HPLC Coupled with Chemical Fingerprinting for Multi-Pattern Recognition for Identifying the Authenticity of Clematidis Armandii Caulis

Summary

Here, we present a protocol to establish high-performance liquid chromatography (HPLC), coupled with chemical fingerprint multi-pattern recognition, which provides a new strategy for effectively identifying the genuine varieties of Clematidis Armandii Caulis and its adulterants.

Abstract

A method for identifying Chinese medicinal materials and their related adulterants was constructed by taking Clematidis Armandii Caulis (Chuanmutong, a universally used traditional Chinese medicine) as an example. Ten batches of genuine Chuanmutong varieties and five batches of related adulterants were analyzed and compared based on the high-performance liquid chromatography (HPLC) fingerprints combined with chemometrics, including cluster analysis (CA), principal component analysis (PCA), and orthogonal partial least-squares discrimination analysis (OPLS-DA). In addition, the content of β-sitosterol was determined. The control chemical fingerprint of Chuanmutong was established, and 12 common peaks were identified. The similarity between the fingerprint of 10 batches of genuine Chuanmutong varieties and the control fingerprint was 0.910-0.989, while the similarity of five batches of adulterants was only 0.133-0.720. Based on the common peaks in the chromatogram, 15 batches of samples were classified into three content levels by PCA, and were aggregated into four categories by CA, achieving a clear distinction between authentic Chuanmutong and adulterants of Chuanmutong. Further, seven differential components that can effectively identify authentic Chuanmutong and adulterants of Chuanmutong were found through OPLS-DA. The β-sitosterol content of 10 batches of genuine Chuanmutong varieties was 97.53-161.56 µg/g, while the β-sitosterol content of the five batches of adulterants varied greatly, among which the β-sitosterol content of Clematis peterae Hand.-Mazz. and Clematis gouriana Roxb. Var. finetii Rehd. et Wils. was significantly lower than that of authentic varieties of Chuanmutong. The HPLC index component content and chemical fingerprint multi-pattern recognition method established in this study provide a new strategy for effectively identifying authentic Chinese medicinal materials and related adulterants.

Introduction

Chuanmutong, dry Caulis of Clematis armandii Franch. or Clematis montana Buch.-Ham., is a traditional Chinese medicine commonly used in clinics^1,2,3. It is used for treating urinary problems, edema, sores on the tongue and mouth, decreased milk secretion, joint stiffness, and muscle pain caused by damp heat⁴. Chuanmutong has always been obtained from wild varieties, mainly distributed in southwest China, especially in Sichuan, where the best quality can be found^5,6. It is difficult to distinguish between authentic varieties and their closely related adulterants due to their similar characteristics^7,8,9,10. The quality standard of Chuanmutong in the 2020 edition of Chinese Pharmacopoeia only stipulates the properties, microscopic identification, and thin-layer identification without content determination, which cannot effectively identify adulterants, and hence has potential risks. Moreover, there are few reports comparing and identifying Chuanmutong and related plants. Consequently, a quality control method to ensure the authenticity of Chuanmutong is worthy of further study.

The chemical constituents of Chuanmutong are mainly composed of oleanane-type pentacyclic triterpenoids and their glycosides, flavonoids, and organic acids^11,12,13,14. Among them, oleanolic acid, β-sitosterol, stigmasterol, and ergosterol have diuretic effects of different intensities, which may be potential pharmacodynamic substances for promoting diuresis and relieving stranguria^15,16. Chemical fingerprints are obtained by separating and detecting many chemical components contained in samples by high-performance liquid chromatography (HPLC), gas chromatography (GC), etc. Adopting appropriate statistical analysis methods to analyze the characteristics of Chuanmutong can determine the overall quality control and scientific identification of traditional Chinese medicine^17,18,19.

In this study, 10 batches of Chuanmutong authentic varieties and five batches of adulterants were collected. Their quality was compared and analyzed by the HPLC fingerprint method combined with multi-pattern recognition, including cluster analysis (CA), principal component analysis (PCA), orthogonal partial least-squares discrimination analysis (OPLS-CA), and content determination of the pharmacodynamic component. This protocol establishes a method for identifying authentic varieties with high specificity, a new strategy for the scientific identification of authentic varieties and adulterants of Chinese medicinal materials.

