Structural Characterization of Chemical Compounds Based on Their Fragmentation Rules in Sophorae Fructus by UPLC-QTOF-MS/MS

This study aims to identify the chemical components in Sophorae Fructus , and explore the mass spectrometric cleavage rules using the UPLC-Q-TOF-MS/MS method. The main characteristic fragments of the compounds were analyzed by electrospray ionization (ESI) ion source under positive and negative ion modes. The compounds were identi ﬁ ed by molecular formula, multistage mass spectrometry, ultraviolet spectrum, and the fragmentation patterns of standards. A total of 142 compounds were identi ﬁ ed, including 68 ﬂ avonoids, 39 saponins, 21 organic acids, and 14 others, of which 43 substitution position, dehydrated nucleus,


Introduction
Sophorae Fructus is the dry and mature fruit of Sophora japonica (L.), a leguminous plant. It has the functions of clearing away heat and toxic material, cooling blood and stopping bleeding, and is usually used for treating intestinal heat, hematochezia, nevus swelling and bleeding, dizziness, as well as red eyes. 1 It also has anticancer and estrogen-like effects, and plays a roles in prevention and treatment of cardiovascular disease, osteoporosis, and female menopause syndrome. 2 The study of the chemical components of S. japonica is of great significance for its quality control and clinical application. The main components of Sophorae Fructus are flavonoids, isoflavones, alkaloids, triterpenoid saponins, amino acids, stearic acids etc., among which isoflavones and their glycosides are the highest. However, up to now, there are few reports on the analysis of the total components of Sophorae Fructus. Sun et al identified and inferred 24 common compounds and 21 variance compounds in Sophorae Fructus from different producing areas by ultra-performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-Q-TOF/MS). 3 Zhang identified 131 compounds from the total extract of Sophorae Fructus in positive and negative ion modes by UPLC-Q-TOF/MS. 4 The structural types include 81 flavonoids, 18 triterpenoid saponins, 5 steroids, 2 anthraquinones, 3 phenols, and 22 others. Triterpenoid saponins were less identified, and flavonoids were mainly reported.
Traditional methods for phytochemical isolation and identification are time-consuming and labor-intensive. In recent years, techniques combining the efficient separation ability of liquid chromatography and strong identification ability of MS have been widely used in the separation and qualitative and quantitative analysis of complex Chinese medicines. 5,6 In this study, UPLC-Q-TOF/MS was used to rapidly identify the chemical composition of Sophorae Fructus. We established a UPLC-Q-TOF/MS qualitative analysis method to analyze the constituents of Sophorae Fructus, which lay a foundation for the study of its pharmacodynamic substance basis and quality control.
At present, the studies on MS pyrolysis are relatively scattered. The electrospray ionization MS of saponin 7,8 and flavonoid 9,10 components has been reported. Flavonoid species can be identified according to the ultraviolet (UV) absorption characteristics of compounds and the characteristic ion fragments of the parent nuclei. But there are few literature reports on determining the connection mode between glycosyl groups in flavonoid glycosides by using the cleavage rule of MS. In this study, through the comparison of the MS data of 21 reference substances, including flavonoid oxyglycosides, flavonoid carbon glycosides, dihydro-flavonoid glycosides, isoflavone glycosides and saponins, and a large number of literature reports, we systematically deduced the cleavage characteristics of these compounds, so as to provide reference for the MS structure identification of such components.
(v/v) (250 W, 40 kHz) for 60 minutes. The sample solution and standard solution were filtered through 0.22 µm microporous filter membrane.

Instrumentation and Conditions
The UPLC-QTOF MS/MS analysis was performed using a Waters Acquity UPLC system coupled with a Xevo G2-XS QTOF mass spectrometer (Waters, United States) with an electrospray ionization ion source in MS E mode.
The chromatographic separation process of flavonoids was performed on an ACQUITY CSH C18 (150 mm Â 2.1 mm, 1.7 μm; Waters, United States) at 35°C, with a mobile phase consisting of methanol (B) and 0.05% formic acid aqueous solution (A). The gradient elution was as follows: 0-9 minutes, 10-20% eluent B; 9-27 minutes, 20 MS conditions were operated in both positive and negative ion modes and applied as the following: solvent gas temperature (nitrogen), 450°C; capillary voltage, 3.0/2.5 KV; an ion source temperature, 120°C; desolvation gas flow, 500 L/h; cone gas flow, 100 L/h; the low collision energy, 6 V; the high collision energy, 25 to 60 V.

