Qualitative and Quantitative Analysis of Eclipta prostrata L. by LC/MS

Eclipta prostrata L. is one of the Chinese medicinal tonics which are usually used for treating loose teeth, dizziness, tinnitus, hemoptysis, hematuria, and uterine bleeding. However, quality control of this herbal medicine has been not satisfactory. This study reported its qualitative and quantitative analyses based on LC/MS method. UHPLC-DAD-Q-TOF-MS fingerprinting and MS fragmentation cleavage pathway were investigated for qualitative analysis. Furthermore, a method for simultaneous quantitative determination of nine compounds, luteolin 7-O-β-D-glucopyranoside, ecliptasaponin C, luteolin, eclalbasaponin IV, apigenin, ecliptasaponin A, echinocystic acid 28-O-β-D-glucopyranoside, echinocystic acid, and 3-oxo-16α-hydroxy-olean-12-en-28-oic acid in E. prostrata, was established. The method was validated for samples of E. prostrata from different habitats. The results showed good linear correlation, precision, accuracy, and repeatability that could be used for contents determination of the nine compounds in E. prostrata from different habitats.


Introduction
Eclipta prostrata L. (Compositae) is one of the Chinese medicinal tonics, widely distributed in the tropical and subtropical regions of the world. It has been used for the treatment of loose teeth, dizziness, tinnitus, hemoptysis, hematuria, and uterine bleeding [1]. Modern pharmacological research has confirmed its biological effects such as in antiosteoporosis [2], anti-inflammatory [3], antihyperlipemia [4,5], and antitumor [6] activities.
Literature on its phytochemical constituents and our previous study showed that it contained triterpenoid saponins, flavonoids, thiophenes, and steroids [6][7][8][9][10]. Several analytical methods including HPLC, UPLC, LC/MS, and GC-MS have been used for quality control analyses of E. prostrata [11][12][13][14][15]; however, the standards which those methods used were finite (one or two flavonoids which could be obtained on the market). Furthermore, analytical time of the reported methods was relatively long in order to obtain ideal resolution. Therefore, we intended to establish a method which is based on our previous phytochemical study to determine main constituents of E. prostrata. Although there was a report on qualitative analysis of E. prostrata through UHPLC-Q-TOF/MS [12], the results lacked to deduce fragmentation pathway since they did not contain MS/MS data.
In this paper, we demonstrated how modern analytical methods could be used for quality control on natural product medicine. We initially analyzed a UHPLC-DAD-Q-TOF-MS fingerprint for rapid profiling of chemical constituents, and eighteen compounds in the extract of E. prostrata were identified or tentatively characterized. A rapid LC-QQQ-MS method was later validated for simultaneous determination of nine major compounds in E. prostrata. The results showed good linear correlation, precision, accuracy, and repeatability that could be used for quality control analysis of E. prostrata from different habitats.   (Figure 1) were isolated from the aerial part of E. prostrata, and the structures elucidated based on 1 H-NMR and 13 C-NMR spectral analyses as previously reported by [10]. The purities of these reference compounds were determined to be above 98% by normalization of the peak areas detected by HPLC-ELSD (Alltech Grace Evaporative Light Scattering Detector 3300 with the following acquisition parameters: temp: 40 ∘ C; gas flow: 1.8 L/min; gain: 10). HPLC grade acetonitrile and methanol were acquired from Fisher Chemicals (Pittsburg, USA). Formic acid (HPLC grade) was obtained from Tedia, USA. The other chemicals and reagents used were of analytical grade and purchased from Tianjin Concord Technology Company (Tianjin, China). Purified water was obtained using a Milli-Q system (Millipore, USA).
Thirteen dried samples (S1-S13) from the aerial part of E. prostrata were collected from different habitats and identified by Professor Lijuan Zhang (College of Traditional Chinese Medicine, Tianjin University of Traditional Chinese Medicine, China). Voucher specimens (20120903) deposited in our laboratory.
The aerial part of E. prostrata was crushed into powder and 1.00 g weighed into a 50 mL flask with 25 mL 50% methanol (v/v). It was extracted in an ultrasonic bath at room temperature for 30 min. The extract was filtered. 1 mL of the filtrate was obtained and diluted to the mark in a 10 mL volumetric flask. It was then centrifuged at 13171 ×g for 10 min for LC/MS analysis.

UHPLC-DAD-Q-TOF-MS Fingerprint.
UHPLC-DAD-Q-TOF-MS fingerprint analysis was performed on an Agilent 1290 UHPLC consisting of a binary pump, a diode-array detector, an autosampler, and a column thermostat connected to an Agilent 6520 Q-TOF spectrometry system via an ESI interface (Agilent Corp., Santa Clara, CA, USA).
Full-scan analyses in negative ionization modes were conducted and the spectra were recorded in the range of m/z 100-1700. Liquid nitrogen was used as the nebulizer and drying gas. High purity nitrogen was used as the collision gas. The major parameters used were as follows: drying gas with a flow rate of 8.0 L/min, drying gas temperature 350 ∘ C, nebulizer 30 psig, capillary voltage 3500 V, fragmentor voltage 175 V, and collision energy 35, 70 V.
The ESI source parameter of drying gas was set at the flow rate of 8.0 L/min, temperature of 350 ∘ C, 45 psig nebulizer, and 4000 V capillary voltage. The MS/MS analysis was conducted on multiple reactions monitoring (MRM) mode; the MRM parameters were shown in Table 1.

