Skip to main content
NCBI home page
Journal of Pharmaceutical Analysis logoLink to Journal of Pharmaceutical Analysis
. 2023 Feb 6;13(3):296–304. doi: 10.1016/j.jpha.2023.01.002

Rapid authentication of different herbal medicines by heating online extraction electrospray ionization mass spectrometry

Zidong Qiu a,∗,1, Chaofa Wei a,1, Xiang Li a, Changjiangsheng Lai a, Zhilai Zhan a, Yan Jin a, Li Zhou a, Qingxiu Hao a, Jian Yang a, Shuanglong Wang b, Liping Kang a,∗∗, Luqi Huang a,∗∗∗
PMCID: PMC10123936  PMID: 37102106

Abstract

The rapid and accurate authentication of traditional Chinese medicines (TCMs) has always been a key scientific and technical problem in the field of pharmaceutical analysis. Herein, a novel heating online extraction electrospray ionization mass spectrometry (H-oEESI-MS) was developed for the rapid and direct analysis of extremely complex substances without the requirement for any sample pretreatment or pre-separation steps. The overall molecular profile and fragment structure features of various herbal medicines could be completely captured within 10–15 s, with minimal sample (<0.5 mg) and solvent consumption (<20 μL for one sample). Furthermore, a rapid differentiation and authentication strategy for TCMs based on H-oEESI-MS was proposed, including metabolic profile characterization, characteristic marker screening and identification, and multivariate statistical analysis model validation. In an analysis of 52 batches of seven types of Aconitum medicinal materials, 20 and 21 key compounds were screened out as the characteristic markers of raw and processed Aconitum herbal medicines, respectively, and the possible structures of all the characteristic markers were comprehensively identified based on Compound Discoverer databases. Finally, multivariate statistical analysis showed that all the different types of herbal medicines were well differentiated and identified (R2X > 0.87, R2Y > 0.91, and Q2 > 0.72), which further verified the feasibility and reliability of this comprehensive strategy for the rapid authentication of different TCMs based on H-oEESI-MS. In summary, this rapid authentication strategy realized the ultra-high-throughput, low-cost, and standardized detection of various complex TCMs for the first time, thereby demonstrating wide applicability and value for the development of quality standards for TCMs.

Keywords: Heating online extraction electrospray ionization mass spectrometry, Rapid authentication, Traditional Chinese medicine

Graphical abstract

Image 1

Open in a new tab

Highlights

  • •

    A novel heating online extraction electrospray ionization mass spectrometry was developed.

  • •

    Rapid analysis (∼10 s) of complex TCMs without sample pretreatment and pre-separation was realized.

  • •

    A rapid differentiation and authentication strategy of TCMs based on H-oEESI-MS was proposed.

  • •

    Characteristic markers of raw and processed Aconitum herbal medicines were screened out.

  • •

    Aconitum herbal medicines were rapid authenticated by H-oEESI-MS coupled with PCA and PLS-DA.

1. Introduction

The rapid and accurate authentication of traditional Chinese medicines (TCMs), which is an important prerequisite for promoting the modernization, standardization, and high-quality development of the TCM industry, has always been a key scientific and technical problem in the field of pharmaceutical analysis [1,2]. TCMs are most often derived from natural sources. As a consequence, the confusion caused by customary substitutes, multiple botanical origins, and multiple processing methods is widespread and inevitable, which is also a key issue affecting the safety and efficacy of TCMs [3,4]. Therefore, the question of how to accurately authenticate complex TCMs in a high-throughput and standardized manner is a key bottleneck issue for relevant research institutions, enterprises, and regulatory authorities. The development and establishment of targeted rapid analysis strategies and special instruments used for complex TCMs to achieve the rapid, accurate and efficient authentication of herbal medicines, ensure the rational use of TCMs, and promote the healthy development of the TCM industry is an urgent problem to be solved [5,6].

With continuous improvement in the quality control system used for TCMs and the advancement of modern analytical science, the techniques and methods used for the rapid detection of TCMs have also undergone substantial innovations in recent years [2]. More specifically, there has been an evolution characterized by the movement away from methods focused on specific substances or indicators (such as chemical colorimetric methods [7], rapid test strips [8], enzyme-linked immunoassay [9], electrochemical analysis [10], and bioluminescence detection [11]) towards non-contact, high-throughput spectroscopic detection techniques (such as infrared and near-infrared spectroscopy [12], Raman spectroscopy [13], and hyperspectral imaging [14]). These spectroscopic techniques are characterized by simple and rapid operation, low input sample requirements, and low sample destructiveness. However, given the complex characteristics of multi-morphology, multi-component, and multi-matrix TCMs, as well as the requirements of multi-scenario and multi-purpose analysis, spectroscopic techniques often have difficulty in achieving accurate analysis and evaluation of complex TCMs in terms of specificity, accuracy, sensitivity, and versatility [15]. Mass spectrometry (MS) is a powerful tool for understanding substances at the molecular level, featuring the highest analytical sensitivity and specificity [16,17]. Chromatography-mass spectrometry (liquid chromatography (LC)-MS/gas chromatography-MS) [18,19] is also currently the most commonly used analytical tool in TCM metabolomics due to its excellent and accurate multi-component qualitative and quantitative capabilities. However, the inevitable pretreatment process and chromatographic separation process for TCMs with complex matrices undermine the ability of chromatography-MS to meet the requirements of the in-situ rapid analysis of TCMs [20]. Therefore, it is necessary and meaningful to develop a novel rapid analysis strategy based on MS. Ambient MS (AMS) [21,22] enables the direct analysis of complex samples in their native state at atmospheric pressure. Compared with classical ionization techniques (such as electrospray ionization), ambient ionization techniques (such as desorption electrospray ionization [23,24] and direct analysis in real-time [25,26]) generally have the advantages of strong resistance to complex matrix interference [20], fast analysis speed [27], and the ability to conduct the direct analysis of all types of samples without any sample pretreatment [28]. In addition, the involved devices are generally simple, convenient to move and easy to be coupled with various types of mass spectrometers [29], and have shown promising applications in the analysis of extremely complex samples, such as TCMs.

