| Marsdenia tenacissima(Roxb.)Wight et Arm.(Asclepiadaceae)is a perennial climber belonging to the Marsdenia genus in the Asclepiadaceae family,which is well known as"Tong-guang-teng" in Chinese folk medicine and has long been used as a remedy to treat cystitis,tracheitis,tonsillitis and pneumonia.Modern pharmacological experiments showed that the aqueous extractions of M.tenacissima(C21 steroidal glycoside)possess a number of pharmacological properties such as antioxidant,anti-aging and anti-cancer activities.In clinical applications,Tong-guang-teng tablets have been used for the treatment of chronic bronchitis and tracheitis,and Xiao-ai-ping preparations are effective in treating gastric carcinoma,liver cancer and lung cancer.For the outstanding anticancer effect,Xiao-ai-ping has become the preferred clinical formulation.M.tenacissima is only native to the karst landform areas in southwest of China,where the rocky desertification is relatively serious and the soil is quite poor.Almost all the raw medicinal materials depended on cutting the wild resources.Because of the difference in weather conditions,soil characteristics and water quality,the quality of M.tenacissima displayed large differences among the different origins,which was likely to pose a security risk for clinical use.Therefore,it is urgent need to establish origin traceability technology and quality evaluation system for M.tenacissima.Origin traceability system of traditional Chinese medicine is effective measure for ensuring the safety of medicinal materials,protecting the geography trademark medicinal materials,and enhancing confidence of patients.If the origins of medicinal plants can’t be distinguish correctly,the concepts,such as "geoherbs","geographical indication medicinal materials",and "high-quality region" were meaningless.Therefore,our study not only of great significance for the origin identification and quality evaluation of M.tenacissima,but also provided a reference of origin traceability for other herbs.To explore the feasibility of different fingerprint technology for tracing the origins of M.tenacissima samples,we investigated the wild resources and collected large number of samples in Yunnan,Guizhou,Guangxi and Sichuan province.We analyzed 27 elements in M.tenacissima samples collected from different regions by inductively coupled plasma-atomic emission spectroscopy,explored the infrared spectrum characteristics of M.tenacissima samples in different regions by FT-NIR/FT-IR,determined the chemical composition of M.tenacissima samples in different regions by HPLC,then combined the pattern recognition methods(support vector machine-grid search,random forest,radial basis function neural network,k-nearest neighbors,support vector machine-particle swarm optimization and support vector machine-genetic algorithms)and spectra preprocessing methods(multiplicative scatter correction,standard normal variate,wavelet domain denoising,derivative)to construct the discrimination models for tracing geographical origins and controlling the quality.Our study gets the following conclusions.(1)M.tenacfissima is one of typical medicinal plants of regional distribution,its wild resources had been severely damaged.The distribution areas and resource reserves of M.tenacissima displayed a more significant decline,which was urgent need for habitat protection and sustainable development.M.tenacissima is mainly grown in karst area with warm climate and good sunshine,the altitude range from 500 m to 1800 m,and have certain selectivity in slope degree and slope direction(suitable for steep and sunny slopes).There is no typical associated plant and climbing plant,and also no clear requirements on growth of soil type and fertility conditions.The biological properties,such as leaf,plant height,growth increment,et al.,exhibit big difference among individuals,however,with the increase of growth period,the various indicators tend to be stable(substantially more than 5 years).Through detailed investigation and estimation,we calculate the resources reserves of M.tenacissima in Xishuangbanna city was 529.2 ton,the resources reserves in Puer city was 175.7 ton,the resources reserves in Baise city was 349.5 ton,the resources reserves in Xichang city was 0.48 ton.Resource reserves of M.tenacissima is relatively scarce,it is hard to meet the increasing market demand.In breeding science,M.tenacissima exists the phenomenon of flower but fruitless,almost no seedlings grown in the vicinity.However,the asexual reproduction ability of M.tenacissima is stronger,and there are many adventitious roots at internode.Therefore,we adviced screening high-quality source and establishing the asexual reproduction system for M.tenacissima.