Chemical Fingerprint and Quantitative Analysis for the Quality Evaluation of Docynia dcne Leaves by High-Performance Liquid Chromatography Coupled with Chemometrics Analysis

Chemical Fingerprint and Quantitative Analysis for the Quality Evaluation of Docynia dcne Leaves... Abstract Docynia dcne leaf from the genus of Docynia Dcne (including three species of Docynia delavayi, Docynia indica and Docynia longiunguis.) is an important raw material of local ethnic minority tea, ethnomedicines and food supplements in southwestern areas of China. However, D. dcne leaves from these three species are usually used confusingly, which could influence the therapeutic effect of it. A rapid and effective method for the chemical fingerprint and quantitative analysis to evaluate the quality of D. dcne leaves was established. The chemometric methods, including similarity analysis, hierarchical cluster analysis and partial least-squares discrimination analysis, were applied to distinguish 30 batches of D. dcne leaf samples from these three species. The above results could validate each other and successfully group these samples into three categories which were closely related to the species of D. dcne leaves. Moreover, isoquercitrin and phlorizin were screened as the chemical markers to evaluate the quality of D. dcne leaves from different species. And the contents of isoquercitrin and phlorizin varied remarkably in these samples, with ranges of 6.41–38.84 and 95.73–217.76 mg/g, respectively. All the results indicated that an integration method of chemical fingerprint couple with chemometrics analysis and quantitative assessment was a powerful and beneficial tool for quality control of D. dcne leaves, and could be applied also for differentiation and quality control of other herbal preparations. Introduction Docynia Dcne is an aiphyllium belongs to the Rosaceae family and distributed widely in China, Japan, Vietnam, India and Thailand (1). The genus of D. Dcne including three species of Docynia delavayi, Docynia indica and Docynia longiunguis, which is widely distributed in the southwest China, such as Diqing Tibetan Autonomous Prefecture, Xishuangbanna Dai Autonomous Prefecture, Liangshan Yi Autonomous Prefecture, Yunnan and Guizhou Province (2–4). It is named “Duoyi” by Dai and Yi people, the local ethnic minorities in the southwestern China, and it has been conventionally used as ethnomedicines for fever, cancer, empyrosis, fracture and rheumatic disease (5). Existing studies demonstrated that flavonoids were the characteristic chemical constituents of D. Dcne, which exhibited significant bioactivities, including antioxidation, antitumor, anti-obesity and antibacterial activities (2, 3, 6–10). Meanwhile, besides the root, leaf and rhizome of D. Dcne as ethnomedicines, the fruit and leaf of it are also the popular local foods in southwest china. Especially, the leaf is widely used as an important local ethnic tea (11). Docynia Dcne leaves from the three species are always used confusingly because of their highly similar morphology and efficacy, which could influence the therapeutic effect of it. However, due to their complex composition, the definite characterization of their therapeutic mechanism is always a great challenge. Nowadays, chemical fingerprint analysis has been approved in many countries as a rational and practical method for the comprehensive quality control of herbal products (12–15). Beyond the fingerprint technique, chemometric methods based on metabolic profiling, such as similarity analysis (SA), hierarchical cluster analysis (HCA) and partial least-squares discrimination analysis (PLS-DA), have exhibited much more powerful capabilities in multivariate analysis (16–18). Our previous report has demonstrated the strategy and performance of high-performance liquid chromatography coupled with ultraviolet detectors (HPLC-UV)-based metabolic profiling, in the quality assessment of herbal extracts are feasible (19). This strategy has shown early promise to be suitable for the quality evaluation of widely used herbal products. In this study, 30 batches of D. Dcne leaves were collected from the major product base, Sichuan and Yunnan province. A HPLC–diode array detector (DAD) method was developed for chemical fingerprint analysis of D. Dcne leaves for the first time. The chemometric analysis including SA, HCA and PLS-DA, were developed to evaluate the differences between these three different species. And two maker components, isoquercitrin and phlorizin, were screened and quantitatively determined simultaneously. The objective of this work was to develop an effective tool to discriminate the three different species of D. Dcne leaves, which will be helpful for the quality control of it. Materials and methods Materials and reagents A total of 30 batches of D. dcne leaves involving three species (D. delavayi, D. indica and D. longiunguis) were collected from the province of Sichuan and Yunnan in China (Table I). All samples were identified and confirmed by Professor Qiang Luo from College of Xichang (Sichuan, China). All samples were dried at 50°C, pulverized and sieved through an 80-mesh stainless-steel sieve before extraction. Table I. Contents (mg/g) of Two Marker Flavonoids in D. Dcne Leaves (30 samples) and Chemometrics Analysis (SA, HCA and PLS-DA) Number Region Phlorizin (mg/g) Isoquercitrin (mg/g) Speciesa Similarity HCA cluster PLS-DA categorization S1 Xishan, Eryuan,Yunnan Mar., 2013 112.03 ± 0.64 8.14 ± 0.23 Sp1 0.996 Cluster 1-2 Group 1 S2 Xishan, Eryuan,Yunnan May, 2013 108.93 ± 0.25 9.96 ± 0.05 Sp1 0.997 Cluster 1-2 Group 1 S3 Dehua, Ninger, Yunnan Mar., 2013 128.80 ± 0.26 33.99 ± 0.05 Sp1 0.992 Cluster 1-2 Group 1 S4 Dehua, Ninger, Yunnan May, 2013 105.90 ± 0.10 29.07 ± 0.05 Sp1 0.991 Cluster 1-2 Group 1 S5 Mohei, Ninger, Yunnan Mar., 2013 174.73 ± 0.47 42.8 ± 0.10 Sp1 0.993 Cluster 1-2 Group 3 S6 Mohei, Ninger, Yunnan May, 2013 146.30 ± 0.10 29.24 ± 0.86 Sp1 0.997 Cluster 1-2 Group 3 S7 Tuguo, Ninger, Yunnan Apr., 2013 120.80 ± 1.00 21.29 ± 0.73 Sp1 0.998 Cluster 1-2 Group 1 S8 Tuguo, Ninger, Yunnan Apr., 2013 113.07 ± 0.15 20.75 ± 0.23 Sp1 0.998 Cluster 1-2 Group 1 S9 Tongxin, Ninger, Yunnan Mar., 2013 95.73 ± 0.45 20.2 ± 0.65 Sp1 0.997 Cluster 1-2 Group 1 S10 Tongxin, Ninger, Yunnan Mar., 2013 106.50 ± 0.26 8.83 ± 0.26 Sp1 0.997 Cluster 1-2 Group 1 S11 Xiaomiao, Xichang, Sichuan Mar., 2012 217.06 ± 0.25 38.84 ± 0.67 Sp2 0.999 Cluster 2 Group 2 S12 Sihe, Xichang, Sichuan Mar., 2012 217.46 ± 0.46 38.78 ± 0.81 Sp2 0.999 Cluster 2 Group 2 S13 Kaiyuan, Xichang, Sichuan Mar., 2012 210.26 ± 0.23 6.47 ± 0.93 Sp2 0.993 Cluster 2 Group 2 S14 Yuehua, Xichang, Sichuan Mar., 2012 205.63 ± 0.58 6.41 ± 0.79 Sp2 0.993 Cluster 2 Group 2 S15 Daxing, Xichang, Sichuan Mar., 2012 217.76 ± 0.55 7.03 ± 0.05 Sp2 0.993 Cluster 2 Group 2 S16 Xiaomiao, Xichang, Sichuan Oct., 2012 210.13 ± 0.32 6.47 ± 0.43 Sp2 0.993 Cluster 2 Group 2 S17 Sihe, Xichang, Sichuan Oct., 2012 213.66 ± 0.83 34.86 ± 0.05 Sp2 0.999 Cluster 2 Group 2 S18 Kaiyuan, Xichang, Sichuan Oct., 2012 212.53 ± 1.51 36.38 ± 0.64 Sp2 0.999 Cluster 2 Group 2 S19 Yuehua, Xichang, Sichuan Oct., 2012 204.23 ± 0.20 34.42 ± 0.73 Sp2 0.999 Cluster 2 Group 2 S20 Daxing, Xichang, Sichuan Oct., 2012 204.36 ± 0.05 34.35 ± 0.25 Sp2 0.999 Cluster 2 Group 2 S21 Xiaomiao, Xichang, Sichuan Mar., 2012 149.53 ± 3.34 23.48 ± 0.10 Sp3 1.000 Cluster 1-1 Group 3 S22 Xiaomiao, Xichang, Sichuan Oct., 2012 145.76 ± 0.81 22.93 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S23 Xiaomiao, Xichang, Sichuan Mar., 2013 153.66 ± 0.31 23.93 ± 0.49 Sp3 1.000 Cluster 1-1 Group 3 S24 Sihe, Xichang, Sichuan Mar., 2012 167.90 ± 0.69 23.6 ± 0.89 Sp3 1.000 Cluster 1-1 Group 3 S25 Sihe, Xichang, Sichuan Oct., 2012 173.83 ± 0.25 24.35 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S26 Sihe, Xichang, Sichuan Mar., 2013 179.07 ± 0.57 24.63 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 S27 Kaiyuan, Xichang, Sichuan Mar., 2012 171.43 ± 0.91 23.86 ± 0.83 Sp3 1.000 Cluster 1-1 Group 3 S28 Kaiyuan, Xichang, Sichuan Oct., 2012 168.86 ± 0.06 23.95 ± 0.46 Sp3 1.000 Cluster 1-1 Group 3 S29 Kaiyuan, Xichang, Sichuan Mar., 2013 167.63 ± 0.31 23.57 ± 0.52 Sp3 1.000 Cluster 1-1 Group 3 S30 Yuehua, Xichang, Sichuan Mar., 2013 143.80 ± 0.69 24.41 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 Number Region Phlorizin (mg/g) Isoquercitrin (mg/g) Speciesa Similarity HCA cluster PLS-DA categorization S1 Xishan, Eryuan,Yunnan Mar., 2013 112.03 ± 0.64 8.14 ± 0.23 Sp1 0.996 Cluster 1-2 Group 1 S2 Xishan, Eryuan,Yunnan May, 2013 108.93 ± 0.25 9.96 ± 0.05 Sp1 0.997 Cluster 1-2 Group 1 S3 Dehua, Ninger, Yunnan Mar., 2013 128.80 ± 0.26 33.99 ± 0.05 Sp1 0.992 Cluster 1-2 Group 1 S4 Dehua, Ninger, Yunnan May, 2013 105.90 ± 0.10 29.07 ± 0.05 Sp1 0.991 Cluster 1-2 Group 1 S5 Mohei, Ninger, Yunnan Mar., 2013 174.73 ± 0.47 42.8 ± 0.10 Sp1 0.993 Cluster 1-2 Group 3 S6 Mohei, Ninger, Yunnan May, 2013 146.30 ± 0.10 29.24 ± 0.86 Sp1 0.997 Cluster 1-2 Group 3 S7 Tuguo, Ninger, Yunnan Apr., 2013 120.80 ± 1.00 21.29 ± 0.73 Sp1 0.998 Cluster 1-2 Group 1 S8 Tuguo, Ninger, Yunnan Apr., 2013 113.07 ± 0.15 20.75 ± 0.23 Sp1 0.998 Cluster 1-2 Group 1 S9 Tongxin, Ninger, Yunnan Mar., 2013 95.73 ± 0.45 20.2 ± 0.65 Sp1 0.997 Cluster 1-2 Group 1 S10 Tongxin, Ninger, Yunnan Mar., 2013 106.50 ± 0.26 8.83 ± 0.26 Sp1 0.997 Cluster 1-2 Group 1 S11 Xiaomiao, Xichang, Sichuan Mar., 2012 217.06 ± 0.25 38.84 ± 0.67 Sp2 0.999 Cluster 2 Group 2 S12 Sihe, Xichang, Sichuan Mar., 2012 217.46 ± 0.46 38.78 ± 0.81 Sp2 0.999 Cluster 2 Group 2 S13 Kaiyuan, Xichang, Sichuan Mar., 2012 210.26 ± 0.23 6.47 ± 0.93 Sp2 0.993 Cluster 2 Group 2 S14 Yuehua, Xichang, Sichuan Mar., 2012 205.63 ± 0.58 6.41 ± 0.79 Sp2 0.993 Cluster 2 Group 2 S15 Daxing, Xichang, Sichuan Mar., 2012 217.76 ± 0.55 7.03 ± 0.05 Sp2 0.993 Cluster 2 Group 2 S16 Xiaomiao, Xichang, Sichuan Oct., 2012 210.13 ± 0.32 6.47 ± 0.43 Sp2 0.993 Cluster 2 Group 2 S17 Sihe, Xichang, Sichuan Oct., 2012 213.66 ± 0.83 34.86 ± 0.05 Sp2 0.999 Cluster 2 Group 2 S18 Kaiyuan, Xichang, Sichuan Oct., 2012 212.53 ± 1.51 36.38 ± 0.64 Sp2 0.999 Cluster 2 Group 2 S19 Yuehua, Xichang, Sichuan Oct., 2012 204.23 ± 0.20 34.42 ± 0.73 Sp2 0.999 Cluster 2 Group 2 S20 Daxing, Xichang, Sichuan Oct., 2012 204.36 ± 0.05 34.35 ± 0.25 Sp2 0.999 Cluster 2 Group 2 S21 Xiaomiao, Xichang, Sichuan Mar., 2012 149.53 ± 3.34 23.48 ± 0.10 Sp3 1.000 Cluster 1-1 Group 3 S22 Xiaomiao, Xichang, Sichuan Oct., 2012 145.76 ± 0.81 22.93 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S23 Xiaomiao, Xichang, Sichuan Mar., 2013 153.66 ± 0.31 23.93 ± 0.49 Sp3 1.000 Cluster 1-1 Group 3 S24 Sihe, Xichang, Sichuan Mar., 2012 167.90 ± 0.69 23.6 ± 0.89 Sp3 1.000 Cluster 1-1 Group 3 S25 Sihe, Xichang, Sichuan Oct., 2012 173.83 ± 0.25 24.35 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S26 Sihe, Xichang, Sichuan Mar., 2013 179.07 ± 0.57 24.63 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 S27 Kaiyuan, Xichang, Sichuan Mar., 2012 171.43 ± 0.91 23.86 ± 0.83 Sp3 1.000 Cluster 1-1 Group 3 S28 Kaiyuan, Xichang, Sichuan Oct., 2012 168.86 ± 0.06 23.95 ± 0.46 Sp3 1.000 Cluster 1-1 Group 3 S29 Kaiyuan, Xichang, Sichuan Mar., 2013 167.63 ± 0.31 23.57 ± 0.52 Sp3 1.000 Cluster 1-1 Group 3 S30 Yuehua, Xichang, Sichuan Mar., 2013 143.80 ± 0.69 24.41 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 aSp1, D. delavayi; Sp2, D. indica; Sp3, D. longiunguis. View Large Table I. Contents (mg/g) of Two Marker Flavonoids in D. Dcne Leaves (30 samples) and Chemometrics Analysis (SA, HCA and PLS-DA) Number Region Phlorizin (mg/g) Isoquercitrin (mg/g) Speciesa Similarity HCA cluster PLS-DA categorization S1 Xishan, Eryuan,Yunnan Mar., 2013 112.03 ± 0.64 8.14 ± 0.23 Sp1 0.996 Cluster 1-2 Group 1 S2 Xishan, Eryuan,Yunnan May, 2013 108.93 ± 0.25 9.96 ± 0.05 Sp1 0.997 Cluster 1-2 Group 1 S3 Dehua, Ninger, Yunnan Mar., 2013 128.80 ± 0.26 33.99 ± 0.05 Sp1 0.992 Cluster 1-2 Group 1 S4 Dehua, Ninger, Yunnan May, 2013 105.90 ± 0.10 29.07 ± 0.05 Sp1 0.991 Cluster 1-2 Group 1 S5 Mohei, Ninger, Yunnan Mar., 2013 174.73 ± 0.47 42.8 ± 0.10 Sp1 0.993 Cluster 1-2 Group 3 S6 Mohei, Ninger, Yunnan May, 2013 146.30 ± 0.10 29.24 ± 0.86 Sp1 0.997 Cluster 1-2 Group 3 S7 Tuguo, Ninger, Yunnan Apr., 2013 120.80 ± 1.00 21.29 ± 0.73 Sp1 0.998 Cluster 1-2 Group 1 S8 Tuguo, Ninger, Yunnan Apr., 2013 113.07 ± 0.15 20.75 ± 0.23 Sp1 0.998 Cluster 1-2 Group 1 S9 Tongxin, Ninger, Yunnan Mar., 2013 95.73 ± 0.45 20.2 ± 0.65 Sp1 0.997 Cluster 1-2 Group 1 S10 Tongxin, Ninger, Yunnan Mar., 2013 106.50 ± 0.26 8.83 ± 0.26 Sp1 0.997 Cluster 1-2 Group 1 S11 Xiaomiao, Xichang, Sichuan Mar., 2012 217.06 ± 0.25 38.84 ± 0.67 Sp2 0.999 Cluster 2 Group 2 S12 Sihe, Xichang, Sichuan Mar., 2012 217.46 ± 0.46 38.78 ± 0.81 Sp2 0.999 Cluster 2 Group 2 S13 Kaiyuan, Xichang, Sichuan Mar., 2012 210.26 ± 0.23 6.47 ± 0.93 Sp2 0.993 Cluster 2 Group 2 S14 Yuehua, Xichang, Sichuan Mar., 2012 205.63 ± 0.58 6.41 ± 0.79 Sp2 0.993 Cluster 2 Group 2 S15 Daxing, Xichang, Sichuan Mar., 2012 217.76 ± 0.55 7.03 ± 0.05 Sp2 0.993 Cluster 2 Group 2 S16 Xiaomiao, Xichang, Sichuan Oct., 2012 210.13 ± 0.32 6.47 ± 0.43 Sp2 0.993 Cluster 2 Group 2 S17 Sihe, Xichang, Sichuan Oct., 2012 213.66 ± 0.83 34.86 ± 0.05 Sp2 0.999 Cluster 2 Group 2 S18 Kaiyuan, Xichang, Sichuan Oct., 2012 212.53 ± 1.51 36.38 ± 0.64 Sp2 0.999 Cluster 2 Group 2 S19 Yuehua, Xichang, Sichuan Oct., 2012 204.23 ± 0.20 34.42 ± 0.73 Sp2 0.999 Cluster 2 Group 2 S20 Daxing, Xichang, Sichuan Oct., 2012 204.36 ± 0.05 34.35 ± 0.25 Sp2 0.999 Cluster 2 Group 2 S21 Xiaomiao, Xichang, Sichuan Mar., 2012 149.53 ± 3.34 23.48 ± 0.10 Sp3 1.000 Cluster 1-1 Group 3 S22 Xiaomiao, Xichang, Sichuan Oct., 2012 145.76 ± 0.81 22.93 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S23 Xiaomiao, Xichang, Sichuan Mar., 2013 153.66 ± 0.31 23.93 ± 0.49 Sp3 1.000 Cluster 1-1 Group 3 S24 Sihe, Xichang, Sichuan Mar., 2012 167.90 ± 0.69 23.6 ± 0.89 Sp3 1.000 Cluster 1-1 Group 3 S25 Sihe, Xichang, Sichuan Oct., 2012 173.83 ± 0.25 24.35 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S26 Sihe, Xichang, Sichuan Mar., 2013 179.07 ± 0.57 24.63 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 S27 Kaiyuan, Xichang, Sichuan Mar., 2012 171.43 ± 0.91 23.86 ± 0.83 Sp3 1.000 Cluster 1-1 Group 3 S28 Kaiyuan, Xichang, Sichuan Oct., 2012 168.86 ± 0.06 23.95 ± 0.46 Sp3 1.000 Cluster 1-1 Group 3 S29 Kaiyuan, Xichang, Sichuan Mar., 2013 167.63 ± 0.31 23.57 ± 0.52 Sp3 1.000 Cluster 1-1 Group 3 S30 Yuehua, Xichang, Sichuan Mar., 2013 143.80 ± 0.69 24.41 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 Number Region Phlorizin (mg/g) Isoquercitrin (mg/g) Speciesa Similarity HCA cluster PLS-DA categorization S1 Xishan, Eryuan,Yunnan Mar., 2013 112.03 ± 0.64 8.14 ± 0.23 Sp1 0.996 Cluster 1-2 Group 1 S2 Xishan, Eryuan,Yunnan May, 2013 108.93 ± 0.25 9.96 ± 0.05 Sp1 0.997 Cluster 1-2 Group 1 S3 Dehua, Ninger, Yunnan Mar., 2013 128.80 ± 0.26 33.99 ± 0.05 Sp1 0.992 Cluster 1-2 Group 1 S4 Dehua, Ninger, Yunnan May, 2013 105.90 ± 0.10 29.07 ± 0.05 Sp1 0.991 Cluster 1-2 Group 1 S5 Mohei, Ninger, Yunnan Mar., 2013 174.73 ± 0.47 42.8 ± 0.10 Sp1 0.993 Cluster 1-2 Group 3 S6 Mohei, Ninger, Yunnan May, 2013 146.30 ± 0.10 29.24 ± 0.86 Sp1 0.997 Cluster 1-2 Group 3 S7 Tuguo, Ninger, Yunnan Apr., 2013 120.80 ± 1.00 21.29 ± 0.73 Sp1 0.998 Cluster 1-2 Group 1 S8 Tuguo, Ninger, Yunnan Apr., 2013 113.07 ± 0.15 20.75 ± 0.23 Sp1 0.998 Cluster 1-2 Group 1 S9 Tongxin, Ninger, Yunnan Mar., 2013 95.73 ± 0.45 20.2 ± 0.65 Sp1 0.997 Cluster 1-2 Group 1 S10 Tongxin, Ninger, Yunnan Mar., 2013 106.50 ± 0.26 8.83 ± 0.26 Sp1 0.997 Cluster 1-2 Group 1 S11 Xiaomiao, Xichang, Sichuan Mar., 2012 217.06 ± 0.25 38.84 ± 0.67 Sp2 0.999 Cluster 2 Group 2 S12 Sihe, Xichang, Sichuan Mar., 2012 217.46 ± 0.46 38.78 ± 0.81 Sp2 0.999 Cluster 2 Group 2 S13 Kaiyuan, Xichang, Sichuan Mar., 2012 210.26 ± 0.23 6.47 ± 0.93 Sp2 0.993 Cluster 2 Group 2 S14 Yuehua, Xichang, Sichuan Mar., 2012 205.63 ± 0.58 6.41 ± 0.79 Sp2 0.993 Cluster 2 Group 2 S15 Daxing, Xichang, Sichuan Mar., 2012 217.76 ± 0.55 7.03 ± 0.05 Sp2 0.993 Cluster 2 Group 2 S16 Xiaomiao, Xichang, Sichuan Oct., 2012 210.13 ± 0.32 6.47 ± 0.43 Sp2 0.993 Cluster 2 Group 2 S17 Sihe, Xichang, Sichuan Oct., 2012 213.66 ± 0.83 34.86 ± 0.05 Sp2 0.999 Cluster 2 Group 2 S18 Kaiyuan, Xichang, Sichuan Oct., 2012 212.53 ± 1.51 36.38 ± 0.64 Sp2 0.999 Cluster 2 Group 2 S19 Yuehua, Xichang, Sichuan Oct., 2012 204.23 ± 0.20 34.42 ± 0.73 Sp2 0.999 Cluster 2 Group 2 S20 Daxing, Xichang, Sichuan Oct., 2012 204.36 ± 0.05 34.35 ± 0.25 Sp2 0.999 Cluster 2 Group 2 S21 Xiaomiao, Xichang, Sichuan Mar., 2012 149.53 ± 3.34 23.48 ± 0.10 Sp3 1.000 Cluster 1-1 Group 3 S22 Xiaomiao, Xichang, Sichuan Oct., 2012 145.76 ± 0.81 22.93 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S23 Xiaomiao, Xichang, Sichuan Mar., 2013 153.66 ± 0.31 23.93 ± 0.49 Sp3 1.000 Cluster 1-1 Group 3 S24 Sihe, Xichang, Sichuan Mar., 2012 167.90 ± 0.69 23.6 ± 0.89 Sp3 1.000 Cluster 1-1 Group 3 S25 Sihe, Xichang, Sichuan Oct., 2012 173.83 ± 0.25 24.35 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S26 Sihe, Xichang, Sichuan Mar., 2013 179.07 ± 0.57 24.63 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 S27 Kaiyuan, Xichang, Sichuan Mar., 2012 171.43 ± 0.91 23.86 ± 0.83 Sp3 1.000 Cluster 1-1 Group 3 S28 Kaiyuan, Xichang, Sichuan Oct., 2012 168.86 ± 0.06 23.95 ± 0.46 Sp3 1.000 Cluster 1-1 Group 3 S29 Kaiyuan, Xichang, Sichuan Mar., 2013 167.63 ± 0.31 23.57 ± 0.52 Sp3 1.000 Cluster 1-1 Group 3 S30 Yuehua, Xichang, Sichuan Mar., 2013 143.80 ± 0.69 24.41 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 aSp1, D. delavayi; Sp2, D. indica; Sp3, D. longiunguis. View Large The methanol and acetonitrile (HPLC grade) were purchased from Tedia Company Inc. (Fairfield, USA). Ultrapure water was supplied by a WSD-UP-III-10 water purification system from Chengdu Weisida Company (Chengdu, China). All other reagents used in the present study were of analytical grade. Phlorizin was isolated from the D. dcne leaves in our laboratory. The purity of it was higher than 98%, as analyzed by the HPLC area normalization method. Phlorizin was confirmed by MS, 1H and 13C NMR spectroscopy as shown in Supplementary 1. Isoquercitrin in the D. dcne leaves was confirmed by its external reference including retention time, UV spectra and MS data (Supplementary 2), which also matched with the reported paper (20). Apparatus and chromatographic conditions For fingerprint analysis HPLC analysis was performed on an Agilent 1200 HPLC system (Agilent, USA) consisting of an autosampler, thermostatted, vacuum degasser, binary pump, column compartment and diode array detector. System control and data analysis were performed on the Chemstation Software program (version A.10.02). The separation was performed on an Agilent ZORBAX Extend-C18 column (4.6 × 250 mm, 5 μm) at 25°C. The mobile phase was composed of acetonitrile (A), methanol (B) and water (C) with gradient elution system (0–40 min, A: 5–15%, B: 0–35%; 40–44 min, A: 15–30%, B: 35–0%; 44–55 min, A: 30–40%, B: 0–0%; 55–60 min, A: 40–100%, B: 0–0%; 60–71 min, A: 100–100%, B: 0–0%) at 1 mL/min. The ultraviolet detector was set at 285 nm during the experiment and the injection volume of each sample and standard solution was set at 10 μL. All solutions were filtered through a 0.45 μm membrane filter before HPLC analysis. For quantitative analysis The instrument for quantitative analysis was consistent with the fingerprint analysis. The mobile phase was composed of acetonitrile (A) and water (B) with gradient elution system (0–10 min, A: 15–50%) at 1 mL/min. The ultraviolet detector was set at 285 nm and the injection volume was 6 μL. All solutions were filtered through a 0.45 μm membrane filter before quantitative analysis. Preparation of sample solutions For fingerprint analysis Two grams of dried powder sample was accurately weighed and extracted with 40 mL methanol by ultrasonic extraction at 25°C for three times (each for 30 min). The extracted solution was concentrated under vacuum at 50°C, and the dried extract was dissolved in 25 mL of methanol. Finally, the extract solution was filtrated through a 0.45 μm membrane filter before HPLC analysis. For quantitative analysis Each of the dried powder samples (0.5 g) were accurately weighed and extracted with 20 mL methanol by ultrasonic extraction at 25°C for twice (each for 30 min). The extracted solution was concentrated under vacuum at 50°C, and the dried extract was dissolved in 25 mL of methanol. Then, 0.6 mL of the extract solution was diluted to 10 mL with methanol, and the sample solution was filtrated through a 0.45 μm membrane filter before quantitative analysis. To obtain the single-analyte standard solutions, phlorizin and isoquercitrin were accurately weighed, dissolved in methanol and the mixed standard solutions were then diluted to generate an appropriate concentration range to establish calibration curves. All calibration curves were constructed by using seven different concentrations of mixed standards in triplicate, and all the standard solutions were filtered through 0.45 μm membrane filters before HPLC analysis. Method validation For fingerprint analysis According to the guidelines of the CFDA (21), the developed HPLC–DAD fingerprint method was validated in terms of its precision, stability and repeatability. For quantitative analysis The data of peak area versus the corresponding concentration were treated using linear least square regression analysis. The working standard solutions were further diluted to a certain concentration to explore the limit of detection (LOD) and quantification (LOQ). The intra- and inter-day precisions were determined by continuously injecting the standard solutions at three levels for six replicates within 1 day and on 5 consecutive days, respectively. The standard solutions at three levels were separately tested at 0, 2, 4, 8, 12, and 24 h for assessing stability. As for the repeatability, the sample solution from identical batch sample (S1) was prepared and detected in six parallels. The recovery test for reflecting accuracy was done by the standard addition approach. Accurate amounts of mixed standard solutions at three levels were added to sample S1 with six parallels. Chemometrics analysis The chemometric analysis was applied to demonstrate the variability of 30 batches of D. Dcne leaves samples. SA was performed using Similarity Evaluation System for Chromatographic Fingerprint of Traditional Chinese Medicine software (Version 2004 A, Chinese Pharmacopoeia Committee), which was recommended by CFDA. The correlation coefficient of similarity for entire chromatographic profiles among samples was calculated by this system using the median method with the time width of 0.1 to conduct SA of different chromatograms. The HCA of 30 samples was performed using SPSS software (IBM SPSS Statistics, Version 20.0, USA) to classify samples with regard to similarities of chemical properties, and the average linkage method and cosine applied in the measurements. PLS-DA procedures were similar to those in our previously published paper (22). Briefly, the chromatographic data used for peak integration were retention time (RT) of 2–70 min, minimum area of 10, advanced baseline calibration mode and vertical shoulder peak mode. No specific peak was excluded. The resulting data set were exported to SIMCA-P+ software 13.0 (Umetrics) for multivariate analysis. SA was initially used to calculate the correlation coefficients of chromatographic profiles of the 30 batches of D. Dcne leaves. HCA, an unsupervised multivariate analysis method, was used to generate a dendrogram of the 30 batches samples based on relative peak areas of those common characteristic peaks calculated by the similarity evaluation system. Thereafter, PLS-DA, a supervised multivariate analysis method, was carried out. Variables with the higher loading values in the PLS‑DA loadings plot may be regarded as marker components which contributed significantly to the categorization of D. Dcne leaves. Results Optimization of the extraction for fingerprint We first compared ultrasonic and reflux extraction methods for sample preparation. Our results suggested that the ultrasonic extraction was better than the reflux extraction. Various extraction conditions, including solvent, volume of solvent, time and repeats were investigated, using the total peak areas of three maximum peak area compound levels as the output. In this study, we chose extraction with ultrasound for three times (each for 30 min) in methanol (40 mL) at 25°C. Optimization of the HPLC conditions To give the comprehensive chemical information and best separation in the chromatograms, the column, detection wavelength, mobile phase and elution condition were investigated. Four different types of LC columns including the Agilent ZORBAX C18 Extend-C18 column (4.6 × 250 mm, 5 μm), AICHROM AichromBond-AQ C18 column (4.6 × 250 mm, 5 μm), TIANHE Kromasil C18 column (4.6 × 250 mm, 5 μm) and AICHROM AichromBond-1 C18 column (4.6 × 250 mm, 5 μm) were analyzed. Agilent ZORBAX Extend-C18 column (4.6 × 250 mm, 5 μm) was found to exhibit best separation efficiency than the other columns. The HPLC mobile phase (acetonitrile–water, methanol–water and acetonitrile–methanol–water) and the flow rate of the mobile phase (0.5, 0.8, 1.0 mL/min) were also examined for optimization, and it was found that acetonitrile–methanol–water with a flow rate of 1.0 mL/min achieved the best separation and suppressed the tailing of the peaks. An added benefit was that more compounds could be eluted within 70 min. The wavelength for the detection of the target compounds was set at 285 nm because more characteristic peaks could be attained. Method validation of fingerprint analysis Sample No. S1 (collected from the Yunnan province) was used for fingerprint method validation. The precision was assessed by six successive injections of the same sample (Sample S1) solution in a day. The relative retention time and