Metabolite Profiling Analysis of the Tongmai Sini Decoction in Rats after Oral Administration through UHPLC-Q-Exactive-MS/MS

Tongmai Sini decoction (TSD), the classical prescriptions of traditional Chinese medicine, consisting of three commonly used herbal medicines, has been widely applied for the treatment of myocardial infarction and heart failure. However, the absorbed components and their metabolism in vivo of TSD still remain unknown. In this study, a reliable and effective method using ultra-performance liquid chromatography coupled with hybrid quadrupole-Orbitrap mass spectrometry (UHPLC-Q-Exactive-MS/MS) was employed to identify prototype components and metabolites in vivo (rat plasma and urine). Combined with mass defect filtering (MDF), dynamic background subtraction (DBS), and neutral loss filtering (NLF) data-mining tools, a total of thirty-two major compounds were selected and investigated for their metabolism in vivo. As a result, a total of 82 prototype compounds were identified or tentatively characterized in vivo, including 41 alkaloids, 35 phenolic compounds, 6 saponins. Meanwhile, A total of 65 metabolites (40 alkaloids and 25 phenolic compounds) were tentatively identified. The metabolic reactions were mainly hydrogenation, demethylation, hydroxylation, hydration, methylation, deoxylation, and sulfation. These findings will be beneficial for an in-depth understanding of the pharmacological mechanism and pharmacodynamic substance basis of TSD.


Introduction
The classical prescriptions of traditional Chinese medicine (TCM) have originated from the fixed combination of certain kinds of herbal medicines recorded in the ancient classics, which have been still widely used in East Asia and exhibited precise clinical efficacy, with obvious characteristics and advantages [1].For a long time, some classic prescriptions have been developed into modern Chinese medicinal preparations through the optimization of preparation technology and drug development research [2,3].Most classical prescriptions have existed for at least hundreds of years, and with time, classic prescriptions may have changed to some extent, while their core characteristics (e.g., composition of herbs, proportion of herbs, etc.) have not changed significantly [4,5].The reasons for the inheritance of classical prescriptions to the present day can be attributed to the high safety of the prescriptions and their proven efficacy due to a large number of clinical applications.
Tongmai Sini decoction (TSD) is a classic formula from the Chinese medical masterpiece "The Treatise on Typhoid Fever", written 1800 years ago.It consists of three herbal concoctions of Radix Aconiti Lateralis Preparata (RALP), Rhizoma Zingiberis (RZ), and Radix Glycyrrhizae Preparata (RGP) and is commonly used in modern times for myocardial infarction and heart failure, atherosclerosis, shock, diarrhea, etc. [6,7].TSD has the effects of raising blood pressure, strengthening the heart, anti-hypoxia, anti-shock, anti-thrombosis, anti-myocardial ischemia, anti-slowing arrhythmia, and so on [8,9].The main chemical constituents of TSD include alkaloids (from RALP), phenolic acids and saponins (from RGP), and volatile oils (from RZ).At present, chemical composition [10], pharmacological, pharmacokinetic [11][12][13], and metabolomics [8,9,14] studies have been preliminarily conducted on TSD, especially on its cardiovascular activities.Most of the studies on TDS concentrated on the pharmacokinetics of diterpene alkaloids after oral administration of TDS, and some of the studies focused on the changes in the in vivo metabolome or lipidome profile against myocardial ischemia, heart failure, hypothyroidism.There is a lack of systematic and in-depth in vivo chemical and metabolite studies of TDS.
The components of different botanicals enter the body and produce metabolites, which exert therapeutic effects through multiple pathways [15].To fully understand the therapeutic components, it is necessary to first analyze the blood-entering components and their metabolites, as well as to study the metabolites distributed in plasma, urine, feces, and tissues, which is conducive to analysis of the potential components and pathways of action in the body [16,17].
High-resolution mass spectrometry (HRMS), in combination with chromatography technology, has provided useful structural information about chemical components, offering strong support for the characterization of in vivo and in vitro metabolic components of botanicals [18][19][20].In recent years, in order to improve the sensitivity and selectivity of obtaining MS/MS or MS n data for trace components in vivo, some acquisition and identification strategies have been developed, combined, and applied, for instance, the extracted ion chromatogram (EIC), mass defect filter (MDF), dynamic background subtraction (DBS) and neutral loss filter (NLF) [21][22][23].
In this paper, the established UHPLC-Q-Exactive-MS/MS methods have great advantages for the qualitative analysis of bioactive samples in rats after oral doses of TSD, and a variety of post-data processing techniques, including EIC, MDF, DBS, and NLF was applied for quickly screen and systematically identify the metabolites.These metabolic studies can provide the chemical foundation and an in-depth understanding of metabolic transformation for further research on effective substances and the action mechanism of TSD.

