Analysis of Chemical Constituents of Melastoma dodecandrum Lour. by UPLC-ESI-Q-Exactive Focus-MS/MS

The ethnic drug Melastoma dodecandrum Lour. (MDL) is widely distributed throughout South China, and is the major component of Gong Yan Ping Tablets/Capsules and Zi Di Ning Xue San. Although the pharmacological effects of MDL have been well documented, its chemical profile has not been fully determined. In this study, we have developed a rapid and sensitive UPLC-ESI-Q-Exactive Focus-MS/MS method to characterize the chemical constituents of MDL in the positive and negative ionization modes. A comparison of the chromatographic and spectrometric data obtained using this method with data from databases, the literature and reference standards allowed us to identify or tentatively characterize 109 compounds, including 26 fatty acids, 26 organic acids, 33 flavonoids, six tannins, 10 triterpenoids, two steroids and six other compounds. Notably, 55 of the compounds characterized in this study have never been detected before in this plant. The information obtained in this study therefore enriches our understanding of the chemical composition of MDL and could be used in quality control, pharmacological research and the development of drugs based on MDL. In addition, this study represents the first reported comprehensive analysis of the chemical constituents of MDL.


Introduction
Traditional Chinese medicine (TCM) has been used for thousands of years to treat a variety of different diseases. Based on its theoretical therapeutic efficacy and wide range of clinical applications, TCM has received considerable interest from healthcare professionals, as well as those working towards the identification of new therapeutic agents for commercialization. In contrast to the pharmacological characteristics of single agent drugs, multicomponent drugs can exhibit synergistic pharmacological effects, through a "network" approach, where multiple compounds interact with multiple targets, pharmacokinetic or physicochemical synergisms in vivo with interdependent activities to achieve an improved optimal effect [1][2][3]. It is therefore essential to evaluate the chemical composition of each TCM, so that this information can be used to support further studies, such as drug effect, toxicity and metabolism studies.
Melastoma dodecandrum Lour. (MDL) is extensively distributed throughout the southern provinces of China, including Guizhou, Fujian, Zhejiang, Jiangxi and Yunnan. This plant is widely used for its medicinal properties by the Yao, Miao and She people, as well as several other minority groups. Modern pharmacological studies have shown that MDL exhibits several biological effects, including antihypoglycemic, hemostatic, analgesic, anti-inflammatory, blood lipid reducing, antioxidant and liver protection properties [4]. MDL has been used to treat a variety of different ailments, including dysmenorrhea, postpartum abdominal pain, metrorrhagia, leucorrhea, hematochezia, dysentery, carbuncle swollen and boils. To date, only 76 compounds have been isolated from MDL, including organic acids, flavonoids, triterpenoids and steroids [5][6][7][8][9][10][11][12][13]. However, much of the chemical composition of MDL remains unknown, making it difficult to rationalize its bioactivity or evaluate the safety of this material as a therapeutic agent. There is therefore an urgent need to develop an analytical method capable of determining the chemical composition of MDL. With this in mind, the aim of the current study was to establish a rapid and sensitive method for identifying the constituents of MDL. In this study, we used a Q-Exactive Focus MS/MS method to obtain high-resolution mass spectra of the different components. This method was proven to be an advanced, accurate and reliable tool for the comprehensive identification of compounds belonging to a wide range of structural classes [14][15][16][17][18][19][20][21]. Using this method, we tentatively identified a total of 109 compounds, highlighting the efficiency and accuracy of this new technique.