Method Validation for Simultaneous Quantification.
Seven serial working solutions were prepared as described above and injected into the LC/MS system. Calibration curves were plotted based on linear regression analysis of the integrated peak areas ( ) versus concentrations ( , g/mL). Each solution was tested in triplicate, with limits of detection (LOD) and quantification (LOQ) for each analyte defined at signal-to-noise ratios (S/N) of 3 and 10, respectively. Intraday and interday precision for each analyte at a specific concentration were performed by six replicates on the same day (intraday) and on three consecutive days (interday). Recovery tests were performed by spiking reference standards into appropriately weighed sample 9. Six different samples were spiked with the reference standards, extracted, and prepared as described above. Three replicates were performed for each analysis. To confirm the repeatability, six replicates of the same sample were extracted and analyzed. Variations were expressed in terms of relative standard deviation (RSD) in all the tests.

Optimization of Extraction Procedure.
In order to obtain satisfactory extraction efficiency, several extraction solvents including water, 50% methanol (v/v), and methanol were examined. The 50% methanol (v/v) solvent was chosen and used as the extraction solvent due to its high yield of target compounds. Reflux and ultrasonic extraction methods were similarly effective in the extraction of the target analytes. The ultrasonic extraction method was ultimately chosen because of its flexibility. Different extraction times (30, 45, and 60 min) were compared, which showed similar percent yields and as such 30 minutes was chosen as the ideal extraction time.

Optimization of LC-QQQ-MS Chromatographic Conditions.
Chromatographic conditions of the mobile phase and gradient elution system were optimized in this study. In order to achieve good resolution and symmetric peak shapes of the nine reference compounds, we chose methanol-water (with and without acid) and acetonitrile-water (with and without acid) to optimize the mobile phase. We also optimized the column temperature from 25 to 45 ∘ C with 5 ∘ C in one step. Finally, acetonitrile-water (with 0.1% formic acid) and column temperature of 35 ∘ C were chosen which showed good resolution of adjacent peaks within a short time. Figure 2 shows the total ion current (TIC) chromatogram of the extract from E. prostrata (S9). A total of 18 compounds were identified and 13 of them confirmed by comparing their MS features and retention times with those of reference compounds. According to the structural characteristics, the 18 compounds can be grouped into three types, namely, flavonoids, triterpenoids, and other types. Three flavone aglycones were identified corresponding to the diagnostic fragment ions 1,3 B, 0,4 A, and 0,4 B-2H. Finally, Compounds 5, 8, and 9 were unambiguously identified as luteolin, apigenin, and 3 -hydroxybiochanin A by comparing their retention times, MS/MS fragment data, and UV spectra with reference compounds.

Identification of Triterpenoids.
The triterpenoids were the major components isolated from E. prostrata and compounds 10-16 were identified as triterpenoid saponins. In order to assist in structural elucidation, compound 10 was chosen as an example to elucidate the nomenclature ( Figure 4). The ions retaining the charges on the main core structures were termed Y representing glycosidic cleavages and X for cross-ring cleavages [16]. Cross ring cleavage ions were designated by superscript numbers indicating cleavage of the two bonds. The oligosaccharide chain at C-3 was defined as the -chain whereas the one at C-28 was thechain.
Since the molecular masses of saponins were usually large, we used two collision energies (CE) to elucidate the structures.  Figure 6 shows the elucidation of the detailed fragment cleavage pathway and similar fragmentation pathways were observed in the MS/MS spectrum of [M+HCOO] − ( Figure 5(d )). Through UV, MS fragmentation pathway, and retention times, compound 13 was unambiguously identified as eclalbasaponin IV.

Identification of Other
Compounds. Two compounds (2 and 6) were unambiguously identified as 3,4-dihydroxybenzoic acid ethyl ester and wedelolactone, respectively, by comparing their UV, exact molecular masses, MS/MS spectra (Table 2), and retention times.

Validation of Quantitative Method
3.4.1. Linearity. A series of standard solutions with seven different concentrations were analyzed by an established method in triplicate. Every calibration curve was plotted based on linear regression analysis of the integrated peak areas ( ) versus concentrations ( , g/mL) as listed in Table 3. Calibration curves were linear with correlation coefficients ( 2 ) above 0.9990 for all analytes.

LOD and LOQ.
The stock solutions containing nine reference compounds were diluted to a series of appropriate concentrations, using 50% methanol (v/v), and injected into LC/MS for analysis. The LOD and LOQ under the chromatographic conditions were determined at approximate signal-noise (S/N) ratios of 3 and 10, respectively. The results were given in Table 3.

Precision and Repeatability.
Intra-and interday precisions were performed by repetitive injections on the same day (intraday) for a total of six injections and on three consecutive days (interday). RSD values for both intra-and interday precision were below 2.5% ( Table 4). The analytic repeatability was examined by the injection of six different samples (S9), which were prepared with the same sample preparation procedure. The repeatability of the solution was less than 2.99% (Table 4).

Accuracy.
The accuracy of the method was determined by spiking an appropriate amount of each crude E. prostrata extract sample (S9) with accurate amounts of the nine reference standards and each sample was analyzed in triplicate. The results showed good accuracy with average recovery from