Simultaneous Determination of Five Components in Aster tataricus by Ultra Performance Liquid Chromatography–Tandem Mass Spectrometry

Abstract

A novel quantitative method using high-performance liquid chromatography coupled to electrospray ionization tandem mass spectrometry was developed for simultaneous determination of the important active constituents including shionone, luteolin, quercetin, kaempferol and ferulic acid in Aster tataricus from different habitats. The separation was performed on a C₁₈ column with acidified aqueous acetonitrile gradients. Quantification of the analytes was achieved by the use of a hybrid quadrupole spectrometer. Multiple reaction monitoring scanning was employed with positive and negative modes at the same time in a single run. The validated results of the method indicated that the method was simple, rapid, specific and reliable. The results demonstrated that the quantitative difference in content of five active compounds was useful not only for chemotaxonomy of numerous samples from different sources but also for the standardization and differentiation of several similar samples. It was the first time to report a UPLC–ESI-MS-MS method for determination of five components in A. tataricus extract. Simultaneous quantification of bioactive components by UPLC–ESI-MS could be a well-acceptable strategy to control the quality of A. tataricus extract comprehensively.

Introduction

Aster tataricus rhizoma (Ziwan in Chinese), the dried rhizome of A. tataricus L. f., is a well-known traditional Chinese medicinal herb officially listed in the Chinese Pharmacopoeia (1). In clinical applications of traditional Chinese medicine (TCM), this herb has been proved to be effective in treating sputum and cough diseases. For that reason, A. tataricus works as an important component of many common Chinese medicine curing cough and asthma (2). In addition, existing clinical reports confirm that A. tataricus can be used as diuretic and laxative (3). Phytochemical studies on A. tataricus reveal that it contains flavonoids, phenolic acids, sterols, triterpenes, etc. (2, 4, 5). Modern pharmacological studies have shown that the triterpenoids shionone in A. tataricus has significant expectorant, antitussive effect, and it is the feature component of A. tataricus. Also luteolin and quercetin are expectorant, antitussive ingredients. Recently, it was discovered that A. tataricus extract had a number of bioactivities, including anti-inflammatory, antihemorrhagic and antioxidant activities. For example, kaempferol and quercetin take significant effect on inhibiting hemolysis, lipid peroxides reactivity and immunomodulatory (6). Ferulic acid has some effect in health care involving free radical scavenging, blood pressure lowering, antithrombotic therapy, anti-bacterial and anti-inflammation (4). Herbal drugs, individually and in combination, contain a myriad of compounds in complex matrices in which no single active constituent is responsible for the overall efficacy. Consequently, simultaneous quantitative analysis of various kinds of active components is the most direct and important method for the quality control of TCM.

Several methods have been developed for the determination of A. tataricus, including thin-layer chromatography (TLC) and high-performance liquid chromatography coupled with ultraviolet detection (HPLC-UV) (2, 4, 5). However, these methods suffered from low resolution and sensitivity (e.g., TLC) and long run time (e.g., HPLC-UV). For instance, shionone in A. tataricus is difficult to be detected using UV due to its poor UV absorption. The components of A. tataricus are complex, and some of the constituents are usually of low contents. It is particularly difficult to simultaneously determine much more active constituents using traditional methods for quality control of A. tataricus. Therefore, a more sensitive, selective and rapid method is demanded. The emergence of ultra performance liquid chromatography connected with tandem mass spectrometry (UPLC–MS-MS) makes the determination possible. First, UPLC can increase the separation efficiency so that the run time is shortened. Second, it not only provides adequate structural information but also performs accurate quantification of multiple compounds. Therefore, it is a powerful approach to solve the problems mentioned above, as well as having the advantages of saving time and solvent.

In this study, we developed and validated a UPLC–ESI-MS-MS method for simultaneous determination of shionone, luteolin, quercetin, kaempferol and ferulic acid in A. tataricus extract. Their structures are listed in Figure 1. Multiple reaction monitoring (MRM) was applied and an ESI source was operated in positive and negative modes at the same time. An information-dependent acquisition method was employed to trigger production scans above the MRM signal threshold so that the five components could be identified through enhanced product ion scans. The method is very suitable for the analysis of TCM and its prescriptions, in particular for low-abundance compounds in complex compounds, which are difficult to obtain by conventional isolation means.

Figure 1.

Chemical structures of the five compounds in A. tataricus.

Experimental

Chemicals, reagents and materials

HPLC-grade acetonitrile (Merck, Germany) was used for UPLC analysis. Double-distilled water was prepared in our own laboratory by Heal Force-PWVF Reagent Water System (Shanghai CanRex Analyses Instrument, Shanghai, China). HPLC-grade methanol (Merck, Germany) was used for sample preparation. HPLC-grade formic acid was purchased from Diamond Technology (Dikma Technology Corporation, USA).

