Résumé
OBJECTIVE To establis h the fingerprint of Cynanchum auriculatum，to conduct its chemical pattern recognition analysis，and to determine the contents of four components at the same time. METHODS High performance liquid chromatography （HPLC）method was adopted. The determination was performed on ACE UExcel C 18 column with mobile phase consisted of acetonitrile-0.1％ phosphoric acid solution （gradient elution ）at the flow rate of 1.2 mL/min. The determination wavelength was set at 210 nm，and the column temperature was 40 ℃. The sample size wa s 10 μL. Taking qingyang shengenin as the reference ，HPLC fingerprints of 16 batches of C. auriculatum medicinal materials were drawn and similarity was evaluated by using the Similarity Evaluation of Chromatographic Fingerprints of Traditional Chinese Medicine （2012 edition）， and the common peaks were determined. SPSS 26.0 software andSIMCA 14.0 software were used for cluster analysis ，principal component analysis and orthogonal partial least squares- discriminant analysis. The differential components affecting the quality of C. auriculatum were screened by taking the value of variable importance in projection （VIP）greater than 1 as the standard ；same HPLC method was used to determine the contents of syringic acid ，acyl asclepiadelenin ，baishouwubenzophenone and qingyang shengenin. RESULTS There were 29 common peaks in 16 batches of C. auriculatum ，with a similarity of 0.723-0.998. Four common peaks were identified ，namely syringic acid （peak 7），acyl asclepiadoidin （peak 9），baishouwubenzophenone（peak 13）and qingyang shengenin（peak 15）. The results of cluster analysis showed that 16 batches of C. auriculatum could be clustered into three categories ，among which S 1 were grouped into one category，S3 were grouped into one c ategory，S2，and S 4-S16 were grouped into one category. The results of principal component analysis showed that the cumulative variance contribution rate of the five principal components was 88.706％，and the classification results were consistent with the results of cluster analysis. The results of orthogonal partial least squares-discriminant analysis showed that the common peaks （from large to small ）with VIP value greater than 1 were peak 20，peak 10，peak 25，peak 12， peak 15（qingyangshengenin），peak 21，peak 14，peak 16，peak 26，peak 22 and peak 17. The linear ranges of syringic acid，acyl asclepterin，baishouwubenzophenone and qingyangshengenin were 0.715 3-45.778 0，2.379 4-152.281 0，0.642 0- 41.085 0，14.541 6- 930.662 0 µg/mL respectively （all R 2＞0.999）. The quantitative limits were 0.357 7，0.475 9，0.642 0 and 2.423 6 μg/mL；the detective limits were 0.146 0，0.164 1，0.248 8 and 0.833 3 μg/mL，respectively. RSDs of precision ，repeatability and stability （24 h）tests were less than 3％；the average recoveries were 99.11％（RSD＝2.00％，n＝9），98.54％（RSD＝2.21％，n＝9）， 96.33％（RSD＝2.54％，n＝9）and 95.96％（RSD＝2.93％，n＝9）；the contents were 17.12-147.80，95.23-583.10（S8 below the quantitative limit ），16.91-210.88 and 211.68-3 587.15（S1 below the quantitative limit ）μg/g，respectively. CONCLUSIONS Established HPLC fingerprint and the method of content determination are stable ，reliable，accurate and reproducible. Combined with analysis of chemical pattern recognition ，it can be used for the quality control of C. auriculatum .
Résumé
OBJECTIV E To establish the method for evaluating the quality o f Plantago asi atica and fried P. asiatica . METHODS The fingerprints of 15 batches of P. asiatica and 15 batches of fried P. asiatica were established by HPLC. The common peaks were identified with the Similarity Evaluation System for Chromatographic Fingerprinting of TCM （2012 edition）， and similarity evaluation was performed. Analysis of chemical pattern recognition was performed by using SPSS 25.0 and SIMCA-P 14.1 software（cluster analysis ，principal component analysis and orthogonal partial least squares regression analysis ）. The markers which affected the difference in the quality between P. asiatica and fried P. asiatica were screened with variable importance projection（VIP）value greater than 1. RESULTS There were 18 common peaks in the fingerprints of 15 batches of P. asiatica and 13 common peaks in the fingerprints of 15 batches of fried P. asiatica . A total of 8 common peaks were found in both of them. Their similarities were greater than 0.920. Two common peaks were identified as geniposidic acid ，acteoside. The results of cluster analysis showed that when the spacing was 10，the 30 batches of samples could be clustered into three categories ，with S 1-S5 as one，S16-S20 as one ，S6-S15 and S 21-S30 as one . The results of the pri ncipal component analysis showed that the cumulative variance contribution rate of the first two principal components was 82.575％ . The results of the orthogonal partial least squares regression analysis showed that the VIP values of the three common peaks were greater than 1，namely peak E（acteoside）， peak D （geniposidic acid ） and peak G. CONCLUSIONS Established fingerprints are stable ，simple sina.com and rapid. It can be used for the quality evaluation of P. asiatica and fried P. asiatica ，by combining with analysis of chemical pattern recognition. Acteoside ，geniposidic acid and the component represented by peak G may be the markers affecting the difference in quality of P. asiatica and fried P. asiatica .