Twelve batches of Fupenzi samples were collected from different regions of China (1-Hangzhou (Zhejiang), 2-Jinhua (Zhejiang), 3-Zhengzhou (Henna), 4-Nanjing (Jiangsu), 5-Guangzhou (Guangdong), 6-Guilin (Guangxi), 7-Nanning (Guangxi), 8-Chengdu (Sichuang), 9-Fuzhou (Fujian), 10-Zhangzhou (Fujian), 11-Changchun (Jilin), 12-shengyang (Jilin)). Rutin, hyperin, quercetin and kaempferol were purchased from National Institute for the Control of Pharmaceutical and Biological Products ( Beijing , China ). DPPH (2,2-diphenyl-1-picrylhy- drazyl) and ABTS (2,2’-azinobis[3-ethylbenzothiazoline-6-sulphonic acid] diammonium salt) were purchased from Sigma chemical Co. ( Sigma , USA ). Acetonitrile and methanol were of HPLC grade (Shield Fine Chemicals Company, Tianjin , China ). Water for HPLC analysis was purified by a Milli-Q water purification system ( Shanghai , China ). Other reagents were all of analytical grade.

Grind the fruit samples of Fupenzi into powder by use of a pestle and mortar. 3.0 g pulverized powder was accurately weighed and ultrasonically extracted with 40 mL of 95% ethanol for 50 min in a calibrated flask at 40˚C, and then filtered through a 0.22 μm membrane. The filtrate was evaporated to dryness at 35˚C and the residue was dissolved into 10 mL methanol. The solution was filtrated through a syringe filter (0.22 μm) and aliquots (20 μL) were subjected to HPLC before analysis.

A standard stock solution containing the five components (rutin, hyperin, quercetin and kaempferol) was prepared in methanol and stored away from light at 4˚C. Working standard solutions containing the four compounds were prepared by appropriate dilution of the stock solution.

The total flavonoid content (TFC) was determined using a colorimetric method [10] with a little modification. 0.1 mL properly diluted sample of the extract was mixed with 0.3 mL of a 5% NaNO₂ solution. The mixture was allowed to stay at room temperature for 6 min; 0.3 mL of a 10% AlCl₃ solution was added for 6 min followed by the addition of a 2 mL 4% NaOH solution. The double-distilled water was added to reach a final volume of 10 ml. The solution was mixed and kept at room temperature for 15 min. Absorbance was measured immediately against the prepared blank at 510 nm using a spectrophotometer (UV-2550, Shimadzu, equipped with 10 mm quartz cuvettes). Quantitative measurements were performed, based on a standard calibration curve of six points from 5.0 to 80 mg∙L⁻¹ of rutin in methanol (y = 0.0105x − 0.0031, R² = 0.9965, y and x represent the absorbance and concentration, respectively). The total phenolic content was expressed as rutin equivalents in mg∙g⁻¹ of extract.

DPPH radical-scavenging capacity of raspberry extracts was evaluated according to the method of Chen and Ho [11] , with a little modification. Briefly, 0.05 mL properly diluted sample of the extract was added to 5.0 mL methanol solution of DPPH radical (final concentration was 0.5 mmol∙L⁻¹). The mixture was shaken vigorously for 1 min by vortexing and left to stand at room temperature in the dark for 30 min. Thereafter, the absorbance of the sample was measured using the UV-vis spectrophotometer at 517 nm against methanol blank. A negative control was taken after adding DPPH solution to 0.05 mL methanol. The ascorbic acid was used as a reference compound [y = 14.032x + 0.4534, R² = 0.9975, y and x represent the scavenging rate and concentration (0.006 - 5.6 mg∙L⁻¹), respectively]. The ascorbic acid equivalent antioxidant capacity was expressed as mg ascorbic acid corresponding to one g dry weight.

The ABTS radical scavenging test was used to determine the antioxidant activity [12] . ABTS radical was obtained by reaction between ABTS and potassium persulfate. Blank sample was prepared from the daily solution by adding 100 mL PBS buffer (5 mmol∙L⁻¹, pH7.4), which gives an absorbance of 0.7 ± 0.02. The radical scavenging activity was assessed by mixing 5.0 mL ABTS solution with 0.05 mL properly diluted sample of the extract. The reactive mixture was allowed to stand at room temperature for 6 min and the absorbance was recorded at 734 nm. The ascorbic acid was used as standard (y = 22.821x + 1.3393, R² = 0.9978, y and x represent the scavenging rate and concentration (0.005 - 4.0 mg∙L⁻¹), respectively). The ascorbic acid equivalent antioxidant capacity was calculated like DPPH assay.

