Study on the biotransformation and activities of three isoflavone glycosides from Radix Astragali by human feces in vitro

Xin Liu; Jiede Ma; Xianming Lan; Shuying Zhao; Wenzhe Yang; Xue Yang; Yuqing Man; Jiayu Zhang; Shaoping Wang

doi:10.1016/j.arabjc.2023.105005

View/Download PDF

Buy Reprints

PDF

Translate this page into:

Original article

08 2023

:16;

105005

doi:

10.1016/j.arabjc.2023.105005

Study on the biotransformation and activities of three isoflavone glycosides from Radix Astragali by human feces in vitro

Xin Liu^{^a,1}, Jiede Ma^{^b,1}, Xianming Lan^{^a,1}, Shuying Zhao^{^a}, Wenzhe Yang^{^d}, Xue Yang^{^d}, Yuqing Man^{^a,^c,⁎}, Jiayu Zhang^{^a,⁎}, Shaoping Wang^{^a,⁎}

a School of Pharmacy, Binzhou Medical University, Yantai 264003, China

b Department of Pharmacy, Yantai Hospital of Traditional Chinese Medicine, Yantai 264003, China

c Department of Pharmacy, Affiliated Hospital of Binzhou Medical University, Binzhou 256603, China

d Key Laboratory of Marine Drugs of Ministry of Education, School of Medicine and Pharmacy, Ocean University of China, Qingdao 266003, China

⁎Corresponding authors. myq388@126.com (Yuqing Man), zhangjiayu0615@163.com (Jiayu Zhang), wsp.0104@163.com (Shaoping Wang)

Received: 2022-12-20, Accepted: 2023-5-13,

Disclaimer:
This article was originally published by Elsevier and was migrated to Scientific Scholar after the change of Publisher.

1 The three authors are equal to this manuscript.

Abstract

Radix Astragali (RA) is used as a traditional spice and food additive, and it is also rich in isoflavone glycosides, including calycosin-7-O-glucoside, ononin and genistin, which have beneficial effects against inflammation and oxidative stress. However, most isoflavone glycosides are hydrolyzed to other metabolites by intestinal bacteria before absorbed into the body. In this study, the levels of the above three isoflavone glycosides and their aglycones in 12 RAs from different regions were analyzed by high performance liquid chromatograph (HPLC). Then, the metabolites of the three isoflavone glycosides transformed by human feces at different times were quantified by HPLC and identified by ultra-high performance liquid chromatograph coupled with high resolution mass spectrometer (UHPLC-HRMS) and nuclear magnetic resonance hydrogen spectrum (¹H-NMR). Furthermore, the anti-inflammatory and antioxidant activities of the metabolites were evaluated and the possible mechanism was analyzed by network pharmacology. As a result, a total of 18 metabolites were identified and the metabolic profiles were constructed. Among them, 4 metabolites, including 3′,4′,7-trihydroxyisoflavone, pratensein, dihydrogenistein and daidzein, were separated and identified, of which the metabolic process of pratensein was discovered for the first time. The anti-inflammatory and antioxidant activity results and network pharmacology analysis showed that there was a close relationship between the activities and the metabolic profiles, especially the formation of aglycones. This study is of significance to reveal the metabolic process of the isoflavones of RA in the intestine, and also provides guidance for the metabolism of other flavonoid glycosides.

Keywords

Radix Astragali

Isoflavone glycoside

Human feces

Anti-inflammatory

Antioxidant

Network pharmacology

Show Related Articles from PubMed

1 Introduction

RA, as the dried root of Astragalus membranaceus (Fisch.) Bge. var. mongholicus (Bge.) Hsiao and Astragalus membranaceus (Fisch.) Bge, is treated as traditional spice and food additives for thousands of years in East Asia and the Middle East (Bi et al., 2020; Chen et al., 2020). RA possesses a variety of therapeutic effects, including anti-inflammation, antioxidation, anti-cancer, immunomodulation, cardio-protection, treatment of diabetes and hypertension, and so on (Yang et al., 2020; Chen et al., 2020; Su et al., 2021; Yang et al., 2021; Tang and Huang, 2022 ). Among them, inflammation and oxidative stress, as a defensive pathological response of the body tissues to infections, tissue injury and other harmful stimulants, are both predisposing factors for other diseases (Adesso et al., 2018). Oxidation stress is an imbalance state in the body between the production of reactive oxygen species (ROS) and the capacity of antioxidant system. ROS include dioxide (O₂), hydrogen peroxide (H₂O₂), nitric oxide (NO), and hydroxide (OH). Excessive ROS caused by environmental factors, physical or chemical stimulation and infection can induce DNA damage, cell death and tissue injury. Oxidative stress is also a propellant of inflammation by promoting the release of inflammatory cytokines. Inflammation is a defensive pathological response of the body tissues to infections, tissue injury and other harmful stimulants. As inflammation progresses, local vascular permeability increases, white blood cells accumulate to injured tissue and inflammatory cytokines like NO, tumor necrosis factor α (TNF-α), interferon γ (IFN-γ) and interleukins (IL) are released.

The modern chemical studies show that isoflavones are the main bioactive compounds of RA (Xiao et al., 2008). In particular, calycosin-7-O-glucoside is considered to be the signature ingredient of RA in the Chinese Pharmacopoeia. Moreover, ononin and genistin were also taken for the major active components of RA in many studies (Lv et al., 2011; Huang et al., 2018). These foodborne isoflavones are natural anti-inflammatory and antioxidant agents. Isoflavones are a kind of phenolic compounds with 2-phenyl chromophorone basic skeleton mainly found in legumes plants, such as RA. These compounds are good nucleophiles due to the strong electron donor effect of the multiple active hydroxyl groups in the structure. They are also easy to lose electrons and be oxidized, thus directly scavenging ROS and reducing the release of inflammatory cytokines. Previous studies have shown that a variety of isoflavones have anti-inflammatory and antioxidant activities, including daidzein, genistein, formononetin, puerarin, biochanin A and so on.