relative peak area of common peaks were lower than 2.00 and 2.94% in relative standard deviation (RSD) (n = 6), respectively. The RSD of relative retention time and relative peak area for sample repeatability were evaluated by analysis of six replicates of Sample S1 and estimated to be no more than 0.66 and 2.98%, respectively. The stability of relative retention time and relative peak area of common peaks were evaluated by the analysis of eight replicates of Sample S1 at 0, 2, 4, 8, 12, 24, 36 and 48 h (in 2 days) and established to be lower than 0.48 and 2.96%, respectively. The results of the precision, repeatability and stability studies met the national standard for TCM fingerprint analysis, and the method was suitable for the fingerprint analysis of D. dcne leaves. Method validation of quantitative analysis The results of method validation for two external standards (phlorizin and isoquercitrin) are shown in Table II. The linearity of the calibration curves was verified and the correlation coefficients were all better than 0.9998. The LODs and LOQs of this method were <0.054 and 0.163 μg/mL, which were determined by a signal-to-noise (S/N) ratio of 3 and 10, respectively. The precision expressed as the RSD of the two flavonoids were below ±0.52% (n = 6). The stability and repeatability of samples expressed as the RSD of six parallel samples of S1 were below ±1.11%. The recoveries of phlorizin and isoquercitrin were between 98.33 and 101.08%. All these indicated that our developed method was precise, accurate and sensitive enough for simultaneous quantitative determination of these two flavonoids in D. dcne leaves. Table II. Linear Range, Regression Equation, R2, LOD, LOQ, Precision, Stability, Repeatability and Recovery of Isoquercitrin and Phlorizin Compounds Linearity ranges (μg/mL) Regression equation R2 LOD (μg/mL) LOQ (μg/mL) Precision (RSD %) (n = 6) Repeatability (RSD %) (n = 6) Stability (RSD %) (n = 6) Recovery rate (Mean ± RSD %) (n = 6) Isoquercitrin 1.76–112.50 Y = 19.776 X + 4.229 0.9998 0.054 0.163 0.47 1.11 0.65 98.33 ± 0.57 Phlorizin 6.45–412.50 Y = 34.788 X + 2.965 0.9999 0.021 0.071 0.52 0.78 0.47 101.08 ± 1.03 Compounds Linearity ranges (μg/mL) Regression equation R2 LOD (μg/mL) LOQ (μg/mL) Precision (RSD %) (n = 6) Repeatability (RSD %) (n = 6) Stability (RSD %) (n = 6) Recovery rate (Mean ± RSD %) (n = 6) Isoquercitrin 1.76–112.50 Y = 19.776 X + 4.229 0.9998 0.054 0.163 0.47 1.11 0.65 98.33 ± 0.57 Phlorizin 6.45–412.50 Y = 34.788 X + 2.965 0.9999 0.021 0.071 0.52 0.78 0.47 101.08 ± 1.03 Table II. Linear Range, Regression Equation, R2, LOD, LOQ, Precision, Stability, Repeatability and Recovery of Isoquercitrin and Phlorizin Compounds Linearity ranges (μg/mL) Regression equation R2 LOD (μg/mL) LOQ (μg/mL) Precision (RSD %) (n = 6) Repeatability (RSD %) (n = 6) Stability (RSD %) (n = 6) Recovery rate (Mean ± RSD %) (n = 6) Isoquercitrin 1.76–112.50 Y = 19.776 X + 4.229 0.9998 0.054 0.163 0.47 1.11 0.65 98.33 ± 0.57 Phlorizin 6.45–412.50 Y = 34.788 X + 2.965 0.9999 0.021 0.071 0.52 0.78 0.47 101.08 ± 1.03 Compounds Linearity ranges (μg/mL) Regression equation R2 LOD (μg/mL) LOQ (μg/mL) Precision (RSD %) (n = 6) Repeatability (RSD %) (n = 6) Stability (RSD %) (n = 6) Recovery rate (Mean ± RSD %) (n = 6) Isoquercitrin 1.76–112.50 Y = 19.776 X + 4.229 0.9998 0.054 0.163 0.47 1.11 0.65 98.33 ± 0.57 Phlorizin 6.45–412.50 Y = 34.788 X + 2.965 0.9999 0.021 0.071 0.52 0.78 0.47 101.08 ± 1.03 Sample analysis Similarity analysis The standard fingerprint of 30 batches of D. dcne samples was analyzed and shown in Figure 1a. The peaks that existed in all 30 samples with reasonable heights and good resolution were assigned as “characteristic peaks” for the identification of the plant, and the 13 characteristic peaks that within 70 min were shown in Figure 1b. The similarities were generated by comparing the 30 D. dcne leaf samples with the standard chromatogram (Figure 1b), which were shown in Table I. The similarity values of the 30 samples were more than 0.990, indicating that various samples shared similar chromatographic patterns and the entire chromatograms of these samples were generally consistent and stable. Especially, the samples of D. longiunguis had the maximal correlation coefficients among these samples. The above results have shown these three species share the similar chemical constituents and the limitation of correlation coefficients in distinguishing these three different species of D. dcne leaves. Figure 1. View largeDownload slide HPLC fingerprints of the 30 batches of D. dcne samples (a) and the simulative mean chromatogram (b). The chromatograms marked with S1-S30 and R represent 30 batches of D. dcne samples and the simulative mean chromatogram, respectively. The peaks marked with peak 2, 14, 16, 31, 32, 35, 38, 40, 61, 70 and 81 in the chromatogram represent the marker compounds in chemical profiling analysis. The peaks marked with 1–13 in the simulative mean chromatogram represent the 13 characteristic peaks in fingerprints analysis. Figure 1. View largeDownload slide HPLC fingerprints of the 30 batches of D. dcne samples (a) and the simulative mean chromatogram (b). The chromatograms marked with S1-S30 and R represent 30 batches of D. dcne samples and the simulative mean chromatogram, respectively. The peaks marked with peak 2, 14, 16, 31, 32, 35, 38, 40, 61, 70 and 81 in the chromatogram represent the marker compounds in chemical profiling analysis. The peaks marked with 1–13 in the simulative mean chromatogram represent the 13 characteristic peaks in fingerprints analysis. Hierarchical clustering analysis To assess the resemblance and differences among these three different species of D. dcne leaf samples, a HCA analysis was further performed using the 13 characteristic peaks identified from the standard fingerprint chromatogram (Figure 1b). The relative peak areas of 13 characteristic peaks of the 30 chromatograms of D. dcne samples formed a matrix of 13 × 30, and the result of HCA was shown in Figure 2 and Table I. The shorter distance between two samples in HCA dendrogram indicated their higher similarity and these samples clustered into the same group were the most similar ones (23). When an appropriate rescaled distance (about 25, Figure 2) was chosen, the samples could be categorized into two quality clusters (Clusters 1 and 2). From the information of Table I, it was interesting that the Cluster 2 including 10 batches of samples all belong to D. Indica, and the Cluster 1 including 20 batches of samples belong to D. delavayi and D. Longiunguis. However, when another rescaled distance (about 22, Figure 2) was chosen, the Cluster 1 was easily to be categorized into two quality clusters (Cluster 1-1 and 1-2). It was easy to find that the categorizations were related to their species, in briefly, Cluster 1-1 and 1-2 were D. Longiunguis and D. delavayi, respectively. The results indicated that HCA could accurately distinguish these three different species of D. dcne leaves. Figure 2. View largeDownload slide HCA of three different species of D. dcne leaves. Figure 2. View largeDownload slide HCA of three different species of D. dcne leaves. Partial least-squares discrimination analysis To confirm whether or not there are difference between three species of D.dcne leaves and identify the characteristic components which have the most influence on the chemical profiling of 30 batches of D. dcne leaves. PLS-DA analysis was further performed. R2X (cumulative), R2Y (cumulative) and Q2 (cumulative) of the PLS-DA model were 0.94, 0.974 and 0.913, respectively. The results indicated that the method of PLS-DA was stable and reliable. PLS-DA scores plot (Figure 3) demonstrated the different clustering pattern as HCA analysis (Figure 2). Group 2 (Figure 3) was same as Cluster 2 (Figure 2) including 10 batches of samples all belong to D. Indica, but the Group 3 (Figure 3) included S21–30, S5 and S6, which was different from Cluster 1-2 (Figure 2). Actually, from the Figure 2, S5 and S6 were categorized into Cluster 1-2 which belong to D. delavayi. Subsequently, PLS-DA loadings plot was analyzed and illustrated in Figure 4. An arbitrary loadings threshold was set on the loadings plot at ±0.10 for w*c (24), ±0.10 for w*c [2]; and at ±0.25 for w*c (24), ±0.20 for w*c [2], highlighted in gray and square, respectively. The selection of the threshold was further verified by a correlation study described below similar to our recently published paper (22). Hence, the variables located outside of the threshold region were regarded as the components contributing most significantly to the categorization of 30 batches of D. dcne leaves. When the loadings threshold was set on the loadings plot at ±0.10 for w*c (24) and ±0.10 for w*c [2] highlighted in gray, 11 components shown in Figure 4 including peak 2, peak 14, peak 16, peak 31, peak 32, peak 38, peak 61, peak 71 and peak 81, peak 35 (isoquercitrin) and peak 40 (phlorizin) were regarded as the marker components. However, when the loadings threshold was set at ±0.25 for w*c (24) and ±0.20 for w*c [2] highlighted in square, there were only two compounds including peak 35 (isoquercitrin) and peak 40 (phlorizin) were shown in Figure 4. This suggests that these two compounds were the biggest contributing factors the categorization of the 30 batches of D. dcne leaves. Combining with the HPLC chromatograms at 285 nm (Figure 1), the preliminary results above showed that the chemical profiling differentiation might be mostly explained by two components including peak 35 (isoquercitrin) and peak 40 (phlorizin). In other words, isoquercitrin and phlorizin are selected as chemical markers to evaluate the quality of D. dcne leaves from the different species, and the quantitative determination of these two chemical markers to evaluate these three different species is necessary. Figure 3. View largeDownload slide PLS-DA scores plot of three different species of D. dcne leaves. Figure 3. View largeDownload slide PLS-DA scores plot of three different species of D. dcne leaves. Figure 4. View largeDownload slide PLS-DA Loading scores plot of three different species of D. dcne leaves. An arbitrary loadings threshold was set on the loadings plot at ±0.10 for w*c [1], ±0.10 for w*c [2]; and at ±0.25 for w*c [1], ±0.20 for w*c [2] highlighted in gray and square, respectively. Figure 4. View largeDownload slide PLS-DA Loading scores plot of three different species of D. dcne leaves. An arbitrary loadings threshold was set on the loadings plot at ±0.10 for w*c [1], ±0.10 for w*c [2]; and at ±0.25 for w*c [1], ±0.20 for w*c [2] highlighted in gray and square, respectively. Quantitative analysis In this study, the content of two flavonoids in 30 batches of D. dcne leaf samples was determined and the data are presented in Table I. The peaks of isoquercitrin and phlorizin in each sample were identified by comparing the retention times and the UV spectra with those of the standards. The content levels of phlorizin and isoquercitrin in 30 batches of D. dcne leaves varied significantly. The results showed that isoquercitrin and phlorizin are the two main chemical markers in the leaf samples, with ranges of 6.41–38.84 and 95.73–217.76 mg/g, respectively. The content of the two flavonoids in different species was significantly different, and the content of phlorizin in all samples was the most abundant, which indicated the phlorizin was the main constitute of D. dcne leaves. Especially, the samples from D. indica had the highest content of phlorizin with the range of 204.23–217.76 mg/g, whereas the samples from D. delavayi and D. longiunguis had the lower content of phlorizin with ranges of 95.73–174.73 and 143.80–179.07 mg/g, respectively. Combining with the PLS-DA loadings plot result (Figure 4) of chemical profiling analysis, it was easily deduced that the difference of the content levels of phlorizin and isoquercitrin was related to the cauterization differences of samples. In this regard, phlorizin and isoquercitrin are the rational marker compounds which represent the comprehensive quality of D. dcne leaves. At the same time, the content levels of isoquercitrin (peak 35) and phlorizin (peak 40) of S5 and S6 were easily similar to the S21–30, which were cauterized into Group 3 (Figure 3). All the results illustrated that the internal quality of 30 batches of D. dcne leaves from different species was variant, and the isoquercitrin and phlorizin should be used as the indicator compounds to evaluate the quality of these different leaves. Discussion In this study, a rapid and efficient fingerprint method and the simultaneous determination of two chemical markers was first developed to evaluate the quality of three different species (D. delavayi, D. Indica and D. Longiunguis) of D. dcne leaves. A total of 30 samples collected from different regions of China were assessed by fingerprint and chemometrics including SA, HCA and PLS-DA. The similarities of