A control ] × 100, where A sample is the absorbance in the presence of the sample and A control is the absorbance of the blank without the test sample. The ABTS + scavenging activity of Vc was assayed for positive control.
Determination of DPPH radical scavenging activity. The DPPH radical scavenging activity of the samples was determined following [14], with minor modifications. Briefly, 400 μL of each sample at different concentrations was added to 400 μL of DPPH solution (0.4 mM). The concentrations of the fractions were 0, 25,75,100,125,150,200, and 400 μg/mL, and of the compounds were 0, 2.5, 5, 12.5, 25, 50, 100, 200, and 400 μg/mL. The mixture was shaken immediately and incubated in the dark at room temperature for 30 min. Absorbance was recorded at 517 nm. The DPPH radical scavenging activity was calculated as follows: DPPH radical scavenging activity (%) = [1 -A sample /A control ] × 100, where A sample is the absorbance in the presence of the sample and A control is the absorbance of the blank without the fraction. The DPPH radical scavenging activity of Vc was used as a positive control.
Evaluation of ferric reducing antioxidant power (FRAP). The ferric reducing antioxidant power (FRAP) of the samples was evaluated according to previous research [15], with some modifications. Briefly, 10 mL of TPTZ solution (10 mM, in 40 mM HCl), 100 mL of acetate buffer (0.3 M, pH 3.6), and 10 mL of ferric chloride (20 mM) were mixed to prepare fresh FRAP working solution, which was warmed at 37˚C prior to testing. We added 200 μL of each sample at different concentrations to the FRAP solution (1 mL), with the mixture then placed in a 37˚C water bath for 20 min. The concentrations of the fractions were 0, 25, 100, 200, 400, 600, 800, and 1,000 μg/mL, and of the compounds were 0, 2.5, 5, 12.5, 25, 50, 100, 200, and 400 μg/mL. Absorbance was read at 593 nm. Different concentrations (10-1,600 μg/mL) of ferrous sulfate were used to prepare a standard curve. Results were expressed as μM Fe (II). FRAP of Vc was also used as a positive control.
Statistical analyses. All experiments were independently conducted in triplicate, and experimental results were expressed as means ± standard deviations or average. One-way analysis of variance (ANOVA) and Duncan's multiple range tests were performed using SPSS version 17.0 (SPSS Inc., Chicago, IL, USA) software. Statistical significance was determined at p < 0.05. Interpolation from linear regression analysis was used to obtain the EC 50 . In order to interpret the relationships between antioxidant activity and total phenolic contents, two-tailed Pearson's correlation coefficient analysis and principal component analysis (PCA) were conducted using SPSS version 17.0 (SPSS Inc., Chicago, IL, USA) software.

Total phenolic content (TPC)
C. nitidissima Chi flowers are used as a popular tea in China. As the main phytochemicals of tea, phenolic compounds play an important role in biological activities [10]. As shown in Fig  1, the CNC flower fractions contained many phenolic compounds. The TPC of the ethyl acetate fraction was highest and that of the water fraction was lowest (345.14 ± 4.05 and 31.69 ± 1.75 mg GAE/g, respectively) among all fractions. The TPC of the n-butanol fraction was 164.19 ± 3.18 mg GAE/g, similar to that of the ethanolic extract (170.74 ± 1.99 mg GAE/g). The TPC of the dichloromethane fraction (85.02 ± 0.88 mg GAE/g) was significantly lower than that of the ethyl acetate fraction, n-butanol fraction, and ethanolic extract [11]. These results indicate that phenolic compounds in this species can be solubilized in medium polar solvents, such as water-saturated ethyl acetate [12,16].