However, most isoflavones in RA are existed as glycosides, which limits the absorption of them in the intestine. The published researches have shown that isoflavone glycosides can be hydrolyzed by intestinal bacteria, and converted to low molecular metabolites with some reactions such as deglycation, demethylation, dehydroxylation their compositions (Murota et al., 2018). The metabolites of isoflavone glycosides possessed more significant activities than their prototypes. For example, naringenin afforded higher treatment effects for antioxidant, anti-inflammatory, metabolic syndrome and obesity than naringin in Citrus fruit extracts (Alam et al., 2014). Genistin and daidzin were transformed into genistein and daidzein through the associated enzyme reactions in soymilk, and the secondary metabolites showed higher potential for anti-inflammatory and treatment of inflammatory bowel disease (Hiramatsu et al., 2020). Therefore, the biotransformation of the isoflavone glycosides in RA by intestinal bacteria or enzymes is extremely essential to improve their bioavailability.

Some researches of the transformation processes of isoflavones in intestinal tract have been reported, which were studied by incubating compounds in vitro with human or animal feces in anaerobic condition, simulating the intestinal microbiota environment in vivo. However, a considerable number studies focus solely on the types of metabolites and their reactions, while the quantitative changes of the metabolites in the transformation were seriously neglected. Therefore, in this study, three isoflavone glycosides from RA, including calycosin-7-O-glucoside, ononin and genistin, were transformed by mixed human feces under anoxic conditions. In order to clarify the metabolism of calycosin-7-O-glucoside, ononin and genistin, the metabolites were qualitatively analyzed by UHPLC-HRMS and ¹H-NMR, and quantitatively analyzed by HPLC at different times. The antioxidative and anti-inflammatory activities of the metabolites were also tested in vitro. Finally, the network pharmacology was used to explain the relationships between anti-inflammatory and antioxidant activities and the metabolites.

2 Materials and methods

2.1

2.1 Chemicals and reagents

Methanol, acetonitrile and formic acid of HPLC grade were purchased from Kemiou (Tianjin, China). General anaerobic medium (GAM) broth, vitamin K1 and hematin chloride were purchased from Hopebio (Qingdao, China). The pure water for the experiment was gained from Wahaha (Hangzhou, China). Reference standards of calycosin-7-O-glucoside, ononin, genistin, calycosin, formononetin, genistein and daidzein were purchased from Must Bio-tech Co., Ltd (Chengdu, China). 3′,4′,7-Trihydroxyisoflavone was from Shenzhen Bamboo Bio-tech Co., Ltd (Shenzhen, China). Pratensein and dihydrodaidzein were from Taopu Biological Co., Ltd (Shanghai, China). Dihydrogenistein and dihydroformononetin were from TRC (Toronto, Canada). All reference standards were detected by HPLC, and their purities were more than 98%.

2.2

2.2 Quantitative analysis of 6 isoflavones in multi-batch RA by HPLC

According to the literature, three isoflavone glycosides (calycosin-7-O-glucoside, ononin and genistin) and their aglycones (calycosin, formononetin and genistein) could be regarded as the representative components of RA. 12 RAs from different regions were collected to evaluate the quantitative differences of 6 isoflavones, and the information of 12 RAs are shown in Table S1 and Fig.S1. 12 RAs were powdered and sieved, and 1 g powder was added into 50 mL methanol for ultrasonic extraction for 1.0 h. 25.0 mL extract was filtered and condensed. The residue was transferred to a 5.0 mL volumetric flask, and diluted with methanol. All solutions were stored at 4℃ until use.

All samples were analyzed using an HPLC, which consisted of two LC-20AT solvent delivery systems, a DGU-20A3R degasser, a CTO-20A column oven, an SIL-20AXR autosampler, an SPD-20A UV detection and a Labsolution workstation (Shimadzu, Kyoto, Japan). Based on the method in the Chinese Pharmacopoeia (2020), a Kromasil 100–5-C18 column (250 mm × 4.6 mm, 5 μm, AKZO NOBEL, Bohus, Sweden) was set as a separating medium. The mobile phase consisted of 0.2% formic acid aqueous solution (A) and acetonitrile (B). The gradient elution program was set as follows: 0 ∼ 20 min, 20%-40% (B); 20 ∼ 30 min, 40% (B); 30 ∼ 35 min, 40%-20% (B). The UV absorption spectra were measured using a flow rate of 1.0 mL/min, a column temperature of 35℃, and the wavelength of 260 nm.

2.3

2.3 Biotransformation in vitro of calycosin-7-O-glucoside, ononin and genistin by human feces

Fresh fecal samples were collected on the day of the experiment from 10 healthy volunteers (6 males and 4 females), who did not have previous intestinal disease and were not treated with antibiotics during 3 months (the ethical committee of Binzhou Medical University: 2022–351). All feces were immediately mixed and stored in the anaerobic bag (Anaero Pack, Mitsubishi Gas Chemical Company, MGC, Tokyo, Japan), then diluted and homogenized with stroke-physiological saline solution (SPSS), and homogenized to obtain a 10% (w/v) slurry, which was used as the biotransformation incubation starter.

49.0 g of GAM broth was dissolved in 1000 mL water, and sterilized under high pressure (0.15 MPa) and temperature (121 °C) for 15 min. 1.0 mg vitamin K1 and 5.0 mg hematin chloride were dissolved in the above solution after being cooled under aseptic conditions (Liu et al., 2017).

The final transformation solution was composed as followed: 50% fecal slurry, 49% GAM broth, and 1% standard compound solution (calycosin-7-O-glucoside, ononin or genistin dissolved in SPSS with 1.0 mg/mL final concentration). The mixture was incubated in anaerobic bags with oxygen indicator (MGC, Tokyo, Japan) at 37 °C, and the samples were collected at 0 h, 2 h, 6 h, 12 h, 18 h and 24 h, respectively. All studies were repeated six times to acquire the accurate results, and the transformation solution without standard (calycosin-7-O-glucoside, ononin or genistin) was included as the negative control. All samples were collected and added cold methanol to precipitate the impurities at 4 °C for 12 h. The supernatant was obtained after centrifugation at 12 000 rpm/min for 5 min, and then determined to 5 mL. Hereafter, these samples were analyzed by HPLC with the method under “2.2”. Certainly, we also classified different metabolites as follows: the metabolites of calycosin-7-O-glucoside were named as series A, these of ononin were as series B, and genistin as series C.

2.4

2.4 Separation and identification analysis of the unknown metabolites

In order to characterize the structure of the unknown metabolites, the metabolites were separated by semi-preparative LC using the previous method in “2.2” with minor modifications. A Shim-pack GIS C18 preparative column (20 mm × 250 mm, 10 μm, Shimadzu, Kyoto, Japan) was used as the separate medium, and the flow rate was increased to 10 mL/min. The unknown peaks with the same retention time were collected after 100 repetitions. They were dried using a lyophilizer, and purities were detected by HPLC.