Data Processing and Compound Identification
Masslynx 4.1 software (Waters, United States) was used to analyze the mass spectra peaks of Sophorae Fructus in positive and negative ion modes. According to the comparison of reference standards or references, the compounds were identified by UV spectrum, retention time, excimer ion peak, molecular formula, fragment ions, and other information combined with Scifinder database.

Results and Discussion
To systematically and qualitatively analyze the chemical components in Sophorae Fructus, the MS behavior of the existing reference standards was studied to summarize their chromatographic retention behavior, UV absorption, cracking rule, and characteristic fragment ions.

The Cracking Rules of the Deglycosylation Group of Flavone-O-diglycoside
Kaempferol-3-O-sophoroside (t R ¼ 26.28 minutes) and kaempferol-3-O-gentiobioside (t R ¼ 28.33 minutes) are iso-mers, their mass spectra in negative and positive ion modes are shown in ►Fig. 1. In the negative ion mode (►Fig. 0568. This shows that whether in the positive or the negative ion mode, the glucose connected with 1!2 at the end was lost first, then the rhamnose connected with 1!6 was lost, and finally the glucose connected with aglycone was lost (►Fig. 2B).
Pharmaceutical Fronts © 2022. The Author(s). Therefore, we come to the conclusion that: (1) the polarity of flavonol glycoside substitution at the C 3 and C 7 sites is greater than that of glycoside substitution at the C 3 site only. (2) Flavonol glycosides replaced by glycogroups at the C 3 and C 7 sites lose the glycogroups on the C 7 position first and then the glycogroups on the C 3 position is lost in turn in the anion mode. In the positive ion mode, the glycogroups at the end of position C 3 were lost successively, and then the glycogroups at the C 7 site were lost, which was consistent with the pyrolysis rule of flavonoids in Herba Epimedii in the positive ion mode described in the literature. 13 (3) The three monosaccharides in flavone-O-triglycoside are connected to each other. Whether in the positive or negative electrode, the glycosyl connected at the end with 1!2 is lost first, then the glycosyl connected with 1!6 is lost, and finally the loss of the glycosyl connected with aglycone.   [14][15][16][17] in the positive ion scanning mode, the continuous dehydration of glycosyl mainly occurred, and the negative ion scanning mode has more obvious mass spectrum characteristics than the positive ion scanning mode.

Cleavage of Dihydroflavonoid Glycosides and Isoflavone Glycosides
Through the secondary mass spectra of naringin (t R ¼ 29.63 minutes), hesperidin (t R ¼ 30.23 minutes), and neohesperidin (t R ¼ 31.42 minutes) (►Table 1), we found that: (1) the polarity of dihydroflavonoid glycosides connected in the way of 1!2 between the monosaccharides substituted on the C 7 position of dihydroflavonoid glycosides is less than that connected in the way of 1!6.  -120]with strong abundance, and the fragment ion [M -H -162]could be observed. Hesperidin did not appear as these special ions. It may be due to that the rhamnose linked to the hydroxyl in the C 2 position of the glucose at the end of aglycone has rearranged and cleaved and the ion [M -H -120]resulted from the loss of a hexose residue in positions 0-3. The conclusion is to be proven by further experiments.
(3) In the positive ion mode, the relative abundances of ions [M þ K -Rha] þ and [M þ K -O -Rha] þ were higher than those of 1!6 connected when the C 7 -substituted rhamnose and glucose are 1!2 connected. In the negative ion mode, the rhamnose linked to the hydroxyl in the C 2 position of the glucose at the end of aglycone is more likely to rearrange, which may be related to the different charge distribution in the positive and negative ion modes. In the positive ion mode, the charge is mainly concentrated on the added sodium ions; and in the negative ion mode, the charge is mainly distributed in the whole sugar chain. 18 By comparing genistin (t R ¼ 24.04 minutes), sophoricoside (t R ¼ 28.72 minutes), and sophorabioside (t R ¼ 30.28 minutes), we found that the polarity of isoflavone glycosyl substitution on the C 7 position is greater than that at position C 4' (►Fig. 4). The glycosyl group of genistein substituted at position C 7 only lost 120 fragment ions at the negative electrode. Both in the positive and negative electrodes, the glycosyl groups of sophoricoside and sophorabioside substituted at position C 4' detected the loss of 120 fragment signal, and the positive signal intensity is higher. However, whether the lost fragment signal (120 U) can be used as the diagnostic fragment of isoflavone glycosides needs further research.