Plant Extract Preparation
According to the documentary records of TSD, the TSD pieces that included RALP (30 g), RZ (20 g), and RGP (30 g) were soaked with 8 times the amount of water for 30 min and decocted to boil (100 • C) for 2 h.The filtrate was collected, and the residue was decocted in 8 times the amount of water for 1.5 h again.The hot filtrate was combined and concentrated to 80 mL (1 g herbal pieces/1 mL aqueous solution).The obtained TSD extract was stored at −20 • C before use.

Animal and Drug Administration
Male Sprague-Dawley rats (220-260 g) were obtained from Guangdong Provincial Medical Laboratory Animal Center (Guangdong, China Six SD rats were randomly divided into two groups (Urine and plasma groups) and adapted to the metabolic cage for a week before the experiment.Blank urine and plasma samples were collected under abrosia state ahead of gastric gavage.The rats were fasted for 14 h with water ad libitum before oral administration of TSD extract and underwent 4 h of water deprivation after that.TSD extract was orally administered to rats of urine and plasma groups twice at an interval of 1 h, and the dosage was 2 mL per 100 g bodyweight per time.

Sample Collection and Pretreatment
Urine samples from 0 to 24 h after the second dosing were collected and stored at −80 • C prior to analysis.Plasma samples were obtained at 1, 2, 4, 8, and 12 h after the second administration in heparinized 1.5 mL polythene tubes under diethyl ether anesthesia, respectively.All plasma samples were centrifuged at 4000 rpm for 10 min, and the plasma supernatants were then merged in equal volume and frozen at −80 • C prior to analysis.
The collected urine and plasma samples (200 µL) were added with 4× the volume of acetonitrile-methanol (3:1) to precipitate protein, respectively.All separate supernatants were dried under N2 flow, and the residues were resuspended in 200 µL acetonitrile and centrifuged at 15,000× g for 8 min.Finally, a 5 µL sample was injected into the UHPLC-Q-Exactive-Orbitrap MS system for further analysis.

Instrumentation and Conditions
LC analyses were conducted on a Thermo UltiMate 3000 UHPLC system (Thermo Fisher Scientific, San Jose, CA, USA) equipped with a quaternary pump, a cooling autosampler, and a thermostatically controlled column oven.An ACQUITY UPLC HSS T3 Column (2.1 × 100 mm, 1.8 µm) was used.The mobile solvents were composed of acetonitrile (A) and water with 0.02% formic acid (B), and the gradient elution profile was employed as follows: 5% A, 0 min; 16% A, 12 min; 55% A, 23 min; 90% A, 35 min; 95% A, 40 min; returning to initial conditions in 4 min at a flow rate of 200 µL/min at room temperature.The injection volume was 5 µL.The temperatures of the sample tray and the column oven were set at 4 and 35 • C, respectively.
A Q-Exactive hybrid quadrupole-orbitrap mass spectrometer was connected to an LC system via an electrospray ionization source as an interface.Data acquisition and processing were calculated using Compound Discoverer 3.2 software.The optimized parameters for MS analysis were as follows: the mass spectrometer parameters were positive (PI) and negative (NI) ion mode; the resolution of the Orbitrap mass analyzer was set as 30,000; ion spray voltage was −3.8 kV; the capillary temperature was 325 • C; the sheath gas flow rate was 40 psi; the auxiliary gas flow rate was 8 psi; and the mass range was m/z 150-1500.The properties of data-dependent MS 2 scanning (DDS) parameters and events were as follows: resolution, 17,500; HCD, 35 eV; repeat count, 2; exclusion list, 50; repeat duration, 5 s; and exclusion duration, 30 s.The mass error for molecular ions of all compounds identified was within ±5 ppm.

