The extraction efficiency of flavonoids and total phenolics in loquat flowers varied considerably between different organic solvents, and the antioxidant capacities of loquat flowers also differed (Table 1). Antioxidant capacities measured by three different methods were in the following order: ABTS⁺ assay > DPPH assay > FRAP assay. Thaipong et al. [23] measured the antioxidant capacity in methanol and dichloromethane extracts of guava fruit by the above three methods, and also found that the value determined by ABTS⁺ assay was highest, but those determined by DPPH and FRAP assay were similar to each other, though the results using DPPH was slightly higher than that of FRAP. The flavonoids, phenolics and antioxidant capacity of loquat flowers were lower than those of Stachys lavandulifolium flowers, but higher than those of Alcea kardica flowers [24]. Although methanol had the highest extraction efficiency, ethanol was used as extraction solvent in the following experiment due to it providing higher safety and lowed residue than other solvents [25].

Bioactive components and antioxidant capacities may be significantly different among different cultivars [26–29]. Although our previous study showed that the contents of oleanolic acid, ursolic acid and amygdalin in the loquat flower did not vary greatly among different cultivars [22], some differences existed in the contents of flavonoids and phenolics in the loquat flowers of five cultivars tested in the present study, where the highest contents were found in “Dayeyangdun” and the lowest in “Jiajiao”. The average content of flavonoids and phenolics of these five cultivars were 1.59 ± 0.24 and 7.86 ± 0.87 mg/g DW, respectively. Accordingly, differences were also found in the antioxidant capacities of loquat flower ethanol extracts from five cultivars determined by FRAP, DPPH and ABTS⁺ assays, where the highest antioxidant capacity was also found in “Dayeyangdun” and the lowest in “Jiajiao” (Table 2).

The composition of bioactive components and antioxidant capacity in the plant were significantly affected by the developmental stages [30–32]. The effects of developmental stages (Figure 1) on the flavonoids, phenolics contents and antioxidant capacity in loquat flower were similar to that on amygdalin content [22]. Flower at stage 3 contained the highest flavonoid and phenolic content (3.01 ± 0.13 and 13.53 ± 0.38 mg/g DW, respectively) and showed strongest antioxidant capacity with different assays. Flavonoid and phenolic content and antioxidant capacity of flowers at stage 3 were significantly different to those at the other three stages, and were followed by flowers at stage 2. Flowers at stage 1 and stage 4 contained lower flavonoid and phenolic content, and had lower antioxidant capacities (Table 3). Therefore, it is suggested that the best harvesting time for loquat flowers may be between stage 2 and stage 3 for utilization of its flavonoids and phenolics.

The bioactive components and antioxidant capacity in different tissues of the same plant organs vary greatly [30,33,34]. Results showed that petals contained the highest content of flavonoids and phenolics (7.45 ± 0.38 and 19.63 ± 2.72 mg/g DW, respectively), and had highest antioxidant capacities with values of 4.24 ± 0.04, 6.73 ± 0.04 and 7.19 ± 0.17 VCEAC mg/g DW as determined by the FRAP, DPPH and ABTS⁺ assays, respectively. Flavonoids and phenolics contents and antioxidant capacity of petal were significantly different from those of the other three flower tissues, followed by pistil plus stamen. Peduncle and sepal contained lower flavonoid and phenolic content, and had weaker antioxidant capacities (Table 4). These results can explain why the flower at stage 3 had the highest flavonoid and phenolic content and strongest antioxidant capacities.

Many studies showed that the contents of flavonoids and phenolics were positively correlated to the antioxidant capacities in plant extracts [35–37]. Similar results were also obtained in this research (Table 5); correlations between flavonoid and phenolic content and antioxidant capacities obtained from three different assays were positively high (0.694 ≤ r² ≤ 0.785; 0.762 ≤ r² ≤ 0.947), especially among flavonoids, phenolics contents and antioxidant capacities based on ABTS⁺ assay (r² = 0.785**; r² = 0.947**). The lowest correlation was found between flavonoid and phenolic content and antioxidant capacities obtained by FRAP assay (r² = 0.694**; r² = 0.762**). Correlations between phenolics content and antioxidant capacities determined by three different methods were higher than those between flavonoids content and antioxidant capacities. Similar results have also been reported previously [35].

As shown in Figure 2, positive correlation (0.874 ≤ r² ≤ 0.968) was found among antioxidant capacities determined by FRAP, DPPH and ABTS⁺ methods. The highest correlation (r² = 0.968**) existed between antioxidant capacities obtained by FRAP and DPPH assays, while the lowest value (r² = 0.874**) was found between antioxidant capacities obtained by FRAP and ABTS⁺ methods. The high relationship between antioxidant capacities measured by these three different methods was also investigated in other crops. Awika et al. [38] observed that antioxidant capacities in sorghums and their products determined by ABTS⁺ and DPPH assays were highly correlated. In the study of guava, Thaipong et al. [23] also found linear correlations between antioxidant capacities obtained from FRAP, DPPH, ABTS⁺ and ORAC (oxygen radical absorbance capacity) assays, with the highest correlation between FRAP and ABTS assays, and lowest between DPPH and ORAC assays.