Protocol

1. Methods for chemical fingerprint detection

1. Chromatographic conditions
   1. Prepare the acetonitrile (A)/water (B) mobile phase. Set a gradient program as follows in the HPLC software: 0-20 min, 3%A-10%A; 20-25 min, 10%A-13%A; 25-65 min, 13%A-25%A; 65-75 min, 25%A-40%A; 75-76 min, 40%A-3%A; 76-85 min, 3%A-3%A.
   2. Maintain the flow rate of the mobile phase at 1.0 mL/min.
   3. Conduct the chromatographic separation on a C18 column (250 mm x 4.6 mm, 5 µm) maintained at 30 °C.
   4. Set the injection volume to 10 µL.
   5. Detect the samples at a wavelength of 205 nm.
      NOTE: For the specific settings of chromatographic conditions, refer to the operating procedures of the working software of high-performance liquid chromatography (Table of Materials).
2. Preparation of the sample solution
   1. Grind the raw materials to uniform particle size by passing them through a nylon mesh with an inner diameter of 850 µm ± 29 µm.
   2. Place 2 g of the ground raw material (accurately weighed) in a 50 mL conical flask with a stopper and add 50 mL of methanol. Place the stopper on the flask and ultrasonicate (600 W, 40 kHz) for 30 min.
   3. Then, cool the flask to room temperature (RT). Weigh the samples again, and make up for the initial weight by replacing the lost extractant.
   4. Pour 4 mL of the methanol solution containing the medicinal extracts into a 10 mL volumetric flask. Add 6 mL of H₂O, mix, and allow it to settle for 10 min.
   5. Finally, filter the supernatant through a 0.45 µm filter membrane and place it on standby.
3. Validation of fingerprint detection methods
   1. Prepare the sample as described above (step 1.2) and subject it to HPLC analysis (step 1.1) six times a day. To evaluate the precision, calculate the relative standard deviation (RSD) of relative retention time and relative peak areas as described in step 1.3.5.
   2. Evaluate the stability of the sample solution by analyzing the same sample solution stored at RT for 0, 2, 4, 6, 8, 12, and 24 h, and calculate the RSD of relative retention time and relative peak areas as described in step 1.3.5.
   3. Take six replicates of the same sample (CMT-4), prepare the sample solution according to the above procedure (step 1.2), and detect its fingerprint in HPLC following step 1.1. Calculate the RSD of relative retention time and relative peak areas, and evaluate its repeatability as described in step 1.3.5.
   4. Then use peak number 10 in Figure 1B as the reference peak and calculate the RSD of the relative retention time and relative peak area of each common peak as described in step 1.3.5.
   5. Use the formulas mentioned below to calculate the relative retention time and relative peak area of each common peak:
      T_re = T_{characteristic}/T_{reference
      Are = Acharacteristic/Areference}
      Where T_re =_{relative retention time,} T_{characteristic} = characteristic peak retention time, T_reference = reference peak retention time, A_re = relative peak area, A_{characteristic} = characteristic peak area, and A_reference = reference peak area.
      ​NOTE: The establishment of traditional Chinese medicine fingerprints generally requires selecting a chromatographic peak that is easy to obtain and has high resolution. This is used as a reference peak to identify the fingerprints and examine their stability and reproducibility.