Herein, based on the basic principles of AMS and our previous study [30], a novel technique called heating online extraction electrospray ionization mass spectrometry (H-oEESI-MS) was creatively developed for the rapid authentication of different TCMs. Compared to the traditional online extraction electrospray ionization (oEESI) device [30], H-oEESI features the innovative addition of a ceramic heating device to the spray capillary cutout, thereby allowing for the instantaneous atomization and direct ionization of solvent at ultra-high flow rates (up to 1000 μL/min), thus significantly improving the on-line extraction efficiency of the target components and reducing the time required for analysis. In this study, seven kinds of Aconitum herbal medicines (52 batches) including raw samples and processed samples were selected as a typical case. All types of herbal medicines could be directly analyzed, whether raw or processed, without any pretreatment or pre-separation. The technique dramatically increased analytical productivity (∼10 s for one sample) and reduced sample (<0.5 mg) and reagent wastage (<20 μL for one sample). Furthermore, based on the overall MS characterization of the key components of different herbal medicines, a rapid screening strategy for characteristic markers was proposed to quickly visualize the variation between herbal medicines. Twenty compounds (e.g., benzoylmesaconine, deoxyaconitine, and mesaconitine) were screened as characteristic markers of raw Aconitum herbal medicines, and 21 compounds (e.g., benzoylhypaconine, epoxomicin, and mesaconine) were screened from processed Aconitum herbal medicines. The possible structures of all the characteristic markers were comprehensively identified based on authentic compounds, Compound Discoverer databases (e.g., mzCloud and mzVault), and literature matching. Finally, multivariate statistical analysis was made and the results showed that different types of herbal medicines were well differentiated and identified (R2X > 0.87, R2Y > 0.91, and Q2 > 0.72), which further verified the feasibility and reliability of the comprehensive strategy of rapid authentication of different TCMs based on H-oEESI-MS. Above all, the H-oEESI-MS-based strategy proposed in this study features significant advantages and application prospects in the rapid authentication of TCMs, and is thus expected to contribute to the development of TCM standardization.

2. Experimental

2.1. Samples, chemicals, and reagents

A total of 52 batches of seven kinds of Aconitum herbal medicines, namely, seven batches of Shengfupian (SFP), 10 batches of Shengcaowu (SCW), six batches of Shengchuanwu (SCHW), eight batches of Heishunpian (HSP), seven batches of Baifupian (BFP), seven batches of Zhicaowu (ZCW), and seven batches of Zhichuanwu (ZCHW), were purchased from Jiangyou Shenqi Zhongyaocai Zhongzhi Co., Ltd. (Mianyang, China). All the herbal medicines were identified by Prof. Zhilai Zhan and the specimens were preserved in the Chinese Medicine Resource Center, Chinese Academy of Traditional Chinese Medicine (Beijing, China). Detailed information on these Aconitum herbal medicines is listed in Table S1. All the authentic alkaloids (e.g., aconitine, mesaconitine, and hypaconitine) were supplied by Beijing Rongcheng Xinde Technology Development Co., Ltd. (Beijing, China; high performance LC (HPLC) purity > 98%). Methanol and dichloromethane (HPLC grade) were purchased from ROE Scientific Inc. (Newark, DE, USA) and used without further purification. Standard solutions for the optimization of analysis conditions were prepared by dissolving the appropriate amounts of each of the three authentic compounds in dichloromethane-methanol solution (50:50, V/V) with 1% acetic acid added [30]; the final concentration was 10 ng/mL for each authentic compound.

2.2. A H-oEESI device for the rapid characterization of TCMs

The H-oEESI-MS experiments were carried out using a novel modified H-oEESI source installed on an Orbitrap Exploris 240 mass spectrometer (Thermo Fisher Scientific Inc., San Jose, CA, USA). The device used for sample extraction and ionization is schematically illustrated in Fig. 1. The sample chamber consisted of a polypropylene housing (i.d. 2.0 mm) with a microporous membrane (0.22 μm) inside to block solid powders or particles from entering the nebulizer. A ring-mounted ceramic heater (adjustable range: 50–400 °C) was added to the tip of the atomization capillary, which can effectively assist in improving the atomization effect of the sample, especially the atomization of high flow rate solvent, as well as improving the ionization efficiency. The atomization auxiliary gas was nitrogen, which was generated by a nitrogen generator (Peak Scientific, Glasgow, UK). The gas pressure of the nitrogen gas, the ionizing charged voltage, the ion transfer tube temperature, and the vaporizer temperature were manually optimized based on the signal intensity using the three authentic diester alkaloids. Other MS tune methods were either automatically optimized or used the default values in the positive ion scan mode.

Fig. 1.

Fig. 1

Open in a new tab

Schematic illustration of heating online extraction electrospray ionization mass spectrometry (H-oEESI-MS). MS: mass spectrometer; HV: high voltage.