The HPLC system Dionex P680 series (Dionex, USA), equipped with the Chromeleon software (Dionex) and comprised a binary pump, an online vacuum degasser, a manual sampler, a thermostated column compartment and a diode array detection (DAD), was used for the chromatographic analysis. The chromatographic analysis was carried out on a Kromasil C₁₈ (250 mm × 4.6 mm, 5.0 µm) column. The mobile phase was composed of acetonitrile (A) and 1% formic acid solution (B) with gradient elution (0 - 5 min, 5% - 20% A; 5 - 20 min, 20% - 45% A; 20 - 25 min, 45% - 90% A; 25 - 30 min, 90% - 100% A, 30 - 40 min, 100% A). The solvent flow rate was 1.0 mL∙min⁻¹, the elute was continuously monitored using a DAD detector at a wavelength of 360 nm and the column temperature was maintained at 30˚C. Aliquots of 20 µL were injected into HPLC system for analysis.

Data analysis was performed using professional software (Analysis System for Chromatographic Fingerprint and Data of Traditional Chinese Medicine version 2004, the Pharmacopoeia Commission of the People’s Republic of China). This software allowed calculation of the correlation coefficients of the sample from different technologies with the reference chromatogram. The hierarchical clustering analysis (HCA) and principal component analysis (PCA) was performed using SPSS for Windows v.18.0 software (SPSS Corporation, Stanford , USA ).

The content of total flavonoid was determined by a colorimetric method and the results were summarized in Table 1. The total flavonoid contents in all samples ranged from 13.17 ± 0.02 to 22.82 ± 0.02 mg rutin g⁻¹ d.w. The values for DPPH and ABTS scavenging activities in all samples ranged from 20.51 ± 0.06 to 31.08 ± 0.03 mg ascorbic acid g⁻¹ d.w. and from 33.93 ± 0.11 to 54.83 ± 0.12 mg ascorbic acid g⁻¹ d.w., respectively (Table 1). A significant correlation between DPPH assay and ABTS assay was observed (r² = 0.821, ρ = 0.001, SPSS for windows 19.0, USA), which suggested the two methods were comparable in the case of Fupenzi (Table 2).

Generally, the total flavonoid contents were positively correlated with the antioxidant activities [13] [14] . In this study a significant correlation between antioxidant activity and the total flavonoid content was also observed for the DPPH assay (r² = 0.758, ρ = 0.004) and the ABTS assay (r² = 0.788, ρ = 0.002) (Table 2).

With the same sample extracts contributed to antioxidant activities, the HPLC chromatograms of 12 batches of samples were obtained under the optimized chromatographic conditions (Figure 1). As shown in Figure 1, these samples showed similar chromatographic profiles and more than 22 peaks were detected in all the 12 chromatograms. After carefully analyzing the chromatographic profiles of these samples, 15 common peaks with acceptable heights and good resolution were selected as characteristic peaks for fingerprinting analysis (Figure 2). Peak 10 was selected as the reference peak. Relative retention times (RRTs) and relative peak areas (RPAs) of the 15 characteristic peaks were calculated as follows: RRT = retention time of characteristic peak/retention time of marker peak, and RPA = peak area of characteristic peak/peak area of marker peak. The results from the 12 samples indicated that the RPAs of the 15 characteristic peaks varied dramatically (Table 3), but the RRTs were invariable for the Fupenzi (Table 4). The results indicated that the RRTs of the 15 characteristic peaks can serve as characteristic peaks for identification of “unknown” samples. A sample with a similar HPLC chromatographic profile and matched RRT values to the typical fingerprint chromatogram can be authenticated as genuine Fupenzi. However, due to several ingredients, such as climate, season of harvest and regions of cultivation, the same types of components were generally existed in the same variety, but their contents were often different

^*Was significant at 0.01 probability level.

[15] .

Due to similarity in appearance, it is impossible to divide these samples from the physical appearances. According to the RRTs and RPAs of 15 characteristic peaks, hierarchical clustering analysis (HCA) was perforemed by SPSS software (SPSS for windows 19.0, SPSS Inc., USA) and the results were shown in Figure 3. It is clear that the samples could be divided into three clusters: with samples 1, 2, 4, 5, 6, 7 and 9 in cluster one, samples 3, 10, 11 and 12 in cluster two and sample 8 in cluster three. The rescaled distance between group three and group one, group two are 25, while the rescaled distance between group one and group two are 10, which could demonstrate that the quality of group one and group two are more similar than with group three.

Flavonols such as rutin, hyperin, quercetin [16] and kaempferol [17] possess antioxidant activity and have been suggested to play an important role in the protective effects of berries and vegetables [18] . In the chromato-

graphic profile, four peaks (5, 6, 9 and 10) were structurally identified as rutin, hyperin, quercetin and kaempferol, respectively, by comparing their retention times (10.12 min for rutin, 10.80 for hyperin, 17.49 min for quercetin and 20.75 min for kaempferol) with those of standard substance. Therefore, the simultaneous determination of four active components in these samples was investigated by the proposed method.