The unsearchable metabolites were determined by UHPLC-HRMS. Firstly, the LC analysis was performed on a DIONEX Ultimate 300 UHPLC system (Thermo Fisher Scientific, MA, USA) with a binary pump, an autosampler and a column oven. Chromatographic separation was performed on an ACQUITY UPLC BEH C18 column (2.1 mm × 100 mm, 17 μm, Waters, Milford, MA, USA). The column temperature was maintained at 35 °C, and 3 μL sample was injected at a flow rate of 0.3 mL/min. The mobile phase consisted of 0.1% formic acid aqueous solution (A) and acetonitrile (B). The elution gradient was set as follow: 0 ∼ 5 min, 5%-30% (B); 5%∼10 min, 30%-50% (B); 10 ∼ 27 min, 50%-90% (B); 27 ∼ 27.1 min, 90%-5% (B); 27.1 ∼ 30 min, 5% (B). Subsequently, HRMS analysis was performed on a Q-Exactive Orbitrap mass spectrometer (Thermo Fisher Scientific, MA, USA). Data acquisition parameters in the positive and negative ion mode were set as follows: spray voltage, 3 000 V (positive) /3 500 V (negative); capillary voltage, −35 V; sheath gas flow rate, 30 arb; auxiliary gas flow rate, 10 arb; capillary temperature, 325℃ (positive) /350℃ (negative); tube lens, + 110 V (positive) /- 110 V (negative). Metabolites were detected using full-scan MS analysis (m/z 70–1 050) at a resolving power of 70 000 FWHM. The resolution of dd-MS² was set as 17 500 FWHM. The collision energy of collision-induced dissociation (CID) was 30 eV. The data were acquired with Thermo Xcalibur software (Version 2.2).

¹H-NMR analysis was performed on an AMX spectrometer equipped with a PATXT Bruker probe (5-mm) (Bruker Corporation, Karlsruhe, Germany). The spectrometer was operated at 400.13 MHz and the temperature was set at 298 K. The unsearchable metabolites were dissolved in CD₃OD containing chloroform in NMR tubes (5-mm). The residual signal of CD₃OD (δ¹H = 3.31 ppm) was used for chemical shift calibration. The frequency sweep was set at 12.33 kHz. 128 scans of 32 K data points each were acquired with a relaxation delay of 1.5 s, acquisition time of 4.00 s, and mixing time of 400 ms. The spectra were Fourier transformed with a line-broadening factor of 0.3 Hz. The NMR spectra were processed by MestReNova software (Version 6.1.0). Baseline correction and phase correction were performed manually.

2.5

2.5 Quantitative analysis of the identified metabolites by HPLC

After the structure identification of the metabolites, their contents would be detected by HPLC at different phases of biotransformation. The HPLC method was conformed with the previous method in “2.2.”.

2.6

2.6 Identification analysis of the other metabolites by LC-MS/MS

Moreover, in order to identify more metabolites and construct a metabolic pathway network, the metabolites of the three transformations were used for UHPLC-HRMS/MS analysis after mixed respectively. The LC-MS/MS condition was as “2.4.”.

2.7

2.7 Antioxidant activity of the metabolites by DPPH radical scavenging assay in vitro

The capacity in scavenging 2,2-diphenyl-1-picrylhydrazyl (DPPH) free radical can be used to determine the antioxidant activity of metabolites. The method for determining DPPH scavenging activity with minor modifications was as follows: 100 μL biotransformation solution at different time were added to the 96-well plate mixed with the same amount of DPPH solution (0.2 mmol/L, dissolved in 75% methanol). Then the mixture in plate was react for 30 min in the dark at 37℃ before measuring absorbance at 517 nm. All measures were repeated three times to acquire the accurate results. The antioxidant activity of the metabolites was expressed by DPPH scavenging rate (%): $DPPH s c a v e n g i n g r a t e (%) = [1 - (A_{i} - A_{j}) / A_{c}] \times 100 %$ where “A_i”, “A_j” and “A_c” were absorbance of the sample with DPPH, sample with 75% methanol, and 75% methanol with DPPH, respectively.

2.8

2.8 Anti-inflammatory and antioxidative activities of the metabolites based on LPS-stimulated RAW 264.7 cell

The mice macrophage cell line RAW264.7 was cultured in Dulbecco’s Modified Eagle Medium (DMEM, Biological Industries, BI, Bet-Haemek, Israel) containing 10% fetal bovine serum (FBS, BI, Bet-Haemek, Israel), 100 U/mL penicillin and 100 μg/mL streptomycin (Gibco-BRL, Grand Island, NY, USA) at 37 °C under a humidified 5% CO₂ atmosphere.

Lipopolysaccharides (LPS) is a unique component in the cell wall of Gram-negative bacteria. LPS-stimulated RAW264.7 is often used as a model for anti-inflammation and antioxidation, whose production of NO increased and reduction capacity of ferric ion decreased. 1.25 × 10⁵ RAW264.7 cells were plated in 24-well plates and divided into the control group, the LPS model group and the treatment groups (treated by the metabolite solutions of calycosin-7-O-glucoside, ononin and genistin of 0 h, 2 h, 6 h, 12 h, 16 h and 24 h, respectively). All cells were incubated for 24 h, and then the treatment groups were pre-treated with the metabolites (1% diluted in complete medium) for another 1 h. After that the LPS model group and treatment groups were challenged with LPS (1 µg/mL, Yuanye, Shanghai, China) for an additional 24 h.

Anti-inflammatory and antioxidative activities were measured using the NO assay Kit (Beyotime, Shanghai, China) and the Total antioxidant capacity (T-AOC) assay Kit (FRAP method) (Jiancheng, Nanjing, China), respectively. In brief, referring to the process in the manual, the culture medium was added to the 96-well plates and mixed with the above Kits. The absorbances at 540 nm and 593 nm were determined with a microplate reader (BioTek, Winooski, VT, USA), respectively.