the 30 samples ranged from 0.991 to 1.000 based on fingerprint peaks, indicating these three different species of D. dcne leaves have similar chromatographic patterns and the entire chromatograms of these samples were generally consistent and stable. Furthermore, HCA analysis according to the 13 characteristic fingerprint peaks was successfully applied to distinguish these three species. In addition, two chemical marker compounds (isoquercitrin and phlorizin) were identified through PLS-DA. These two markers were quantitatively determined and showed that contents of the two flavonoids in D. dcne leaves displayed notable differences in samples collected from three species. Especially, D. dcne leaves rich in phlorizin with a range from 9.57 to 21.71%. It was found that D. dcne leaves showed higher content levels of phlorizin than the ones in Malus Hupehensis (0.21–1.55%) (25), Lithocarpus Polystachyus (0.78–6.24%) (26) and apple leaves (1.13–4.03%) (27, 28). However, as many reports have revealed that the bioactivity of the herbs is not always consistent with its main peak intensity in the fingerprints, and the “marker” components may be cannot be used in quality control for the herbs (29). For example, various low-content compounds in the herbs may have the synergetic effect for its bioactivities and the effect of these non-common chromatographic peaks could not be excluded (30). D. Dcne leaves have been conventionally used as an important local ethnic tea for obesity or diabetes and ethnomedicines for anti-inflammation medicine (11). In this study, two components including isoquercitrin and phlorizin were selected as chemical markers to evaluate the quality of the D. dcne leaves from different species. Specially, isoquercitrin (31, 32) and phlorizin (33, 34) both have the obviously anti-inflammatory activity in vivo and in vitro. As for anti-obesity or anti-diabetic activity, phlorizin has the significant anti-diabetic activity (34–36) and its analog such as empagliflozin or canagliflozin has been used in clinical application, while the isoquercitrin not. So, we think these two components contribute to its anti-inflammatory activity, and the phlorizin is response for its anti-diabetic activity. To determine the efficacy of other components, more accurate analysis means such as spectrum–effect relationships method are needed, which will be the focus of our further research. Conclusion All the results indicate that the HPLC combination of fingerprint and quantitative analysis of maker components (isoquercitrin and phlorizin) is a powerful and practical tool for evaluating the quality of D. dcne leaves. Briefly, 30 batches of D. dcne leaves could be successfully divided into three groups, and showed good similarity on chemical constituents according to the results of chemometric analysis. In addition, isoquercitrin and phlorizin could be selected as chemical markers to evaluate the quality of D. dcne leaves with different sources. Our work also demonstrated that the D. dcne leaves, especially D. indica leaves are the potential natural resources of phlorizin and could be applied in functional food and medicine in the future. Supplementary Data Supplementary material is available at Journal of Chromatographic Science online. References 1 Commission of Chinese academy of sciences ; Flora of China, Vol. 9 , Science Press , Beijing , ( 2003 ), pp. 170 – 171 . 2 Deng , X. , Zhao , X. , Lan , Z. , Jiang , J. , Yin , W. , Chen , L. ; Anti-tumor effects of flavonoids from the ethnic medicine Docynia delavayi (Franch.) Schneid. and its possible mechanism ; Journal of Medicinal Food , ( 2014 ); 17 ( 7 ): 787 – 794 . Google Scholar CrossRef Search ADS PubMed 3 Sharma , P.B. , Handique , P.J. , Devi , H.S. ; Antioxidant properties, physico-chemical characteristics and proximate composition of five wild fruits of Manipur, India ; Journal of Food Science and Technology , ( 2015 ); 52 ( 2 ): 894 – 902 . Google Scholar CrossRef Search ADS PubMed 4 Luo , Q. , Liu , J.L. , Cai , G.Z. ; Docynia longiunguis Q.Luo et J.L.Liu, a new species of Docynia Dcne. from China ; Bulletin of Botanical Research , ( 2011 ); 31 ( 31 ): 389 – 391 . 5 Lin , Y.F. , Yi , Z. , Zhao , Y.H. ; Chinese Dai medicine colorful illustrations . Yunnan National Press, Kunming , ( 2003 ); p. 380 . 6 Guan , Z. , Tan , S.D. , Su , J.Y. , Zeng , L.M. ; Flavones from the bark of Docynia delavayi (Franch.) Schneid ; Natural Product Research and Development , ( 2000 ); 12 ( 12 ): 34 – 37 . 7 Nguyen , T.T.L. , Hoang , T.M.T. , Vu , T.H.T. , Chu , L.L. , Le , M.H. , Do , N.L. ; Anti-obesity and body weight reducing effect of Docynia indica(Wall.) Decne fruit extract fractions in experimentally obese mice ; VNU Journal of Science, Natural Sciences and Technology , ( 2011 ); 27 ( 2 ): 125 – 133 . 8 Hazarika , T.K. , Lalramchuana , Nautiyal , B.P. ; Studies on wild edible fruits of Mizoram, India used as ethno-medicine ; Genetic Resources and Crop Evolution , ( 2012 ); 59 ( 8 ): 1767 – 1776 . Google Scholar CrossRef Search ADS 9 Zhao , X. , Shu , G. , Chen , L. , Mi , X. , Mei , Z. , Deng , X. ; A flavonoid component from Docynia delavayi (Franch.) Schneid represses transplanted H22 hepatoma growth and exhibits low toxic effect on tumor-bearing mice ; Food and Chemical Toxicology , ( 2012 ); 50 ( 9 ): 3166 – 3173 . Google Scholar CrossRef Search ADS PubMed 10 Liu , G. , Wu , J. , Zhu , M.J. , Zhang , X.Y. , Zhang , H. ; On evaluation of antioxidant activities and screening of potential antioxidant compounds from Docynia delavayi (Franch.) Schneid. leaves ; Journal of Southwest China Normal University (Natural Science Edition) , ( 2014 ); 39 ( 39 ): 73 – 81 . 11 Yuan , W. , Li , W.X. , Lin , Q. , Yang , Z.S. ; Studies on resource and utilization of Docynia in Yunnan ; Journal of Yunnan Agrieultural University , ( 1995 ); 10 ( 3 ): 256 – 258 . 12 Liang , X.M. , Jin , Y. , Wang , Y.P. , Jin , G.W. , Fu , Q. , Xiao , Y.S. ; Qualitative and quantitative analysis in quality control of traditional Chinese medicines ; Journal of Chromatography A , ( 2009 ); 1216 ( 11 ): 2033 – 2044 . Google Scholar CrossRef Search ADS PubMed 13 Jin , X.F. , Lu , Y.H. , Wei , D.Z. , Wang , Z.T. ; Chemical fingerprint and quantitative analysis of Salvia plebeia R.Br. by high-performance liquid chromatography ; Journal of Pharmaceutical and Biomedical Analysis , ( 2008 ); 48 ( 1 ): 100 – 104 . Google Scholar CrossRef Search ADS PubMed 14 Li , W. , Deng , Y. , Dai , R. , et al. . ; Chromatographic fingerprint analysis of Cephalotaxus sinensis from various sources by high-performance liquid chromatography-diodearray detection-electrospray ionization-tandem mass spectrometry ; Journal of Pharmaceutical and Biomedical Analysis , ( 2007 ); 45 ( 1 ): 38 – 46 . Google Scholar CrossRef Search ADS PubMed 15 Alaerts , G. , Matthijs , N. , Smeyers-Verbeke , J. , Vander Heyden , Y. ; Chromatographic fingerprint development for herbal extracts: a screening and optimization methodology on monolithic columns ; Journal of Chromatography A , ( 2007 ); 1172 ( 1 ): 1 – 8 . Google Scholar CrossRef Search ADS PubMed 16 Liu , H. , Liu , J. , Zhang , J. , et al. . ; Simultaneous quantitative and chemical fingerprint analysis of Receptaculum nelumbinis based on HPLC-DAD-MS combined with chemometrics ; Journal of Chromatographic Science , ( 2016 ); 54 ( 4 ): 618 – 624 . Google Scholar CrossRef Search ADS PubMed 17 Shu , Z. , Li , X. , Rahman , K. , Qin , L. , Zheng , C. ; Chemical fingerprint and quantitative analysis for the quality evaluation of Vitex negundo seeds by reversed-phase high-performance liquid chromatography coupled with hierarchical clustering analysis ; Journal of Separation Science , ( 2016 ); 39 ( 2 ): 279 – 286 . Google Scholar CrossRef Search ADS PubMed 18 Cheng , T.F. , Jia , Y.R. , Zuo , Z. , et al. . ; Quality assessment of traditional Chinese medicine herb couple by high-performance liquid chromatography and mass spectrometry combined with chemometrics ; Journal of Separation Science , ( 2016 ); 39 ( 7 ): 1223 – 1231 . Google Scholar CrossRef Search ADS PubMed 19 Lan , K. , Zhang , Y. , Yang , J. , Xu , L. ; Simple quality assessment approach for herbal extracts using high performance liquid chromatography-UV based metabolomics platform ; Journal of Chromatography A , ( 2010 ); 1217 ( 8 ): 1414 – 1418 . Google Scholar CrossRef Search ADS PubMed 20 Zhang , Y. , Peng , M. , Liu , L. , Shi , S. , Peng , S. ; Screening, identification and potential interaction of active compounds from Eucommia ulmodies leaves binding with bovine serum albumin ; Journal of Agricultural and Food Chemistry , ( 2012 ); 60 ( 12 ): 3119 – 3125 . Google Scholar CrossRef Search ADS PubMed 21 Drug Administration Bureau of China . , Requirements for Studying Fingerprint of Traditional Chinese Medicine Injections (Draft). ( 2000 ). 22 Wang , Z. , Hu , H. , Chen , F. , et al. . ; Metabolic profiling assisted quality assessment of Rhodiola rosea extracts by high-performance liquid chromatography ; Planta Medica , ( 2012 ); 78 ( 7 ): 740 – 746 . Google Scholar CrossRef Search ADS PubMed 23 Kong , W. , Li , Z. , Xiao , X. , Zhao , Y. ; Quality control for Coptidis rhizoma through the determination of five alkaloids by HPLC-ELSD coupled with chemometrics ; Natural Product Research , ( 2010 ); 24 ( 17 ): 1616 – 1629 . Google Scholar CrossRef Search ADS PubMed 24 Chase , H.P. , Boulware , D. , Rodriguez , H. , et al. . ; Effect of docosahexaenoic acid supplementation on inflammatory cytokine levels in infants at high genetic risk for type 1 diabetes ; Pediatric Diabetes , ( 2015 ); 16 ( 4 ): 271 – 279 . Google Scholar CrossRef Search ADS PubMed 25 Guo , D.Y. , Li , J. , Shi , Y.Q. , Wang , J.Z. ; Study on the flavon ingredients of Malus hupehensis ; Chinese Herbal Medicine , ( 2011 ); 34 ( 34 ): 1026 – 1029 . 26 He , C. , Peng , Y. , Xiao , W. , Hu , Y. , Xiao , P. ; Quick determination of five sweet constituents in Duosuike tiancha by RSLC ; China Journal of Chinese Materia Medica , ( 2012 ); 37 ( 7 ): 961 – 965 . Google Scholar PubMed 27 Zhao , Y.N. , Wang , J. , Song , G.M. , Song , X.B. , Liu , D.N. , Xie , W.L. ; Study on the dynamic changs of phlordzin and phloretin in leave and branches of apple tree ; Food Research and Development , ( 2013 ); 34 ( 34 ): 95 – 98 . 28 Li , R.T. , Liu , J.C. , Jiao , Z.G. , Wang , S.X. , Du , X.F. ; Determination of phlorhizin in apple tree branch and leaf by RP-HPLC ; Food Industry Reseach , ( 2009 ); 30 ( 30 ): 385 – 388 . 29 Xu , G.L. , Xie , M. , Yang , X.Y. , et al. . ; Spectrum-effect relationships as a systematic approach to traditional Chinese medicine research: current status and future perspectives ; Molecules , ( 2014 ); 19 ( 11 ): 17897 – 17925 . Google Scholar CrossRef Search ADS PubMed 30 Du , W.J. , Ji , J. , Wang , L. , et al. . ; Relationship between the UPLC-Q-TOF-MS fingerprinted constituents from Daphne genkwa and their anti-inflammatory, anti-oxidant activities ; Biomedical Chromatography , ( 2017 ); 31 ( 12 ): 1 – 9 . Google Scholar CrossRef Search ADS 31 Rogerio , A.P. , Kanashiro , A. , Fontanari , C. , et al. . ; Anti-inflammatory activity of quercetin and isoquercitrin in experimental murine allergic asthma ; Inflammation Research , ( 2007 ); 56 ( 10 ): 402 – 408 . Google Scholar CrossRef Search ADS PubMed 32 Xie , W. , Wang , M. , Chen , C. , Zhang , X. , Melzig , M.F. ; Hepatoprotective effect of isoquercitrin against acetaminophen-induced liver injury ; Life Sciences , ( 2016 ); 152 : 180 – 189 . Google Scholar CrossRef Search ADS PubMed 33 Chang , W.T. , Huang , W.C. , Liou , C.J. ; Evaluation of the anti-inflammatory effects of phloretin and phlorizin in lipopolysaccharide-stimulated mouse macrophages ; Food Chemistry , ( 2012 ); 134 ( 2 ): 972 – 979 . Google Scholar CrossRef Search ADS PubMed 34 Shin , S.K. , Cho , S.J. , Jung , U.J. , Ryu , R. , Choi , M.S. ; Phlorizin supplementation attenuates obesity, inflammation, and hyperglycemia in diet-induced obese mice fed a high-fat diet ; Nutrients , ( 2016 ); 8 ( 2 ): 92 . Google Scholar CrossRef Search ADS PubMed 35 Rossetti , L. , Smith , D. , Shulman , G.I. , Papachristou , D. , DeFronzo , R.A. ; Correction of hyperglycemia with phlorizin normalizes tissue sensitivity to insulin in diabetic rats ; Journal of Clinical Investigation , ( 1987 ); 79 ( 5 ): 1510 – 1515 . Google Scholar CrossRef Search ADS PubMed 36 Mei , X. , Zhang , X. , Wang , Z. , et al. . ; Insulin sensitivity-enhancing activity of phlorizin is associated with lipopolysaccharide decrease and gut microbiota changes in obese and type 2 diabetes (db/db) mice ; Journal of Agricultural and Food Chemistry , ( 2016 ); 64 ( 40 ): 7502 – 7511 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Chromatographic Science Oxford University Press

Chemical Fingerprint and Quantitative Analysis for the Quality Evaluation of Docynia dcne Leaves by High-Performance Liquid Chromatography Coupled with Chemometrics Analysis

Loading next page...
