An extractive nanoelectrospray ionization-mass spectrometry method for Chinese herbal medicine authentication.

Herein, we describe a rapid, sensitive, and nondestructive method-extractive nanoelectrospray ionization-mass spectrometry (EnESI-MS)-for traditional Chinese medicine (TCM) authentication. The mass-spectral fingerprints of volatile compounds released from various TCMs can be rapidly acquired using EnESI-MS without sample pretreatment. EnESI-MS was applied to successfully differentiate between two commonly used medicinal herbs, Schisandra chinensis and Schisandra sphenanthera, which are morphologically similar but exhibit different therapeutic effects. Specific volatile compounds of each herb in a ten-component Chinese herbal product, Jia Wei Xiao Yao San, were also identified, and the method was applied to discriminate between the commercial product and a substandard version.

1. Introduction

Traditional Chinese medicine (TCM) has been practiced for thousands of years. In the 21st century, TCM has been widely studied for drug discovery [1-4]. Since the quality of a medicine may influence its efficacy and safety, quality control is essential. However, a major concern for quality control in TCM is the misidentification of herbal species, which can occur inadvertently because of misidentification due to morphological similarity, multiple sources, or confusing naming conventions (differing regional names for the same herbs, or different herbs with the same names from different regions), or intentionally, through product counterfeiting. Consequently, the authentication of TCMs is crucial to ensure their efficacy and safety. Several identification methods are widely used to authenticate TCMs, based on morphological, microscopic, chemical, and/or genetic examination [5]. More recently, other approaches have been developed for rapid TCM identification, such as cysteine-rich peptide (CRP) fingerprinting [6] and electronic nose (E-nose) systems [7]. E-nose technology benefits from non-destructive sample preparation, relatively fast assessment, and real-time information, which facilitates the monitoring of diverse volatile components. Because odor can be an important property of TCMs, E-noses have also been applied their analyses [7,8]. However, an E-nose system may be limited in TCM identification because of its unsatisfactory recognition accuracy and lack of chemical identification information [8].

Electrospray ionization-mass spectrometry (ESI-MS) is a sensitive, specific, and highly reliable analytical technique for detecting chemical compounds in the solution phase. In ESI-MS analysis, dissolved analytes are infused directly into the instrument to generate ions, which may cause significant ion suppression effects and lead to a low tolerance against sample matrixes. Therefore, samples with complex matrices must undergo pretreatments such as purification, extraction, and chromatographic separation before MS analysis, to lower the ion suppression effect and increase sensitivity. However, tedious sample preparation usually makes the analytical work laborious and time-consuming. In 2006, Cooks et al. developed another ionization method, extractive electrospray ionization (EESI) [9], that uses an ESI sprayer and a sample sprayer to produce charged microdroplets of solvent and nebulize the sample, respectively. During the EESI process, sample droplets from the sample sprayer are initially aerosolized and then ionized by the charged solvent droplets sprayed by the ESI spray when they intersect [10]. In 2007, the EESI interface was slightly modified by Zenobi et al. by using air as a carrier gas to introduce gaseous analytes into the electrospray beam to monitor the maturity of fruits [11]. This made EESI an alternative ionization method for the analysis of samples with complex matrices, and has been applied in the analysis of breath aerosols [12], reaction mixtures [13], and toxic substances from biological samples [14]. However, EESI-MS has not been applied for the analysis of the volatile components of TCMs.