2.3. Direct analysis of Aconitum herbal medicines using H-oEESI-MS

All types of solid or liquid samples can be directly analyzed by H-oEESI-MS. In order to improve the reproducibility and stability of the samples, the herbal samples used in these experiments were in powdered form (0.5 mg) and did not undergo any other pretreatment or pre-separation steps. Since the key components of the samples were alkaloids, the extraction solvents were dichloromethane and methanol (50:50, V/V), and 1% acetic acid was added to enhance ionization while inhibiting the degradation of diester alkaloids in samples [30]. Taking into account the extraction efficiency and solvent consumption, the flow rate of the extraction solvent in this experiment was 100 μL/min. The MS spectra were collected in the mass range of m/z 200–1500 in the positive ion mode with a resolution of 120,000. Compared to the upper flow rate limit of 10–20 μL/min for oEESI, the extraction efficiency was greatly improved, and the online extraction and data acquisition process was sufficient for the rapid identification and characterization of samples within 20 s.

2.4. Rapid screening and identification of characteristic markers of herbal medicines

The primary fingerprint of each sample obtained by H-oEESI-MS was exported to Microsoft Excel 2019 to obtain the m/z and corresponding intensity data of each MS peak. According to the order of the peak intensities, the m/z of the top 100 intensities was screened to construct candidate characteristic markers groups. Duplicate peaks (such as isotopic peaks and impurity peaks contained in blanks) were removed. Finally, characteristic markers for each herbal medicine were obtained. Further, these characteristic markers were rapidly identified based on the precise molecular masses and MS2 fragment ions using Compound Discoverer 3.2 (CD, Thermo Fisher Scientific Inc.) and then verified by the corresponding authentic compound or by literature matching. The ions [M+H]+ and [M+Na]+ were set as the base peaks in CD, and the minimum peak intensity threshold was set to 50,000 to collect the MS data. The minimum number of isotopic peaks was 1, the minimum scan point was 3, and the MS Delta was 10 ppm. Structure identification was primarily based on the mzCloud database, which features the highest accuracy, while other databases such as mzVault and local libraries were used as auxiliaries.

2.5. Multivariate statistical analysis for the rapid authentication and structural identification of key components

In this study, principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA) were performed to extract molecular information related to Aconitum samples from MS spectral fingerprints, as well as to establish recognition modes for Aconitum herbal medicines of different species and processing methods. The resulting three-dimensional data set comprising sample names, screening characteristic markers, and ion intensities was subjected to PLS-DA and PCA using SIMCA 14.1 software.

3. Results and discussion

3.1. Improvements on the H-oEESI-MS device

First, a novel and more efficient H-oEESI ion source device (Fig. 1) was built and coupled with an Orbitrap-MS for the experiments. In addition to the higher ionization efficiency for the target analytes, by benefiting from the ceramic heating device close to the sample capillary outlet, the H-oEESI ion source significantly improved the tolerance to the extraction solvent flow (up to 1000 μL/min). This greatly increased the extraction efficiency and was important for ensuring the rapid extraction of target analytes in complex samples. Although the samples analyzed herein were crude herbal medicines, which should be more difficult to extract the analytes compared to the proprietary Chinese medicines, the time consumption for the complete extraction of a sample was significantly shortened to about 3–5 min (Fig. S1) compared to 10–15 min for oEESI [30]. As shown in Fig. S1, the concentration of the target analyte (m/z 590.29) peaked at approximately 0.27 min after the sampling, after which the signal gradually decreased. The signal of the target analyte fluctuated around the baseline after 3.5 min, indicating that the sample was completely extracted. The purpose of this study was the rapid differentiation and authentication of TCMs, and to this end, the full-scan MS fingerprint and MS2 spectra could be obtained in approximately 10−15 s, which can then be used for further multivariate statistical analysis and the structural identification of characteristic components. Moreover, one of the features of H-oEESI is its capacity for various solvents, including some that cannot be used in traditional LC-MS. With this feature, a wide range of different solvents can be selected to choose the optimal solvent for each substance, thereby allowing TCM-derived substances with different physicochemical properties to ionize in sequence according to our needs, and ultimately minimizing the undesirable effects of ionization discrimination in simultaneous ionization.

Key parameters of the H-oEESI source, including ionization voltage, nebulizer gas pressure, ion transfer tube temperature, and vaporizer temperature, were optimized based on the signal intensity of the targeted alkaloids in order to achieve more effective ionization and determination (Fig. 2). As a result, the optimal ionization voltage of 4.0 kV, the optimal atomizing gas pressure of 20 arb, the optimal ion transfer tube temperature of 320 °C, and the optimal vaporizer temperature of 300 °C were selected for subsequent experiments (detailed in Supplementary data). Other conditions in the experiment were referred to the alkaloid analysis method that we have reported [30], without further optimization. In summary, compared with classical AMS technologies like extractive electrospray ionization (EESI) [[31], [32], [33]], H-oEESI shares their common characteristics, including simplicity of the direct MS analysis of complex matrix samples under atmospheric pressure without any sample pre-treatment or pre-separation. It also offers higher ionization efficiency and faster analysis rates for a wide diversity of target analytes and is highly versatile, thereby allowing the direct analysis of complex samples of any solid or liquid. All of these features are particularly suitable for applications in the rapid analysis and authentication of herbal medicines.

Fig. 2.

Fig. 2

Open in a new tab

Optimization of heating online extraction electrospray ionization mass spectrometry (H-oEESI-MS) analytical conditions. (A) Total ion chromatography (TIC) spectrum for the optimization of ionization voltage. (B) TIC spectrum for the optimization of nebulizer gas pressure. (C) Optimization results of the ion transfer tube temperature. (D) Optimization results of the vaporizer temperature. The mass spectrometry (MS) signals in the figures were all detected from the compound aconitine (m/z 646.3190), which was used as a representative of the main components in Aconitum herbal medicines.