Results and Discussion
Melastoma dodecandrum Lour. was analyzed in the positive and negative ionization modes using a Q Exactive Focus mass spectrometer, and the base peak chromatogram (BPC) chromatograms for both of these ESI modes are shown in Figure 1. Some of the compounds found in this study were identified based on a comparison of their analytical data (i.e., retention times and high-resolution mass spectra) with those of several reference standards. Thus compounds 10, 34, 35, 46, 68, 79, and 84 were unambiguously identified as gallic acid, luteolin, kaempferide, quercetin, oleanic acid, asiatic acid, and rutin, repectively. Moreover, the fragmentation patterns and pathways of the standards helped further confirm the structures of the derivatives of the reference compounds. Compounds without reference standards were identified by determining the elemental compositions of the precursor and product ions. The molecular formula and rational fragmentation patterns and pathways of these compounds were then identified based on a comparison of these data with chemical databases and the literature as described below in Section 3.5. In this way, we used a UPLC-ESI-Q-Exactive Focus-MS/MS method in combination with available standards, databases and literature data to characterize 109 compounds from MDL. Seven of these compounds were unambiguously identified based on a comparison with the corresponding reference standards. Data for all of these compounds are summarized in Table 1.
Two different fragmentation patterns and pathways were observed for the flavonoids (I and II, shown in Figure 3). The retro-Diels-Alder (RDA) fragmentation (I) would result in the formation of A 1, 3 and B 1, 3 as the main fragment ions of the flavonoid moiety (aglycones) because of the X 1, 3 cleavage of the C ring. The substituent groups on the parent compounds were determined based on the compositions of the A 1, 3 and B 1, 3 fragments. The fragmentation of the parent compound according to Pattern I resulted in high-intensity fragment ions, whereas the fragmentation according to pattern II result in low-intensity fragment ions. Furthermore, the main fragmentation patterns of flavonoids    Two different fragmentation patterns and pathways were observed for the flavonoids (I and II, shown in Figure 3)          Table 2.  Two different fragmentation patterns and pathways were observed for the flavonoids (I and II, shown in Figure 3). The retro-Diels-Alder (RDA) fragmentation (I) would result in the formation of A 1, 3 and B 1, 3 as the main fragment ions of the flavonoid moiety (aglycones) because of the X 1, 3 cleavage of the C ring. The substituent groups on the parent compounds were determined based on the compositions of the A 1, 3 and B 1, 3 fragments. The fragmentation of the parent compound according to Pattern I resulted in high-intensity fragment ions, whereas the fragmentation according to pattern II result in low-intensity fragment ions. Furthermore, the main fragmentation patterns of flavonoids  Table 2.   Pentacyclic triterpenes are wildly distributed in Nature and consist of five rings, which are typically referred to as the A, B, C, D and E rings. Pentacyclic triterpenes can be divided into several different categories, depending on the nature of their E ring. In this study, we characterized two different types of pentacyclic triterpenes, including ursane-and oleanane-type pentacyclic triterpenes. The endo-double bond in pentacyclic triterpenes can readily undergo a RDA reaction during MS analysis (shown in Figure 4). Dehydration and decarboxylation are also observed as common fragmentation pathways in these systems during MS analysis [21][22][23]. Depending on the different substituents attached to their endo-double bonds and their accurate mass measurements, we were able to fully characterize all of the pentacyclic triterpenes found in MDL. As shown in Figure 5, we characterized a total of nine pentacyclic triterpenes using high-resolution MS 2 mass spectrometry, including four ursane-type and five oleanane-type compounds pentacyclic triterpenes.  Pentacyclic triterpenes are wildly distributed in Nature and consist of five rings, which are typically referred to as the A, B, C, D and E rings. Pentacyclic triterpenes can be divided into several different categories, depending on the nature of their E ring. In this study, we characterized two different types of pentacyclic triterpenes, including ursane-and oleanane-type pentacyclic triterpenes. The endo-double bond in pentacyclic triterpenes can readily undergo a RDA reaction during MS analysis (shown in Figure 4). Dehydration and decarboxylation are also observed as common fragmentation pathways in these systems during MS analysis [21][22][23]. Depending on the different substituents attached to their endo-double bonds and their accurate mass measurements, we were able to fully characterize all of the pentacyclic triterpenes found in MDL. As shown in Figure 5, we characterized a total of nine pentacyclic triterpenes using high-resolution MS 2 mass spectrometry, including four ursane-type and five oleanane-type compounds pentacyclic triterpenes. Pentacyclic triterpenes are wildly distributed in Nature and consist of five rings, which are typically referred to as the A, B, C, D and E rings. Pentacyclic triterpenes can be divided into several different categories, depending on the nature of their E ring. In this study, we characterized two different types of pentacyclic triterpenes, including ursane-and oleanane-type pentacyclic triterpenes. The endo-double bond in pentacyclic triterpenes can readily undergo a RDA reaction during MS analysis (shown in Figure 4). Dehydration and decarboxylation are also observed as common fragmentation pathways in these systems during MS analysis [21][22][23]. Depending on the different substituents attached to their endo-double bonds and their accurate mass measurements, we were able to fully characterize all of the pentacyclic triterpenes found in MDL. As shown in Figure 5, we characterized a total of nine pentacyclic triterpenes using high-resolution MS 2 mass spectrometry, including four ursane-type and five oleanane-type compounds pentacyclic triterpenes.    Pentacyclic triterpenes are wildly distributed in Nature and consist of five rings, which are typically referred to as the A, B, C, D and E rings. Pentacyclic triterpenes can be divided into several different categories, depending on the nature of their E ring. In this study, we characterized two different types of pentacyclic triterpenes, including ursane-and oleanane-type pentacyclic triterpenes. The endo-double bond in pentacyclic triterpenes can readily undergo a RDA reaction during MS analysis (shown in Figure 4). Dehydration and decarboxylation are also observed as common fragmentation pathways in these systems during MS analysis [21][22][23]. Depending on the different substituents attached to their endo-double bonds and their accurate mass measurements, we were able to fully characterize all of the pentacyclic triterpenes found in MDL. As shown in Figure 5, we characterized a total of nine pentacyclic triterpenes using high-resolution MS 2 mass spectrometry, including four ursane-type and five oleanane-type compounds pentacyclic triterpenes.