In this study, to develop a simple, sensitive, practical MS platform for the rapid detection of volatile compounds and apply it to TCM authentication, a nanoESI source was modified as an extractive nanoelectrospray ionization source and coupled to a high-resolution mass spectrometer (EnESI-MS). Comparing with a hard ionization technique of EIMS, the soft ionization technique of EnESI-MS can produce the intact molecular ions in MS scan mode without generating fragmented ions, which cause ambiguous and complex mass spectrum. In addition, without the use of heating injector and GC, EnESI-MS is more feasible to analyze thermal labile compounds and is a more rapid and easily handling method compared to GC-EI-MS. To the best of our knowledge, it is the first study to develop an extractive nanoelectrospray ionization source (EnESI) with higher sensitivity compared to EESI. The EnESI-MS was applied to analyze two TCMs, Wu Wei Zi and Jia Wei Xiao Yao San. Wu Wei Zi consists of two closely related species, Bei Wu Wei (Schisandra chinensis (S. chinensis)) and Nan Wu Wei (Schisandra Sphenanthera (Schisandra sphenanthera)), which have quite similar morphologies; a well-experienced examiner is needed for morphological identification. Because they have dissimilar therapeutic effects, differentiation between the S. chinensis and S. sphenanthera fruits is important for clinical efficacy. Jia Wei Xiao Yao San, composed of ten herbs, is a traditional formula to soothe the liver and fortify the spleen [15] and is the most frequently prescribed formula in Taiwan for the treatment of agitation, functional dyspepsia, insomnia and mood disorders [16]. Therefore, the EnESI-MS approach was applied to discover marker peaks for Wu Wei Zi and the ten herbs of Jia Wei Xiao Yao San, to evaluate its potential use in TCM authentication.

2. Materials and methods

2.1. Reagents and chemicals

Acetonitrile (ACN) was purchased from J.T. Baker (Phillipsburg, NJ, USA). Deionized water (Milli-Q, Millipore, USA) was used. Fruits of S. chinensis and S. sphenanthera, Moutan Radicis Cortex, Angelicae Sinensis Radix, Menthae herba, Atractylodis Macrocephalae Rhizoma, Gardeniae Fructus, Bupleuri Radix, Paeoniae Alba Radix, Poria, Zingiberis Rhizoma, and Glycyrrhizae Preparata Radix were all purchased from three different stores. (Lian-He, Taichung, Taiwan; Yi-De, Taichung, Taiwan; Jiing-Han-Tang, New Taipei city, Taiwan) Jia Wei Xiao Yao San were purchased from Sheng Chang Pharmaceutical Co., Ltd. (Taoyuan, Taiwan) and Sun-Ten Pharmaceutical Co., Ltd. (Taichung, Taiwan). Home-made Jia Wei Xiao Yao San powder was prepared according to the method for concentrated TCMs from the Taiwan Herbal Pharmacopeia(3rd Edition) [15]. The ten herbs with the following proportions were mixed: Moutan Radicis Cortex (1.25 g), Angelicae Sinensis Radix (2 g), Menthae herba (1 g), Atractylodis Macrocephalae Rhizoma (2 g), Gardeniae Fructus (1.25 g), Bupleuri Radix (2 g), Paeoniae Alba Radix (2 g), Poria (2 g), Zingiberis Rhizoma (2 g), and Glycyrrhizae Preparata Radix (1 g). The herbal mixture was grinded using an electrical blade grinder to have a tiny powder. The powder of herb mixture was stored in a tightly closed bottle to prevent the loss of volatile compounds, and a tea bag was used to packed the powder (2g) for EnESI-MS analysis.

2.2. EESI and EnESI-MS analysis

Samples were placed in a clean, dry glass container to avoid potential chemical contamination. We modified our ESI (Bruker, Germany) and nanoESI sources (CaptiveSpray, Bruker, Germany) as EESI and an EnESI sources, respectively. For either EESI or EnESI, a tube, originally used as a nebulization gas tube in ESI or nanoESI, was connected to the sample container. For EESI, N2 (~10 psi) was infused into the sample container and the volatile compounds were delivered from the samples into the EESI source. An 80% ACN/water mixture was infused as the ESI solution at 4 μL/min using a syringe pump (KD Scientific Inc, MA). The spray voltage was operated at 3600 V. For EnESI, N2 (~4 psi) was infused into the sample container and the volatile compounds were delivered from the sample into the EnESI source. An 80% ACN/water mixture was infused as the nanoESI solution at 300 nL/min by a nanoflow pump (Ultimate 3000, Dionex). The spray voltage was operated at 2100 V. EESI and EnESI were performed using a QTOF-MS instrument (Maxis Impact QTOF mass spectrometer, Bruker Daltonics). Data were collected in the positive ion mode. Precursor ions were isolated with 3–4 Da. Spectra were collected for 1.5 min and processed (DataAnalysis, Bruker, Germany).