The standards including shionone (1,11,581–2,00,302), quercetin (1,00,081–2,00,406) and ferulic acid (1,10,773–2,01,313) were purchased from the China Pharmaceutical Biological Products Analysis Institute (Beijing, China). Luteolin (11,042,524) was provided by Winherb Medical Technology Co. Ltd (Shanghai, China). Kaempferol (1,01,127) was obtained from Tauto Biotech Co. Ltd (Shanghai, China). Purity was >98%.

Six batches of the raw material samples of A. tataricus from different places were collected (Table I). All the voucher specimens identified by Professor Zengke Kong were deposited in the herbarium of School of Pharmacy, Hebei Medical University.

Table I.

The List of A. tataricus Samples

Sample no.	Code	Source
1	HB-1	Hebei-anguo
2	HB-2	Hebei-anguo
3	HB-3	Hebei-anguo
4	HB-4	Hebei-anguo
5	HBQC	Hubei-qichun
6	AHBZ	Anhui-bozhou

Sample preparation

The dry herb samples were ground to fine powder by a pulverizer, filtered through a 40-mesh sieve and 1.0 g of powder was accurately weighed and ultrasonically extracted with 100 mL methanol for 40 min. The extracted solution was adjusted to the original weight by adding methanol, and then the aliquot of the supernatant was filtered through a 0.22-μm microporous membrane before UPLC injection of 10 μL.

Instrumentation and conditions

For quantitative determination of the five compounds, all experiments were carried out in positive and negative modes simultaneously using a Waters ultra performance liquid chromatography, equipped with a vacuum degasser, a binary high-pressure pump and an autosampler, connected to a MassLynx 4.1 data system, a Xevo™ TQ-S system from Waters (USA) and a hybrid triple quadrupole mass spectrometer. The chromatographic separation was achieved at 40°C on a Waters UPLC BEH C₁₈ column (50 × 2.1 mm, 1.7 μm). The mobile phases consisted of acetonitrile (A) and water (B) both containing 0.1% formic acid. The elution program was optimized as follows: 0–3 min, linear change from A–B (5 : 95, v/v) to A–B (45 : 55, v/v); 3–10 min, isocratic elution A–B (95 : 5, v/v); then quickly returned to initial A–B (5 : 95, v/v). This was followed by the equilibration period of 4 min prior to the injection of each sample. The flow rate of mobile phase was set at 0.3 mL/min and the injection volume was 10 μL. The operating conditions for the ESI interface were as follows: the ion spray voltage was set to +3,500 and –4,500 V, respectively; the turbo spray temperature was 500°C and nebulizer gas (Gas 1) and heater gas (Gas 2) were set at 60 and 65 arbitrary units, respectively. The curtain gas was kept at 25 arbitrary units and the interface heater was on. Nitrogen was used in all cases. The full-scan mass covered the range from m/z 100 to 1,000. MRM mode was employed for quantification. All the peaks of target compounds in the solution of A. tataricus samples were unambiguously identified by the comparison of retention time, parent and product ions with standards. The retention time, characteristic MS–MS fragmentation data, precursor-to-production pair, cone voltage (CV) and collision energy (CE) for each analyte are described in Table II, and the product ions scan spectra are shown in Figure 2. All instrumentations were controlled and synchronized by MassLynx data systems (version 4.1) from Waters.

Table II.

UPLC–ESI-MS Data of the Five Components from A. tataricus

Compounds	MW	tR (min)	MS1 (m/z)	MS2 (m/z)	CV (V)	CE (eV)
Shionone	426.40	7.75	427.4a	95.1a	14	30
				109.1		24
Luteolin	286.23	2.74	285.0a	133.0a	−60	−32
				151.0		−26
Quercetin	302.23	2.75	301.0a	151.0a	−54	−20
				179.0		−18
Kaempferol	286.24	3.12	285.1a	93.0a	−62	−36
				151.0		−18
Ferulic acid	194.19	2.08	193.0a	134.0a	−34	−16
				178.0		−14

^aMonitored MRM transitions.

Figure 2.

The product ions scan spectra of the five compounds.

Preparation of standard solutions and calibration curves

The stock solutions of shionone, luteolin, quercetin, kaempferol and ferulic acid were individually prepared in methanol. The stock solutions of the standards were further diluted in methanol to produce combined standard working solutions. At least six concentrations of the solution were analyzed, and then the calibration curves were constructed by plotting the peak area versus the concentration of each analyte. All solutions were stored at 4°C before analysis.

