Identification and Characterization of Key Chemical Constituents in Processed Gastrodia elata Using UHPLC-MS/MS and Chemometric Methods

Gastrodia elata Blume belongs to the Orchidaceae family. G. elata is often processed when used in traditional Chinese medicine (TCM). In the current study, a traditional processing method, known as “Jianchang Bang,” was applied. Steamed and dried (S&D) G. elata was processed with ginger juice for up to 5 days (GEP5D). An UHPLC-MS/MS combined with a chemometric method was developed for the analysis of processed G. elata along with the raw material as well as steamed and dried G. elata. As a result, the primary marker compounds were identified with the aid of TOF-MS and MS/MS analyses. Compared with the raw material of G. elata with GEP5D, three new parishin-type compounds were identified according to their retention time, accurate mass, and fragmentation patterns. The chromatographic peak areas for marker compounds, including S-(gastrodin)-glutathione, S-(4-hydroxybenzylamine)-glutathione, and parishin-type compounds, changed significantly. This result indicated that by applying the “Jianchang Bang” method, changes in chemical composition in G. elata contents were observed. The study also demonstrated that chemometric analysis is helpful in understanding the processing mechanism and will provide scientific support for the clinical application of G. elata.


Introduction
e dry tuber of G. elata, commonly called as Tianma, Chi Jian, or Ming Tianma, has been officially listed in the Chinese Pharmacopeia [1] to treat symptoms of hyperactive liver, arrest endogenous wind, and stop tetany according to the theory of traditional Chinese medicine (TCM). G. elata is used for the prevention and treatment of neuralgic and nervous diseases, such as headaches, migraine, dizziness, rheumatism, convulsion, and epilepsy, for centuries in TCM [2][3][4][5]. Natural products have been widely accepted as important sources for nutritional supplements, dietary products, and drug discoveries due to their remarkable and multiple medical benefits for human health. On this account, herbal plants and their bioactive ingredients, especially those with new compounds, have been studied to highlight some of their potential therapeutic values and subjected to extensive research.
ese structurally identified compounds were derived from organic solvent extracts of the rhizome which were produced from either fresh raw materials or from steamed and dried materials using different analytical techniques. In recent years, pharmaceutical studies suggested that phenolic glucoside gastrodin, gastrodigenin (p-hydroxybenzylalcohol), and parishins were identified as secondary metabolites responsible for biological activities [9,10]. Other compounds with relatively low concentrations, such as nucleosides, polysaccharides, and phospholipids, have also been reported in the literature [11,12]. e bioactive compounds in G. elata have increasingly become the focus of this research for important potential sources in developing new drug products [13].
G. elata has to be processed before clinical application according to the Chinese pharmacopeia. e processing of G. elata can cause changes in the content, potentially in the chemical property of the ingredients, or both of them, and influence its clinical efficacy. Processing methods, including mixing G. elata with wine, honey, Tribulus terrestris, wheat bran, or serofluid, have been recorded in some ancient books of TCM [14][15][16][17], but the most commonly used processing method is to steam and dry the tuber and then cut it into slices [18]. Among the main genres of TCM, "Jianchang Bang" is a genre located in the Nancheng county of Jiangxi Province in China. In this, steamed and sliced G. elata is mixed with ginger juice and then dried at room temperature. e processing method has been believed to be a special traditional processing means because of local clinical application of the theory of "toxicity reducing and efficacy enhancing." After processing with ginger juice, the plain property of G. elata would be turned to warm, according to the theory of TCM. Compared with the pharmacopeia processing method, the product processed with ginger juice by the "Jianchang Bang" method might have more power to relieve headache and reduce side effects of vomiting caused by the smell of G. elata. Although the processed product of G. elata with ginger has shown good efficacy practically, there is limitation for its clinical application owing to the lack of scientific evidences. erefore, finding out the differences in the chemical composition of G. elata obtained before and after processing by the "Jianchang Bang" method is the key to revealing the concoction mechanism that can be used for providing clinical application.
Many studies have been carried out and focused on the differences in composition changes in processed G. elata products [19][20][21][22]. Li et al. used the UHPLC/Q-TOF-MS/MS method for the identification of 64 chemical components in G. elata by comparing the retention time, accurate mass measurements, and characteristic fragmentation patterns [23]. Yet, these studies were only based on the conventional processing methods, such as steaming and drying or sulfur fumigation, rather than processing G. elata with accessories including ginger or wine. It is hard to determine and characterize the key chemical constituents and markers in processed G. elata using the "Jianchang Bang" method.
In order to understand the chemical composition of G. elata processed with ginger juice, a UHPLC-MS/MS combined with multivariate statistical analysis method was developed. e method facilitated the elucidation of characteristic fragmentation pathways of targeted and nontargeted constituents, especially of key chemical ones, in processed G. elata using the "Jianchang Bang" method by offering accurate mass measurement and significant molecular ions with different ionization modes, and then confirmed their molecular structures using the MS/MS fragmentation pattern. is analysis can also offer an excellent resolution of chromatographic separation of main components from different processed products of G. elata.
is is the first time that UHPLC-MS/MS and chemometric methods are being used for the identification of G. elata processed with ginger using the "Jianchang Bang" method.