 
/lp/ou_press/chemical-fingerprint-and-quantitative-analysis-for-the-quality-mluj5IvtvB
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
0021-9665
eISSN
1945-239X
D.O.I.
10.1093/chromsci/bmy007
Publisher site
See Article on Publisher Site

Abstract

Abstract Docynia dcne leaf from the genus of Docynia Dcne (including three species of Docynia delavayi, Docynia indica and Docynia longiunguis.) is an important raw material of local ethnic minority tea, ethnomedicines and food supplements in southwestern areas of China. However, D. dcne leaves from these three species are usually used confusingly, which could influence the therapeutic effect of it. A rapid and effective method for the chemical fingerprint and quantitative analysis to evaluate the quality of D. dcne leaves was established. The chemometric methods, including similarity analysis, hierarchical cluster analysis and partial least-squares discrimination analysis, were applied to distinguish 30 batches of D. dcne leaf samples from these three species. The above results could validate each other and successfully group these samples into three categories which were closely related to the species of D. dcne leaves. Moreover, isoquercitrin and phlorizin were screened as the chemical markers to evaluate the quality of D. dcne leaves from different species. And the contents of isoquercitrin and phlorizin varied remarkably in these samples, with ranges of 6.41–38.84 and 95.73–217.76 mg/g, respectively. All the results indicated that an integration method of chemical fingerprint couple with chemometrics analysis and quantitative assessment was a powerful and beneficial tool for quality control of D. dcne leaves, and could be applied also for differentiation and quality control of other herbal preparations. Introduction Docynia Dcne is an aiphyllium belongs to the Rosaceae family and distributed widely in China, Japan, Vietnam, India and Thailand (1). The genus of D. Dcne including three species of Docynia delavayi, Docynia indica and Docynia longiunguis, which is widely distributed in the southwest China, such as Diqing Tibetan Autonomous Prefecture, Xishuangbanna Dai Autonomous Prefecture, Liangshan Yi Autonomous Prefecture, Yunnan and Guizhou Province (2–4). It is named “Duoyi” by Dai and Yi people, the local ethnic minorities in the southwestern China, and it has been conventionally used as ethnomedicines for fever, cancer, empyrosis, fracture and rheumatic disease (5). Existing studies demonstrated that flavonoids were the characteristic chemical constituents of D. Dcne, which exhibited significant bioactivities, including antioxidation, antitumor, anti-obesity and antibacterial activities (2, 3, 6–10). Meanwhile, besides the root, leaf and rhizome of D. Dcne as ethnomedicines, the fruit and leaf of it are also the popular local foods in southwest china. Especially, the leaf is widely used as an important local ethnic tea (11). Docynia Dcne leaves from the three species are always used confusingly because of their highly similar morphology and efficacy, which could influence the therapeutic effect of it. However, due to their complex composition, the definite characterization of their therapeutic mechanism is always a great challenge. Nowadays, chemical fingerprint analysis has been approved in many countries as a rational and practical method for the comprehensive quality control of herbal products (12–15). Beyond the fingerprint technique, chemometric methods based on metabolic profiling, such as similarity analysis (SA), hierarchical cluster analysis (HCA) and partial least-squares discrimination analysis (PLS-DA), have exhibited much more powerful capabilities in multivariate analysis (16–18). Our previous report has demonstrated the strategy and performance of high-performance liquid chromatography coupled with ultraviolet detectors (HPLC-UV)-based metabolic profiling, in the quality assessment of herbal extracts are feasible (19). This strategy has shown early promise to be suitable for the quality evaluation of widely used herbal products. In this study, 30 batches of D. Dcne leaves were collected from the major product base, Sichuan and Yunnan province. A HPLC–diode array detector (DAD) method was developed for chemical fingerprint analysis of D. Dcne leaves for the first time. The chemometric analysis including SA, HCA and PLS-DA, were developed to evaluate the differences between these three different species. And two maker components, isoquercitrin and phlorizin, were screened and quantitatively determined simultaneously. The objective of this work was to develop an effective tool to discriminate the three different species of D. Dcne leaves, which will be helpful for the quality control of it. Materials and methods Materials and reagents A total of 30 batches of D. dcne leaves involving three species (D. delavayi, D. indica and D. longiunguis) were collected from the province of Sichuan and Yunnan in China (Table I). All samples were identified and confirmed by Professor Qiang Luo from College of Xichang (Sichuan, China). All samples were dried at 50°C, pulverized and sieved through an 80-mesh stainless-steel sieve before extraction. Table I. Contents (mg/g) of Two Marker Flavonoids in D. Dcne Leaves (30 samples) and Chemometrics Analysis (SA, HCA and PLS-DA) Number Region Phlorizin (mg/g) Isoquercitrin (mg/g) Speciesa Similarity HCA cluster PLS-DA categorization S1 Xishan, Eryuan,Yunnan Mar., 2013 112.03 ± 0.64 8.14 ± 0.23 Sp1 0.996 Cluster 1-2 Group 1 S2 Xishan, Eryuan,Yunnan May, 2013 108.93 ± 0.25 9.96 ± 0.05 Sp1 0.997 Cluster 1-2 Group 1 S3 Dehua, Ninger, Yunnan Mar., 2013 128.80 ± 0.26 33.99 ± 0.05 Sp1 0.992 Cluster 1-2 Group 1 S4 Dehua, Ninger, Yunnan May, 2013 105.90 ± 0.10 29.07 ± 0.05 Sp1 0.991 Cluster 1-2 Group 1 S5 Mohei, Ninger, Yunnan Mar., 2013 174.73 ± 0.47 42.8 ± 0.10 Sp1 0.993 Cluster 1-2 Group 3 S6 Mohei, Ninger, Yunnan May, 2013 146.30 ± 0.10 29.24 ± 0.86 Sp1 0.997 Cluster 1-2 Group 3 S7 Tuguo, Ninger, Yunnan Apr., 2013 120.80 ± 1.00 21.29 ± 0.73 Sp1 0.998 Cluster 1-2 Group 1 S8 Tuguo, Ninger, Yunnan Apr., 2013 113.07 ± 0.15 20.75 ± 0.23 Sp1 0.998 Cluster 1-2 Group 1 S9 Tongxin, Ninger, Yunnan Mar., 2013 95.73 ± 0.45 20.2 ± 0.65 Sp1 0.997 Cluster 1-2 Group 1 S10 Tongxin, Ninger, Yunnan Mar., 2013 106.50 ± 0.26 8.83 ± 0.26 Sp1 0.997 Cluster 1-2 Group 1 S11 Xiaomiao, Xichang, Sichuan Mar., 2012 217.06 ± 0.25 38.84 ± 0.67 Sp2 0.999 Cluster 2 Group 2 S12 Sihe, Xichang, Sichuan Mar., 2012 217.46 ± 0.46 38.78 ± 0.81 Sp2 0.999 Cluster 2 Group 2 S13 Kaiyuan, Xichang, Sichuan Mar., 2012 210.26 ± 0.23 6.47 ± 0.93 Sp2 0.993 Cluster 2 Group 2 S14 Yuehua, Xichang, Sichuan Mar., 2012 205.63 ± 0.58 6.41 ± 0.79 Sp2 0.993 Cluster 2 Group 2 S15 Daxing, Xichang, Sichuan Mar., 2012 217.76 ± 0.55 7.03 ± 0.05 Sp2 0.993 Cluster 2 Group 2 S16 Xiaomiao, Xichang, Sichuan Oct., 2012 210.13 ± 0.32 6.47 ± 0.43 Sp2 0.993 Cluster 2 Group 2 S17 Sihe, Xichang, Sichuan Oct., 2012 213.66 ± 0.83 34.86 ± 0.05 Sp2 0.999 Cluster 2 Group 2 S18 Kaiyuan, Xichang, Sichuan Oct., 2012 212.53 ± 1.51 36.38 ± 0.64 Sp2 0.999 Cluster 2 Group 2 S19 Yuehua, Xichang, Sichuan Oct., 2012 204.23 ± 0.20 34.42 ± 0.73 Sp2 0.999 Cluster 2 Group 2 S20 Daxing, Xichang, Sichuan Oct., 2012 204.36 ± 0.05 34.35 ± 0.25 Sp2 0.999 Cluster 2 Group 2 S21 Xiaomiao, Xichang, Sichuan Mar., 2012 149.53 ± 3.34 23.48 ± 0.10 Sp3 1.000 Cluster 1-1 Group 3 S22 Xiaomiao, Xichang, Sichuan Oct., 2012 145.76 ± 0.81 22.93 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S23 Xiaomiao, Xichang, Sichuan Mar., 2013 153.66 ± 0.31 23.93 ± 0.49 Sp3 1.000 Cluster 1-1 Group 3 S24 Sihe, Xichang, Sichuan Mar., 2012 167.90 ± 0.69 23.6 ± 0.89 Sp3 1.000 Cluster 1-1 Group 3 S25 Sihe, Xichang, Sichuan Oct., 2012 173.83 ± 0.25 24.35 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S26 Sihe, Xichang, Sichuan Mar., 2013 179.07 ± 0.57 24.63 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 S27 Kaiyuan, Xichang, Sichuan Mar., 2012 171.43 ± 0.91 23.86 ± 0.83 Sp3 1.000 Cluster 1-1 Group 3 S28 Kaiyuan, Xichang, Sichuan Oct., 2012 168.86 ± 0.06 23.95 ± 0.46 Sp3 1.000 Cluster 1-1 Group 3 S29 Kaiyuan, Xichang, Sichuan Mar., 2013 167.63 ± 0.31 23.57 ± 0.52 Sp3 1.000 Cluster 1-1 Group 3 S30 Yuehua, Xichang, Sichuan Mar., 2013 143.80 ± 0.69 24.41 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 Number Region Phlorizin (mg/g) Isoquercitrin (mg/g) Speciesa Similarity HCA cluster PLS-DA categorization S1 Xishan, Eryuan,Yunnan Mar., 2013 112.03 ± 0.64 8.14 ± 0.23 Sp1 0.996 Cluster 1-2 Group 1 S2 Xishan, Eryuan,Yunnan May, 2013 108.93 ± 0.25 9.96 ± 0.05 Sp1 0.997 Cluster 1-2 Group 1 S3 Dehua, Ninger, Yunnan Mar., 2013 128.80 ± 0.26 33.99 ± 0.05 Sp1 0.992 Cluster 1-2 Group 1 S4 Dehua, Ninger, Yunnan May, 2013 105.90 ± 0.10 29.07 ± 0.05 Sp1 0.991 Cluster 1-2 Group 1 S5 Mohei, Ninger, Yunnan Mar., 2013 174.73 ± 0.47 42.8 ± 0.10 Sp1 0.993 Cluster 1-2 Group 3 S6 Mohei, Ninger, Yunnan May, 2013 146.30 ± 0.10 29.24 ± 0.86 Sp1 0.997 Cluster 1-2 Group 3 S7 Tuguo, Ninger, Yunnan Apr., 2013 120.80 ± 1.00 21.29 ± 0.73 Sp1 0.998 Cluster 1-2 Group 1 S8 Tuguo, Ninger, Yunnan Apr., 2013 113.07 ± 0.15 20.75 ± 0.23 Sp1 0.998 Cluster 1-2 Group 1 S9 Tongxin, Ninger, Yunnan Mar., 2013 95.73 ± 0.45 20.2 ± 0.65 Sp1 0.997 Cluster 1-2 Group 1 S10 Tongxin, Ninger, Yunnan Mar., 2013 106.50 ± 0.26 8.83 ± 0.26 Sp1 0.997 Cluster 1-2 Group 1 S11 Xiaomiao, Xichang, Sichuan Mar., 2012 217.06 ± 0.25 38.84 ± 0.67 Sp2 0.999 Cluster 2 Group 2 S12 Sihe, Xichang, Sichuan Mar., 2012 217.46 ± 0.46 38.78 ± 0.81 Sp2 0.999 Cluster 2 Group 2 S13 Kaiyuan, Xichang, Sichuan Mar., 2012 210.26 ± 0.23 6.47 ± 0.93 Sp2 0.993 Cluster 2 Group 2 S14 Yuehua, Xichang, Sichuan Mar., 2012 205.63 ± 0.58 6.41 ± 0.79 Sp2 0.993 Cluster 2 Group 2 S15 Daxing, Xichang, Sichuan Mar., 2012 217.76 ± 0.55 7.03 ± 0.05 Sp2 0.993 Cluster 2 Group 2 S16 Xiaomiao, Xichang, Sichuan Oct., 2012 210.13 ± 0.32 6.47 ± 0.43 Sp2 0.993 Cluster 2 Group 2 S17 Sihe, Xichang, Sichuan Oct., 2012 213.66 ± 0.83 34.86 ± 0.05 Sp2 0.999 Cluster 2 Group 2 S18 Kaiyuan, Xichang, Sichuan Oct., 2012 212.53 ± 1.51 36.38 ± 0.64 Sp2 0.999 Cluster 2 Group 2 S19 Yuehua, Xichang, Sichuan Oct., 2012 204.23 ± 0.20 34.42 ± 0.73 Sp2 0.999 Cluster 2 Group 2 S20 Daxing, Xichang, Sichuan Oct., 2012 204.36 ± 0.05 34.35 ± 0.25 Sp2 0.999 Cluster 2 Group 2 S21 Xiaomiao, Xichang, Sichuan Mar., 2012 149.53 ± 3.34 23.48 ± 0.10 Sp3 1.000 Cluster 1-1 Group 3 S22 Xiaomiao, Xichang, Sichuan Oct., 2012 145.76 ± 0.81 22.93 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S23 Xiaomiao, Xichang, Sichuan Mar., 2013 153.66 ± 0.31 23.93 ± 0.49 Sp3 1.000 Cluster 1-1 Group 3 S24 Sihe, Xichang, Sichuan Mar., 2012 167.90 ± 0.69 23.6 ± 0.89 Sp3 1.000 Cluster 1-1 Group 3 S25 Sihe, Xichang, Sichuan Oct., 2012 173.83 ± 0.25 24.35 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S26 Sihe, Xichang, Sichuan Mar., 2013 179.07 ± 0.57 24.63 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 S27 Kaiyuan, Xichang, Sichuan Mar., 2012 171.43 ± 0.91 23.86 ± 0.83 Sp3 1.000 Cluster 1-1 Group 3 S28 Kaiyuan, Xichang, Sichuan Oct., 2012 168.86 ± 0.06 23.95 ± 0.46 Sp3 1.000 Cluster 1-1 Group 3 S29 Kaiyuan, Xichang, Sichuan Mar., 2013 167.63 ± 0.31 23.57 ± 0.52 Sp3 1.000 Cluster 1-1 Group 3 S30 Yuehua, Xichang, Sichuan Mar., 2013 143.80 ± 0.69 24.41 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 aSp1, D. delavayi; Sp2, D. indica; Sp3, D. longiunguis. View Large Table I. Contents (mg/g) of Two Marker Flavonoids in D. Dcne Leaves (30 samples) and Chemometrics Analysis (SA, HCA and PLS-DA) Number Region Phlorizin (mg/g) Isoquercitrin (mg/g) Speciesa Similarity HCA cluster PLS-DA categorization S1 Xishan, Eryuan,Yunnan Mar., 2013 112.03 ± 0.64 8.14 ± 0.23 Sp1 0.996 Cluster 1-2 Group 1 S2 Xishan, Eryuan,Yunnan May, 2013 108.93 ± 0.25 9.96 ± 0.05 Sp1 0.997 Cluster 1-2 Group 1 S3 Dehua, Ninger, Yunnan Mar., 2013 128.80 ± 0.26 33.99 ± 0.05 Sp1 0.992 Cluster 1-2 Group 1 S4 Dehua, Ninger, Yunnan May, 2013 105.90 ± 0.10 29.07 ± 0.05 Sp1 0.991 Cluster 1-2 Group 1 S5 Mohei, Ninger, Yunnan Mar., 2013 174.73 ± 0.47 42.8 ± 0.10 Sp1 0.993 Cluster 1-2 Group 3 S6 Mohei, Ninger, Yunnan May, 2013 146.30 ± 0.10 29.24 ± 0.86 Sp1 0.997 Cluster 1-2 Group 3 S7 Tuguo, Ninger, Yunnan Apr., 2013 120.80 ± 1.00 21.29 ± 0.73 Sp1 0.998 Cluster 1-2 Group 1 S8 Tuguo, Ninger, Yunnan Apr., 2013 113.07 ± 0.15 20.75 ± 0.23 Sp1 0.998 Cluster 1-2 Group 1 S9 Tongxin, Ninger, Yunnan Mar., 2013 95.73 ± 0.45 20.2 ± 0.65 Sp1 0.997 Cluster 1-2 Group 1 S10 Tongxin, Ninger, Yunnan Mar., 2013 106.50 ± 0.26 8.83 ± 0.26 Sp1 0.997 Cluster 1-2 Group 1 S11 Xiaomiao, Xichang, Sichuan Mar., 2012 217.06 ± 0.25 38.84 ± 0.67 Sp2 0.999 Cluster 2 Group 2 S12 Sihe, Xichang, Sichuan Mar., 2012 217.46 ± 0.46 38.78 ± 0.81 Sp2 0.999 Cluster 2 Group 2 S13 Kaiyuan, Xichang, Sichuan Mar., 2012 210.26 ± 0.23 6.47 ± 0.93 Sp2 0.993 Cluster 2 Group 2 S14 Yuehua, Xichang, Sichuan Mar., 2012 205.63 ± 0.58 6.41 ± 0.79 Sp2 0.993 Cluster 2 Group 2 S15 Daxing, Xichang, Sichuan