2.3. Processing of MS spectral data

The raw data from the mass peak intensity lists were exported manually as .csv files and compressed to .zip files. The compressed files were loaded into MetaboAnalyst (www.metaboanalyst.ca) for PCA analysis. Mass tolerance was set at 0.025 m/z, and no normalization or data transformation methods were used.

3. Results and discussion

3.1. EnESI optimization

Fig. 1 shows the flow direction of the gas and spray solution inside the EnESI. Volatile compounds from the sample bottle are swept by N2 into the EnESI for analysis. An EESI sprayer ion source is generated by introducing volatile analytes into the electrospray beam through the desolvation gas line. The sensitivities of EESI and EnESI were compared for the volatiles in S. chinensis. As shown in Fig. 2a, EESI detects volatile compounds with m/z 97.03, 111.04, 121.10, 147.04, 205.19, and 219.17 from S. chinensis (~1.5 g), compared to the background signals from an empty bottle (Fig. 2b). The peak of m/z 147.04 could be coumarin due to their ESI-MS/MS spectra (Fig. S1) in a reported study [17]. EnESI also detects the aforementioned volatile compounds from S. chinensis (Fig. 2c); background signals (an empty bottle) are presented in Fig. 2d.

Fig. 1

Schematic diagram of EnESI.

Fig. 2

Sensitivity comparison between EESI and EnESI using S. chinensis (~1.5 g). (a) Volatile compound fingerprints of S. chinensis by EESI. (b) Blank signals obtained from an empty bottle with EESI. (c) Schematic diagram of EESI. (d) Volatile compound fingerprints of S. chinensis by EnESI. (e) Blank signals obtained from an empty bottle with EnESI. (f) Schematic diagram of EnESI.

However, compared to EESI, EnESI shows a significant signal improvement for these volatile compounds and has a 119 folds higher peak intensity of m/z 111.04 (intensity 2.5 × 106, S/N 7733 in EnESI vs. intensity 2.1 × 104, S/N 33.5 in EESI), 97 folds higher peak intensity of m/z 121.01(intensity 8.0 × 105, S/N 2406.7 in EnESI vs. intensity 8.2 × 103, S/N 17.9 in EESI), 110 folds higher peak intensity of m/z 147.04 (intensity 1.1 × 106, S/N 2965.8 in EnESI vs. intensity 1.0 × 104, S/N 15.6 in EESI) and 340 folds higher peak intensity of m/z 205.19 (intensity 1.7 × 106, S/N 3774.4 in EnESI vs. intensity 5.0 × 103, S/N11.1 in EESI). In EESI, a Taylor cone is formed at the spray tip and emits a fine mist of droplets into an open field (Fig. 2c). Analytes are scattered so that not all are extracted in the solvent and carried into the MS inlet. In EnESI, gaseous analytes sweep around the spray tip and focus the Taylor cone into the MS in a small, closed space (Fig. 2f). This design may increase the extraction efficiency for volatile compounds by the spray solution, resulting in high sensitivity for volatile compound analysis. Because volatile organic compounds are compounds that have a high vapor pressure and high nonpolar property, a high percentage of ACN may be suitable for the extraction of these non-polar ions in EnESI. Different percentages (5%, 50% and 80%) of ACN were tested, and the use of 80% ACN were found to have more peaks in a larger m/z range (~m/z 280-m/z 460). (Fig. S2) Therefore, 80% ACN was selected for the spray solvent composition in EnESI.

3.2. Concentration-dependent signal intensity

As shown in Fig. 3, the volatiles in S. chinensis samples as small as 1 grain (~65 mg) can be detected by EnESI, and the intensities of the m/z 111.04 peak increase in a linear fashion when the sample loading increases from 1 to 9 (~550 mg) grains. However, the intensity of the m/z 111.04 peak decreases for 12 (~728 mg) and 15 grains (~910 mg) of S. chinensis. This may be due to the ion suppression effect of the m/z 111.04 peak by other ions, because the m/z 147.04. and 167.07 peaks are of higher intensity than m/z 111.04 at 12 and 15 grains (data not shown). This may indicate that the signals for volatile compounds in EnESI are concentration-dependent and may be applicable for material quantification.