3.2. Rapid analysis of Aconitum herbal medicines without any preparation or pre-separation based on H-oEESI-MS

Under the optimal conditions, direct MS analysis was performed on 52 batches of seven species of Aconitum herbal medicines. All the herbal samples were directly analyzed in their powdered form by H-oEESI-MS without any other pretreatment. Benefiting from the ultra-high resolution (120,000) and sensitivity of Orbitrap-MS and the high ionization efficiency of H-oEESI, the key components in each herbal sample could be extensively characterized. Fig. 3 shows the typical characteristic H-oEESI-MS fingerprints of the seven different herbal medicines. Among them, Figs. 3A−D exhibit the MS spectra of the four kinds of processed TCMs (HSP, BFP, ZCHW, and ZCW), while Figs. 3E−G show the MS spectra of the three kinds of raw TCMs (SFP, SCHW, and SCW). Obviously, these results showed significant differences in the characteristic fingerprint profiles of the processed and raw groups from an overall perspective. Although the processing techniques of different Aconitum herbal medicines were different, steaming or boiling at high temperatures was essential for detoxification. The highly toxic diester alkaloids (e.g., aconitine m/z 646.3244, mesaconitine m/z 632.3073, and hypaconitine m/z 616.3094) were mostly converted to the corresponding less-toxic monoester alkaloids (e.g., benzoylaconine m/z 604.3106, benzoylmesaconine m/z 590.2949, and benzoylhypaconine m/z 574.2995). Thus, on the whole, the MS signals of processed Aconitum herbs were mostly concentrated in the m/z 550–610 and m/z 400–500 ranges, which represent monoester-type alkaloids and alkanolamine-type alkaloids, respectively.

Fig. 3.

Fig. 3

Open in a new tab

Mass spectra of different types of Aconitum herbal medicines analyzed by heating online extraction electrospray ionization mass spectrometry (H-oEESI-MS). (A) Heishunpian (HSP), (B) Baifupian (BFP), (C) Zhicaowu (ZCW), (D) Zhichuanwu (ZCHW), (E) Shengfupian (SFP), (F) Shengchuanwu (SCHW), (G) Shengcaowu (SCW), and (H) Blank.

The fingerprint profiles of the raw herbal medicines were markedly different from those of the processed herbal medicines. As can be seen in Figs. 3E−G, the base peaks in the three medicinal samples were all diester alkaloids, such as m/z 616.3094 (hypaconitine) in SFP and m/z 632.3073 (mesaconitine) in SCHW, suggesting that the content of diester alkaloids was relatively high in the raw samples. Interestingly, a series of peaks with high abundances of lipo-alkaloids (m/z 852.5245, 836.5333, 854.5306, etc.) were detected in the SCW samples we collected. Lipo-alkaloids also contain two ester bonds, but due to the introduction of different long-chain fatty acids, the toxicity of different lipo-alkaloids is also quite different. The specific toxicity of lipo-alkaloids has rarely been reported; therefore, further in-depth studies on lipo-alkaloids are urgently needed to effectively promote the toxicity evaluation and quality control of SCW.

3.3. Screening and identification of characteristic markers of different Aconitum herbal medicines based on Compound Discoverer

The accurate screening of characteristic markers is a prerequisite for the simple, direct, and rapid authentication and differentiation of herbal medicines. The whole profile of metabolites is complex and difficult to assess intuitively. Therefore, the characteristic markers of Aconitum herbal medicines were first screened for further evaluation and analysis (Fig. 4). Fig. 4A outlines the rapid screening process, including the rapid acquisition of the whole profile of metabolites by H-oEESI-MS (>10,000 compounds of each sample), screening of the top 100 intensities to form candidate marker groups, and final determination of characteristic markers. The key common components of different batches of the same herbal medicines were selected as characteristic markers, while the isotopic peaks and impurity peaks in the blank were selectively excluded. As a result, 13, 7, 11, 7, 12, 12, and 10 characteristic markers were screened from SFP, SCHW, SCW, HSP, BFP, ZCHW, and ZCW, respectively (Figs. 4B and C). In terms of grouping, there were five markers (m/z 438.2850, 454.2802, 590.2949, 616.3094, and 632.3073) in common among the three kinds of raw herbal medicines (Fig. 4B) and five common markers (m/z 438.2850, 454.2806, 574.2996, 590.2948, and 604.3106) among the four kinds of processed herbal medicines (Fig. 4C).

Fig. 4.

Fig. 4

Open in a new tab

Screening characteristic markers of Aconitum herbal medicines. (A) The screening strategy of the characteristic markers. (B) The Venn diagram of characteristic marker numbers for raw samples. (C) The Venn diagram of characteristic marker numbers for processed samples. H-oEESI-MS: heating online extraction electrospray ionization mass spectrometry; SCHW: Shengchuanwu; SCW: Shengcaowu; SFP: Shengfupian; HSP: Heishunpian; BFP: Baifupian; ZCHW: Zhichuanwu; ZCW: Zhicaowu.