Ethanolic extract analyses by HPLC Triple TOF MS/MS
Twenty-one phenolic compounds in the ethanolic extract of the CNC flower were identified ( Table 2) by HPLC Triple TOF MS/MS analysis (Fig 3, S1 Fig).
The extract ion chromatogram at m/z 289.0712 showed two peaks at RT 4.60 and 5.91 min. These two peaks showed fragments at m/z 245, 205, 203, and 137 (  [24]. Thus, the peak at 4.60 min was assigned to catechin and at 5.91 min was assigned to (epi)catechin [25,26]. The extract ion chromatogram at m/z 577.1352 showed a peak at RT 4.29 min, which produced fragments at m/z 451, 425, 407, and 289 (  [24]. The extract ion chromatogram at m/z 441.0827 showed a peak at RT 8.14 min, which produced fragment ions at m/z 289 and 169 (Table 2) corresponding to the deprotonated ions of (epi)catechin and gallic acid, respectively; thus, the compound was identified as (epi)catechin-gallate [24]. The extract ion chromatogram at m/z 729.1461 showed a peak at RT 6.49 min, which produced a fragment ion at m/z 577 through the loss of one galloyl group and fragments at m/z 559, 441, and 407 ( Table 2); thus, the compound was identified as procyanidin-gallate [24]. The extract ion chromatogram at m/z 457.0776 showed a peak at RT 5.99 min, producing fragments at m/z 305 and 169 ( Table 2) corresponding to the deprotonated ions of gallocatechin and gallic acid, respectively, hence it was assigned as gallocatechin-gallate [24]. These results showed that the CNC flowers were rich in catechins and their derivatives, which are regarded as effective antioxidants due to their ability to scavenge ROS [27]. In addition, it has been reported that catechin, (epi)catechin, catechin dimer, catechin-gallate, procyanidin-gallate, and gallocatechingallate are all antioxidants that contribute to beneficial effects on human health [28][29][30][31]. The extract ion chromatogram at m/z 301.0354 showed a peak at RT 12.17 min, with the fragments at m/z 273, 255, 179, and 151 ( Table 2) corresponding to the loss of CO, CH 2 O 2 , C 7 H 6 O 2 , and C 8 H 6 O 3 , respectively. The compound was identified as quercetin [32].  Table 2), and were thus identified as isoquercitrin [34], rutin [35], and quercetin-glucosyl-rhamnosyl-glucoside [24], respectively. Previous research has shown that quercetin and quercetin glycosides, such as isoquercitrin and rutin, exhibit antioxidant ability in many teas and foods [36][37][38].
The extract ion chromatogram at m/z 479.0831 showed a peak at RT 7.36 min, and fragments at m/z 317 (Y 0 -) and 316 (Y 0 -1) [33] ( Table 2) corresponding to the loss of 162 and 163 Da, which is consistent with the cleavage of a hexosyl group. The compound was therefore identified as myricitrin-glucoside [24], which is regarded as an antioxidant [42]. The extract ion chromatogram at m/z 563.1396 showed a peak at RT 7.31 min and fragments at m/z 545, 503, 473, 443, 383, and 353 ( Table 2). This compound was identified as apigenin-pentosyl-glucoside [43], a flavone glycoside with known antioxidant activity [44]. The extract ion chromatogram at m/z 431.0976 showed a peak at RT 8.21 min and fragments at m/z 311 and 341 ( Table 2). This compound was assigned to vitexin [45], which is regarded as a good antioxidant [46].
The extract ion chromatogram at m/z 315.05133 showed a peak at RT 14.21 min, and fragments at m/z 201 and 229 ( Table 2). This compound was identified as pollenitin [47], a phenolic compound with good antioxidant activity [48]. The extract ion chromatogram at m/z 169.0143 demonstrated a peak at RT 1.42 min. The peak displayed a fragment at m/z 125 ( Table 2) corresponding to the loss of one CO 2 . Thus, it was identified as gallic acid [24], a known antioxidant [49]. ( Table 2); as such, this compound was identified as syringin [50], which plays an antioxidant role in some plants [51].

Antioxidant activity
Antioxidant activity is influenced by many factors, and a single antioxidant property model cannot fully reflect the antioxidant capacity of all samples [15]. Therefore, more than one antioxidant activity measurement was performed to consider the various mechanisms of antioxidant action. In this study, we carried out three antioxidant models to reflect the antioxidant capacity of CNC flowers: ABTS radical cation scavenging activity, DPPH radical scavenging activity, and ferric reducing antioxidant power (FRAP).

ABTS radical cation scavenging activity
Results showed that all five fractions exhibited scavenging activity for the ABTS radical cation in concentration-dependent manners (Fig 4A) and the differences between the fractions were significant (p < 0.05). The ethyl acetate fraction, with an EC 50 of 64.24 ± 1.80 μg/mL (Table 3), exhibited significantly higher (p < 0.05) ABTS radical cation scavenging activity than that of the other four fractions, the EC 50 values were 137.40 ± 4.61, 363.90 ± 1.51, and 127.46 ± 5.00 μg/mL, for ethanolic extract dichloromethane and n-butanol fraction while the EC 50 of water fraction was not detected. Interestingly, the trend of the scavenging activity for the ABTS was consistent with the TPC. So the results indicated that the phenolic compounds in the CNC flower played a vital role in scavenging ABTS radical cations.

DPPH radical scavenging activity
The DPPH radical scavenging activity results were shown in Fig 5A. These results were similar to those of the ABTS radical cation radical scavenging activity. The efficacies were concentration-dependent, and the ethyl acetate fraction showed the highest DPPH radical scavenging activity. The EC 50 value of the ethyl acetate fraction was 78.80 ± 0.34 μg/mL, and the values of the ethanolic extract and n-butanol fraction were 142.60 ± 1.46 and 162.60 ± 2.33 μg/mL, respectively. The EC 50 values the of dichloromethane and water fractions were not detected.
Thus, the order of DPPH radical scavenging activity of the five fractions was ethyl acetate fraction > ethanolic extract > n-butanol fraction > dichloromethane fraction > water fraction, and was consistent with the TPC results, which suggested that the phenolic compounds were the main bioactive components in the scavenging of DPPH radicals in the CNC flowers. It has been reported in various studies that higher TPC can lead to significant increases in DPPH radical scavenging activity [8,13,52].