Systematic Analytical Strategy for Online Metabolite Analysis
Based on our previous research on the cleavage patterns of components in RALP and RGP and a review of the literature [24][25][26][27][28], the metabolite profiling of TSD was systematically investigated by UHPLC-Q-Exactive-MS/MS methods.The workflow of the analytic procedure was carried out and shown in Figure 1.Figures S1 and S2 (Supplementary Materials) displayed the detailed workflow for the identification of prototype components and metabolites, respectively.systematically investigated by UHPLC-Q-Exactive-MS/MS methods.The workflow of the analytic procedure was carried out and shown in Figure 1. Figure S1 and S2 (Supplementary Materials) displayed the detailed workflow for the identification of prototype components and metabolites, respectively.The strategy consisted of the following steps: (1) First, the chemical database (Table S1, Supplementary Materials) was constructed, including mass weights, elemental compositions, and structure information of chemical compositions originating from RALP, RGP, and RZ based on our previous research and the related literature [24][25][26][27][28][29].(2) Then, an online full-scan and MS/MS data acquisition was processed in both negative and positive modes based on the DBS and DDS techniques for potential metabolite detection.(3) Next, the data files were imported into the Compound Discoverer 3.2 software, and the data-mining tools of EIC, NLF, and MDF were applied to screen the possible metabolites of TSD.Table S2 (Supplementary Materials) showed the detailed parameters of data processing.The main compounds with mass spectral peak areas greater than 10 8 in the decoction (shown in Table 1) were used as parent compound templates for MDF data screening (±50 mDa) (4) Next, based on the chemical database, acquired accurate mass data, retention time, and characteristic fragment ions, the identification of prototype components was elucidated (shown in Table 2).In addition, the Clog p values calculated by ChemDraw 14.0 were used to distinguish isomers at different retention times.( 5) Finally, the mass information of potential metabolites, as well as their possible biotransformation pathways and composition change given by Compound Discoverer 3.2, were compared by the data of prototype components and the related literature to verify the metabolites and their metabolic pathways (shown in Table 3).The strategy consisted of the following steps: (1) First, the chemical database (Table S1, Supplementary Materials) was constructed, including mass weights, elemental compositions, and structure information of chemical compositions originating from RALP, RGP, and RZ based on our previous research and the related literature [24][25][26][27][28][29].(2) Then, an online full-scan and MS/MS data acquisition was processed in both negative and positive modes based on the DBS and DDS techniques for potential metabolite detection.(3) Next, the data files were imported into the Compound Discoverer 3.2 software, and the data-mining tools of EIC, NLF, and MDF were applied to screen the possible metabolites of TSD.Table S2 (Supplementary Materials) showed the detailed parameters of data processing.The main compounds with mass spectral peak areas greater than 10 8 in the decoction (shown in Table 1) were used as parent compound templates for MDF data screening (±50 mDa) (4) Next, based on the chemical database, acquired accurate mass data, retention time, and characteristic fragment ions, the identification of prototype components was elucidated (shown in Table 2).In addition, the Clog p values calculated by ChemDraw 14.0 were used to distinguish isomers at different retention times.( 5) Finally, the mass information of potential metabolites, as well as their possible biotransformation pathways and composition change given by Compound Discoverer 3.2, were compared by the data of prototype components and the related literature to verify the metabolites and their metabolic pathways (shown in Table 3).