Results

Method validation

Linearity

The linearity of the method was examined through the series of standard working solutions. Calibration curves were obtained by plotting the peak areas versus its concentrations of each analyte. As shown in Table III, the linear regression results indicated good linear correlation by the correlation coefficients of r² > 0.9970 for all the compounds in the concentration range.

Table III.

Regression Data, LODs and LOQs for the Investigated Compounds

Analytes	Regression equationa	Linear range (ng/mL)	r2	LODb (pg/mL)	LOQc (pg/mL)
Shionone	Y = 1.0429 × 103X + 2.7018 × 104	15.62–1,000	0.9973	12.21	244.1
Luteolin	Y = 1.4417 × 103X − 32.655	0.1855–11.88	0.9991	1.850	9.270
Quercetin	Y = 9.7646 × 102X + 2.6082 × 104	9.375–600.0	0.9970	1.830	7.320
Kaempferol	Y = 74.811X − 64.364	10.44–668.0	0.9971	54.36	163.1
Ferulic acid	Y = 6.3776 × 102X − 1.0821 × 102	0.2050–13.12	0.9977	2.050	10.25

^aY, peak area and X, concentration (ng/mL).

^bS/N = 3.

^cS/N = 10.

Limit of detection and limit of quantification

Limit of detection (LOD) and limit of quantification (LOQ) under the chromatographic conditions used were separately determined at S/N of 3 and 10, respectively. For each target constituent, the stock solutions mentioned above were diluted to a series of appropriate concentrations with methanol, and an aliquot of the diluted solutions was injected into UPLC for analysis to determine the LOD and LOQ. The LOD and LOQ values of all compounds are also summarized in Table III. It can be seen that the LOQ for the components is much lower than that obtained in preceding studies (2, 4, 5), which indicates that the analytical method was sufficiently sensitive.

Precision and stability

Measurement of intra- and interday variability was utilized to determine the precision of the method. The intraday precision was performed with six replications prepared from the A. tataricus sample 1 (HB-1) within 1 day, while the interday precision was performed over three consecutive days. The quantitation of five investigated ingredients was determined from the corresponding calibration curve. The relative standard deviation (RSD) was taken as a measure of precision. The overall intraday and interday precisions (RSD) for the investigated components were <2.84 and 2.99%, respectively. Stability of sample solution was tested at room temperature. The sample solution was analyzed in triplicate every 8 h within 48 h. All analytes were found to be stable within 48 h (RSD < 2.64%). These detailed results are summarized in Table IV.

Table IV.

Precision, Accuracy and Stability of the Five Active Components

Compounds	Precision (n = 6)		Accuracy (n = 3)					Stability (48 h, n = 3)
	Intraday RSD%	Interday RSD%	Original (ng)	Spiked (ng)	Found (ng)	Recovery (%)	RSD% (n = 3)	RSD (%)
Shionone	2.27	2.85	69.41	31.25	101.30	102.0	2.77	1.81
				62.50	131.68	99.62	2.29
				125.00	191.06	97.32	2.85
Luteolin	1.51	2.20	0.39	0.18	0.58	99.87	2.03	2.49
				0.37	0.76	100.3	2.57
				0.74	1.13	99.01	1.27
Quercetin	1.60	2.32	16.81	9.38	26.14	99.29	1.91	1.14
				18.95	35.56	98.92	2.04
				37.50	54.11	99.46	2.41
Kaempferol	2.84	2.75	60.33	20.88	81.36	99.25	1.78	2.14
				41.75	101.58	98.79	1.91
				83.50	142.32	98.19	1.54
Ferulic acid	2.52	2.99	3.77	1.64	5.37	98.58	2.13	2.63
				3.28	6.97	97.72	1.55
				6.56	10.36	100.6	1.69

Accuracy

Recovery test was used to evaluate the accuracy of this method. Known amounts of standard solutions were mixed with ∼1.0 g of A. tataricus (HB-1). Then, the resultant samples were extracted and analyzed with the above-established method, and triplicate experiments were repeated at three different concentration levels (high, middle and low). The percentage of recoveries was calculated by the formula:

$$\text{Recovery}(\mathrm{\%})=\frac{(\text{amount found})-(\text{amount original})}{\text{amount spiked}}\times 100$$

The average recovery was in the range of 97.32–102.0% with RSD ranging from 1.27 to 2.85% (Table IV). The results indicated that the method is accurate and reproducible.