Tannins
The main structural features of tannins include their gallic acid ester moieties (or polymer) and glucose core (or other polyols). The MS 2 spectra of the tannins revealed fragment ions with m/z values of 1091, 939, 769, 617, 599, 447 and 277. The m/z differences between these fragment ions were 152 and 170 atomic mass units (amu), which indicated that the fragments ions were produced by the successive removal of O-galloylhyperin and gallic acid anions. The fragment ion observed with an m/z value of 169 was characteristic of the fragment ions derived from O-galloylhyperin, whereas the fragment ion observed with an m/z value of 125 was attributed to the loss of CO 2 from gallic acid [17]. A total of six other tannins were characterized in this way using high-resolution MS 2 mass spectrometry. The structure of tannins characterized were shown in Figure 6.

Tannins
The main structural features of tannins include their gallic acid ester moieties (or polymer) and glucose core (or other polyols). The MS 2 spectra of the tannins revealed fragment ions with m/z values of 1091, 939, 769, 617, 599, 447 and 277. The m/z differences between these fragment ions were 152 and 170 atomic mass units (amu), which indicated that the fragments ions were produced by the successive removal of O-galloylhyperin and gallic acid anions. The fragment ion observed with an m/z value of 169 was characteristic of the fragment ions derived from O-galloylhyperin, whereas the fragment ion observed with an m/z value of 125 was attributed to the loss of CO2 from gallic acid [17]. A total of six other tannins were characterized in this way using high-resolution MS 2 mass spectrometry. The structure of tannins characterized were shown in Figure 6.  The main structural features of tannins include their gallic acid ester moieties (or polymer) and glucose core (or other polyols). The MS 2 spectra of the tannins revealed fragment ions with m/z values of 1091, 939, 769, 617, 599, 447 and 277. The m/z differences between these fragment ions were 152 and 170 atomic mass units (amu), which indicated that the fragments ions were produced by the successive removal of O-galloylhyperin and gallic acid anions. The fragment ion observed with an m/z value of 169 was characteristic of the fragment ions derived from O-galloylhyperin, whereas the fragment ion observed with an m/z value of 125 was attributed to the loss of CO2 from gallic acid [17]. A total of six other tannins were characterized in this way using high-resolution MS 2 mass spectrometry. The structure of tannins characterized were shown in Figure 6.