Further, the characteristic markers were rapidly identified based on authentic compounds, CD databases (mzCloud-based), and literature matching verification. CD is an intelligent small molecule compound discovery and identification tool that features a variety of advanced processing technologies to extract and analyze related substances. The tool also combines multiple local and online databases to give relatively accurate identification results. Here, a dedicated CD workflow was designed according to the molecular characteristics of Aconitum herbal medicines and the signal response characteristics of H-oEESI-MS (Fig. S2). The workflow functions to analyze the H-oEESI-MS data of all imported files, search various databases (such as mzVault, mzCloud, and local libraries), simultaneously match primary and MS2 fragment ions, and perform structure prediction for unknown compounds. For example, Fig. 5 shows the matching results of aconitine ([M+H]+, m/z 646.3190) and its fragment information obtained from the mzCloud database. Below the 0-axis shows the standard MS2 spectrum in the mzCloud library, and above the 0-axis shows the MS2 spectrum of m/z 646.3190 in the sample. It can be clearly seen that all major fragments obtained excellent mutual matching (marked in green), and their molecular structure and the structure of each fragment could be determined quickly and accurately, such as the [M+H–AcOH]+ ion at m/z 586.3011 and the [M+H–CH3OH−AcOH]+ ion at m/z 554.2748. For compounds that could not be matched in the mzCloud database, other libraries (such as mzVault and local libraries) were referred for supplementation, and the structure inference was also carried out by consulting literature reports. Ultimately, the possible structures of all the screened 33 characteristic markers were quickly identified, of which 19 compounds were corroborated by the reference substance or standard MS/MS spectral data in mzCloud, while the others were identified for possible structures based on literature reports and self-built databases (Table S2).

Fig. 5.

Fig. 5

Open in a new tab

Assignment and identification of m/z 646.3190 based on the mzCloud database in Compound Discoverer (CD). HCD: higher-energy collisional dissociation; MS: mass spectrometry.

The skeletal structures of aconitine-type compounds are similar. Therefore, in the positive ion mode, most of these compounds generate corresponding fragment patterns through the successive loss of −CH3OH and −AcOH groups. For example, Fig. S3A shows the higher-energy collisional dissociation (HCD) MS2 spectrum of m/z 616.3105, where the main fragments were identified as the [M+H−CH3OH]+ ion at m/z 584.2830, [M+H–AcOH]+ ion at m/z 556.2895, [M+H–CH3OH−AcOH]+ ion at m/z 524.2630, [M+H–2CH3OH−AcOH]+ ion at m/z 492.2374, and [M+H–CH3OH−2AcOH]+ ion at m/z 464.2426; hence, the structure was identified as hypaconitine. Benzoyl-derived monoester-type alkaloids (e.g., benzoylaconine and benzoylmesaconine) were determined to be the main active substances in Aconitum herbal medicines. Fig. S3B shows the HCD MS2 spectrum of m/z 590.2949. The main fragments were identified as the [M+H–H2O]+ ion at m/z 572.2847, while m/z 558.2680 was a typical [M+H–CH3OH]+ ion, the highest response was for the [M+H–H2O–CH3OH]+ ion at m/z 540.2583, and the continuous loss of the CH3OH group could form [M+H–H2O–2CH3OH]+ ion at m/z 508.2321, which also confirmed that the compound was benzoylmesaconine. Lipo-alkaloids are another large class of substances present in Aconitum herbal medicines. The MS2 spectrum of the ion at m/z 852.5241 is shown in Fig. S3C. The direct loss of the linoleate group led to the formation of a strong m/z 572.2834 peak, and the existence of fragments such as m/z 540.2582 and m/z 508.2311 also confirmed that the compound was a derivative of benzoylmesaconine and linoleic acid. Combined with the available literature, it was concluded that its structure was 8-lino-14-benzoylmesaconine.

3.4. Multivariate data analysis

To further characterize and visualize the chemical differences and correlations between different kinds of Aconitum herbal medicines more simply and accurately, the chemometrics multivariate statistical analysis techniques, PCA and PLS-DA, were used to verify the characteristic markers and establish authentication models. Due to the significant differences in the composition and appearance between the raw and processed herbal medicines, all samples were divided into two groups, the raw herbal medicines group (i.e., SFP, SCHW, and SCW) and the processed herbal medicines group (i.e., HSP, BFP, ZCHW, and ZCW). The scatter plots of the raw and processed herbal medicines were established by PCA for overall trend analysis, while the supervised PLS-DA model was used to analyze the differential metabolites between different raw or processed products of Aconitum herbal medicines. After deduplication, there were 20 characteristic markers for the raw herbal medicines group and 21 characteristic markers for the processed herbal medicines group.

Figs. 6A and B show the PCA score plots obtained using the MS spectra from 23 batches of raw herbal medicines (i.e., 7 batches of SFP, 6 batches of SCHW, and 10 batches of SCW). A total of 23 data points of the samples were clustered regularly in the PCA plots. The different samples were clearly separated and each clustered into one group, indicating that the 20 rapid screened characteristic markers could distinguish the three raw herbal medicines well. The first principal component (PC1) accounted for 51.7% of the variance, whereas PC2 and PC3 accounted for 28.9% and 7.01%, respectively (Fig. 6B). Fig. 6C shows the PLS-DA score scatter plot of all samples. Similar to the PCA plots, each group of samples was still well differentiated in the PLS-DA plots, as the results showed that the parameters R2X, R2Y, and Q2 of the newly established PLS-DA model were 0.908, 0.985, and 0.939, respectively. The variable important in projection (VIP) plots of the differential chemical markers of raw Aconitum herbal medicines are shown in Fig. 6D. It can be seen that the peaks at m/z 836.5265 (8-lino-14-benzoylhypaconine), m/z 590.2919 (benzoylmesaconine), m/z 852.5245 (8-lino-14-benzoylmesaconine), m/z 828.5259 (8-pal-14-benzoylmesaconine), m/z 630.3298 (deoxyaconitine), m/z 632.3054 (mesaconitine), m/z 574.3011 (benzoylhypaconine), m/z 854.5331 (8-ole-14-benzoylmesaconine), m/z 618.3139 (8-methoxyl-14-benzoylaconine), m/z 646.3206 (aconitine), and m/z 616.3061 (hypaconitine) were the main contributors to the differentiation (Fig. 6D). Thus, these results suggested that the difference in the content of 8-lino-14-benzoylhypaconine (m/z 836.5265), benzoylmesaconine (m/z 590.2919), and 8-lino-14-benzoylmesaconine (m/z 852.5245) in samples can be used as key characteristic indicators to discriminate raw Aconitum samples.