Mar., 2012 217.76 ± 0.55 7.03 ± 0.05 Sp2 0.993 Cluster 2 Group 2 S16 Xiaomiao, Xichang, Sichuan Oct., 2012 210.13 ± 0.32 6.47 ± 0.43 Sp2 0.993 Cluster 2 Group 2 S17 Sihe, Xichang, Sichuan Oct., 2012 213.66 ± 0.83 34.86 ± 0.05 Sp2 0.999 Cluster 2 Group 2 S18 Kaiyuan, Xichang, Sichuan Oct., 2012 212.53 ± 1.51 36.38 ± 0.64 Sp2 0.999 Cluster 2 Group 2 S19 Yuehua, Xichang, Sichuan Oct., 2012 204.23 ± 0.20 34.42 ± 0.73 Sp2 0.999 Cluster 2 Group 2 S20 Daxing, Xichang, Sichuan Oct., 2012 204.36 ± 0.05 34.35 ± 0.25 Sp2 0.999 Cluster 2 Group 2 S21 Xiaomiao, Xichang, Sichuan Mar., 2012 149.53 ± 3.34 23.48 ± 0.10 Sp3 1.000 Cluster 1-1 Group 3 S22 Xiaomiao, Xichang, Sichuan Oct., 2012 145.76 ± 0.81 22.93 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S23 Xiaomiao, Xichang, Sichuan Mar., 2013 153.66 ± 0.31 23.93 ± 0.49 Sp3 1.000 Cluster 1-1 Group 3 S24 Sihe, Xichang, Sichuan Mar., 2012 167.90 ± 0.69 23.6 ± 0.89 Sp3 1.000 Cluster 1-1 Group 3 S25 Sihe, Xichang, Sichuan Oct., 2012 173.83 ± 0.25 24.35 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S26 Sihe, Xichang, Sichuan Mar., 2013 179.07 ± 0.57 24.63 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 S27 Kaiyuan, Xichang, Sichuan Mar., 2012 171.43 ± 0.91 23.86 ± 0.83 Sp3 1.000 Cluster 1-1 Group 3 S28 Kaiyuan, Xichang, Sichuan Oct., 2012 168.86 ± 0.06 23.95 ± 0.46 Sp3 1.000 Cluster 1-1 Group 3 S29 Kaiyuan, Xichang, Sichuan Mar., 2013 167.63 ± 0.31 23.57 ± 0.52 Sp3 1.000 Cluster 1-1 Group 3 S30 Yuehua, Xichang, Sichuan Mar., 2013 143.80 ± 0.69 24.41 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 Number Region Phlorizin (mg/g) Isoquercitrin (mg/g) Speciesa Similarity HCA cluster PLS-DA categorization S1 Xishan, Eryuan,Yunnan Mar., 2013 112.03 ± 0.64 8.14 ± 0.23 Sp1 0.996 Cluster 1-2 Group 1 S2 Xishan, Eryuan,Yunnan May, 2013 108.93 ± 0.25 9.96 ± 0.05 Sp1 0.997 Cluster 1-2 Group 1 S3 Dehua, Ninger, Yunnan Mar., 2013 128.80 ± 0.26 33.99 ± 0.05 Sp1 0.992 Cluster 1-2 Group 1 S4 Dehua, Ninger, Yunnan May, 2013 105.90 ± 0.10 29.07 ± 0.05 Sp1 0.991 Cluster 1-2 Group 1 S5 Mohei, Ninger, Yunnan Mar., 2013 174.73 ± 0.47 42.8 ± 0.10 Sp1 0.993 Cluster 1-2 Group 3 S6 Mohei, Ninger, Yunnan May, 2013 146.30 ± 0.10 29.24 ± 0.86 Sp1 0.997 Cluster 1-2 Group 3 S7 Tuguo, Ninger, Yunnan Apr., 2013 120.80 ± 1.00 21.29 ± 0.73 Sp1 0.998 Cluster 1-2 Group 1 S8 Tuguo, Ninger, Yunnan Apr., 2013 113.07 ± 0.15 20.75 ± 0.23 Sp1 0.998 Cluster 1-2 Group 1 S9 Tongxin, Ninger, Yunnan Mar., 2013 95.73 ± 0.45 20.2 ± 0.65 Sp1 0.997 Cluster 1-2 Group 1 S10 Tongxin, Ninger, Yunnan Mar., 2013 106.50 ± 0.26 8.83 ± 0.26 Sp1 0.997 Cluster 1-2 Group 1 S11 Xiaomiao, Xichang, Sichuan Mar., 2012 217.06 ± 0.25 38.84 ± 0.67 Sp2 0.999 Cluster 2 Group 2 S12 Sihe, Xichang, Sichuan Mar., 2012 217.46 ± 0.46 38.78 ± 0.81 Sp2 0.999 Cluster 2 Group 2 S13 Kaiyuan, Xichang, Sichuan Mar., 2012 210.26 ± 0.23 6.47 ± 0.93 Sp2 0.993 Cluster 2 Group 2 S14 Yuehua, Xichang, Sichuan Mar., 2012 205.63 ± 0.58 6.41 ± 0.79 Sp2 0.993 Cluster 2 Group 2 S15 Daxing, Xichang, Sichuan Mar., 2012 217.76 ± 0.55 7.03 ± 0.05 Sp2 0.993 Cluster 2 Group 2 S16 Xiaomiao, Xichang, Sichuan Oct., 2012 210.13 ± 0.32 6.47 ± 0.43 Sp2 0.993 Cluster 2 Group 2 S17 Sihe, Xichang, Sichuan Oct., 2012 213.66 ± 0.83 34.86 ± 0.05 Sp2 0.999 Cluster 2 Group 2 S18 Kaiyuan, Xichang, Sichuan Oct., 2012 212.53 ± 1.51 36.38 ± 0.64 Sp2 0.999 Cluster 2 Group 2 S19 Yuehua, Xichang, Sichuan Oct., 2012 204.23 ± 0.20 34.42 ± 0.73 Sp2 0.999 Cluster 2 Group 2 S20 Daxing, Xichang, Sichuan Oct., 2012 204.36 ± 0.05 34.35 ± 0.25 Sp2 0.999 Cluster 2 Group 2 S21 Xiaomiao, Xichang, Sichuan Mar., 2012 149.53 ± 3.34 23.48 ± 0.10 Sp3 1.000 Cluster 1-1 Group 3 S22 Xiaomiao, Xichang, Sichuan Oct., 2012 145.76 ± 0.81 22.93 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S23 Xiaomiao, Xichang, Sichuan Mar., 2013 153.66 ± 0.31 23.93 ± 0.49 Sp3 1.000 Cluster 1-1 Group 3 S24 Sihe, Xichang, Sichuan Mar., 2012 167.90 ± 0.69 23.6 ± 0.89 Sp3 1.000 Cluster 1-1 Group 3 S25 Sihe, Xichang, Sichuan Oct., 2012 173.83 ± 0.25 24.35 ± 0.26 Sp3 1.000 Cluster 1-1 Group 3 S26 Sihe, Xichang, Sichuan Mar., 2013 179.07 ± 0.57 24.63 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 S27 Kaiyuan, Xichang, Sichuan Mar., 2012 171.43 ± 0.91 23.86 ± 0.83 Sp3 1.000 Cluster 1-1 Group 3 S28 Kaiyuan, Xichang, Sichuan Oct., 2012 168.86 ± 0.06 23.95 ± 0.46 Sp3 1.000 Cluster 1-1 Group 3 S29 Kaiyuan, Xichang, Sichuan Mar., 2013 167.63 ± 0.31 23.57 ± 0.52 Sp3 1.000 Cluster 1-1 Group 3 S30 Yuehua, Xichang, Sichuan Mar., 2013 143.80 ± 0.69 24.41 ± 0.23 Sp3 1.000 Cluster 1-1 Group 3 aSp1, D. delavayi; Sp2, D. indica; Sp3, D. longiunguis. View Large The methanol and acetonitrile (HPLC grade) were purchased from Tedia Company Inc. (Fairfield, USA). Ultrapure water was supplied by a WSD-UP-III-10 water purification system from Chengdu Weisida Company (Chengdu, China). All other reagents used in the present study were of analytical grade. Phlorizin was isolated from the D. dcne leaves in our laboratory. The purity of it was higher than 98%, as analyzed by the HPLC area normalization method. Phlorizin was confirmed by MS, 1H and 13C NMR spectroscopy as shown in Supplementary 1. Isoquercitrin in the D. dcne leaves was confirmed by its external reference including retention time, UV spectra and MS data (Supplementary 2), which also matched with the reported paper (20). Apparatus and chromatographic conditions For fingerprint analysis HPLC analysis was performed on an Agilent 1200 HPLC system (Agilent, USA) consisting of an autosampler, thermostatted, vacuum degasser, binary pump, column compartment and diode array detector. System control and data analysis were performed on the Chemstation Software program (version A.10.02). The separation was performed on an Agilent ZORBAX Extend-C18 column (4.6 × 250 mm, 5 μm) at 25°C. The mobile phase was composed of acetonitrile (A), methanol (B) and water (C) with gradient elution system (0–40 min, A: 5–15%, B: 0–35%; 40–44 min, A: 15–30%, B: 35–0%; 44–55 min, A: 30–40%, B: 0–0%; 55–60 min, A: 40–100%, B: 0–0%; 60–71 min, A: 100–100%, B: 0–0%) at 1 mL/min. The ultraviolet detector was set at 285 nm during the experiment and the injection volume of each sample and standard solution was set at 10 μL. All solutions were filtered through a 0.45 μm membrane filter before HPLC analysis. For quantitative analysis The instrument for quantitative analysis was consistent with the fingerprint analysis. The mobile phase was composed of acetonitrile (A) and water (B) with gradient elution system (0–10 min, A: 15–50%) at 1 mL/min. The ultraviolet detector was set at 285 nm and the injection volume was 6 μL. All solutions were filtered through a 0.45 μm membrane filter before quantitative analysis. Preparation of sample solutions For fingerprint analysis Two grams of dried powder sample was accurately weighed and extracted with 40 mL methanol by ultrasonic extraction at 25°C for three times (each for 30 min). The extracted solution was concentrated under vacuum at 50°C, and the dried extract was dissolved in 25 mL of methanol. Finally, the extract solution was filtrated through a 0.45 μm membrane filter before HPLC analysis. For quantitative analysis Each of the dried powder samples (0.5 g) were accurately weighed and extracted with 20 mL methanol by ultrasonic extraction at 25°C for twice (each for 30 min). The extracted solution was concentrated under vacuum at 50°C, and the dried extract was dissolved in 25 mL of methanol. Then, 0.6 mL of the extract solution was diluted to 10 mL with methanol, and the sample solution was filtrated through a 0.45 μm membrane filter before quantitative analysis. To obtain the single-analyte standard solutions, phlorizin and isoquercitrin were accurately weighed, dissolved in methanol and the mixed standard solutions were then diluted to generate an appropriate concentration range to establish calibration curves. All calibration curves were constructed by using seven different concentrations of mixed standards in triplicate, and all the standard solutions were filtered through 0.45 μm membrane filters before HPLC analysis. Method validation For fingerprint analysis According to the guidelines of the CFDA (21), the developed HPLC–DAD fingerprint method was validated in terms of its precision, stability and repeatability. For quantitative analysis The data of peak area versus the corresponding concentration were treated using linear least square regression analysis. The working standard solutions were further diluted to a certain concentration to explore the limit of detection (LOD) and quantification (LOQ). The intra- and inter-day precisions were determined by continuously injecting the standard solutions at three levels for six replicates within 1 day and on 5 consecutive days, respectively. The standard solutions at three levels were separately tested at 0, 2, 4, 8, 12, and 24 h for assessing stability. As for the repeatability, the sample solution from identical batch sample (S1) was prepared and detected in six parallels. The recovery test for reflecting accuracy was done by the standard addition approach. Accurate amounts of mixed standard solutions at three levels were added to sample S1 with six parallels. Chemometrics analysis The chemometric analysis was applied to demonstrate the variability of 30 batches of D. Dcne leaves samples. SA was performed using Similarity Evaluation System for Chromatographic Fingerprint of Traditional Chinese Medicine software (Version 2004 A, Chinese Pharmacopoeia Committee), which was recommended by CFDA. The correlation coefficient of similarity for entire chromatographic profiles among samples was calculated by this system using the median method with the time width of 0.1 to conduct SA of different chromatograms. The HCA of 30 samples was performed using SPSS software (IBM SPSS Statistics, Version 20.0, USA) to classify samples with regard to similarities of chemical properties, and the average linkage method and cosine applied in the measurements. PLS-DA procedures were similar to those in our previously published paper (22). Briefly, the chromatographic data used for peak integration were retention time (RT) of 2–70 min, minimum area of 10, advanced baseline calibration mode and vertical shoulder peak mode. No specific peak was excluded. The resulting data set were exported to SIMCA-P+ software 13.0 (Umetrics) for multivariate analysis. SA was initially used to calculate the correlation coefficients of chromatographic profiles of the 30 batches of D. Dcne leaves. HCA, an unsupervised multivariate analysis method, was used to generate a dendrogram of the 30 batches samples based on relative peak areas of those common characteristic peaks calculated by the similarity evaluation system. Thereafter, PLS-DA, a supervised multivariate analysis method, was carried out. Variables with the higher loading values in the PLS‑DA loadings plot may be regarded as marker components which contributed significantly to the categorization of D. Dcne leaves. Results Optimization of the extraction for fingerprint We first compared ultrasonic and reflux extraction methods for sample preparation. Our results suggested that the ultrasonic extraction was better than the reflux extraction. Various extraction conditions, including solvent, volume of solvent, time and repeats were investigated, using the total peak areas of three maximum peak area compound levels as the output. In this study, we chose extraction with ultrasound for three times (each for 30 min) in methanol (40 mL) at 25°C. Optimization of the HPLC conditions To give the comprehensive chemical information and best separation in the chromatograms, the column, detection wavelength, mobile phase and elution condition were investigated. Four different types of LC columns including the Agilent ZORBAX C18 Extend-C18 column (4.6 × 250 mm, 5 μm), AICHROM AichromBond-AQ C18 column (4.6 × 250 mm, 5 μm), TIANHE Kromasil C18 column (4.6 × 250 mm, 5 μm) and AICHROM AichromBond-1 C18 column (4.6 × 250 mm, 5 μm) were analyzed. Agilent ZORBAX Extend-C18 column (4.6 × 250 mm, 5 μm) was found to exhibit best separation efficiency than the other columns. The HPLC mobile phase (acetonitrile–water, methanol–water and acetonitrile–methanol–water) and the flow rate of the mobile phase (0.5, 0.8, 1.0 mL/min) were also examined for optimization, and it was found that acetonitrile–methanol–water with a flow rate of 1.0 mL/min achieved the best separation and suppressed the tailing of the peaks. An added benefit was that more compounds could be eluted within 70 min. The wavelength for the detection of the target compounds was set at 285 nm because more characteristic peaks could be attained. Method validation of fingerprint analysis Sample No. S1 (collected from the Yunnan province) was used for fingerprint method validation. The precision was assessed by six successive injections of the same sample (Sample S1) solution in a day. The