Fig. 3

EnESI-MS peak signals of m/z 111.04 for 1 to 9 grains of S. chinensis (n = 3).

3.3. Differentiation between S. chinensis and S. sphenanthera

Because Bei Wu Wei (S. chinensis) and Nan Wu Wei (S. sphenanthera) are difficult to distinguish based on appearance, the two herbal species were analyzed by EnESI-MS to identify specific identifying markers. Samples of both species were purchased from three different stores and subjected to analysis (~2 g); the raw data sets were imported to MetaboAnalyst for PCA analysis. As shown in Fig. 4a, the MS signals from S. chinensis and S. sphenanthera can be separated into two distinct groups, indicating some differences in chemical constituents. The corresponding loading plot (Fig. 4b) shows that the peaks at m/z 111.04 and 153.13 are the most abundant peaks in S. chinensis, whereas that at m/z 167.07 is predominant in S. sphenanthera. The most abundant peaks of S. chinensis and S. sphenanthera from PCA analysis are consistent with the peaks observed in their EnESI-QTOF-MS spectra (Fig. 4c, d). Therefore, peaks at m/z 111.04, 153.13, and 167.07 can be used as marker peaks to differentiate between the two species. The peak of m/z 167.07 can be identified as paeonol based on their matched ESI-MS/MS spectra (Fig. S3) in a reported study [18]. However, the other two peaks cannot be identified due to the absence of matched MS/MS spectra in mass spectral databases.

Fig. 4

(a) PCA results for S. chinensis (n = 3) and S. sphenanthera (n = 3). (b) Loading plot of PC1 and PC2 components. EnESI-MS spectra of (c) S. chinensis and (d) S. sphenanthera.

3.4. Marker peak determination for ten herbs in Jia Wei Xiao Yao San

To determine markers for the identification of Jia Wei Xiao Yao San, ten commonly used raw herbal components were bought from three different stores and analyzed using EnESI-MS. Their morphologies are shown in Fig. S4. Principal component analysis (PCA) and a loading plot were used to analyze the possible marker peaks of Angelicae Sinensis Radix. As shown in Fig. 5a, EnESI-MS signals from this herb and the blank (empty bottle) can be separated into two groups by PCA. Peaks at m/z 191.11, 173.10, 163.11, and 145.10 are unique markers in both the loading plot (Fig. 5b) and MS spectrum (Fig. 5c). The MS/MS spectrum of m/z 191.11 (Fig. 5d) shows four fragment ion peaks at m/z 173.10, 163.11, and 145.10, enabling identification of m/z 191.11 as butylphthalide, a known constituent of Angelicae Sinensis Radix [19,20].

Fig. 5

(a) PCA result for Angelicae Sinensis Radix (n = 3) and (b) loading plot of PC1 and PC2 components. (c) EnESI-MS spectrum of Angelicae Sinensis Radix and (d) MS/MS spectrum of the marker peak at m/z 191.11. The fragments of structural elucidation were obtained by using software Mass Frontier 8.0 (Thermo Fisher Scientific, Waltham, United States).

The other nine raw herbs (Moutan Radicis Cortex (Fig. S5), Menthae Herba (Fig. S6), Atractylodis Macrocephalae Rhizoma (Fig. S7), Gardeniae Fructus (Fig. S8), Bupleuri Radix (Fig. S9), Paeoniae Alba Radix (Fig. S10), Poria (Fig. S11), Zingiberis Rhizoma (Fig. S12), and Glycyrrhizae Preparata Radix (Fig. S13)) were also analyzed by EnESI-MS. Some raw herbs did not have strong marker signals and may have similar spectra with the blanks, which results in an undistinguished PCA result between the sample and blank group. However, the loading plot allow us to quickly find a specific marker for the raw herbs.