Fig. 6.

Fig. 6

Open in a new tab

Multivariate statistical analysis results of three different types of raw Aconitum samples. (A) Score scatter plot of 2D principal component analysis (PCA) model. (B) Score scatter plot of 3D PCA model. (C) Score scatter plot of partial least squares discriminant analysis (PLS-DA) model. (D) Bar plot with PLS-DA model of variable important in projection (VIP). SCW: Shengcaowu; SCHW: Shengchuanwu; SFP: Shengfupian.

Figs. 7A and B show the PCA score plots obtained using the mass spectra from 29 batches of processed herbal medicines (i.e., 8 batches of HSP, 7 batches of BFP, 7 batches of ZCHW, and 7 batches of ZCW). A total of 29 data points of samples were clustered regularly in the PCA plots. The different samples were clearly separated and each clustered into one group, indicating that the 21 rapidly screened characteristic markers could distinguish the three processed herbal medicines well. The PC1 accounted for 62.7% of the variance whereas PC2 and PC3 accounted for 13.8% and 6.48%, respectively (Fig. 7B). Fig. 7C shows the PLS-DA score scatter plot of all samples. The parameters R2X, R2Y, and Q2 of the newly established PLS-DA model were 0.872, 0.910, and 0.725, respectively, indicating that the model also had a good fitting effect and prediction ability. The VIP plots of the differential chemical markers of processed Aconitum herbal medicines are shown in Fig. 7D. It can be seen that the peaks at m/z 523.2435 (NP-015114), m/z 574.2995 (benzoylhypaconine), m/z 554.3704 (epoxomicin), m/z 486.2653 (mesaconine), m/z 438.2827 (neoline), m/z 632.3025 (mesaconitine), m/z 588.3126 (14-benzoyldeoxyaconine), m/z 358.2358 (songorine), m/z 606.2928 (14-benzoyl-10-OH-mesaconine), and m/z 408.2723 (isotalatizidine) were the main contributors to the differentiation (Fig. 7D). Thus, these results suggested that the difference in the content of NP-015114 (m/z 523.2435), benzoylhypaconine (m/z 574.2995), and epoxomicin (m/z 554.3704) in samples can be used as key characteristic indicators to discriminate processed Aconitum samples. In conclusion, the characteristic markers based on H-oEESI-MS rapid screening combined with chemometric multivariate statistical analysis can effectively distinguish and identify different Aconitum herbal medicines.

Fig. 7.

Fig. 7

Open in a new tab

Multivariate statistical analysis results of four different types of processed Aconitum samples. (A) Score scatter plot of 2D principal component analysis (PCA) model. (B) Score scatter plot of 3D PCA model. (C) Score scatter plot of partial least squares discriminant analysis (PLS-DA) model. (D) Bar plot with PLS-DA model of variable important in projection (VIP). HSP: Heishunpian; BFP: Baifupian; ZCHW: Zhichuanwu; ZCW: Zhicaowu.

4. Conclusion

Based on the AMS strategy, this study independently developed a novel H-oEESI-MS technology to realize ultra-high-throughput in-situ direct MS analysis of complex TCMs. By this method, the overall molecular profile and fragment structure features of various herbal medicines could be completely captured within 10–15 s with minimal sample (<0.5 mg) and solvent consumption (<20 μL for one sample). Furthermore, a rapid differentiation and authentication strategy for TCMs based on H-oEESI-MS was proposed, including metabolic profile characterization, characteristic marker screening and identification, and multivariate statistical analysis model validation. In the analysis of 52 batches of seven types of Aconitum herbal medicines, 20 compounds (e.g., benzoylmesaconine, deoxyaconitine, and mesaconitine) were screened as characteristic markers of raw Aconitum herbal medicines, and 21 (e.g., benzoylhypaconine, epoxomicin, and mesaconine) compounds were screened from processed Aconitum herbal medicines. The possible structures of all the characteristic markers were comprehensively identified based on authentic compounds, Compound Discoverer databases (e.g., mzCloud andmzVault), and literature matching. Finally, multivariate statistical analysis (PCA and PLS-DA) was performed, and the results showed that different types of herbal medicines were well differentiated and identified (R2X > 0.87, R2Y > 0.91, and Q2 > 0.72), which further verified the feasibility and reliability of our comprehensive strategy for the rapid authentication of different TCMs based on H-oEESI-MS. In short, this study developed a novel complete rapid differentiation and authentication strategy of herbal medicines based on H-oEESI-MS to realize the direct, standardized, and low-cost detection of various complex TCMs without any pretreatment and pre-separation, thereby featuring wide applicability and promotion value for further developing the quality standardization of TCMs.

CRediT author statement

Zidong Qiuand Chaofa Wei: Conceptualization, Methodology, Investigation, Writing - Reviewing and Editing; Xiang Li: Investigation, Formal analysis; Changjiangsheng Lai: Data curation; Zhilai Zhan and Yan Jin: Resources; Li Zhou, Qingxiu Hao, and Jian Yang: Software; Shuanglong Wang: Writing - Original draft preparation; Liping Kang: Validation, Writing - Reviewing and Editing; Luqi Huang: Conceptualization, Funding acquisition.