Identification of Prototype Components
An in-house database has been established for each compound involved in RALP, RGP, and RZ based on our previous experimental data and the related literature for the investigation of their chemical constituents.The database consisted of the compound name, molecular formula, accurate molecular mass, chemical structure, MS 2 mass spectra, and related product ion information.The total ion chromatograms (BPIs) of TSD and the urine and plasma samples after oral administration by UHPLC-Q-Exactive-MS/MS in positive and negative ion modes are presented in Figure 2. It is found that the majority of alkaloids responded well in the positive mode, and the majority of phenolic compounds and saponins responded well in the negative mode.A total of 82 prototype compounds were identified or tentatively characterized, including 41 alkaloids, 35 phenolic compounds, and 6 saponins (shown in Table 2) by comparing the EICs among TSD, drugged, and blank samples and by comparison with reference standards, internal database, and the literature.Figure S3 (Supplementary Materials) displayed MS/MS spectra of major prototype compounds in the urine samples.

Identification of Prototype Components
An in-house database has been established for each compound involved in RALP RGP, and RZ based on our previous experimental data and the related literature for th investigation of their chemical constituents.The database consisted of the compound name, molecular formula, accurate molecular mass, chemical structure, MS 2 mass spectra and related product ion information.The total ion chromatograms (BPIs) of TSD and th urine and plasma samples after oral administration by UHPLC-Q-Exactive-MS/MS in pos itive and negative ion modes are presented in Figure 2. It is found that the majority o alkaloids responded well in the positive mode, and the majority of phenolic compound and saponins responded well in the negative mode.A total of 82 prototype compound were identified or tentatively characterized, including 41 alkaloids, 35 phenolic com pounds, and 6 saponins (shown in Table 2) by comparing the EICs among TSD, drugged and blank samples and by comparison with reference standards, internal database, an the literature.Figure S3 (Supplementary Materials) displayed MS/MS spectra of majo prototype compounds in the urine samples.

Identification of Alkaloid Components
Metabolites for alkaloids obtained in this study could be classified into three subtypes, namely, diester-diterpenoid alkaloids (DDAs), monoester-diterpenoid alkaloids (MDAs), and amine-diterpenoid alkaloids (ADAs) [30].We conducted an in-depth study of the chemical constituents of alkaloids of Aconitum carmichaeli in previous research [24,29], in which we carried out detailed mass fragmentation analysis of DDAs, MDAs, and ADAs, and a total of 42 DDAs and 120 diterpenoid alkaloids were identified, respectively.
In the MS 2 spectra of DDAs, the most abundant ion yielded from the loss of a molecule of AcOH at the C 8 site, which could be a diagnostic neutral loss for the differentiation of DDAs from MDAs and ADAs [29] were tentatively identified as 10-OH-mesaconitine, dehydrohypaconitine, secoyunaconitine, 3-deoxyaconitine, and chasmaconitine by comparing their acquired accurate mass data, characteristic fragment ions with those of compounds in our previous research [29].
In the MS spectra for the urine sample, by extraction of NLF for both 32 Da and 18 Da with limitation of molecular weight ranging from 500 to 620 Da, ten peaks were found.Neutral losses of 32, 18, and 122 Da, corresponding to the elimination of acetic acid, methanol, and benzoic acid, or combinations of these, could be considered diagnostic fragment ions for MDAs [31].However, fragment peaks formed by the loss of the typical substituent group as BzOH (122 Da) were hardly detected for MDAs in this study.Thus, Compounds 22-28 and 30-32 were identified as MDAs accordingly by comparing the accurate mass data and diagnostic fragment ions with those of the compounds in our previous research [24].
A total of 21 prototype compounds were identified as ADAs, most of which possessed molecular weight between 390 and 500 Da and were eluted within the initial 16 min.The substitutions of C 1 and C 3 sites of ADAs were relatively active sites and could be easily cleaved, yielding major peaks [M+H-H 2 O] + or [M+H-CH 3 OH] + in MS 2 spectra as the diagnostic ion accordingly.Fragmentation pathways of differently substituted ADAs included different diagnostic ions.Compounds 1, 4, 6, 7, 8, 13, 14, as ADAs with C 1 -OH substitution, firstly fragmented into [M+H-H 2 O] + as diagnostic fragment ions and followed by losses of typical substituent groups (CH 3 OH and H 2 O) in their MS 2 spectra.By comparing their accurate mass data with our chemical database and the literature [24,32], they were identified as karakolidine, senbusine A, senbusine B, karakoline, isotalatizidine, fuziline, and neoline, respectively.For Compounds 18 (talatizamine), the most prominent fragmentation ions were designated as 390.2696 ([M+H-CH 3 OH] + ), suggesting its C 1 site with -OCH 3 substitutions.It also yielded 372.