(2)27elements(Sn、Be、AS、Na、Mg、Al、K、Ca、V、Cr、Mn、Fe、Co、Bi、Ni、Hg、Cu、Zn、Se、Si、Cd、Sb、Ba、Ti、Pb、Zr and Sr)in 128 samples from four provinces of China were analyzed by inductively coupled plasma-atomic emission spectroscopy(ICP-AES).Pattern recognition techniques,including principal component analysis(PCA),stepwise linear discriminant analysis(SLDA),k-nearest neighbor analysis(KNN),radial basis function neural network(RBFNN),support vector machine(SVM)and random forest(RF)were used for tracing the geographical source of these samples.Results showed that most of elements in the M.tenacissima samples from different regions showed significant differences,which carrying geographical features.The order of successful identification rates is as follows:SVM-grid>SVM-ga>RF>RBF-NN>SLDA>SVM-pso.The results revealed that the SVM-grid model was superior to all the other mathematical methods with an average discrimination rate of 100%for the test set.(3)A combination of FT-IR with chemometric tools provided an approach for studying M.tenacissima according to its geographical origin.A total of 128 M.tenacissima samples from four provinces in China were analyzed with FTIR.Six pattern recognition methods(support vector machine-grid search,random forest,radial basis function neural network,K-nearest neighbors,support vector machine-particle swarm optimization and support vector machine-genetic algorithms)and spectra preprocessing methods(multiplicative scatter correction,standard normal variate,wavelet domain denoising,1st derivative and 2nd derivative)were used to construct the discrimination models used to trace geographical origins.Experimental results showed K-nearest neighbors was superior to other mathematical algorithms after data were preprocessed with wavelet denoising,with a discrimination rate of 100%in both the training and prediction sets.(4)FT-NIR analysis was used to analyze the characteristics of NIR spectra of M.tenacissima samples from different origins.A combination of data mining techniques(characteristic peaks analysis,principal component analysis,support vector machine,et al.)and spectra preprocessing methods(standard normal variate correction,wavelet domain de-noising,derivative,et al.)provided an approach for analyzing the difference of NIR spectra,and investigated the possibility of using FT-NIR spectroscopy together with different chemometric methods to build a rapid,simple and non-invasive analytical procedure for tracing the geographical origin of M.tenacissima samples.The results revealed that there were some differences in NIR spectra of M.tenacissima samples from different origins.The order of identification rate as follows:SVM-grid>KNN>SVM-ga>SVM-pso>RF>RBF-NN.SVM-grid model was superior to all the other mathematical methods with an average discrimination rate of 96.4%for the test set.(5)The contents of tenacissoside G,tenacissoside H and chlorogenic acid in M.tenacissima were determined with HPLC-EDSD/DAD.We utilized principal component analysis,regression analysis and correlation analysis to reveal the relationship among chemical composition,environmental factors and soil properties.Results indicates that the difference of tenacissoside H was most obvious,the highest was Sichuan Dechang city,with the content of 0.1410%;while,the lowest was Yunnan Mengla city,with the content of 0.0159%,which showing a difference of nearly tenfold.The difference of tenacissoside G was also obvious,the highest was Sichuan Dechang city,with the content of 0.0980%;while,the lowest was Guizhou Xingyi city,with the content of 0.0334%,which showing a difference of nearly triple.The content of chlorogenic acid was ranged from 0.2800%to 0.7427%,the highest was Sichuan Dechang city,while,the lowest was Yunnan Ninger city.Correlation analysis showed that the content of tenacissoside G was significant negative correlation with MAP;the content of tenacissoside H was significant negative correlation with MAP and RH,and significant positive correlation with latitude;the content of chlorogenic acid was significant positive correlation with Cu.Stepwise regression analysis showed that MAP was the forecast variables for the content of tenacissoside G;RH and TMainW was the forecast var:iables for the content of tenacissoside H;the content of Cu was the forecast variables for chlorogenic acid.(6)Detection methods of the content of flavonoids,polysaccharides and total phenolic in M.tenacissima were established in this study,moreover,principal component analysis,regression analysis and correlation analysis were applied for revealing the relationship among chemical composition,environmental factors and soil properties.Results indicates that the contents of flavonoids,polysaccharides and total phenolic were 3.34%～9.80%,1.59%～9.10%and 2.80%～7.43%,respectively,which illustrated the differences between regions were obvious.Correlation analysis showed that the content of flavonoids was significant negative correlation with MAP;the content of polysaccharides was significant negative correlation with MAP and RH,and significant positive correlation with latitude and the content of Cu.Stepwise regression analysis showed that MAP was the forecast variables for the content of flavonoids and polysaccharides;the content of Cu was the forecast variables for total phenolic. |