2. Establishment of Chuanmutong fingerprint and similarity analysis

1. Use 10 batches of authentic samples and five batches of adulterants such as Clematis argentilucida (Levl. et Vant.) W. T. Wang (CC), Clematis apiifolia var. obtusidentata Rehd. et Wils. (DC), Clematis peterae Hand.-Mazz. (DE), Clematis gouriana Roxb. Var. finetii Rehd. et Wils (XS), and Clematis finetiana Levl. et Vaniot. (SMT) as samples for fingerprint analysis.
2. Prepare the sample solutions as described in step 1.2. Perform fingerprint analysis of all sample solutions by HPLC according to the conditions described under step 1.1.
3. Import the relevant data into the similarity evaluation system of chromatographic fingerprints of traditional Chinese medicine (SESCF-TCM, 2012 version).The system will designate the peaks with reasonable height and good resolution in the chromatograms of all samples as common peaks.
   NOTE: The SESCF-TCM software can be downloaded after registration on the website of the Chinese Pharmacopoeia Commission (http://114.247.108.158:8888/login).
   1. In the software, click the Set Reference Spectrum button in the menu.
   2. Then in the Parameter Settings window, set the Time Window Width to 0.5 and select Control Spectrum Generation Method as the Median Method.
   3. Click on Multi-point Calibration in the main menu, then select Peak Matching as Full Spectrum Peak Matching.
   4. Finally, click on Generate Control to generate the reference chromatographic fingerprint of the authentic species of Chuanmutong.
4. Import the retention time and peak area of 10 batches of authentic Chuanmutong samples and five batches of adulterants into SESCF-TCM for analysis. The specific operations are as follows:
   1. In the software, click on the Set Reference Spectrum button in the main menu.
   2. In the Parameter Settings window, set the reference chromatographic fingerprint of the authentic species of Chuanmutong as the reference, select the Control Spectrum Generation Method as the Median Method, and set the Time Window Width to 0.5.
   3. Click on Multi-point Calibration in the main menu, then select Peak Matching as Full Spectrum Peak Matching.
   4. Finally, click on Calculate Similarity to calculate the similarity based on the reference chromatogram fingerprints of Chuanmutong. Finally, calculate the similarity of fingerprints using the Chinese Medicine Chromatographic Fingerprint Evaluation System (2012 version).
      ​NOTE: For the specific operations, refer to the operating specifications for the Chinese Medicine Chromatographic Fingerprint Evaluation System (2012 version).

3. Multi-pattern recognition analysis of Chuanmutong fingerprint

1. Cluster analysis (CA)
   1. Use the peak areas of 12 common peaks in the fingerprints of 10 batches of authentic Chuanmutong samples and their five batches of adulterants as variables, and input them into statistical analysis software for systematic cluster analysis (CA).
   2. Choose the Between-Groups method and use the Pearson correlation coefficient as the classification basis to draw a cluster analysis diagram of Chuanmutong and its adulterants. The specific operations are as follows:
      1. In the statistical analysis software, click on File to import data.
      2. Click on Analysis in the menu and then click on System Clustering in Classification.
      3. Select the common peak area as a variable, and set the number of clusters to four.
      4. Click on Method, select the clustering method as Inter-Group Connection, select the measurement interval as Pearson Correlation, and click on OK to draw the CA map.
2. Principal component analysis (PCA)
   1. Import the relative common peak area of the authentic varieties and their adulterants into the analysis software for PCA analysis, and use the PCA score map to evaluate the score matrix map of sample differences. The specific operations are as follows:
      1. Open the data analysis software, click on File on the menu and create a new regular project. Import the peak area of 12 common peaks in a spreadsheet (e.g., excel format) from the HPLC system. Then click on Finish to complete the data import.
      2. Click on New to create a new model to set the model type with PCA. Click on Autofit and Add to fit the data, then click on Scores to get the PCA score map.
3. Orthogonal partial least-squares discrimination analysis (OPLS-DA)
   1. Use the Orthogonal partial least-squares discrimination analysis method with supervision mode to further analyze the relative common peak area peaks of the authentic Chuanmutong varieties and adulterants and draw an OPLS-DA classification score map of all samples. The specific operations are as follows:
      1. In data analysis software, click on File in the menu to import a file and create a new regular project. Import the peak area of 12 common peaks in a spreadsheet from the HPLC system, then click on Finish to complete the data import.
      2. Click on New to create a new model to set the model type with PCA. Click on Autofit and Add to fit the data. Then click on Scores to get the PCA score map.
      3. Click on New and choose New as Model One to set the model type with OPLS-DA.
      4. Click on Scale and set type with Par for All. Click on Autofit first and then click on Scores to get the OPLS-DA score map.
   2. In order to determine the influence of each common peak in Chuanmutong on its classification results and the difference between authentic Chuanmutong materials and related adulterants, use the variable importance in the projection (VIP) for analysis.
   3. Draw the VIP map of the different components of Chuanmutong. Use the resulting VIP map to assess the impact of each variable on the classification and to screen out components that contribute significantly to the differences between groups. The specific operations are as follows:
      1. In the data analysis software, click on Analyse in the menu and click on Permutations, set the number of permutations to 200, and get the R² and Q² of the OPLS-DA score map.
      2. Click on VIP and choose VIP Predictive to get the VIP map.