Table S1 presents the marker peaks for the ten herbs in Jia Wei Xiao Yao San detected by EnESI-MS, as well as their fragment ions with possible compound identities. Only two marker peak compounds, butylphthalide in Angelicae Sinensis Radix and paeonol in Moutan Radicis Cortex [18,21], could be identified based on their matched ESI-MS/MS spectra in reported studies [18-21]. Because most volatile compound were analyzed with GC–MS, these volatile compounds only have EI-MS spectra in MS spectra databases. Thurman, E. M. et al. compared EI with ESI spectra, and found that although EI and ESI spectra are different, they may still have some fragmented ions in common [22]. Therefore, the other marker peaks with no available ESI-MS/MS spectrum in database were referred to their correspond EI spectra in NIST database or reported studies. The potential structures identified from EI-MS spectra were further matched to the theoretical ESI-MS fragments elucidated by Mass Frontier software (version 8.0, Thermo Fisher Scientific, Waltham, United States). Finally, atractylon was identified as the marker peak (m/z 217.16) of Atractylodis Macrophalae Rhizoma; 2,6,6-trimethyl-1,3-cyclohexadiene-1-carboxaldehyde was identified as the marker peak (m/z 151.11) of Gardeniae Fructus; Eucalyptol was identified as the marker peak (m/z 155.10) of Bupleuri Radix [23]; 2-pentylfuran was identified as the marker peak (m/z 139.11) of Poria. The marker peak of m/z 153.13 were identified as four possible compounds (Cyclohexanone, 5-methyl-2-(1-methylethylidene)-; (E)-3(10)-Caren-4-ol; 2-Cyclohexen-1-ol, 1-methyl-4-(1-methylethenyl)-, trans-; thujone), because they have the same theoretical [M+H]+ value and the same elucidated fragmented ions, and have been identified in a GC–MS study of Mentha haplocalyx [24] (Table S1). Because the volatile compounds were not separated in EnESI-MS, the MS/MS spectrum may contain all the fragment ions from the 4 possible compounds. However, the structures of marker peaks of 121.1 (paeoniae alba radix), 137.13 (zingiberis rhizome) and 100.07 (glycyrrhizae preparata radiz) cannot be identified in this study.

3.5. EnESI-MS analysis of Jia Wei Xiao Yao San raw herb powder mixture and commercial products

To evaluate the performance of EnESI-MS in the authenticity of Jia Wei Xiao Yao San, the homemade and commercial Jia Wei Xiao Yao San products were analyzed and compared. The marker peaks of each herb (Table S1) are found in the MS spectrum of the home-made preparation (Fig. 6a): Atractylodis Macrocephalae Rhizoma (marker peak: m/z 217.16), Angelicae Sinensis Radix (m/z 191.11), Moutan Radicis Cortex (m/z 167.07), Menthae Herba (m/z 153.13), Gardeniae Fructus (m/z 151.11), Bupleuri Radix (m/z 155.10), Poria (m/z 139.11), Paeoniae Alba Radix (m/z 121.10), Zingiberis Rhizoma (m/z 137.13), and Glycyrrhizae Preparata Radix (m/z 100.07).

Fig. 6

EnESI-MS spectra of Jia Wei Xiao Yao San from (a) the homemade powder and (b) a commercial product.

The aforementioned marker peaks of the herbs in the home-made sample are also observed in the MS spectrum of a commercial Jia Wei Xiao Yao San product purchased from a TCM company (Sheng Chang Pharmaceutical Co., Fig. 6b). The EnESI-MS analysis of another commercial Jia Wei Xiao Yao San product, purchased from another supplier (Sun-Ten Pharmaceutical Co.), is shown in Fig. S14. The EnESI-MS spectra of the two commercial products are similar to our homemade powder mixture. The LOD of Jia Wei Xiao Yao San by EnESI-MS was found to be ~20 mg by evaluating the two lowest marker peaks of m/z 100.07 (S/N: 3.9) and m/z 155.10 (S/N: 3.6). Therefore, EnESI-MS could be a rapid and attractive method for the quality control of commercial concentrated products.