Chemicals and Reagents.
Methanol, acetonitrile, and formic acid were of HPLC grade and purchased from Fisher Scientific (Fair Lawn, NJ, USA). Water used as the UHPLC mobile phase was purified using a Millipore Synergy UV Water Purification System (Millipore SAS, Molsheim, France). Reference standards parishin A (No: wkq17110609), parishin B (No: wkq17052501), and parishin C (No: wkq17120605) were purchased from Sichuan Vikki Biotechnology Co., Ltd. (Xi'an, China), and their purities were ≥98%.
Dried rhizomes of G. elata were obtained from Zhaotong county, Yunnan Province, China (Jiangxi Wuzhou Pharmaceutical Co., Ltd.) and identified by Dr. Lingyun Zhong (Jiangxi University of Traditional Chinese Medicine). Fresh ginger was purchased from a local supermarket in Mississippi, USA.

Raw and Processed Plant
Samples. G. elata (400 g) was steamed, cut into slices, and dried. Ginger juice (50 mL) was obtained by grinding fresh ginger using a mortar and pestle. Dried slices of G. elata (30 g) were mixed with the ginger juice (3.75 mL), which is the "Jianchang Bang" method.
ree different materials, such as the raw material, dried slices of G. elata after steaming, and the one processed by the "Jianchang Bang" method (steamed and dried slices of G. elata processed in ginger juice for 5 days) were frozen at − 20°C for a day and then taken out for freeze-drying (LABCONCO).

Sample Preparation.
After grinding to a fine powder using a pulverizing machine (Chart 5 Retsch. MM400), each sample (2.0 g), such as the raw material of G. elata, dried slices of G. elata after steaming, and samples processed by the "Jianchang Bang" method, was weighed in a centrifuge tube, then 3.5 mL of 50% methanol (methanol/water � 1 : 1, v/v) was added, and the mixture was sonicated in a water bath for 30 min. After that, the sample was centrifuged for 10 min at 959 ×g. e supernatant was transferred to a 10-ml volumetric flask. e procedure was repeated for a total of three times and the respective supernatants were combined. e final volume was adjusted to 10 ml with 50% methanol and mixed thoroughly. Prior to injection, an adequate sample (ca. 2 ml) was passed through a 0.45-μm PTFE membrane filter and collected in an LC vial.

UHPLC Chromatographic Conditions.
e UHPLC analyses were performed on a Waters Acquity UPLC ™ system (Waters Corp., Milford, MA, USA) that included a binary solvent manager, a sample manager, a heated column compartment, and a photodiode array (PDA) detector [24].