relative retention time and relative peak area of common peaks were lower than 2.00 and 2.94% in relative standard deviation (RSD) (n = 6), respectively. The RSD of relative retention time and relative peak area for sample repeatability were evaluated by analysis of six replicates of Sample S1 and estimated to be no more than 0.66 and 2.98%, respectively. The stability of relative retention time and relative peak area of common peaks were evaluated by the analysis of eight replicates of Sample S1 at 0, 2, 4, 8, 12, 24, 36 and 48 h (in 2 days) and established to be lower than 0.48 and 2.96%, respectively. The results of the precision, repeatability and stability studies met the national standard for TCM fingerprint analysis, and the method was suitable for the fingerprint analysis of D. dcne leaves. Method validation of quantitative analysis The results of method validation for two external standards (phlorizin and isoquercitrin) are shown in Table II. The linearity of the calibration curves was verified and the correlation coefficients were all better than 0.9998. The LODs and LOQs of this method were <0.054 and 0.163 μg/mL, which were determined by a signal-to-noise (S/N) ratio of 3 and 10, respectively. The precision expressed as the RSD of the two flavonoids were below ±0.52% (n = 6). The stability and repeatability of samples expressed as the RSD of six parallel samples of S1 were below ±1.11%. The recoveries of phlorizin and isoquercitrin were between 98.33 and 101.08%. All these indicated that our developed method was precise, accurate and sensitive enough for simultaneous quantitative determination of these two flavonoids in D. dcne leaves. Table II. Linear Range, Regression Equation, R2, LOD, LOQ, Precision, Stability, Repeatability and Recovery of Isoquercitrin and Phlorizin Compounds Linearity ranges (μg/mL) Regression equation R2 LOD (μg/mL) LOQ (μg/mL) Precision (RSD %) (n = 6) Repeatability (RSD %) (n = 6) Stability (RSD %) (n = 6) Recovery rate (Mean ± RSD %) (n = 6) Isoquercitrin 1.76–112.50 Y = 19.776 X + 4.229 0.9998 0.054 0.163 0.47 1.11 0.65 98.33 ± 0.57 Phlorizin 6.45–412.50 Y = 34.788 X + 2.965 0.9999 0.021 0.071 0.52 0.78 0.47 101.08 ± 1.03 Compounds Linearity ranges (μg/mL) Regression equation R2 LOD (μg/mL) LOQ (μg/mL) Precision (RSD %) (n = 6) Repeatability (RSD %) (n = 6) Stability (RSD %) (n = 6) Recovery rate (Mean ± RSD %) (n = 6) Isoquercitrin 1.76–112.50 Y = 19.776 X + 4.229 0.9998 0.054 0.163 0.47 1.11 0.65 98.33 ± 0.57 Phlorizin 6.45–412.50 Y = 34.788 X + 2.965 0.9999 0.021 0.071 0.52 0.78 0.47 101.08 ± 1.03 Table II. Linear Range, Regression Equation, R2, LOD, LOQ, Precision, Stability, Repeatability and Recovery of Isoquercitrin and Phlorizin Compounds Linearity ranges (μg/mL) Regression equation R2 LOD (μg/mL) LOQ (μg/mL) Precision (RSD %) (n = 6) Repeatability (RSD %) (n = 6) Stability (RSD %) (n = 6) Recovery rate (Mean ± RSD %) (n = 6) Isoquercitrin 1.76–112.50 Y = 19.776 X + 4.229 0.9998 0.054 0.163 0.47 1.11 0.65 98.33 ± 0.57 Phlorizin 6.45–412.50 Y = 34.788 X + 2.965 0.9999 0.021 0.071 0.52 0.78 0.47 101.08 ± 1.03 Compounds Linearity ranges (μg/mL) Regression equation R2 LOD (μg/mL) LOQ (μg/mL) Precision (RSD %) (n = 6) Repeatability (RSD %) (n = 6) Stability (RSD %) (n = 6) Recovery rate (Mean ± RSD %) (n = 6) Isoquercitrin 1.76–112.50 Y = 19.776 X + 4.229 0.9998 0.054 0.163 0.47 1.11 0.65 98.33 ± 0.57 Phlorizin 6.45–412.50 Y = 34.788 X + 2.965 0.9999 0.021 0.071 0.52 0.78 0.47 101.08 ± 1.03 Sample analysis Similarity analysis The standard fingerprint of 30 batches of D. dcne samples was analyzed and shown in Figure 1a. The peaks that existed in all 30 samples with reasonable heights and good resolution were assigned as “characteristic peaks” for the identification of the plant, and the 13 characteristic peaks that within 70 min were shown in Figure 1b. The similarities were generated by comparing the 30 D. dcne leaf samples with the standard chromatogram (Figure 1b), which were shown in Table I. The similarity values of the 30 samples were more than 0.990, indicating that various samples shared similar chromatographic patterns and the entire chromatograms of these samples were generally consistent and stable. Especially, the samples of D. longiunguis had the maximal correlation coefficients among these samples. The above results have shown these three species share the similar chemical constituents and the limitation of correlation coefficients in distinguishing these three different species of D. dcne leaves. Figure 1. View largeDownload slide HPLC fingerprints of the 30 batches of D. dcne samples (a) and the simulative mean chromatogram (b). The chromatograms marked with S1-S30 and R represent 30 batches of D. dcne samples and the simulative mean chromatogram, respectively. The peaks marked with peak 2, 14, 16, 31, 32, 35, 38, 40, 61, 70 and 81 in the chromatogram represent the marker compounds in chemical profiling analysis. The peaks marked with 1–13 in the simulative mean chromatogram represent the 13 characteristic peaks in fingerprints analysis. Figure 1. View largeDownload slide HPLC fingerprints of the 30 batches of D. dcne samples (a) and the simulative mean chromatogram (b). The chromatograms marked with S1-S30 and R represent 30 batches of D. dcne samples and the simulative mean chromatogram, respectively. The peaks marked with peak 2, 14, 16, 31, 32, 35, 38, 40, 61, 70 and 81 in the chromatogram represent the marker compounds in chemical profiling analysis. The peaks marked with 1–13 in the simulative mean chromatogram represent the 13 characteristic peaks in fingerprints analysis. Hierarchical clustering analysis To assess the resemblance and differences among these three different species of D. dcne leaf samples, a HCA analysis was further performed using the 13 characteristic peaks identified from the standard fingerprint chromatogram (Figure 1b). The relative peak areas of 13 characteristic peaks of the 30 chromatograms of D. dcne samples formed a matrix of 13 × 30, and the result of HCA was shown in Figure 2 and Table I. The shorter distance between two samples in HCA dendrogram indicated their higher similarity and these samples clustered into the same group were the most similar ones (23). When an appropriate rescaled distance (about 25, Figure 2) was chosen, the samples could be categorized into two quality clusters (Clusters 1 and 2). From the information of Table I, it was interesting that the Cluster 2 including 10 batches of samples all belong to D. Indica, and the Cluster 1 including 20 batches of samples belong to D. delavayi and D. Longiunguis. However, when another rescaled distance (about 22, Figure 2) was chosen, the Cluster 1 was easily to be categorized into two quality clusters (Cluster 1-1 and 1-2). It was easy to find that the categorizations were related to their species, in briefly, Cluster 1-1 and 1-2 were D. Longiunguis and D. delavayi, respectively. The results indicated that HCA could accurately distinguish these three different species of D. dcne leaves. Figure 2. View largeDownload slide HCA of three different species of D. dcne leaves. Figure 2. View largeDownload slide HCA of three different species of D. dcne leaves. Partial least-squares discrimination analysis To confirm whether or not there are difference between three species of D.dcne leaves and identify the characteristic components which have the most influence on the chemical profiling of 30 batches of D. dcne leaves. PLS-DA analysis was further performed. R2X (cumulative), R2Y (cumulative) and Q2 (cumulative) of the PLS-DA model were 0.94, 0.974 and 0.913, respectively. The results indicated that the method of PLS-DA was stable and reliable. PLS-DA scores plot (Figure 3) demonstrated the different clustering pattern as HCA analysis (Figure 2). Group 2 (Figure 3) was same as Cluster 2 (Figure 2) including 10 batches of samples all belong to D. Indica, but the Group 3 (Figure 3) included S21–30, S5 and S6, which was different from Cluster 1-2 (Figure 2). Actually, from the Figure 2, S5 and S6 were categorized into Cluster 1-2 which belong to D. delavayi. Subsequently, PLS-DA loadings plot was analyzed and illustrated in Figure 4. An arbitrary loadings threshold was set on the loadings plot at ±0.10 for w*c (24), ±0.10 for w*c [2]; and at ±0.25 for w*c (24), ±0.20 for w*c [2], highlighted in gray and square, respectively. The selection of the threshold was further verified by a correlation study described below similar to our recently published paper (22). Hence, the variables located outside of the threshold region were regarded as the components contributing most significantly to the categorization of 30 batches of D. dcne leaves. When the loadings threshold was set on the loadings plot at ±0.10 for w*c (24) and ±0.10 for w*c [2] highlighted in gray, 11 components shown in Figure 4 including peak 2, peak 14, peak 16, peak 31, peak 32, peak 38, peak 61, peak 71 and peak 81, peak 35 (isoquercitrin) and peak 40 (phlorizin) were regarded as the marker components. However, when the loadings threshold was set at ±0.25 for w*c (24) and ±0.20 for w*c [2] highlighted in square, there were only two compounds including peak 35 (isoquercitrin) and peak 40 (phlorizin) were shown in Figure 4. This suggests that these two compounds were the biggest contributing factors the categorization of the 30 batches of D. dcne leaves. Combining with the HPLC chromatograms at 285 nm (Figure 1), the preliminary results above showed that the chemical profiling differentiation might be mostly explained by two components including peak 35 (isoquercitrin) and peak 40 (phlorizin). In other words, isoquercitrin and phlorizin are selected as chemical markers to evaluate the quality of D. dcne leaves from the different species, and the quantitative determination of these two chemical markers to evaluate these three different species is necessary. Figure 3. View largeDownload slide PLS-DA scores plot of three different species of D. dcne leaves. Figure 3. View largeDownload slide PLS-DA scores plot of three different species of D. dcne leaves. Figure 4. View largeDownload slide PLS-DA Loading scores plot of three different species of D. dcne leaves. An arbitrary loadings threshold was set on the loadings plot at ±0.10 for w*c [1], ±0.10 for w*c [2]; and at ±0.25 for w*c [1], ±0.20 for w*c [2] highlighted in gray and square, respectively. Figure 4. View largeDownload slide PLS-DA Loading scores plot of three different species of D. dcne leaves. An arbitrary loadings threshold was set on the loadings plot at ±0.10 for w*c [1], ±0.10 for w*c [2]; and at ±0.25 for w*c [1], ±0.20 for w*c [2] highlighted in gray and square, respectively. Quantitative analysis In this study, the content of two flavonoids in 30 batches of D. dcne leaf samples was determined and the data are presented in Table I. The peaks of isoquercitrin and phlorizin in each sample were identified by comparing the retention times and the UV spectra with those of the standards. The content levels of phlorizin and isoquercitrin in 30 batches of D. dcne leaves varied significantly. The results showed that isoquercitrin and phlorizin are the two main chemical markers in the leaf samples, with ranges of 6.41–38.84 and 95.73–217.76 mg/g, respectively. The content of the two flavonoids in different species was significantly different, and the content of phlorizin in all samples was the most abundant, which indicated the phlorizin was the main constitute of D. dcne leaves. Especially, the samples from D. indica had the highest content of phlorizin with the range of 204.23–217.76 mg/g, whereas the samples from D. delavayi and D. longiunguis had the lower content of phlorizin with ranges of 95.73–174.73 and 143.80–179.07 mg/g, respectively. Combining with the PLS-DA loadings plot result (Figure 4) of chemical profiling analysis, it was easily deduced that the difference of the content levels of phlorizin and isoquercitrin was related to the cauterization differences of samples. In this regard, phlorizin and isoquercitrin are the rational marker compounds which represent the comprehensive quality of D. dcne leaves. At the same time, the content levels of isoquercitrin (peak 35) and phlorizin (peak 40) of S5 and S6 were easily similar to the S21–30, which were cauterized into Group 3 (Figure 3). All the results illustrated that the internal quality of 30 batches of D. dcne leaves from different species was variant, and the isoquercitrin and phlorizin should be used as the indicator compounds to evaluate the quality of these different leaves. Discussion In this study, a rapid and efficient fingerprint method and the simultaneous determination of two chemical markers was first developed to evaluate the quality of three different species (D. delavayi, D. Indica and D. Longiunguis) of D. dcne leaves. A total of 30 samples collected from different regions of China were assessed by fingerprint