Loquat flowers of five cultivars, i.e., Ruantiaobaisha, Dahongpao, Jiajiao, Baozhu, and Dayeyangdun, were collected in winter from the Yuhang District, Hangzhou, Zhejiang Province, China. Flowers at the four developmental stages (i.e., stage 1, stage 2, stage 3, stage 4) and four tissues (i.e., petal, sepal, pistil plus stamen, peduncle) of Ruantioabaisha cultivar were used to study the effect of different developmental stages and flower tissues on flavonoid and phenolic content and antioxidant capacities. All the plant materials were dried in a microwave oven (Panasonic) because of its high efficiency and low cost until the weight was unchangeable, and then ground into powder and stored at −20 °C until analysis.

Plant powder of 0.5 g was solubilized in 20 mL anhydrous ethanol and other organic solvents (AR) purchased from Eage Chemical Reagent Co. Ltd. (Shanghai) for 2 h followed by 30 min ultrasonic extraction by TBT/C-YCL 500Tt/3P (D) ultrasonic machine (Sinobest electronic Co. Ltd., Jining, Shangdong Province, China) under the selected frequency and power (47 kHz and 500 W) combination. The samples were extracted twice and both extracts were combined and evaporated to dryness at 35 °C The residue was dissolved in 1 mL methanol and transferred to an Eppendorf tube, and then centrifuged at 8000 × g for 10 min. After centrifugation, the supernatant was diluted with ethanol for the determination of flavonoids, total phenolic content and antioxidant capacity.

Total flavonoids were determined by the procedure of Zhou et al. [39] with some modification. Aliquots (1 mL) of loquat flower extracts were placed in two test tubes, respectively. 7 mL methanol was added to one tube. In the other tube, 1 mL 2% ZrOCl₂·8H₂O and 6 mL methanol was added. The solution was mixed again and placed into water bath at 30 °C for 1 h. The absorbance was measured at 420 nm with DU-8000 UV-Vis spectrophotometer (Beckman Coulter, USA) and ΔOD was calculated. The amount of total flavonoids was calculated as a rutin equivalent from the standard curve, and expressed as mg rutin/g dry plant material (mg/g DW). All measurements were repeated three times.

Total phenolics were determined by Folin-Ciocalteu procedure [40]. Aliquots (0.5 mL) of loquat flower extracts were transferred into the test tubes and their volumes were made up to 4.5 mL with distilled water. After addition of 0.5 mL Folin-Ciocalteu reagent and 1 mL 7% aqueous sodium carbonate solution, tubes were vortexed and placed into water bath at 30 °C for 1 h. Then absorbance of mixtures was recorded at 760 nm with DU-8000 UV-Vis spectrophotometer (Beckman Coulter, USA) against a blank containing 0.5 mL of extraction solvent. The amount of total phenolics was calculated as a chlorogenic acid equivalent from the standard curve, and expressed as mg chlorogenic acid/g dry plant material (mg/g DW). All measurements were done in triplicate.

The FRAP (ferric reducing antioxidant power) assay was carried out by the methods of Deighton et al. [41] and Kim et al. [42] with some modification. The FRAP reagent was freshly prepared by mixing 25 mL of 300 mmol/L sodium acetate buffer (pH 3.6), 2.5 mL of 10 mmol/L TPTZ (2,4,6-tripyridyl-s-triazine) solution, and 2.5 mL of 20 mmol/L ferric chloride solution. The absorbance at 595 nm was measured 10 min after the mixing of 100 μL of extract (or antioxidant standard) with 900 μL of FRAP reagent, using a spectrophotometer (Beckman Coulter, USA), ethanol was used as control and Vc (Vitamin C) as antioxidant standard. The antioxidant capacity was expressed as Vitamin C equivalent antioxidant capacity (VCEAC, mg/g DW). All determinations were repeated three times.

The DPPH (2,2-diphenyl-1-picrylhydrazyl) assay was measured according to the procedures of Bao et al. [43] and Kim et al. [42] with some modification. The absorbance at 515 nm was measured 6 h after the mixing of 100 μL of extract (or antioxidant standard) with 3.9 mL DPPH (60 μmol/L) solution, using a spectrophotometer (Beckman Coulter, USA), ethanol used as control and Vc as antioxidant standard. The antioxidant capacity was expressed as VCEAC (mg/g DW). All measurements were done in triplicate.

The ABTS [2,2′-azinobis(3-ethylbenzothiazoline-6-sulfonic acid)] assay was measured by the procedures of Delgado-Andrade et al. [44] and Kim et al. [42] with some modification. ABTS⁺ was produced by reaction of 7 mmol/L ABTS stock solution with 2.45 mmol/L potassium persulfate and allowing the mixture to stand in the dark at room temperature for 12–16 h before use. The ABTS⁺ solution (stable for 2 day) was diluted with 5 mmol/L phosphate-buffered saline (pH 7.4) to an absorbance of 0.70 ± 0.02 at 730 nm. The absorbance at 730 nm was investigated 10 min after the mixing of 100 μL of fruit extract (or antioxidant standard) with 3.9 mL ABTS⁺ solution, using a spectrophotometer (Beckman Coulter, USA), ethanol used as control and Vc as antioxidant standard. The antioxidant capacity was expressed as VCEAC (mg/g DW). All measurements were repeated in triplicate.

Standard deviations (SD) were calculated by Origin (Microcal Software Inc., Northampton, MA, USA). Duncan’s new multiple range method test (DPS version 3.11) was calculated for mean separation. Differences were considered significant at P < 0.05.