Identification of Phenolic Compounds
In addition to alkaloids from RALP, the main prototype compounds identified in vivo included flavonoids, isoflavonoids, coumarins, and saponins from RGP, and volatile oils from RZ, as shown in Table 2.The MS data of these compounds were compared with those of reference standards, internal databases, and the literature, while isomers could be initially identified by comparing their ClogP.
Flavonoids are important active components of RGP, among which four components, namely liquiritigenin, isoliquiritigenin, iquiritin, and isoliquiritin, have the highest content and are regarded as the indicator components of RGP, which were identified by comparing mass data with those of the reference standards.

Identification of Phenolic Compounds
In addition to alkaloids from RALP, the main prototype compounds identified in vivo included flavonoids, isoflavonoids, coumarins, and saponins from RGP, and volatile oils from RZ, as shown in Table 2.The MS data of these compounds were compared with those of reference standards, internal databases, and the literature, while isomers could be initially identified by comparing their ClogP.
Flavonoids are important active components of RGP, among which four components, namely liquiritigenin, isoliquiritigenin, iquiritin, and isoliquiritin, have the highest content and are regarded as the indicator components of RGP, which were identified by comparing mass data with those of the reference standards.Compound 50 formed the [M+H] + molecular ion peak at m/z 431.13280 and further removed one molecule of glucose residues to form the aglycone at m/z 269.08121, which was identified as ononin, the main isoflavone of RGP.Its aglycone formed the same ion at m/z 269.08170 at the retention time of 26.58 min and was fragmented into the fragments of m/z 253.0497, 237.0554, and 213.0911, which is identified as formononetin, and the two prototypes are the most important isoflavonoid components in RGP.
The elemental compositions of other types of licorice flavonoid constituents determined by LC-MS were compared with the data of existing database compounds.Compounds 44, 53, 54, 67, and 68 were preliminarily identified as 5-hydroxyliquiritin, licochalcone B, dihydroxyflavone, licoflavone A, and isolicoflanonol.Similarly, other types of phenolic compounds, such as coumarins, were identified or preliminarily identified, including Compounds 60, 66, 64, 70, 73, and 74, which were identified as glycycoumarim, licocoumarione, licopyranocoumarin, glycyrin isoglycyrol, and glycyrol, correspondingly.A few other phenolic components observed in vivo of TSD were derived from RZ, while compounds 56, 63, and 75 were tentatively identified as 6-gingerol, 6-shogaol, and 10-shogaol, respectively, with fragment ions m/z 177.09 and 137.06 as their characteristic fragment ions in PI mode, which is consistent with the literature [33].

Identification of Saponins
From the LC-MS/MS profiles, six saponin components were found as absorbed prototype components, all of which were derived from RGP.The saponins (Compounds 77, 78, 79, 81, and 82) were within the retention time of 14-21 min and had both mass spectral response in NI and in PI mode.
As a general rule for triterpenoid saponins in MS/MS spectra, the fragmentation reactions undergone by activated saponin ions almost occur within the glycan part of the saponin ions, and the sugar chains can be eliminated successively from end to inner and finally to obtain an aglycone ion [34].Through glycosidic cleavages or cross-ring cleavages, the parent ion obtained a series of ions retaining the charge at the reducing terminus were termed Y and Z (glycosidic cleavages) and X (cross-ring cleavages), whereas those ions retaining the charge at the non-reducing terminus are termed B, C (glycoside cleavages), and A (cross-ring cleavages) [35].
The MS cleavage pathways of saponins from RGP, however, were incompletely abided by this rule.Take glycyrrhizic acid as an example; in MS spectra of PI mode, the ions of [M-H] − were obtained, accompanied by the fragment ions of m/z 647.37744 [M+H-β-Dglucuronopyronosyl (glcA)] + and m/z 453.33554 [aglycottne (agl)+H-H 2 O] + , which were similarly for the other detected saponins and has not been reported up to present.More interestingly, in the MS/MS spectra of the detected saponins, the ions of [agl+H-H 2 O] + rather than [agl+H] + were observed as the base peaks, namely, m/z 453.34  ) were observed, corresponding to the successive loss of two glucuronopyranosyls.Thus, the identification information for aglycone s and sugar chains of licorice saponins can be obtained from PI and NI ion modes, respectively.