4. Determination of β-sitosterol in Chuanmutong by HPLC

1. Chromatographic conditions (refer to step 1.1)
   1. Prepare the mobile phase: methanol-water (97:3).
   2. Set the flow rate of the mobile phase to 1.0 mL/min.
   3. Conduct the chromatographic separation on a C18 column (250 mm x 4.6 mm, 5 µm) maintained at 30 °C.
   4. Set the injection volume to 10 µL.
   5. Detect the component at a wavelength of 204 nm.
2. Preparation of the sample solution
   1. Prepare the stock standard solution of β-sitosterol (0.1 mg/mL) by dissolving an accurately weighed quantity of the corresponding reference standard in methanol.
   2. Grind the analysis sample of the raw material to uniform particle size by passing the sample through the nylon mesh with an inner diameter of 180 µm ± 7.6 µm.
   3. Place 2 g of the ground raw material (accurately weighed) in a round-bottomed flask and add 50 mL of chloroform to it.
   4. Connect the flask to a reflux condenser and heat it in a boiling water bath (moderate boiling) for 60 min. Filter the extraction solution with a 15-20 µm filter paper.
   5. Evaporate the filtrate to near dryness on a boiling water bath (moderate boiling) for about 10 min.
   6. Dissolve the residue and make up the volume to 5 mL using methanol. Finally, pass the supernatant through a 0.45 µm filter membrane, and place it on standby.
3. Method validation
   1. Take the stock solution of β-sitosterol prepared in sub-step 4.2.1, dilute it with 100% methanol, and prepare solutions with 100 µg/mL, 80 µg/mL, 60 µg/mL, 50 µg/mL, 40 µg/mL, 30 µg/mL, and 20 µg/mL concentrations.
   2. Inject the samples under the chromatographic conditions described in step 4.1 to determine the peak area, perform regression analysis with the peak area to the injection volume, and obtain the regression equation and correlation coefficient to evaluate its linearity.
   3. Prepare the samples as described above (step 4.2) and subject them to HPLC analysis (step 4.1) six times on the same day. Then calculate the RSD of peak areas to evaluate the precision.
   4. Evaluate the stability of the sample solution by analyzing the same sample solutions stored at RT for 0, 2, 4, 6, 8, 12, and 24 h, as described in step 4.1. Then calculate the RSD of peak areas as described in step 1.3.5.
   5. Examine the repeatability by dissolving the same sample (CMT-4) in sextuplicate, prepared as described in step 4.2, and subjecting them to HPLC analysis as described in step 4.1. Then calculate the RSD of β-sitosterol content in six samples.
   6. Assess the method's accuracy by employing the standard addition method. For this, add β-sitosterol reference solutions to the samples at 80%, 100%, and 120% of the β-sitosterol content and repeat each condition three times as described in step 4.1. Evaluate the method's accuracy by calculating the average recovery and RSD.
      NOTE: The calculation formula of the rate of recovery (RR) is as follows:
      RR % = [(M_t - M₀) / M_s] × 100
      Where M_t = the quality of β-sitosterol after adding the standard, M₀ = the quality of the sample solution, and M_s = the quality of β-sitosterol added.
4. Determination of β-sitosterol content of samples
   1. Take 10 batches of authentic Chinese medicinal materials and five batches of related adulterants to prepare sample solutions according to step 4.2.
   2. Then inject each sample solution and β-sitosterol reference solution to determine the peak area under the conditions described in step 4.1, and calculate the β-sitosterol content of each sample by using the external standard one-point method.

Results

Chromatographic fingerprint of Chuanmutong and similarity analysis (SA)
The RSD values of the relative retention time of precision, repeatability, and stability were below 0.46%, 1.65%, and 0.53%, respectively; the RSD values of the relative peak area were below 4.23%, 3.56%, and 3.96%, respectively. As shown in Figures 1A,B, there were 12 distinct common peaks (from peak 1 to peak 12) in the HPLC fingerprints in the 10 authentic Chuanmutong samples. Sin...