ESI-QTOF-MS.
e high-resolution mass spectrometric analyses were performed using electrospray ionization (ESI) in the negative mode on a Waters Xevo G2-S QTOF mass spectrometer (Waters Corporation). e MS instrument was operated in the following conditions: mass scan range of 100-1500 Da, capillary voltage of 2.0 kV, cone voltage of 40 V, source temperature of 80°C, desolvation temperature of 450°C, desolvation gas flow of 900 L/hr, cone gas flow of 50 L/hr, and collision energy of 6 eV. Leucine-enkephalin was used for lock mass correction at a concentration of 5 μg/ mL and a flow rate of 10 μL/min. Ions [M-H] − (m/z 554.2615 Da) and fragments at m/z 276.6268 Da of leucineenkephalin were applied to ensure mass accuracy during the MS analysis. e lock spray interval was set at 30 s, and the data were averaged over three scans. e mass spectrometer was programmed to switch between two steps: between low (10 V) and elevated (30-50 V) collision energies on the gas cell, using a scan time of 0.1 s per function.

Statistical
Analysis. PCA is one of the statistical analysis techniques used for finding the difference between different sample groups [25,26]. In the study, data were achieved using Waters MassLynx ™ Software v. 4.1 connected to MarkerLynx ™ XS Application Manager (Waters Corporation, Milford, USA). A series of procedures in the software were used to handle a retention time of 0.1 min, the chromatogram from 0.2 to 25 min, the mass range of 100-1500 Da, the mass window of 0.02 Da, and the noise elimination level of 10.00%. All compounds identified in G. elata samples before and after processing using ginger juice were distinguished via PCA and Orthogonal Partial Least Squares-Discriminant Analysis (OPLS-DA) through the MarkerLynx XS Application Manager, as shown in Figure 1. e model has 3 score components for the samples: G. elata raw material, dried G. elata, and G. elata processed with GEP5D. ese scores display a good separation for the three groups. e scores t[1] (x-axis) and t [2] (y-axis) are key variables in summarizing and separating the data. Each point in the plot corresponds to an observation. e groups are shown in different colors. is variation was assuredly caused by the use of the "Jiangchang Bang" method on G. elata in this study.

Results
Parishin A, B, and C ( Figure 2) are major components in G. elata and eluted at 14.19, 12.64, and 13.01 min, respectively, under optimized conditions (Figure 3). eir structures in G. elata were unambiguously confirmed by comparing the retention time, accurate mass, and fragmentation pattern with that of reference standards. For instance, parishin A is a conjugate comprising three gastrodin units based on a citric acid unit by ester linkage.  (Figure 4). e fragmentation pathway is proposed in Figure 5.
Parishin B and C are isomers that the gastrodin unit connects at 2′ and 3′, respectively ( Figure 2). ey showed different retention times but had fragmentation pattern identical to that of parishin A.
Compound 9 is eluted at 12.92 min and found to be a deprotonated molecular ion at 889.  20 ] − corresponding to the loss of 132 Da. is indicates that compound 9 contains a pentose. Meanwhile, fragments at m/z 757 Da are 30 Da higher than key fragments at m/z 727 Da in parishin A, B, or C, which suggests that the fragment at m/z 757 Da exists as a methoxy group. In G. elata, reported compounds parishin L and H contain a methoxy unit on their molecules [23]. erefore, the fragmentation pattern of compound 9 is proposed in Figure 6.
is compound is identified as a pentose derivative of parishin H or its isomer, a new marker compound identified only in the processed sample by the "Jianchang Bang" method.
Compound 16 is a new mark in GEP5D that eluted at 13.9 min and appeared to be deprotonated molecular ion at  Table 1.
Peak areas of many key marker compounds in Table 1 are different in raw, S&D, and GEP5D samples. ese compounds are mainly parishin-type compounds, except for compounds 1 and 5 which are derivatives of glutathione. As seen in Table 1, the peak areas of maker compounds in S&D samples are usually decreased when compared with that of raw samples. Among them, the most obvious decline is found for compound 12: the peak area of this compound in the raw sample is more than sixfold that of the dried sample at the same optimized analysis conditions. e peak area of compound 3 also experiences a great decrease after  processing G. elata by the drying method. Yet, there are few marker compounds, such as compounds 5, 8, and 11, whose peak areas increase slightly when compared with raw samples and dried samples. However, the peak areas of some compounds, such as compounds 1, 2, 3, 12, 15, and 19, in GEP5D samples have been found to be lower than that of raw samples but higher than that of S&D samples. Moreover, the peak areas of compounds 6 and 11 in GEP5D samples are greater than that of raw samples. On the basis of MS and MS/ MS data, it has been found that pentose or hexose derivatives of parishin-type compounds can be accumulated in G. elata samples processed by the "Jianchang Bang." is observation has been proved by the identification of new marker compounds 9, 14, and 16.
In Table 1, parishin-type compounds are characterized following the fragmentation pattern of parishin A, B, and C according to MS and MS/MS data. Two derivatives of glutathione are determined as follows. Deprotonated     S] − correspond to the loss of the 4-hydroxy benzyl group. erefore, compound 5 is tentatively identified as S-(4-hydroxybenzyl)glutathione [25]. Compound 1 at 9.37 min (m/z 574.1700, [C 23 H 32 N 3 O 12 S] − , calc. 574.1712) is 162 Da higher than compound 5, which suggests that compound 1 is a hexose derivative of compound 5. Considering gastrodin as one of the main components, compound 1 is characterized as S-gastrodin glutathione, and its fragmentation pathway is shown in Figure 8.