Mass Spectrometric Cleavage of Saponins
Full scan and mass spectrometric cleavage analysis were performed for saponin standard under positive and negative ion modes. The analysis results of characteristic fragments are shown in ►Table 2, and the mass spectrometric cleavage pathway of asperosaponin VI is shown in ►Fig. 5. The summary rules are as follows: (1) In the negative ion mode, the saponin parent nucleus fragments are not obvious, mainly the deglycosylated fragments and [M þ Cl] -, [M -H]excimer ion peaks; in the positive ion mode, a series of dehydrated fragments and [M þ Na] þ excimer ion peaks in the mother nucleus were mainly detected, while the response of deglycosylated fragments was weak. (2) The glycosyl group at C 3 position in asperosaponin VI, mogroside V, and ginsenoside Re is the last to fall off. Asperosaponin VI first lost the glycosylation at position C 28 , ginsenoside Re first lost the glycosylation at position C 20 , and mogroside V first lost the glycosylation at position C 23 . This may be due to the ester bond and the ether bond on the straight chain is easier to break than the ether bond on the C 3 ring. (3) In the positive ion mode, the dehydration reaction of saponin parent nucleus fragments is only related to the number of hydroxyl groups carried on the mother nucleus, not related to the type of saponin, sugar chain

Identification of Flavonoids
When analyzed the test sample according to chromatographic conditions (1), under the negative (►Fig.
can be observed. Combined with the above flavone cleavage rules, rhamnose was bound to a phenolic hydroxyl at position C 7 , dihexosides with interglycosidic linkage 1!2 was substituted at position C 3 . Therefore, compound 35 was proposed to be isorhamnetin-3-O-sophoroside-7-O-rhamnoside by comparison with the literature. 11

Identification of Saponins
After analyzing the test sample according to chromatographic conditions (2), the total ion flow diagram is shown in ►Fig. 6C (negative ion scanning mode) and ►Fig. 6D (positive ion scanning mode). A total of 58 compounds were identified by using the above saponin cleavage rules in combination with relevant literature and reference standards, including 39 saponins, 10 phenolic acids, 3 fatty acids, 2 phenylpropanoids, 1 flavonol, and 3 others (►Table 4).
Triterpenoid saponins are mainly contained in Sophorae Fructus, and the structure is mostly oleanene type. The sugar chain structure in saponins is easy to be removed during cracking. If it is a branched glycosyl group and the two terminal glycosyl groups are different, the fragment peaks that lose the two terminal glycosyl groups will appear, so it is easy to distinguish between branched glycosyl groups and straight chain glycosyl groups. 21 Compound 104 in ►Table 4 is taken as an example to derive the cracking rule of these compounds.        Structural Characterization of Chemical Compounds Based on Their Fragmentation Rules He et al.          22 it was speculated that the compound may be azukisaponin V.

Conclusion
In this experiment, the UPLC-Q-TOF-MS/MS method was used to quickly characterize the chemical components of Sophorae Fructus in positive and negative ion modes. The cracking rules of main flavone glycosides and saponins, which were preliminarily discussed, were helpful to improve the structural analysis efficiency and provide reference for the rapid screening and identification of flavonoids and saponins. From the data presented, 142 compounds were analyzed and inferred, including 67 flavonoids, 39 saponins, 18 organic acids, 10 amino acids and sugars, 2 phenylpropanes, 3 fatty acids, and 3 other types. A total of 43 components were first reported from the genus Sophora. This work will be helpful for the further study of pharmacodynamic material basis and quality evaluation of Sophorae Fructus.
Through the application of LC-MS technology, the repeated identification of known compounds by traditional separation and purification methods is avoided, which is conducive to saving resources, increasing the discovery probability of new compounds, and effectively improving work efficiency. It provides ideas and methods for the basic research and new drug development of traditional Chinese medicine and other complex substrates.

Supporting Information
The chemical structures of the 32 reference substances can be seen in the Supporting Information (►Fig. S1 [online only]).

Conflicts of Interests
None declared.