Declaration of competing interest

The authors declare that there are no conflicts of interest.

Acknowledgments

This work was supported by the CACMS Innovation Fund, China (Grant Nos.: CI2021A04504 and CI2021A05206), the National Natural Science Foundation of China (Grant Nos.: 82104380, 81891010, 81891013, and 82074012), the Fundamental Research Funds for the Central Public Welfare Research Institutes, China (Grant Nos.: ZZ14-YQ-047 and ZZXT202105), the Key Project at Central Government Level (Grant No.: 2060302-2201-26), and the Beijing Nova Program.

Footnotes

Peer review under responsibility of Xi'an Jiaotong University.

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.jpha.2023.01.002.

Contributor Information

Zidong Qiu, Email: qiuzidong@126.com.

Liping Kang, Email: kang_liping21@163.com.

Luqi Huang, Email: huangluqi01@126.com.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Multimedia component 1
mmc1.docx (293KB, docx)

References

  • 1.Zhao Z., Liang Z., Guo P. Macroscopic identification of Chinese medicinal materials: Traditional experiences and modern understanding. J. Ethnopharmacol. 2011;134:556–564. doi: 10.1016/j.jep.2011.01.018. [DOI] [PubMed] [Google Scholar]
  • 2.Wang L., Kong W., Yang M., et al. Safety issues and new rapid detection methods in traditional Chinese medicinal materials. Acta Pharm. Sin. B. 2015;5:38–46. doi: 10.1016/j.apsb.2014.12.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Song J., Yao H., Li Y., et al. Authentication of the family Polygonaceae in Chinese pharmacopoeia by DNA barcoding technique. J. Ethnopharmacol. 2009;124:434–439. doi: 10.1016/j.jep.2009.05.042. [DOI] [PubMed] [Google Scholar]
  • 4.Kang T., Dou D., Xu L. Establishment of a quality marker (Q-marker) system for Chinese herbal medicines using burdock as an example. Phytomedicine. 2019;54:339–346. doi: 10.1016/j.phymed.2018.04.005. [DOI] [PubMed] [Google Scholar]
  • 5.Zheng X., Shi P., Cheng Y., et al. Rapid analysis of a Chinese herbal prescription by liquid chromatography-time-of-flight tandem mass spectrometry. J. Chromatogr. A. 2008;1206:140–146. doi: 10.1016/j.chroma.2008.08.038. [DOI] [PubMed] [Google Scholar]
  • 6.Zeng S., Wang L., Chen T., et al. Direct analysis in real time mass spectrometry and multivariate data analysis: A novel approach to rapid identification of analytical markers for quality control of traditional Chinese medicine preparation. Anal. Chim. Acta. 2012;733:38–47. doi: 10.1016/j.aca.2012.04.025. [DOI] [PubMed] [Google Scholar]
  • 7.Chen G., Chen C., Chen P. Colorimetric method for determination of carboxymethyl pachyman. Zhongguo Zhong Yao Za Zhi. 1995;20:543–576. [PubMed] [Google Scholar]
  • 8.Huang X., Huang T., Li X., et al. Flower-like gold nanoparticles-based immunochromatographic test strip for rapid simultaneous detection of fumonisin B1 and deoxynivalenol in Chinese traditional medicine. J. Pharm. Biomed. 2020;177 doi: 10.1016/j.jpba.2019.112895. [DOI] [PubMed] [Google Scholar]
  • 9.Cui Y., Zhao J., Zhou J., et al. Development of a sensitive monoclonal antibody-based indirect competitive enzyme-linked immunosorbent assay for analysing nobiletin in citrus and herb samples. Food Chem. 2019;293:144–150. doi: 10.1016/j.foodchem.2019.04.101. [DOI] [PubMed] [Google Scholar]
  • 10.Zhang T., Liang Y., Yuan B., et al. Determining method and conditional factors of electrochemical fingerprint of Chinese traditional medicine. Chin. Sci. Bull. 2007;52:2190–2202. [Google Scholar]
  • 11.Miller-Martini D.M., Chan R.Y.K., Ip N.Y., et al. A reporter gene assay for the detection of phytoestrogens in traditional Chinese medicine. Phytother. Res. 2001;15:487–492. doi: 10.1002/ptr.767. [DOI] [PubMed] [Google Scholar]
  • 12.Zhang C., Su J. Application of near infrared spectroscopy to the analysis and fast quality assessment of traditional Chinese medicinal products. Acta Pharm. Sin. B. 2014;4:182–192. doi: 10.1016/j.apsb.2014.04.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Mao J., Xu J. Discrimination of herbal medicines by molecular spectroscopy and chemical pattern recognition. Spectrochim. Acta A Mol. Biomol. Spectrosc. 2006;65:497–500. doi: 10.1016/j.saa.2005.11.030. [DOI] [PubMed] [Google Scholar]
  • 14.Tankeu S., Vermaak I., Chen W., et al. Differentiation between two "fang ji" herbal medicines, Stephania tetrandra and the nephrotoxic Aristolochia fangchi, using hyperspectral imaging. Phytochemistry. 2016;122:213–222. doi: 10.1016/j.phytochem.2015.11.008. [DOI] [PubMed] [Google Scholar]
  • 15.Qiu Z.-D., Wei X.-Y., Chen Z.-Y., et al. Discovery of the directionally detoxification effect and chemical mechanism of Ginseng-Fuzi co-decoction based on real-time online filtration electrospray ionization mass spectrometry. Phytomedicine. 2022;100 doi: 10.1016/j.phymed.2022.154059. [DOI] [PubMed] [Google Scholar]