Discussion
For the UHPLC/Q-TOF-MS/MS analysis, different UHPLC columns, column temperatures, mobile phases, gradient elutions, and flow rates were evaluated to achieve the optimal separation. As a result, acetonitrile-water with 0.05% formic acid combined with the optimized gradient elution on an UPLC HSS T3 column at 45°C offered an appropriate chromatographic separation and MS responses in the negative ion scan mode. e OPLS-DA was used for the identification and differentiation of potential markers based on their peak  Figure 8: Proposed fragmentation pathway for compound 1.
areas [27]. As the most important observation, three new markers, such as compounds 9, 14, and 16, along with at least 17 known compounds in G. elata were identified by the differences in compounds obtained from raw and GEP5D samples. In total, 20 key marker compounds were obtained from raw, S&D, and GEP5D samples based on differences identified by the comparison of their peak areas under the same optimized analysis conditions (Table 1). An UHPLC-MS/MS method combined with multivariate statistical analysis was conducted to determine and characterize key marker compounds in G. elata raw material and processed samples by steaming or the "Jianchang Bang" method. Because of the significant evaluating means which reflected the changes in major compounds in different samples of G. elata processed by ginger, new marker compounds have been found and many important constituents experienced great changes between raw and GEP5D or raw and dried samples in peak areas.

Conclusions
In this study, unsupervised PCA can differentiate G. elata raw material, steamed samples, and samples processed by the "Jianchang Bang" method. OPLS-DA of raw material vs. GEP5D samples processed by the "Jianchang Bang" method provides a tool to select and determine marker compounds. A total of 20 marker compounds were characterized based on the accurate mass of MS and fragment ions of MS/MS data as well as by comparing with corresponding reference standards and literature. Compounds 9, 14, and 16 are new markers of parishin-type compounds in G. elata samples processed with ginger by the "Jianchang Bang" method. Except for two derivatives of glutathione (1 and 5), other marker compounds are parishin-type constituents. ese results suggest that the developed UHPLC-MS/MS method combined with PCA and OPLS-DA can characterize key components in G. elata processed with ginger. is method will support the clinical application of G. elata processed by the "Jianchang Bang" method in a scientific way.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
No potential conflicts of interest were reported by the authors.