3.6. EnESI-MS for the determination of substandard TCM products

Substandard TCM products have been a serious issue that influence safety and efficacy. Substandard products can arise for a variety of reasons, such as improper storage methods that cause drug deterioration, the inclusion of contaminants, or out-of-specification drug quality or quantities. Therefore, the ability of EnESI-MS to identify substandard TCM products was next assessed. A modified Jia Wei Xiao Yao San formula was prepared by decreasing the amounts of Angelicae Sinensis Radix and Atractylodis Macrocephalae Rhizoma to a half and a quarter of their original quantities, respectively. As shown in Fig. 7, compared to the original formula (Fig. 7a), the marker peak intensities for Angelicae Sinensis Radix and Atractylodis Macrocephalae Rhizoma significantly decrease in the modified formula (Fig. 7b). The degrees of decrease were quantified by comparing the peak intensities of the markers for these two herbs to the most abundant peak for the standard formula, m/z 167.07. The m/z 191.11/167.07 value decreases from 0.094 to 0.021 when the amount of Angelicae Sinensis Radix is reduced by 50%. The m/z 217.16/167.07 value changes from 0.041 to 0.008 for the 75% reduction in Atractylodis Macrocephalae Rhizoma content. Therefore, EnESI-MS may be able to detect proportional changes in the constituents of an herbal mixture and be applied in the quality control of commercial concentrated products.

Fig. 7

EnESI-MS spectrum of (a) a standard preparation of Jia Wei Xiao Yao San powder and a (b) modified formula.

4. Conclusion

In this study, a sensitive EnESI-MS method was developed for the rapid analysis of volatile compounds in traditional Chinese medicines. The EnESI-MS signals for the volatile compounds increased with the amount of material. EnESI-MS was applied to successfully differentiate S. chinensis from S. sphenanthera. Marker peaks of ten herbs in Jia Wei Xiao Yao San were also identified and applied for the identification of commercial Jia Wei Xiao Yao San and a substandard formulation. EnESI-MS could be a practical and attractive method for the rapid authentication of TCMs and the quality control of commercial concentrated TCM products.

MS/MS spectrum of the marker peak at m/z 147.04. This peak was identified as coumarin and the elucidated fragment structures were obtained by Mass Frontier 8.0.

EnESI-MS analysis of Schisandra chinensis using 5% ACN, 50% ACN and 80% ACN as the spray solution.

MS/MS spectrum of the marker peak at m/z 167.07. This peak was identified as paeonol and the elucidated fragment structures were obtained by Mass Frontier 8.0.

Morphologies of ten herbs in Jia Wei Xiao Yao San.

(a) PCA result for Mountan Radicis Cortex (n = 3). (b) Loading plot of PC1 and PC2 components. (c) EnESI-MS spectra of Mountan Radicis Cortex from three different stores. (d) MS/MS spectrum of the marker peak at m/z 167.07. This peak was identified as paeonol, and the elucidated fragment structures were obtained by Mass Frontier 8.0.

(a) PCA result for Menthae Herba(n = 3). (b) Loading plot of PC1 and PC2 components. (c) EnESI-MS spectra of Menthae Herba from three different stores. (d) MS/MS spectrum of the marker peak at m/z 153.13. (e)The possible identified structures and their fragmented ions deduced by Mass Frontier 8.0.

(a) PCA result for Atractylodis Macrocephalae Rhizoma(n = 3). (b) Loading plot of PC1 and PC2 components. (c) EnESI-MS spectra of Atractylodis Macrocephalae Rhizoma from three different stores. (d) MS/MS spectrum of the marker peak at m/z 217.16. This peak was identified as atractylon, and the elucidated fragment structures were obtained by Mass Frontier 8.0.

(a) PCA result for Gardeniae Fructus (n = 3). (b) Loading plot of PC1 and PC2 components. (c) EnESI-MS spectra of Gardeniae Fructus from three different stores. (d) MS/MS spectrum of the marker peak at m/z 151.11. This peak was identified as 2, 6, 6-trimethyl-1, 3-cyclohexadiene-1-carboxaldehyde, and the elucidated fragment structures were obtained by Mass Frontier 8.0.