Identification of Metabolites
Prototypes and metabolites exist simultaneously in plasma and urine samples.Thirtytwo major prototypes, including 11 alkaloids from RALP, as well as 21 phenolic and saponin compounds from RGP and RZ, were selected as MDF templates for metabolite screening.The 32 compounds contained a wide range of chemical structure types with relatively high content in TDS.A total of 40 alkaloids and 25 phenolic compounds were identified or tentatively characterized by comparing the mass data with those of prototype compounds and metabolic pathways reported by the literature [36][37][38][39][40].
After prototypes are absorbed into the body, some of them are excreted as prototypes, and some of them can be converted into other metabolites.DDAs were ester hydrolyzed to MDAs in rats; for example, MA, HA, and AC could be ester hydrolyzed to 14-Benzoylmesaconine (BM), 14-Benzoylhypaconine (BH), and 14-Benzoylaconitine (BA) during the process of metabolism in rat, while BM, BH, and BA themselves could be metabolized to mesaconine, hypaconine, and aconine [36].Therefore, certain prototypes are themselves metabolites and metabolized from other prototypes in rats.

Identification of Alkaloid Metabolites
For diterpenoid alkaloids, most metabolites from hydroxylation, deoxylation, demethylation, deethylation, dehydrogenation, ester hydrolysis, and demethylation with deoxylation have been found in vivo.Metabolites of alkaloids were identified or tentatively identified based on their metabolic pathways, as reported in the literature [37].
After oral administration of TSD, eight related metabolites of talatizamine (18) were identified in urine samples.Metabolite M18 and M19 showed [M+H] + ion at m/z 408.27386 and 408.27393 (giving formula C 23 H 37 NO 5 ), 14 Da (CH2) less than the parent compound.In the MS 2  M4 was confirmed as hydroxylated talatizamine for the [M+H] + ion at m/z 438.28433 (formula C 24 H 39 NO 6 ), 16 Da (O) more than talatizamine, and the fragment ions at m/z 406.2588, 388.2476, 374.230 and 356.2226 were all 16 Da less than those of talatizamine.Therefore, M4 was deduced as 10-Hydroxy Talatizamine, as for the C 10 site in diterpenoid alkaloids prone to be hydroxylated by the literature [38].
Apart from these three metabolites, other metabolites (M13, M26, M36, M39, and M40) of talatizamine were produced through the reaction of dehydrogenation, demethylation, N-deethylation, and deoxidation.The proposed metabolic pathways of talatizamine are shown in Figure 4.The other metabolites of alkaloids were deduced accordingly by their acquired accurate mass data, retention time, and characteristic fragment ions, as well as the Clog p values, and biotransformation pathways information and composition change calculated by ChemDraw 14.0 and Compound Discoverer 3.2.

Identification of Phenolic Compound Metabolites
Metabolites of phenolic compounds, mainly from hydroxylation, oxylation, methylation, dehydrogenation, hydration, methylation with oxylation, dehydrogenation with oxylation, and sulfation, have been observed in vivo.They were identified or tentatively identified by comparing their accurate mass data with prototypes and their metabolic pathways reported by the literature [39,40].The metabolites for major phenolic compounds found in the urine and plasma samples were exhibited in Table 4.
Metabolites of phenolic compounds observed in vivo were mainly derived from the metabolism of liquiritigenin, isoliquiritigenin, and 6-gingerol, which were the most important aglycones from RGP and RZ in TSD.According to the MS data and the metabolic pathways reported in the literature, eleven related metabolites were identified in urine and plasma samples after the absorption of liquiritigenin and isoliquiritigenin.