Quantitative analysis of the investigated compounds in A. tataricus

The proposed UPLC–MS method was applied to analyze the five analytes in six batches of A. tataricus. The analysis time was reduced to 9 min by coupling the positive and negative modes in a single chromatographic run. Moreover, MRM scanning mode offered good sensitivity because it significantly decreased the noise levels and accordingly enhanced the response of analytes. Thus, some minor constituents in A. tataricus could also be accurately measured. The target compounds were identified on the basis of comparison of retention time, parent and product obtained from UPLC–MS-MS analysis of the standard compounds. The quantitative analysis was performed by means of the external standard methods. The data are summarized in Table V.

Table V.

Content of the Five Active Components in A. tataricus (n = 3)

Code	Content (mg/g)					Total
	1a	2	3	4	5
HB-1	3.63b	8.08 × 10−2	2.08 × 10−3	0.30	2.08 × 10−2	4.03
HB-2	2.69	5.94 × 10−2	1.65 × 10−3	0.17	1.58 × 10−2	2.94
HB-3	3.53	8.91 × 10−2	1.08 × 10−3	0.14	2.18 × 10−2	3.78
HB-4	4.35	1.04 × 10−1	6.16 × 10−4	0.20	1.34 × 10−2	4.67
HBQC	4.20	2.22 × 10−2	4.85 × 10−4	0.12	5.16 × 10−3	4.35
AHBZ	4.64	4.67 × 10−2	3.68 × 10−4	0.15	7.63 × 10−3	4.84

^aThe compounds are the same as in Figure 1.

^bAverage of duplicates.

Discussion

Optimization of the extraction procedure

To achieve the optimal extraction conditions, three important factors, namely, extraction methods, extraction solvents and extraction time that might influence the extraction efficiency of the target constituents, were optimized. The different levels of each factor including extraction methods (ultrasonic extraction versus heat reflux extraction), extraction solvents (50, 70 and 100% methanol) and extraction time (30, 40 and 60 min) were investigated individually using univariate approach. Ultrasonic bath extraction is a commonly used extraction method for the quantitative analysis of TCM. It has many advantages compared with heat reflux extraction, such as convenience, rapidness and the use of less solvent. Therefore, ultrasonic bath extraction was preferentially chosen as the extraction method. Based on the literature and our experience, a single-factor experiment was used. To evaluate the optimal solvent and duration of extraction, the total areas of the characteristic peaks in each chromatogram obtained using different conditions were compared. As shown in Figures 3 and 4, a comparative study on different extraction solvents of 50, 70 and 100% methanol (ethanol) and the duration of extraction for 30, 40 and 60 min were conducted at ambient conditions using ultrasonic bath extraction. Figure 3 indicates that the total peak area of 100% methanol was significantly higher than others. So we chose 100% methanol as the extraction solvent. As shown in Figure 4, there is no obvious difference in total peak area between 40 and 60 min, and they were both higher than 30 min. So we chose 40 min in this study to reduce the experimental duration. The interaction between different variables was ignored in our study. Therefore, the optimal condition for extraction of A. tataricus was as follows: 1.0 g powder of each dried sample was extracted with 100 mL of 100% methanol in ultrasonic bath for 40 min. We have used the same extraction method to extract the residue after extracting the sample. We found that each peak area difference of extract detection was <5% of the first extract from the herb. Therefore, the components that we quantified had been extracted completely.

Figure 3.

Extraction efficiency of different concentrations of solvents.

Figure 4.

Extraction efficiency of different extraction time of methanol.

Optimization of UPLC–MS-MS conditions

Full-scan and collision-activated dissociation tests were operated to set up an appropriate MRM method. The electrospray interface was used and good sensitivity fragmentation was obtained. The mass spectra showed that the ionization of shionone was more efficient in positive-ion mode than negative-ion mode, whereas others responded much better in negative-ion mode. So monitoring with positive- and negative-ion modes together should be selected. The optimized mass transition ion pairs (m/z) for quantization were 427.4/95.1 for shionone, 285.0/133.0 for luteolin, 301.0/151.0 for quercetin, 285.1/93.0 for kaempferol and 193.0/134.0 for ferulic acid. The fragmentation mechanism of precursor → product ion pairs was assumed to be the following: m/z 427.4 [M+H]⁺ → 95.1 [M+H–C₂₄H₄₄]⁺ for shionone, m/z 285.1 [M−H]⁻ → 93.0 [M−H–C₉H₄O₅]^– for kaempferol and m/z 193.0 [M−H]⁻ → 134.0 [M−H–CH₃–CO₂]^– for ferulic acid. Moreover, most of the quasi-molecule [M−H]⁻ ions for the flavone aglycones underwent a specific retro Diels–Alder reaction, which contributed to the product ions such as m/z 285.0 [M−H]⁻ → 133.0 [M−H–C₇H₄O₄]^– for luteolin and m/z 301.0 [M−H]⁻ → 151.0 [M−H–C₈H₆O₃]^– for quercetin.