2.9

2.9 Network pharmacology analysis of the metabolites and the targets of inflammation and oxidation

Network pharmacology was applied to reveal the anti-inflammatory and antioxidant mechanisms of all metabolites. Firstly, the targets of the metabolites were obtained from the SwissTargetPrediction database (http://www.swisstargetprediction.ch/) when the species were selected as homo sapiens. Then, the GeneCards database (http://www.genecards.org/) was consulted to search for targets associated with “inflammation” and “oxidation”. The targets of metabolites and inflammation / oxidation were imported into Cytoscape (Version 3.7.2) for visualization to construct the network. Next, the DAVID database (https://david.ncifcrf.gov/summary.jsp) was used as the functional annotation of Gene ontologies (GO) and KyotoEncyclopedia of Genes and Genomes (KEGG) pathways. The bubble diagrams of GO and KEGG enrichment were plotted by an online platform for data analysis and visualization (https://www.bioinformatics.com.cn). In order to better evaluate the relationship between the metabolites in every transformation series and the targets, the targets of the metabolites in series A, B and C were respectively imported into the String database (https://string-db.org/) to obtain the protein–protein interaction (PPI) network with the species of homo sapiens and confidence data ≥ 0.9. Finally, the pivotal targets were screened out based on Degree centrality (DC), Eigenvector centrality (EC), Betweenness centrality (BC) and Closeness centrality (CC) to construct the interaction networks with the biotransformation metabolites in every series.

3 Results

3.1

3.1 Quantitative analysis of 6 isoflavones in multi-batch RA by HPLC

The HPLC method of the 6 isoflavones was used based on the Chinese Pharmacopoeia (2020), and the results of the methodological investigation are shown in Table S2. As shown in Table 1 and Fig. 1, six isoflavones all could be detected in 12 RAs from A. membranaceus (Fisch.) Bge. var. mongholicus (Bge.) Hsiao (AMMH) and A. membranaceus (Fisch.) Bge. (AM). Calycosin-7-O-glucoside, ononin and genistin were eluted at 3.9 min, 5.8 min and 13.8 min, respectively, calycosin, formononetin and genistein were eluted at 17.5 min, 21.8 min and 27.7 min. Unsurprisingly, the levels of calycosin-7-O-glucoside of 12 RAs all conformed to the minimum limit of the Chinese Pharmacopoeia (2020) (0.02%), while its content in AMMH was 1.90 times higher than that in AM. At the same time, the contents of calycosin, ononin and formononetin in AMMH were 1.88 times, 2.81 times and 2.11 times higher than those in AM, respectively. On the contrary, genistin and genistein in AM were 1.42 times and 1.41 times higher than those in AMMH, respectively. For all RA, the fourth batch (from Hunyuan, Shanxi Province, 20211107) showed better quality than other RAs with higher levels of calycosin-7-O-glucoside, calycosin, ononin and formononetin.

Table 1 The contents of 6 isoflavones in 12 RAs (mg/g).

	AMMH*						Mean	SD	AM*						Mean	SD
	1	2	3	4	5	6	Mean	SD	7	8	9	10	11	12	Mean	SD
Calycosin-7-O-glucoside	0.627387	0.662410	0.660534	0.876403	0.832411	0.739145	0.733048	0.1019	0.375758	0.384194	0.393817	0.370189	0.397613	0.394772	0.386057	0.0112
Genistin	0.004973	0.004764	0.004898	0.004901	0.004829	0.005031	0.004899	0.0001	0.003755	0.008401	0.010354	0.007862	0.003770	0.007667	0.006968	0.0027
Ononin	0.263020	0.270929	0.267357	0.283622	0.267100	0.223401	0.262572	0.0204	0.090639	0.092237	0.096607	0.090301	0.094365	0.095785	0.093322	0.0027
Calycosin	0.252286	0.248427	0.242568	0.252587	0.24044	0.146292	0.230433	0.0415	0.120003	0.119373	0.129414	0.119073	0.120611	0.127037	0.122585	0.0045
Genistein	0.003936	0.003652	0.003555	0.003721	0.003137	0.002916	0.003486	0.0004	0.005136	0.005582	0.005068	0.004403	0.004540	0.004850	0.004930	0.0004
Formononetin	0.136863	0.130156	0.130316	0.137912	0.129385	0.120893	0.130921	0.0061	0.060152	0.060457	0.066875	0.062170	0.056991	0.066155	0.062133	0.0038

Note: *AMMH: A. membranaceus (Fisch.) Bge. var. mongholicus (Bge.) Hsiao; AM: A.membranaceus (Fisch.) Bge.

A: the structure of calycosin-7-O-glucoside, calycosin, ononin, formononetin, genistin and genistein; B: contents of the six isoflavones in two different varieties of RA (n = 6, data are presented as mean ± SD); C: HPLC chromatograms of the six isoflavones in two different varieties of RA.

3.2

3.2 Biotransformation results of calycosin-7-O-glucoside, ononin and genistin by human feces

3.2.1

3.2.1 HPLC analysis results of the metabolites of calycosin-7-O-glucoside, ononin and genistin

The metabolites in vitro of calycosin-7-O-glucoside, ononin and genistin were analyzed by HPLC, and the chromatograms are shown in Fig. 2. Meanwhile, their metabolites were named as series A, B and C, respectively. For series A (Fig. 2A), the content of calycosin-7-O-glucoside was 0.2533 ± 0.0077 mg/mL at 0 h, and decreased to 0.0006 ± 4.47 × 10^-6 mg/mL at 6 h, and then could not be detected after 12 h. Surprisingly, calycosin, the aglycone of calycosin-7-O-glucoside, appeared at the beginning of transformation with the content of 0.0101 ± 0.0004 mg/mL, and the maximum level of calycosin occurred at 2 h with 0.0422 ± 0.0028 mg/mL, suggesting efficient deglycosylation of calycosin-7-O-glucoside. However, its content was not constant. As the transformation went on, the level of calycosin decreased within 2 to 24 h, persistently. This might be attributed to the transformation of calycosin to other metabolites. Coincidentally, 2 unknown compounds were obviously eluted at 10.8 min and 26.5 min, which were named as A1 and A2. Similarly, for series B (Fig. 2B), the content of ononin decreased continuously from 0.2344 ± 0.0183 mg/mL at 0 h to 0.0122 ± 0.0007 mg/mL at 2 h, and then could no longer be detected after 6 h. As for the aglycon of ononin, formononetin increased from 0.0063 ± 0.0006 mg/mL at 0 h to 0.0517 ± 0.0035 mg/mL at 2 h, and then decreased continuously. At the same time, the unknown metabolites named as B1 and B2 were found at 20.6 min and 26.5 min, respectively. As shown in Fig. 2C for series C, genistin was reduced from 0.1710 ± 0.0074 mg/mL at 0 h to 0.0005 ± 1.46 × 10^-5 mg/mL at 24 h, while genistein was increased from 0.0012 ± 4.20 × 10^-5 mg/mL at 0 h to 0.0412 ± 0.0019 mg/mL at 12 h, and then continuously decreased in the remaining time. The unknown metabolites with the retention times of 15.3 min, 20.6 min and 26.5 min were named as C1, C2 and C3, respectively. Coincidentally, compared with their retention times, A2, B2 and C3 should be the same metabolites, and B1 should be the same as C2.