  • 16.Song L., Chingin K., Wang M., et al. Polarity-specific profiling of metabolites in single cells by probe electrophoresis mass spectrometry. Anal. Chem. 2022;94:4175–4182. doi: 10.1021/acs.analchem.1c03997. [DOI] [PubMed] [Google Scholar]
  • 17.Wu D., Yang S., Li F., et al. Online sequential fractionation analysis of arsenic adsorbed onto ferrihydrite by ICP-MS. Anal. Chem. 2020;92:14309–14313. doi: 10.1021/acs.analchem.0c03516. [DOI] [PubMed] [Google Scholar]
  • 18.Kang D., Ding Q., Xu Y., et al. Comparative analysis of constitutes and metabolites for traditional Chinese medicine using IDA and SWATH data acquisition modes on LC-Q-TOF MS. J. Pharm. Anal. 2020;10:588–596. doi: 10.1016/j.jpha.2019.11.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Wei W.-L., Li H.-J., Yang W.-Z., et al. An integrated strategy for comprehensive characterization of metabolites and metabolic profiles of bufadienolides from Venenum Bufonis in rats. J. Pharm. Anal. 2022;12:136–144. doi: 10.1016/j.jpha.2021.02.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Qiu Z.-D., Chen J.-L., Zeng W., et al. Real-time toxicity prediction of Aconitum stewing system using extractive electrospray ionization mass spectrometry. Acta Pharm. Sin. B. 2020;10:903–912. doi: 10.1016/j.apsb.2019.08.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Cooks R.G., Ouyang Z., Takats Z., et al. Ambient mass spectrometry. Science. 2006;311:1566–1570. doi: 10.1126/science.1119426. [DOI] [PubMed] [Google Scholar]
  • 22.Takáts Z., Wiseman J.M., Gologan B., et al. Mass spectrometry sampling under ambient conditions with desorption electrospray ionization. Science. 2004;306:471–473. doi: 10.1126/science.1104404. [DOI] [PubMed] [Google Scholar]
  • 23.Tong Q., Zhang C., Tu Y., et al. Biosynthesis-based spatial metabolome of Salvia miltiorrhiza Bunge by combining metabolomics approaches with mass spectrometry-imaging. Talanta. 2022;238 doi: 10.1016/j.talanta.2021.123045. [DOI] [PubMed] [Google Scholar]
  • 24.Gao L., Zhang Z., Wu W., et al. Quantitative imaging of natural products in fine brain regions using desorption electrospray ionization mass spectrometry imaging (DESI-MSI): Uncaria alkaloids as a case study. Anal. Bioanal. Chem. 2022;414:4999–5007. doi: 10.1007/s00216-022-04130-3. [DOI] [PubMed] [Google Scholar]
  • 25.Zhou F., Zhu H., Liu S., et al. In situ analysis for herbal pieces of Aconitum plants by using direct analysis in real time mass spectrometry. Chin. J. Chem. 2015;33:241–246. [Google Scholar]
  • 26.Wang Y., Li C., Huang L., et al. Rapid identification of traditional Chinese herbal medicine by direct analysis in real time (DART) mass spectrometry. Anal. Chim. Acta. 2014;845:70–76. doi: 10.1016/j.aca.2014.06.014. [DOI] [PubMed] [Google Scholar]
  • 27.Zhang X., Chen Z.-Y., Qiu Z.-D., et al. Molecular differentiation of Panaxnotoginseng grown under different conditions by internal extractive electrospray ionization mass spectrometry and multivariate analysis. Phytochemistry. 2022;194 doi: 10.1016/j.phytochem.2021.113030. [DOI] [PubMed] [Google Scholar]
  • 28.Qiu Z.-D., Wei X.-Y., Wang Y.-N., et al. Quality tracing evaluation strategies of compatible materials in Aconitum proprietary Chinese medicines. J. Pharm. Biomed. 2021;192 doi: 10.1016/j.jpba.2020.113654. [DOI] [PubMed] [Google Scholar]
  • 29.Xu J., Li F., Xia F., et al. High throughput online sequential extraction of natural rare earth elements and determination by mass spectrometry. Sci. China Chem. 2021;64:642–649. [Google Scholar]
  • 30.Qiu Z.-D., Wei X.-Y., Sun R.-Q., et al. Limitation standard of toxic aconitines in Aconitum proprietary Chinese medicines using on-line extraction electrospray ionization mass spectrometry. Acta Pharm. Sin. B. 2020;10:1511–1520. doi: 10.1016/j.apsb.2019.12.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Xue A.H., Liu Y.Q., Li H.X., et al. Early detection of Huanglongbing with EESI-MS indicates a role of phenylpropanoid pathway in citrus. Anal. Biochem. 2022;639 doi: 10.1016/j.ab.2021.114511. [DOI] [PubMed] [Google Scholar]
  • 32.Liu Y.Q., Xue A.H., Wang S., et al. Metabolic response of Citrus limon to Asian citrus psyllid infestation revealed by EESI-MS and HPLC. Anal. Biochem. 2020;609 doi: 10.1016/j.ab.2020.113973. [DOI] [PubMed] [Google Scholar]
  • 33.Wu D., Cui M., Hao Y.B., et al. In situ study of metabolic response of Arabidopsis thaliana leaves to salt stress by neutral desorption-extractive electrospray ionization mass spectrometry. J. Agric. Food Chem. 2019;67:12945–12952. doi: 10.1021/acs.jafc.9b05339. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Multimedia component 1
mmc1.docx (293KB, docx)

Articles from Journal of Pharmaceutical Analysis are provided here courtesy of Xi'an Jiaotong University

RESOURCES