Identification of Phenolic Compound Metabolites
Metabolites of phenolic compounds, mainly from hydroxylation, oxylation, methylation, dehydrogenation, hydration, methylation with oxylation, dehydrogenation with oxylation, and sulfation, have been observed in vivo.They were identified or tentatively identified by comparing their accurate mass data with prototypes and their metabolic pathways reported by the literature [39,40].The metabolites for major phenolic compounds found in the urine and plasma samples were exhibited in Table 4.
Metabolites of phenolic compounds observed in vivo were mainly derived from the metabolism of liquiritigenin, isoliquiritigenin, and 6-gingerol, which were the most important aglycones from RGP and RZ in TSD.According to the MS data and the metabolic pathways reported in the literature, eleven related metabolites were identified in urine and plasma samples after the absorption of liquiritigenin and isoliquiritigenin.Note: +, response area below 10 6 ; ++, response area between 10 6 and 10 7 ; +++, response area between 10 7 and 10 8 ; ++++, response area above 10 8 .

Difference between Urine and Plasma Samples
Xenobiotics usually vary at trace levels and are interfered with endogenous components.Comparative analysis of metabolites between plasma and urine samples was carried out by the same LC-MS/MS method.Most prototype components and metabolites possessed suitable signal responses in urine samples, mainly as metabolites from phase I metabolism referring to dehydrogenation, demethylation, hydroxylation, deoxylation, and deethylation.A few phase II metabolites were detected in the urine, including sulfate conjugates of liquiritigenin, isoliquiritigenin, and formononetin.
Metabolites of TSD detected in the plasma samples are fewer than those in the urine samples.As for plasma samples, 10 prototype components (eight phenolic compounds and two alkaloids) were detected and tentatively identified, most of which were flavonoid aglycones.Fifteen metabolites derived from neoline, talatizamine, karakoline songorine, and fuziline, as well as sixteen metabolites derived from liquiritigenin, isoliquiritigenin, formononetin, gancaonin M, and 6-gingerol, respectively, were found in plasma samples, which indicated there were fewer metabolites identified in plasma samples.These results are reasonable due to their relatively lower concentration and higher matrix interference in plasma than in urine samples.
In the present study, ADAs and their metabolites from RALP were mainly detected in rats after oral administration of TSD.DDAs are the most toxic but chemically unstable alkaloids in RALP, and the alkaloidal composition changed during concocting and decocting, with DDAs changing to MDAs, and both transformed further to ADAs while the toxicity gradually diminished.ADAs, such as fuziline and neoline, showed activity against pentobarbital sodium-induced cardiomyocyte damage by obviously recovering beating rhythm and increasing the cell viability [41].Mesaconine and hypaconine showed strong cardiac actions on the isolated perfused bullfrog heart.Moreover, mesaconine has protective effects, including improved inotropic effect and left ventricular diastolic function, on myocardial ischemia-reperfusion injury in rats [42].
Metabolites of licorice flavonoids and 6-gingerol were also mainly detected.Liquiritigenin offers cytoprotective effects against various cardiac injuries, and it could protect against myocardial ischemic injury by antioxidation, antiapoptosis, counteraction mitochondrial dysfunction, and damping intracellular Ca 2+ [43].6-Gingerol was identified as a novel angiotensin II type 1 receptor antagonist for cardiovascular disease by highthroughput screening, which partially clarified the mechanism of ginger regulating blood pressure and strengthening the heart [44].6-gingerol administration protected I/R-induced cardiomyocyte apoptosis via the JNK/NF-κB pathway in the regulation of HMGB2 [45].
The results of the in vivo metabolite study of TSD in this study suggested that in the following pharmacokinetic, pharmacological, and efficacy studies, attention should be paid primarily to the ADAs alkaloids, licorice flavonoids, gingerol-6, and their metabolites

Conclusions
A total of 82 compounds, including 41 alkaloids, 35 phenolic compounds, and 6 saponins, were identified or tentatively characterized in TSD by UHPLC-Q-Exactive-MS/MS.Among them, 32 representative compounds with relatively high mass spectral peak areas and different core structures were selected as parent compound templates for further investigation of their metabolic profiles in rats.In total, 65 metabolites were screened out and tentatively characterized in rats' urine and plasma based on their MS characteristic fragmentation patterns and information.The main metabolic reactions involved hydrogenation, demethylation, hydroxylation, hydration, methylation, deoxylation, and sulfation.This is a systematic study of in vivo metabolism of TSD, and it will be beneficial for further understanding of the pharmacological and pharmacokinetic study of TSD.