HPLC analysis of the biotransformation products of A: calycosin-7-O-glucoside, B: ononin and C: genistin (n = 6, data are presented as mean ± SD).

3.2.2

3.2.2 Separation and identification of the unknown metabolites

Firstly, calycosin-7-O-glucoside, ononin and genistin were named as A00, B00 and C00, respectively, and their aglycons, calycosin, formononetin and genistein, were named as A0, B0 and C0. As the prototypes of the above unknown metabolites, they were first analyzed by UHPLC-HRMS, and the results are shown in Fig. S2. Subsequently, the other unknown metabolites were separated by semi-preparative LC and their purities were more than 95% according to HPLC analysis. Additionally, they were identified by UHPLC- HRMS and ¹H NMR, and the results are shown in Fig. 3 and Fig. S3.

Mass spectra in the negative ion mode and chemical shift in 1H NMR of the metabolites A: 3′,4′,7-trihydroxyisoflavone, B: pratensein, C: dihydrogenistein and D: daidzein; E: HPLC analysis of the metabolites of i: calycosin-7-O-glucoside, ii: ononin and iii: genistin (n = 6, data are presented as mean ± SD).

A1 was eluted at 6.19 min with [M−H]^- ion at m/z 269.04554, whose formula was calculated as C₁₅H₉O₅. It was 14 Da less than that of calycosin, suggesting that A1 might be a demethylation metabolite of calycosin. In its MS/MS spectrum, the fragment ions at m/z 241 ([M−H−CO]^-) and m/z 251 ([M−H−H₂O]^-) were generated due to the neutral loss of CO (28 Da) and H₂O (18 Da) respectively, suggesting the presence of hydroxyl and carbonyl group. The ¹H NMR data was as follows: δ 8.04 (d, J = 8.8 Hz, H-5), 7.92 (s, H-2), 7.42 (s, H-6′), 6.99 (s, H-2′), 6.89 (dd, J = 8.8, 2.3 Hz, H-6), 6.83–6.82 (m, H-5′), 6.80 (d, J = 2.3 Hz, H-8). Consequently, according the literature (Zhang et al., 2014), A1 should be identified as 3′,4′,7-trihydroxyisoflavone based on the comparison of the retention time, ESI-MS/MS spectra and ¹H NMR data of its standard (Fig. 3A).

A2, B2 and C3, which generated [M−H]^- ions at m/z 299.05611, were eluted at 8.93 min with formula of C₁₆H₁₁O₆. They were 16 Da more than that of calycosin (m/z 283), indicating that a hydroxylation reaction has been occurred. Additionally, they were obtained not only from the metabolites of calycosin, but also from that of genistein, which has a hydroxyl group at C-5. Thus, the hydroxyl group should be added at C-5 position. In their MS/MS spectrums, the fragment ion at m/z 284 [M−H−CH₃]^- was generated, suggesting the presence of methyl group. The fragment ions at m/z 255 [M−H−C₂H₃−OH]^- and m/z 227 [M−H−C₂(OH)OCH₃]^- were generated due to B-ring cleavage. These results proved that the hydroxyl group and methoxy group should both stabilized on the B ring. Above all, their ¹H NMR data was as follows: δ 7.97 (s, H-2), 7.73 (s, H-2′), 7.02 (d, J = 1.7 Hz, H-6′), 6.98–6.92 (m, H-3′), 6.33 (d, J = 2.2 Hz, H-8), 6.23 (d, J = 2.2 Hz, H-6), 4.77 (s), 3.88 (s, CH₃). Consequently, compared with the standard, A2, B2 and C3 were unambiguously identified as pratensein, which was a hydroxylation metabolite of calycosin and formononetin, and a methylation and hydroxylation metabolite of genistein (Fig. 3B).

B1 and C2 generated the deprotonated molecular [M−H]- ions at m/z 271.06119 (C₁₈H₁₅O₁₁) with the retention time at 7.96 min. They were 2 Da more than that of genistein (m/z 269), suggesting that they might be the hydrogenation metabolite of genistein. The fragment ion at m/z 242 ([M−H−COH]^-) was detected based on the C-ring cleavage. This result illustrated that there was no double bond between C-2 and C-3. Their ¹H NMR were as follows: δ 7.34–7.29 (m, H-2′/H-6′), 6.76 (dd, J = 6.5, 2.1 Hz, H-3′/H-5′), 5.92 (d, J = 0.5 Hz, H-6), 5.89 (d, J = 1.0 Hz, H-8), 4.52–4.40 (m, H-2), 3.80 (dd, J = 8.1, 5.2 Hz, H-3). Therefore, B1 and C2 were speculated to be dihydrogenistein by comparing with the standard (Fig. 3C) (Peiroten et al., 2020).

C1 was eluted at 7.02 min, and afforded [M−H]^- ion at m/z 253.05063 with the formula of C₁₅H₉O₄. It was 16 Da less than that of genistein (m/z 269), suggesting that it might be a dehydroxylation metabolite of genistein, who has three hydroxyl groups on its structure (one on the B ring and two on the A ring). The fragment ions at m/z 225 [M−H−C₂H₄]^- and m/z 223 [M−H−CHOH]^- were generated based on the B-ring cleavage, prompting that there was one hydroxyl group on the B ring, while another hydroxyl group was on the A ring alone. The ¹H NMR was as follows: δ 8.05 (d, J = 8.8 Hz, H-5), 7.93 (s, H-2), 7.35–7.31 (m, H-2′/H-6′), 6.90 (dd, J = 8.8, 2.3 Hz, H-6), 6.87–6.83 (m, H-3′/H-5′), 6.81 (d, J = 2.3 Hz, H-8). And thus, C1 could be attributed to be daidzein by comparing the standard (Fig. 3D).