and chemometrics including SA, HCA and PLS-DA. The similarities of the 30 samples ranged from 0.991 to 1.000 based on fingerprint peaks, indicating these three different species of D. dcne leaves have similar chromatographic patterns and the entire chromatograms of these samples were generally consistent and stable. Furthermore, HCA analysis according to the 13 characteristic fingerprint peaks was successfully applied to distinguish these three species. In addition, two chemical marker compounds (isoquercitrin and phlorizin) were identified through PLS-DA. These two markers were quantitatively determined and showed that contents of the two flavonoids in D. dcne leaves displayed notable differences in samples collected from three species. Especially, D. dcne leaves rich in phlorizin with a range from 9.57 to 21.71%. It was found that D. dcne leaves showed higher content levels of phlorizin than the ones in Malus Hupehensis (0.21–1.55%) (25), Lithocarpus Polystachyus (0.78–6.24%) (26) and apple leaves (1.13–4.03%) (27, 28). However, as many reports have revealed that the bioactivity of the herbs is not always consistent with its main peak intensity in the fingerprints, and the “marker” components may be cannot be used in quality control for the herbs (29). For example, various low-content compounds in the herbs may have the synergetic effect for its bioactivities and the effect of these non-common chromatographic peaks could not be excluded (30). D. Dcne leaves have been conventionally used as an important local ethnic tea for obesity or diabetes and ethnomedicines for anti-inflammation medicine (11). In this study, two components including isoquercitrin and phlorizin were selected as chemical markers to evaluate the quality of the D. dcne leaves from different species. Specially, isoquercitrin (31, 32) and phlorizin (33, 34) both have the obviously anti-inflammatory activity in vivo and in vitro. As for anti-obesity or anti-diabetic activity, phlorizin has the significant anti-diabetic activity (34–36) and its analog such as empagliflozin or canagliflozin has been used in clinical application, while the isoquercitrin not. So, we think these two components contribute to its anti-inflammatory activity, and the phlorizin is response for its anti-diabetic activity. To determine the efficacy of other components, more accurate analysis means such as spectrum–effect relationships method are needed, which will be the focus of our further research. Conclusion All the results indicate that the HPLC combination of fingerprint and quantitative analysis of maker components (isoquercitrin and phlorizin) is a powerful and practical tool for evaluating the quality of D. dcne leaves. Briefly, 30 batches of D. dcne leaves could be successfully divided into three groups, and showed good similarity on chemical constituents according to the results of chemometric analysis. In addition, isoquercitrin and phlorizin could be selected as chemical markers to evaluate the quality of D. dcne leaves with different sources. Our work also demonstrated that the D. dcne leaves, especially D. indica leaves are the potential natural resources of phlorizin and could be applied in functional food and medicine in the future. Supplementary Data Supplementary material is available at Journal of Chromatographic Science online. References 1 Commission of Chinese academy of sciences ; Flora of China, Vol. 9 , Science Press , Beijing , ( 2003 ), pp. 170 – 171 . 2 Deng , X. , Zhao , X. , Lan , Z. , Jiang , J. , Yin , W. , Chen , L. ; Anti-tumor effects of flavonoids from the ethnic medicine Docynia delavayi (Franch.) Schneid. and its possible mechanism ; Journal of Medicinal Food , ( 2014 ); 17 ( 7 ): 787 – 794 . Google Scholar CrossRef Search ADS PubMed 3 Sharma , P.B. , Handique , P.J. , Devi , H.S. ; Antioxidant properties, physico-chemical characteristics and proximate composition of five wild fruits of Manipur, India ; Journal of Food Science and Technology , ( 2015 ); 52 ( 2 ): 894 – 902 . Google Scholar CrossRef Search ADS PubMed 4 Luo , Q. , Liu , J.L. , Cai , G.Z. ; Docynia longiunguis Q.Luo et J.L.Liu, a new species of Docynia Dcne. from China ; Bulletin of Botanical Research , ( 2011 ); 31 ( 31 ): 389 – 391 . 5 Lin , Y.F. , Yi , Z. , Zhao , Y.H. ; Chinese Dai medicine colorful illustrations . Yunnan National Press, Kunming , ( 2003 ); p. 380 . 6 Guan , Z. , Tan , S.D. , Su , J.Y. , Zeng , L.M. ; Flavones from the bark of Docynia delavayi (Franch.) Schneid ; Natural Product Research and Development , ( 2000 ); 12 ( 12 ): 34 – 37 . 7 Nguyen , T.T.L. , Hoang , T.M.T. , Vu , T.H.T. , Chu , L.L. , Le , M.H. , Do , N.L. ; Anti-obesity and body weight reducing effect of Docynia indica(Wall.) Decne fruit extract fractions in experimentally obese mice ; VNU Journal of Science, Natural Sciences and Technology , ( 2011 ); 27 ( 2 ): 125 – 133 . 8 Hazarika , T.K. , Lalramchuana , Nautiyal , B.P. ; Studies on wild edible fruits of Mizoram, India used as ethno-medicine ; Genetic Resources and Crop Evolution , ( 2012 ); 59 ( 8 ): 1767 – 1776 . Google Scholar CrossRef Search ADS 9 Zhao , X. , Shu , G. , Chen , L. , Mi , X. , Mei , Z. , Deng , X. ; A flavonoid component from Docynia delavayi (Franch.) Schneid represses transplanted H22 hepatoma growth and exhibits low toxic effect on tumor-bearing mice ; Food and Chemical Toxicology , ( 2012 ); 50 ( 9 ): 3166 – 3173 . Google Scholar CrossRef Search ADS PubMed 10 Liu , G. , Wu , J. , Zhu , M.J. , Zhang , X.Y. , Zhang , H. ; On evaluation of antioxidant activities and screening of potential antioxidant compounds from Docynia delavayi (Franch.) Schneid. leaves ; Journal of Southwest China Normal University (Natural Science Edition) , ( 2014 ); 39 ( 39 ): 73 – 81 . 11 Yuan , W. , Li , W.X. , Lin , Q. , Yang , Z.S. ; Studies on resource and utilization of Docynia in Yunnan ; Journal of Yunnan Agrieultural University , ( 1995 ); 10 ( 3 ): 256 – 258 . 12 Liang , X.M. , Jin , Y. , Wang , Y.P. , Jin , G.W. , Fu , Q. , Xiao , Y.S. ; Qualitative and quantitative analysis in quality control of traditional Chinese medicines ; Journal of Chromatography A , ( 2009 ); 1216 ( 11 ): 2033 – 2044 . Google Scholar CrossRef Search ADS PubMed 13 Jin , X.F. , Lu , Y.H. , Wei , D.Z. , Wang , Z.T. ; Chemical fingerprint and quantitative analysis of Salvia plebeia R.Br. by high-performance liquid chromatography ; Journal of Pharmaceutical and Biomedical Analysis , ( 2008 ); 48 ( 1 ): 100 – 104 . Google Scholar CrossRef Search ADS PubMed 14 Li , W. , Deng , Y. , Dai , R. , et al. . ; Chromatographic fingerprint analysis of Cephalotaxus sinensis from various sources by high-performance liquid chromatography-diodearray detection-electrospray ionization-tandem mass spectrometry ; Journal of Pharmaceutical and Biomedical Analysis , ( 2007 ); 45 ( 1 ): 38 – 46 . Google Scholar CrossRef Search ADS PubMed 15 Alaerts , G. , Matthijs , N. , Smeyers-Verbeke , J. , Vander Heyden , Y. ; Chromatographic fingerprint development for herbal extracts: a screening and optimization methodology on monolithic columns ; Journal of Chromatography A , ( 2007 ); 1172 ( 1 ): 1 – 8 . Google Scholar CrossRef Search ADS PubMed 16 Liu , H. , Liu , J. , Zhang , J. , et al. . ; Simultaneous quantitative and chemical fingerprint analysis of Receptaculum nelumbinis based on HPLC-DAD-MS combined with chemometrics ; Journal of Chromatographic Science , ( 2016 ); 54 ( 4 ): 618 – 624 . Google Scholar CrossRef Search ADS PubMed 17 Shu , Z. , Li , X. , Rahman , K. , Qin , L. , Zheng , C. ; Chemical fingerprint and quantitative analysis for the quality evaluation of Vitex negundo seeds by reversed-phase high-performance liquid chromatography coupled with hierarchical clustering analysis ; Journal of Separation Science , ( 2016 ); 39 ( 2 ): 279 – 286 . Google Scholar CrossRef Search ADS PubMed 18 Cheng , T.F. , Jia , Y.R. , Zuo , Z. , et al. . ; Quality assessment of traditional Chinese medicine herb couple by high-performance liquid chromatography and mass spectrometry combined with chemometrics ; Journal of Separation Science , ( 2016 ); 39 ( 7 ): 1223 – 1231 . Google Scholar CrossRef Search ADS PubMed 19 Lan , K. , Zhang , Y. , Yang , J. , Xu , L. ; Simple quality assessment approach for herbal extracts using high performance liquid chromatography-UV based metabolomics platform ; Journal of Chromatography A , ( 2010 ); 1217 ( 8 ): 1414 – 1418 . Google Scholar CrossRef Search ADS PubMed 20 Zhang , Y. , Peng , M. , Liu , L. , Shi , S. , Peng , S. ; Screening, identification and potential interaction of active compounds from Eucommia ulmodies leaves binding with bovine serum albumin ; Journal of Agricultural and Food Chemistry , ( 2012 ); 60 ( 12 ): 3119 – 3125 . Google Scholar CrossRef Search ADS PubMed 21 Drug Administration Bureau of China . , Requirements for Studying Fingerprint of Traditional Chinese Medicine Injections (Draft). ( 2000 ). 22 Wang , Z. , Hu , H. , Chen , F. , et al. . ; Metabolic profiling assisted quality assessment of Rhodiola rosea extracts by high-performance liquid chromatography ; Planta Medica , ( 2012 ); 78 ( 7 ): 740 – 746 . Google Scholar CrossRef Search ADS PubMed 23 Kong , W. , Li , Z. , Xiao , X. , Zhao , Y. ; Quality control for Coptidis rhizoma through the determination of five alkaloids by HPLC-ELSD coupled with chemometrics ; Natural Product Research , ( 2010 ); 24 ( 17 ): 1616 – 1629 . Google Scholar CrossRef Search ADS PubMed 24 Chase , H.P. , Boulware , D. , Rodriguez , H. , et al. . ; Effect of docosahexaenoic acid supplementation on inflammatory cytokine levels in infants at high genetic risk for type 1 diabetes ; Pediatric Diabetes , ( 2015 ); 16 ( 4 ): 271 – 279 . Google Scholar CrossRef Search ADS PubMed 25 Guo , D.Y. , Li , J. , Shi , Y.Q. , Wang , J.Z. ; Study on the flavon ingredients of Malus hupehensis ; Chinese Herbal Medicine , ( 2011 ); 34 ( 34 ): 1026 – 1029 . 26 He , C. , Peng , Y. , Xiao , W. , Hu , Y. , Xiao , P. ; Quick determination of five sweet constituents in Duosuike tiancha by RSLC ; China Journal of Chinese Materia Medica , ( 2012 ); 37 ( 7 ): 961 – 965 . Google Scholar PubMed 27 Zhao , Y.N. , Wang , J. , Song , G.M. , Song , X.B. , Liu , D.N. , Xie , W.L. ; Study on the dynamic changs of phlordzin and phloretin in leave and branches of apple tree ; Food Research and Development , ( 2013 ); 34 ( 34 ): 95 – 98 . 28 Li , R.T. , Liu , J.C. , Jiao , Z.G. , Wang , S.X. , Du , X.F. ; Determination of phlorhizin in apple tree branch and leaf by RP-HPLC ; Food Industry Reseach , ( 2009 ); 30 ( 30 ): 385 – 388 . 29 Xu , G.L. , Xie , M. , Yang , X.Y. , et al. . ; Spectrum-effect relationships as a systematic approach to traditional Chinese medicine research: current status and future perspectives ; Molecules , ( 2014 ); 19 ( 11 ): 17897 – 17925 . Google Scholar CrossRef Search ADS PubMed 30 Du , W.J. , Ji , J. , Wang , L. , et al. . ; Relationship between the UPLC-Q-TOF-MS fingerprinted constituents from Daphne genkwa and their anti-inflammatory, anti-oxidant activities ; Biomedical Chromatography , ( 2017 ); 31 ( 12 ): 1 – 9 . Google Scholar CrossRef Search ADS 31 Rogerio , A.P. , Kanashiro , A. , Fontanari , C. , et al. . ; Anti-inflammatory activity of quercetin and isoquercitrin in experimental murine allergic asthma ; Inflammation Research , ( 2007 ); 56 ( 10 ): 402 – 408 . Google Scholar CrossRef Search ADS PubMed 32 Xie , W. , Wang , M. , Chen , C. , Zhang , X. , Melzig , M.F. ; Hepatoprotective effect of isoquercitrin against acetaminophen-induced liver injury ; Life Sciences , ( 2016 ); 152 : 180 – 189 . Google Scholar CrossRef Search ADS PubMed 33 Chang , W.T. , Huang , W.C. , Liou , C.J. ; Evaluation of the anti-inflammatory effects of phloretin and phlorizin in lipopolysaccharide-stimulated mouse macrophages ; Food Chemistry , ( 2012 ); 134 ( 2 ): 972 – 979 . Google Scholar CrossRef Search ADS PubMed 34 Shin , S.K. , Cho , S.J. , Jung , U.J. , Ryu , R. , Choi , M.S. ; Phlorizin supplementation attenuates obesity, inflammation, and hyperglycemia in diet-induced obese mice fed a high-fat diet ; Nutrients , ( 2016 ); 8 ( 2 ): 92 . Google Scholar CrossRef Search ADS PubMed 35 Rossetti , L. , Smith , D. , Shulman , G.I. , Papachristou , D. , DeFronzo , R.A. ; Correction of hyperglycemia with phlorizin normalizes tissue sensitivity to insulin in diabetic rats ; Journal of Clinical Investigation , ( 1987 ); 79 ( 5 ): 1510 – 1515 . Google Scholar CrossRef Search ADS PubMed 36 Mei , X. , Zhang , X. , Wang , Z. , et al. . ; Insulin sensitivity-enhancing activity of phlorizin is associated with lipopolysaccharide decrease and gut microbiota changes in obese and type 2 diabetes (db/db) mice ; Journal of Agricultural and Food Chemistry , ( 2016 ); 64 ( 40 ): 7502 – 7511 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Journal of Chromatographic ScienceOxford University Press

Published: May 24, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off