To get the richest relative abundance of precursor and product ions, the parameters of fragmentation energy and collision were optimized. The protonated molecular ions [M+H]⁺ and deprotonated molecular ions [M−H]⁻ were considered stable and highly abundant; thus, they were chosen as the precursor ions for MS–MS fragmentation analysis of Compound 1 and Compounds 2, 3, 4 and 5, respectively. CV and CE are both necessary mass spectrometric parameter affecting ion response. Therefore, by automatic optimization, the maximum sensitivity and relevant ion pairs of compounds were acquired.

The UPLC conditions were optimized using a standard mixture of the compounds. It was confirmed that acidified aqueous acetonitrile mobile phase should be used in the analysis. In addition, acetonitrile displayed higher resolving power than methanol in the test. It was also found that an acidic eluent A (acetonitrile containing 0.1% formic acid, v/v) and B (0.1% aqueous formic acid, v/v) were beneficial for enhancing the ionization of compounds detected in positive electrospray interface mode and ensured sharp peak shape and reproducible retention time. Although ionization of compounds detected in negative electrospray interface mode was suppressed owing to existing of formic acid in mobile phase, their quantification was not influenced, which could be proved by good sensitivity and accuracy of analysis.

Furthermore, other chromatographic variables were also optimized, including analytical columns, column temperatures and flow rates. Eventually, the optimal condition of the Waters UPLC BEH C₁₈ column (50 × 2.1 mm, 1.7 μm) at the column temperature of 40°C with the flow rate of 0.3 mL/min was beneficial for enhancing the ionization of compounds detected both in positive and negative electrospray interface modes, and it could give good peak separation, sharp peaks and short analysis time. Figure 5 shows the typical extraction chromatograms of standards and samples.

Figure 5.

Representative extract ion chromatograms of the five compounds: (A) standards and (B) A. tataricus samples.

Sample analysis

The developed analytical method was applied to determine five compounds in six batches of A. tataricus. Quantitative analysis was performed by means of the external standard methods. Table V shows that the total contents of six batches of samples ranged from 2.94 to 4.84 mg/g. HB-4 and AHBZ had relatively higher total contents (4.66, 4.84 mg/g), while HB-2 and HB-3 had lower total contents (2.94, 3.78 mg/g). Shionone was the dominant compound in A. tataricus. The contents of shionone in these six batches of samples were all >2.5 mg/g and accorded with the requirement in the Chinese pharmacopoeia (2010), which should be >1.5 mg/g. Results were obtained to show that A. tataricus collected from Hebei (Samples 1–4), Hubei (5) and Anhui (6) province are reliable and high-quality herbs. On the other hand, the contents of luteolin, quercetin, kaempferol and ferulic acid were relatively low. The relative content of each composition was relatively constant in A. tataricus.

The six batches were further compared using Kruskal–Wallis H test. As shown in Table VI, it showed that five components exhibited a statistically significant difference in six batches when P-value was set to < 0.05. The greatest difference among the components was observed for quercetin, followed by luteolin, ferulic acid, kaempferol and shionone. As the major component, the content of shionone in AHBZ (4.64 mg/g) was ∼1.7 times higher than HB-2 (2.69 mg/g). In addition, the content of quercetin in HB-1 (2.08 × 10⁻³ mg/g) was ∼5.6 times higher than HB-2 (3.68 × 10⁻⁴ mg/g). However, multiple active components, including macro- and microcomponents, are frequently considered to be responsible for the therapeutic effects, and thus, the analysis of multiple components is more reasonable for quality control of TCM. The present work will be beneficial for quality control of A. tataricus.

Table VI.

Kruskal–Wallis H test of the five components in six batches

Compounds	χ2	df	P-value
Shionone	16.205	5	0.006
Luteolin	16.064	5	0.007
Quercetin	16.579	5	0.005
Kaempferol	15.596	5	0.008
Ferulic acid	16.251	5	0.006

Conclusion

An efficient, rapid and sensitive UPLC–ESI-MS method operating both positive and negative scanning modes in single analysis process was established for the quantification of five constituents in A. tataricus extract, and it is firstly reported. The recovery and repeatability of the method were appropriate. Compared with other methods, this method is proved to be more efficient, selective, sensitive and reliable for analysis of the five constituents. The result of this study may be helpful for studies on the pharmacokinetics of A. tataricus and may be beneficial for the application of clinical therapy.

Funding

We thank the financial support from the National Natural Science Foundation of China (no. 81473180).