3.2.3

3.2.3 The quantitative analysis of the identified metabolites by HPLC

The unknown metabolites of calycosin-7-O-glucoside, ononin and genistin were quantified using HPLC (Fig. 3E), and the linear functions based on their standards are shown in Table S3. Meanwhile, the metabolites of calycosin-7-O-glucoside, ononin and genistin were named as series A, B and C, respectively. For series A (Fig. 3Ei), 3′,4′,7-trihydroxyisoflavone (A1) was appeared at 12 h with the content of 0.0062 ± 0.0004 mg/mL, and then increased continuously to 0.0210 ± 0.0005 mg/mL at 24 h. Meanwhile, pratensein (A2) was appeared at 0 h with 0.0004 ± 8.31 × 10^-5 mg/mL. Then its level was dramatically increased from 0.0019 ± 0.0001 mg/mL at 12 h to 0.0188 ± 0.0015 mg/mL at 18 h. Afterwards, this growth was slowed down and the content at 24 h was increased to 0.0218 ± 0.0012 mg/mL. Coincidentally, the level trend of calycosin was contrary to that of the above metabolites, suggesting that A1 and A2 seemed more likely to be transformed from calycosin. For series B (Fig. 3Eii), dihydrogenistein (B1) was appeared at 12 h with 0.0118 ± 0.0014 mg/mL, and increased continuously to 0.0228 ± 0.0015 mg/mL at 24 h. Moreover, pratensein (B2) was appeared at 0 h with the level of 0.0010 ± 0.0003 mg/mL, and then its content was significantly increased from 0.0027 ± 0.0001 mg/mL at 6 h to 0.0817 ± 0.0046 mg/mL at 24 h. In contrast, for series C (Fig. 3Eiii), the level of daidzein (C1) was appeared at 2 h with 0.0003 ± 9.31 × 10^-6 mg/mL. Next it was increased to the maximum level of 0.0012 ± 3.87 × 10^-5 mg/mL at 12 h. After that, this level of daidzein was decreased until could not be detected by HPLC after 18 h. This situation might be mainly related to the further degradation of daidzein. Subsequently, dihydrogenistein (C2) was appeared at 6 h with the level of 0.0003 ± 2.89 × 10^-5 mg/mL, and then its content was increased to 0.0028 ± 8.09 × 10^-5 mg/mL at 24 h. Similar to that in series A, pratensein (C3) was appeared at 0 h with 0.0002 ± 3.87 × 10^-5 mg/mL. Next, it was significantly elevated to from 0.0005 ± 0.0001 mg/mL at 6 h to 0.0117 ± 0.0002 mg/mL at 18 h. However, its content seemed to be very stable in the next 6 h, that the content at 24 h was 0.0118 ± 0.0003 mg/mL.

In addition, the level trends of dihydrogenistein in series B and series C were consistent, and the levels of pratensein showed the same trend in three transformation series. Meanwhile, the levels of these metabolites were negative correlated with that of calycosin, formononetin and genistein, illustrating that these metabolites should be transformed from calycosin, formononetin and genisteins. In particular, the level of daidzein began to decrease at 12 h with further transformation, it also showed that daidzein should be further transformed into other metabolites.

3.2.4

3.2.4 Identification analysis of the other metabolites by LC-MS/MS

Except for the above metabolites, other metabolites were further discovered, which were not detected in HPLC, while in LC-MS/MS. As shown in Table 2, 10 metabolites of calycosin-7-O-glucoside, 10 of ononin, and 11 of genistin were identified within error 10 ppm. The metabolic network was constructed from a total of 18 metabolites of them by human feces (Fig. 4). The network indicated that deglycosylation, hydrogenation, B-ring removal, C-ring cleavage, hydroxylation or dihydroxylation, and O-methylation or O-demethylation reactions were the major metabolism.

Table 2 LC-MS/MS data of the identified metabolites.