(a) PCA result for Bupleuri Radix (n = 3). (b) Loading plot of PC1 and PC2 components. (c) EnESI-MS spectra of Bupleuri Radix from three different stores. (d) MS/MS spectrum of the marker peak at m/z 155.10. This peak was identified as eucalyptol, and the elucidated fragment structures were obtained by Mass Frontier 8.0.

(a) PCA result of Paeoniae Alba Radix (n = 3). (b) Loading plot of PC1 and PC2 components. (c) EnESI-MS spectra of Paeoniae Alba Radix from three different stores. (d) MS/MS spectrum of the marker peak at m/z 121.10.

(a) PCA result for Poria (n = 3). (b) Loading plot of PC1 and PC2 components. (c) EnESI-MS spectra of Poria from three different stores. (d) MS/MS spectrum of the marker peak at m/z 139.11. This peak was identified as 2-pentylfuran, and the elucidated fragment structures were obtained by Mass Frontier 8.0.

(a) PCA result for Zingiberis Rhizoma(n = 3). (b) Loading plot of PC1 and PC2 components. (c) EnESI-MS spectra of Zingiberis Rhizoma from three different stores. (d) MS/MS spectrum of the marker peak at m/z 137.13.

(a) PCA result for Glycyrrhizae Preparata Radix (n = 3). (b) Loading plot of PC1 and PC2 components. (c) EnESI-MS spectra of Glycyrrhizae Preparata Radix from three different stores. (d) MS/MS spectrum of the marker peak at m/z 100.07.

EnESI-MS spectrum of commercial concentrated Jia Wei Xiao Yao San from store B (Sun-Ten Pharmaceutical Co., Taichung, Taiwan).

Supplementary Table 1

Marker peaks for ten herbs in Jia Wei Xiao Yao San.

	Herb	[M+H]+ of marker peak	Predominant peaks in MS/MS (relative intensity)	Possible identity
1.	Angelicae Sinensis Radix	191.11	191.11 (100)	Butylphthalide
			173.10 (78)
			163.11 (20)
			145.10 (34)
			117.07 (25)
			91.06 (18)
2.	Moutan Radicis Cortex	167.07	167.07 (100)	Paeonol
			149.06 (64)
			121.06 (58)
3.	Atractylodis Macrocephalae Rhizoma	217.16	217.16 (100)	Atractylon
			199.15 (11.3)
			161.10 (10)
			147.12 (9.3)
			133.10 (6)
			121.10 (5)
			107.10 (4)
			95.05 (6)
4.	Gardeniae Fructus	151.11	151.11 (100)	2,6,6-Trimethyl-1, 3-Cyclohexadiene-1-Carboxaldehyde
			135.12 (20)
			123.1 (15)
			109.10 (23)
			95.1 (24)
			93.07 (19)
5.	Bupleuri Radix	155.10	139.08 (10)	Eucalyptol
			137.09 (10)
			109.10 (100)
			95.08 (29)
			93.06 (23.8)
6.	Poria	139.11	139.11 (100)	2-Pentylfuran
			121.10 (9.3)
			97.06 (9.0)
			95.08 (9.0)
7.	Menthae Herba	153.13	153.13 (100)	① Cyclohexanone, 5-methyl-2-(1-methylethylidene)-② (E)-3 (10)-Caren-4-ol ③ 2-Cyclohexen-1-ol, 1-methyl-4-(1-methylethenyl)-, trans-④ Thujone
			135.1 (20)
			111.08 (20)
			109.11 (30)
			107.08 (30)
			97.06 (38)
			95.07 (20)
			93.07 (47)
8.	Paeoniae Alba Radix	121.10	121.10 (100)	Not identified
			91.05 (31.8)
9.	Zingiberis Rhizoma	137.13	137.13 (100)	Not identified
			107.09 (99)
			95.09 (98)
10.	Glycyrrhizae Preparata Radix	100.07	100.07 (100)	Not identified
			97.06 (4)