Optimization of Analytical Conditions
To obtain better chromatographic separation and mass spectrometric detection, we evaluated three different mobile phase systems, including aqueous methanol, aqueous acetonitrile and aqueous acetonitrile-formic acid solutions. The aqueous acetonitrile solution resulted in the best separation of the major components of MDL. Furthermore, the addition of 0.1% formic acid to this mobile phase resulted in a considerable improvement in the symmetry properties of most of the chromatographic peaks. We also varied the flow rate (0.25, 0.3, and 0.35 mL/min), column temperature (30, 35, and 40 • C) and injection volume (2, 3, and 5 µL) during method development. The results of these optimization experiments established the following conditions for the chromatographic separation of the different components of MDL: mobile phase, aqueous acetonitrile containing 0.1% formic; flow rate, 0.3 mL/min; column temperature, 40 • C; and injection volume, 5 µL.

Structure Analysis Procedure
In the positive and negative scan mode, based on the high-accuracy precursor ions and product ions obtained from Q-Exactive Focus-MS/MS, the elemental compositions were calculated when the maximum tolerance of mass error for all the precursor ions and product ions was set at 1.5 ppm, which can satisfy the requirements for positive identification. Based on the elemental compositions of the precursors, the most rational molecular formula was sought in different chemical databases such as the Spectral Database for Organic Compounds SDBS (http://sdbs.db.aist.go.jp), m/z cloud (https: //www.mzcloud.org) and ChemSpider (http://www.chemspider.com). Meanwhile by searching literature sources, such as PubMed of the U.S. National Library Medicine and the National Institutes of Health, Scifinder Scholar of the American Chemical Society, Science Direct of Elsevier and Chinese National Knowledge Infrastructure (CNKI) of Tsinghua University, all components reported in the literatures on MDL and plants of the same family were summarized in a Microsoft Office Excel table to establish an in-house library [5][6][7][8][9][10][11][12][13] for searching the most rational molecular formula. When several matching compounds with the same formula were found, the fragmentation patterns and pathways of the compounds were analyzed and then validated by Mass Frontier 7.0 (Thermo Scientific) for positive identification.

Conclusions
A new UPLC-ESI-Q-Exactive Focus-MS/MS method was developed to analyze the chemical constituents of MDL based on their mass spectral fragmentation patterns. This new method resulted in the characterization of 109 compounds. The results of this study therefore provide an important reference to improve our understanding of the composition of MDL. We found that flavonoids are the main components of MDL, especially the flavonols, which possess a wide range of interesting pharmacological activities, such as anticancer, antibacterial, and antiviral activities. In terms of their structural characteristics, the triterpenoids found in MDL were ursane-and oleanane-type systems. Several tannins and steroids were also found in MDL. In addition to the fatty acids found in MDL, we found 55 other compounds that have never been reported in MDL besides fatty acids. Further studies pertaining to the chemical constituents in Melastoma dodecandrum Lour. are currently underway in our laboratory. Moreover, The study shows that, with the application of the UPLC-ESI-Q-Exactive Focus-MS/MS to characterizing the constituents of MDL, this method offers a rapid, sensitive and high throughput methodology for the identification of constituents of TCM prescriptions and herbal medicines.
Acknowledgments: This research was financially supported by Beijing Key Laboratory for Quality Evaluation of Traditional Chinese Medicine (No. BZ0386).
Author Contributions: Jinfeng Wang and Ruichao Lin designed the research, performed the experimental work, analyzed data and wrote the manuscript. Ziyao Jia, Yutong Wang and Xi Liu made substantial contributions to acquisition of data and interpretation of data. Zhihao Zhang gave suggestions for writing. Linghua Wang prepared the samples and drew the chemical structures. All authors read and approved the final manuscript.