A. the metabolites of Calycosin-7-O-glucoside
No.	identification	RT (min)	formula	ion mode	m/zcalculated	m/z measured	error (ppm)	MS/MS fragment ions
A00	Calycosin-7-O-glucoside	5.33	C₂₂H₂₂O₁₀	[M + H]⁺	447.12857	447.12735	−2.736	MS² [4 4 7]: 285(1 0 0), 270(54.69), 253(20.97), 225(17.69), 137(13.08)
A0	Calycosin	7.49	C₁₆H₁₂O₅	[M + H]⁺	285.07575	285.07510	−2.280	MS² [2 8 5]: 270(1 0 0), 225(47.10), 285(45.73), 137(41.87), 253(37.95)
				[M−H]^-	283.06119	283.06110	3.533	MS² [2 8 3]: 268(1 0 0), 211(24.91), 239(16.24), 240(15.66), 184(10.56)
A1	3′,4′,7-trihydroxyisoflavone	6.19	C₁₅H₁₀O₅	[M + H]⁺	271.06009	271.05945	−2.398	MS² [2 7 1]: 271(1 0 0), 225(56.44), 215(50.41), 253(29.45), 243(5.02)
				[M−H]^-	269.04554	269.04562	4.349	MS² [2 6 9]: 269(1 0 0), 213(87.58), 241(70.07), 211(13.78), 225(7.81)
A2	Pratensein	8.92	C₁₆H₁₂O₆	[M + H]⁺	301.07066	301.07056	−0.347	MS² [3 0 1]: 301(1 0 0), 286(36.72), 269(21.23), 153(16.71), 229(13.45)
				[M−H]^-	299.05611	299.05615	3.797	MS² [2 9 9]: 284(1 0 0), 299(31.52), 227(5.86), 255(4.69), 200(1.34)
A3	Benzopyran-4,7-diol	1.17	C₉H₈O₃	[M + H]⁺	165.05462	165.05457	−0.307	MS² [1 6 5]: 123(1 0 0), 95(88.73), 119(46.07), 147(7.00), 165(3.16)
A4	2-(3,4-dihydroxyphenyl)-propionic acid	3.14	C₉H₁₀O₄	[M−H]^-	181.05063	181.04997	2.401	MS² [1 8 1]: 91(1 0 0), 153(8.49), 137(2.89), 123(1.13)
A5	2-(4-hydroxyphenyl)-propanoic acid	5.37	C₉H₁₀O₃	[M−H]^-	165.05571	165.05467	0.299	MS² [1 6 5]: 147(1 0 0), 119(77.33), 165(10.21), 103(5.64), 121(4.77)
A6	Daidzein	7.01	C₁₅H₁₀O₄	[M−H]^-	253.05063	253.05049	3.773	MS² [2 5 3]: 253(1 0 0), 225(4.89), 135(3.71), 133(3.41), 223(1.79)
A7	Genistein	8.22	C₁₅H₁₀O₅	[M−H]^-	269.04554	269.04562	4.349	MS² [2 6 9]: 269(1 0 0), 133(19.43), 181(8.12), 224(7.17), 107(6.79)
A8	Formononetin	9.54	C₁₆H₁₂O₄	[M−H]^-	267.06628	267.06631	4.211	MS² [2 6 7]: 252(1 0 0), 251(15.41), 223(14.52), 132(8.32), 267(5.35)
B. the metabolites of Ononin
No.	identification	RT (min)	formula	ion mode	m/z calculated	m/z measured	error (ppm)	MS/MS fragment ions
B00	Ononin	6.94	C₂₂H₂₂O₉	[M + H]⁺	431.13366	431.13272	−2.177	MS² [4 3 1]: 269(1 0 0), 254(12.34), 213(9.59), 237(6.00), 253(3.73)
B0	Formononetin	9.57	C₁₆H₁₂O₄	[M + H]⁺	269.08083	269.08029	−2.027	MS² [2 6 9]: 269(1 0 0), 254(34.02), 237(28.65), 226(19.75), 107(10.26)
				[M−H]^-	267.06628	267.06622	3.874	MS² [2 6 7]: 252(1 0 0), 223(16.64), 251(12.90), 132(8.29), 224(6.92)
B1	Dihydrogenistein	7.96	C₁₅H₁₂O₅	[M−H]^-	271.06119	271.06116	3.911	MS² [2 7 1]: 135(1 0 0), 199(21.81), 227(15.51), 271(7.94), 243(4.08)
B2	Pratensein	8.91	C₁₆H₁₂O₆	[M + H]⁺	301.07066	301.07062	−0.148	MS² [3 0 1]: 301(1 0 0), 286(39.24), 153(33.07), 241(31.10), 269(26.45)
				[M−H]^-	299.05611	299.05627	4.198	MS² [2 9 9]: 284(1 0 0), 299(20.78), 227(7.04), 255(5.50), 200(1.40)
B3	Benzopyran-4,7-diol	1.17	C₉H₈O₃	[M + H]⁺	165.05462	165.05453	−0.549	MS² [1 6 5]: 123(1 0 0), 95(85.06), 119(48.61), 147(8.27), 165(7.20)
B4	2-(4-hydroxyphenyl)-propanoic acid	5.36	C₉H₁₀O₃	[M−H]^-	165.05571	165.05478	0.965	MS² [1 6 5]: 147(1 0 0), 119(68.88), 165(10.19), 103(3.64), 121(1.49)
B5	5-hydroxyequol	7.24	C₁₅H₁₄O₄	[M−H]^-	257.08193	257.08182	3.830	MS² [2 5 7]: 221(1 0 0), 219(58.50), 135(12.06), 121(3.54), 137(3.00)
B6	Calycosin	7.49	C₁₆H₁₂O₅	[M + H]⁺	285.07575	285.0752	−1.929	MS² [2 8 5]: 270(1 0 0), 285(44.82), 137(42.84), 225(40.86), 253(39.81)
B7	Genistein	8.22	C₁₅H₁₀O₅	[M−H]^-	269.04554	269.04553	4.015	MS² [2 6 9]: 269(1 0 0), 133(19.15), 107(8.98), 181(8.46), 224(3.84)
B8	Dihydroformononetin	9.67	C₁₆H₁₄O₄	[M−H]^-	269.08193	269.08185	3.771	MS² [2 6 9]: 135(1 0 0), 269(76.49), 254(18.45), 239(1.48), 121(1.16)
C. the metabolites of Genistin
No.	identification	RT (min)	formula	ion mode	m/z calculated	m/z measured	error (ppm)	MS/MS fragment ions
C00	Genistin	5.57	C₂₁H₂₀O₁₀	[M + H]⁺	433.11292	433.11200	−2.132	MS² [4 3 3]: 271(1 0 0), 153(9.19), 215(8.32), 243(7.11), 145(2.33)
				[M−H]^-	431.09837	431.09793	1.524	MS² [4 3 1]: 268(1 0 0), 269(53.56), 239(14.89), 224(5.99), 133(1.91)
C0	Genistein	8.24	C₁₅H₁₀O₅	[M + H]⁺	271.06010	271.05960	−1.844	MS² [2 7 1]: 271(1 0 0), 153(37.65), 215(31.74), 243(21.89), 145(9.60)
				[M−H]^-	269.04554	269.04562	4.349	MS² [2 6 9]: 269(1 0 0), 133(18.95), 181(9.79), 224(6.65), 107(6.65)
C1	Daidzein	7.00	C₁₅H₁₀O₄	[M + H]⁺	255.06518	255.06465	−2.099	MS² [2 5 5]: 255(1 0 0), 199(42.35), 137(27.50), 227(19.52), 181(9.02)
				[M−H]^-	253.05063	253.05052	3.891	MS² [2 5 3]: 253(1 0 0), 225(4.92), 133(4.20), 135(3.75), 223(2.24)
C2	Dihydrogenistein	7.97	C₁₅H₁₂O₅	[M−H]^-	271.06119	271.06122	4.132	MS² [2 7 1]: 165(1 0 0), 137(11.65), 107(7.30), 151(4.42), 271(1.51)
C3	Pratensein	8.93	C₁₆H₁₂O₆	[M + H]⁺	301.07066	301.07059	−0.248	MS² [3 0 1]: 301(1 0 0), 286(34.18), 269(26.45), 153(26.31), 229(23.35)
				[M−H]^-	299.05611	299.05621	3.997	MS² [2 9 9]: 284(1 0 0), 299(21.62), 227(7.06), 255(5.39), 200(2.37)
C4	Benzopyran-4,7-diol	1.17	C₉H₈O₃	[M + H]⁺	165.05462	165.05444	−0.181	MS² [1 6 5]: 123(1 0 0), 95(89.13), 119(49.29), 147(8.11), 165(3.53)
C5	Daidzin	4.74	C₂₁H₂₀O₉	[M + H]⁺	417.11800	417.11740	−1.459	MS² [4 1 7]: 255(1 0 0), 199(12.05), 137(6.00), 227(5.56), 181(1.33)
C6	2-(4-hydroxyphenyl)-propanoic acid	5.36	C₉H₁₀O₃	[M−H]^-	165.05571	165.05476	0.844	MS² [1 6 5]: 147(1 0 0), 119(86.68), 136(18.15), 121(16.88), 165(8.95)
C7	Dihydrodaidzein	7.11	C₁₅H₁₂O₄	[M−H]^-	255.06628	255.06615	3.782	MS² [2 5 5]: 149(1 0 0), 91(29.48), 135(21.56), 119(2.13), 255(1.08)
C8	5-hydroxyequol	7.22	C₁₅H₁₄O₄	[M−H]^-	257.08193	257.08185	3.947	MS² [2 5 7]: 221(1 0 0), 219(48.56), 151(2.79), 107(2.36)
C9	O-desmethylangolensin	9.33	C₁₅H₁₄O₄	[M−H]^-	257.08193	257.08185	3.947	MS² [2 5 7]: 109(1 0 0), 135(26.04), 257(18.29), 239(16.73), 213(5.93)