Discussion

The sample collection for research is the first key step to constructing multi-pattern recognition in identifying the authenticity of Chinese medicinal materials. Through market research, we found that Sichuan Ya'an, Liangshan, and Leshan are the main production areas of wild resources of Chuanmutong. The related varieties of the same genus also have the same geographical distribution^6,20; CC, DC, DE, XS, and SMT are often misused as Chuangmutong

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Acetic acid	Zhiyuan Chemical Reagent Co., Ltd., Tianjin, China	2017381038
Acetonitrile	Sigma-Aldrich  Trading Co., Ltd., Shanghai, China	WXBD5243V
β-Sitosterol	Meisai Biological Technology Co., Ltd., Chongqing, China	20210201
C18 column	Yuexu Material Technology Co., Ltd., Shanghai, China	Welch Ultimate LP
Chuanmutong	Guoqiang Chinese Herbal Pieces Co., Ltd., Sichuan, China	19020103	CMT-1
Chuanmutong	Hongya Wawushan Pharmaceutical Co., Ltd., Sichuan, China	200701	CMT-2
Chuanmutong	Hongpu Pharmaceutical Co., Ltd., Sichuan, China	200701	CMT-3
Chuanmutong	Hongpu Pharmaceutical Co., Ltd., Sichuan, China	200901	CMT-4
Chuanmutong	Xinrentai Pharmaceutical Co., Ltd., Sichuan, China	210701	CMT-5
Chuanmutong	Haobo Pharmaceutical Co., Ltd., Sichuan, China	210401	CMT-6
Chuanmutong	Xinrentai Pharmaceutical Co., Ltd., Sichuan, China	200901	CMT-7
Chuanmutong	Wusheng Pharmaceutical Group Co., Ltd., Sichuan, China	201201	CMT-8
Chuanmutong	Limin Chinese Herbal Pieces Co., Ltd., Sichuan, China	201001	CMT-9
Chuanmutong	Yuhetang Pharmaceutical Co., Ltd., Sichuan, China	210501	CMT-10
Clematis argentilucida (Levl. et Vant.) W. T. Wang	Madzi Bridge, Sanlang Township, Tianquan County, Sichuan, China	-	CC
Clematis apiifolia var. obtusidentata Rehd. et Wils.	Heilin Village, Qiliping Township, Hongya County, Sichuan, China	-	DC
Clematis peterae Hand.-Mazz.	Huangmu Village, Huangmu Township, Hanyuan County, Sichuan, China	-	DE
Clematis gouriana Roxb. Var. finetii Rehd. et Wils	Mixedang Mountain, Huangwan Township, Emei County, Sichuan, China	-	XS
Clematis finetiana Levl. et Vaniot.	Wannian Village, Huangwan Township, Emei County, Sichuan, China	-	SMT
Electronic balance	Haozhuang Hengping Scientific Instrument Co., Ltd., Shanghai,  China	FA1204
Ergosterol	Meisai Biological Technology Co., Ltd, Chongqing, China	20210201
Ethanol	Kelon Chemical Co., Ltd., Chengdu, China	2021112602
Ethyl acetate	Zhiyuan Chemical Reagent Co., Ltd., Tianjin, China	2017042043
Formic acid	Kelon Chemical Co., Ltd., Chengdu, China	2016062901
High performance liquid chromatography	Agilent, USA.	1260
IBM SPSS Statistics version 26.0	International Business Machines Corporation, USA	-
Methanol	Sigma-Aldrich  Trading Co., Ltd., Shanghai, China	WXBD6409V
Methanol	Kelon Chemical Co., Ltd., Chengdu, China	202010302
n-butyl alcohol	Zhiyuan Chemical Reagent Co., Ltd., Tianjin, China	2020071047
Petroleum ether	Zhiyuan Chemical Reagent Co., Ltd., Tianjin, China	2020090125
Phosphoric acid	Comeo Chemical Reagent Co., Ltd., Tianjin, China	20200110
SESCF-TCM version 2012	National Pharmacopoeia Commission, China	-	http://114.247.108.158:8888/login
Stigmasterol	Meisai Biological Technology Co., Ltd., Chongqing, China	20210201
Trichloromethane	Sinopharm Group Chemical Reagent Co., Ltd., Shanghai, China	20200214
Umetrics SIMCA version 14.1.0.2047	Umetrics, Sweden	-	https://www.sartorius.com/en/products/process-analytical-technology/data-analytics-software/mvda-software/simca/simca-free-trial-download
Ultrapure water machine	Youpu Ultrapure Technology Co., Ltd., Sichuan, China	UPH-II-10T
Ultrasonic cleaner	Kunshan Hechuang Ultrasound Instrument Co., Ltd., Jiangsu, China	KH3200E