Ferric reducing antioxidant power (FRAP)
Results (Fig 6A) showed that the FRAP of the CNC flower fractions increased with their dosage, which was consistent with the ABTS + and DPPH radical scavenging activity. Among the five fractions, the ethyl acetate and water fractions showed the highest and lowest FRAP, respectively, at concentrations ranging from 25 μg/mL to 1,000 μg/mL and FRAP values ranging from 93. 49 [53,54].

Correlation analysis between antioxidant capacity and the total phenolic content (TPC)
Previous studies have reported that the higher phenolic content in the extracts resulted in a higher antioxidant activity, which was in agreement with a positive correlation between TPC   and antioxidant activity [1,15]. So in order to obtain the detail correlations between antioxidant capacity and the TPC of the five fractions, the correlation analyses were conducted. As shown in Table 4, the significant correlations (p < 0.01) between the antioxidant properties and TPC of the five fractions were found. The TPC was highly associated with scavenging ability against ABTS (r = 0.890, 0.983, 0.745, 0.859, and 0.992, for ethanolic extract, dichloromethane fraction, ethyl acetate fraction, n-butanol fraction, and water fraction, respectively), DPPH (r = 0.979, 0.897, 0.893, 0.973, and 0.694, for the five fractions, respectively) and FRAP (r = 0.946, 0.991, 0.823, 0.933, and 0.995, for the five fractions, respectively). So the data indicated that the phenolic compounds in the five fractions of CNC flowers were considered responsible for effective antioxidant properties.

Principal component analysis (PCA)
To investigate the interrelationships between the different variables and to find the optimum number of extracted principal components, principal component analysis (PCA) was applied to reduce the original variables (TPC, ABTS, DPPH, and FRAP) in a smaller number of underlying variables (principal component) [15]. The principal component analysis (PCA) and their correlations were shown in Fig 7 and Table 5. Two principal components together of the ethanolic extract, dichloromethane fraction, ethyl acetate fraction, n-butanol fraction, and water fraction were 99.4%, 98.5%, 99.5%, 99.3%, and 98.0%, respectively. The first principal component (PC1) correlated well with TPC, ABTS, DPPH, and FRAP. In addition, TPC, ABTS, DPPH, and FRAP were significantly correlated with each other in the five fractions of CNC flowers. So the strong correlations among TPC, ABTS, DPPH, and FRAP suggested that the contents of phenolic compounds and antioxidant properties were highly correlated with each other. https://doi.org/10.1371/journal.pone.0195508.g006

Antioxidant activity of 12 flavonoids and their potential structure-activity relationship
To evaluate the antioxidant capacity of 12 flavonoids isolated from CNC flowers, ABTS radical cation scavenging activity, DPPH radical scavenging activity, and ferric reducing antioxidant power (FRAP) analyses were conducted. As seen in Figs 4-6, significant differences in antioxidant capacity were investigated among the 12 flavonoids. As shown in Table 3, the EC 50 values of compounds 1 and 3 were significantly lower than that of compound 5 in DPPH and ABTS + radical scavenging activities, with the FRAP results similar to the other two models. This indicated that the antioxidant capacity of flavonoids of different classes was positively correlated with the number of hydroxyl groups. From further analysis, catechin and quercetin both possess the o-catechol group in the B-ring, whereas kaempferol only possesses one hydroxyl group in the B-ring; thus, the o-catechol group is considered the major group for the antioxidant capacity. In addition, the results proved that 2,3-double bond in conjugation with the 4-oxo group in the C ring is not a determinant structural feature for the antioxidant capacity of flavonoids [55].
The antioxidant capacity results of compounds 1 and 2 (Figs 4B, 5B and 6B) showed that although compound 2 had more hydroxyl groups, both compounds had the same antioxidant capacity because compound 2 had a large substituent with steric hindrance that reduced activity [56,57]. As seen in Figs 4C, 5C and 6C, there were significant differences (p < 0.05) between the antioxidant capacity of compounds 3 and 4. The EC 50 values of the DPPH and ABTS + radical scavenging activities of compound 3 were significantly (p < 0.05) lower than those of compound 4 (Table 3). Similarly, the antioxidant capacity of compound 5 was significantly (p < 0.05) stronger than that of compounds 6-12 (Figs 4D, 5D and 6D), which were all kaempferol glycosides, with the antioxidant capacity decreasing with increasing number of  glycosides ( Table 3). Our results indicated that glycosides could reduce the antioxidant capacity of flavonoids, such as quercetin and kaempferol, due to their production of steric hindrance [56].