Figure 1 .
Figure 1.Workflow of the analytic strategy for the metabolite identification of TSD.

Figure 1 .
Figure 1.Workflow of the analytic strategy for the metabolite identification of TSD.
Compound 47, as reference compound liquiritin, formed the [M-H] − -based peak at m/z 417.11890 (C21H21O9 − ), for which furtherly formed fragmentation ion m/z 255.0662 [M-H-glu] − of the aglycone element in the MS/MS spectrum, accompanied by three characteristic fragments at m/z 135.0074 (C7H3O3 − ), 119.0488 (C8H7O − ), and 91.0173 (C6H3O − ), which can be used for the identification of the same type of licorice flavonoids.
Compound 47, as reference compound liquiritin, formed the [M-H] − -based peak at m/z 417.11890 (C 21 H 21 O 9 − ), for which furtherly formed fragmentation ion m/z 255.0662 [M-H-glu] − of the aglycone element in the MS/MS spectrum, accompanied by three characteristic fragments at m/z 135.0074 (C 7 H 3 O 3 − ), 119.0488 (C 8 H 7 O − ), and 91.0173 (C 6 H 3 O − ), which can be used for the identification of the same type of licorice flavonoids.
spectra, characteristic ions at m/z 376.25 ([M+H-CH 3 OH] + ), 358.24 ([M+H-CH 3 OH-H 2 O] + ), and 326.21 ([M+H-CH 3 OH-H 2 O-CH 3 OH] + ), suggesting its C 1 site with -OCH 3 substitutions.Those characteristic ions were different from the characteristic ions of the prototype component, isotalatizidine (Compound 8), although they shared the same elemental composition (C 23 H 37 NO 5 ).Isotalatizidine, with -OH substitutions at the C 1 site, first yielded 390.2631 ([M+H-H 2 O] + ) by loss of H 2 O at the C 1 site.The fragmentation pathways of demethyl talatizamine and isotalatizidine can be compared in Figure 3.The methyl group of the C 16 site or C 18 site could easily be metabolized instead of that of the C 1 site for M18 and M19.The Clog p values of 18-O-demethyl talatizamine and 16-O-demethyl talatizamine were −0.78 and −0.74, calculated by ChemDraw 14.0.Hence, M18 and M19 were tentatively determined as 18-O-demethyl talatizamine and 16-O-demethyl talatizamine.
).All animal experiments were performed at the SPF animal laboratory [experimental animals license number SYXK (Guangdong, China) 2008-0094].The Institutional Animal Ethics Committee of Guangdong Provincial Hospital of Chinese Medicine approved all experimental protocols (No. 2023131).

Table 1 .
Main prototype components as parent compound templates for MDF data screening.(mass spectral peak areas greater than 10 8 in the decoction).

Table 1 .
Main prototype components as parent compound templates for MDF data screening.(mass spectral peak areas greater than 10 8 in the decoction).

Table 2 .
Prototype compounds identified or tentatively characterized in the urine and plasma samples after oral administration of TSD.
Note: * Compounds identified by comparing with reference standards; glcA: β-D-glucuronopyronosyl; agl: aglycone; +, The produced ions obtained in NI mode were quite different from those in PI mode.The fragment ions of glycosidic cleavages or cross-ring cleavages, as well as the aglycone, were hardly detected in NI mode.The ions of m/z 351.05 (C 12 H 15 O 12

Table 3 .
Metabolites of major alkaloids found in the urine and plasma samples.

Table 4 .
Metabolites of phenolic compounds found in the urine and plasma samples.