Metabolic pathways of calycosin-7-O-glucoside (orange line), ononin (blue line) and genistin (black line) by human intestinal bacteria; red-labeled metabolites were quantified by HPLC.

A3, B3 and C4 showed the same [M + H]⁺ ion at m/z 165.05462, and their formula were calculated as C₉H₉O₃. They all yielded the fragment ion at m/z 147 ([M + H-H₂O]⁺) due to the neutral loss of H₂O (18 Da), suggesting the presence of hydroxyl group. In addition, the fragment ion at m/z 95 suggested that there was a hydroxyl group on the benzene ring. Therefore, according to the literature (Matthies et al., 2009), A3, B3 and C4 were tentatively characterized as benzopyran-4,7-diol, as a B-ring removal product of calycosin, formononetin and genistein.

A5, B4 and C6 yielded the same [M−H]^- ion at m/z 165.05571 with formula of C₉H₉O₃. The [M−H−CO₂]^- ion at m/z 121 reminded the presence of carboxyl group. Meanwhile, the [M−H−H₂O]^- ion at m/z 147 was detected due to the neutral loss of H₂O (18 Da), suggesting the presence of hydroxyl group. Therefore, according to the literature (Peiroten et al., 2020), A5, B4 and C6 were tentatively identified as 2-(4-hydroxyphenyl)-propanoic acid.

A4 afforded [M−H]^- ion at m/z 181.05063, and their formula were calculated as C₉H₉O₄. It was 16 Da more than that of A5, B4 and C6, suggesting the existence of an extra hydroxyl group. Similar to A5, B4 and C6, the fragment ion at m/z 91 proved the presence of benzene ring. The [M−H−CO₂]^- ion at m/z 137 was detected due to the loss of carboxyl group. Moreover, the [M−H−CO₂−CH₃]^- ion at m/z 123 was 14 Da less than the ion at m/z 137, illustrating the existence of methyl group. Compared with the structures of calycosin and 3′,4′,7-trihydroxyisoflavone, A4 was tentatively characterized as 2-(3,4-dihydroxyphenyl)-propionic acid.

B8 with the [M−H]^- ion at m/z 269.08193 (C₁₆H₁₃O₄) was found in the transformation of ononin. It was 2 Da more than that of formononetin (m/z 267), indicating that B8 might be the hydrogenation metabolite of formononetin. Moreover, the [M−H−CH₃]^- ion at m/z 254 suggested the presence of methyl group. Therefore, according to the literature (Park et al., 2011) and comparing with the transformation of dihydrogenistein, B8 was tentatively identified as dihydroformononetin.

C7 with the [M−H]^- ion at m/z 255.06628 (C₁₅H₁₁O₄) was appeared in the transformation of genistin. It was 2 Da more than that of daidzein (m/z 253), while daidzein was the dehydroxylation product of genistein, suggesting that C7 might be the hydrogenation metabolite of daidzein. Meanwhile, the fragment ion at m/z 149 ([M−H−C₆H₄(-C)–OH]^-) could be characterized as the product of B-ring removal, illustrating that there was a hydroxyl group on the B ring. Therefore, C7 was tentatively identified as dihydrodaidzein (Shimada et al., 2010; Park et al., 2011).

C5 could generate the [M + H]⁺ ion at m/z 417.11800, and was 162 Da more than that of daidzein (m/z 255). Meanwhile, its cleavage behavior was highly similar to daidzein. This proved that it might be the glucoside of daidzein. Thus, C5 was identified as daidzin.

B5, C8 and C9 showed the same [M−H]^- ion at m/z 257.08193 (C₁₅H₁₄O₄). B5 and C8 were observed at 7.22 min, while C9 was at 9.33 min. They were all 16 Da less than that of dihydrogenistein (m/z 271), and 2 Da more than that of dihydrodaidzein (m/z 255). The fragment ions [M−H−2H₂O]^- at m/z 221 were found in both B5 and C8, and the [M−H−H₂O]^- ion at m/z 239 was in C9, suggesting the presence of hydroxyl groups. While in C9, the [M−H−CO₂]^- ion at m/z 213 was found due to the loss of carbonyl group. With the results in the literature, B5 and C8 were identified as 5-hydroxyequol (Yanjing and Xiumei 2015), which was the deoxidation metabolite of dihydrogenistein. Furthermore, C9 was characterized as O-desmethylangolensin (Maruo et al., 2008; Murota et al., 2018), the product of C-ring cleavage of dihydrodaidzein.

3.3

3.3 The anti-inflammatory and antioxidative activity results of the metabolites

The removal capacity of DPPH free radical is considered to be commonly associated with the antioxidant activity in vitro of plant extracts, and the DPPHs of series A, B and C (the metabolites of calycosin-7-O-glucoside, ononin and genistin, respectively) are shown in Fig. 5. For series A, the metabolites before 6 h had effective antioxidant activities with DPPHs＞90%. The levels of DPPH were 91.96 ± 0.24% at 0 h, 94.59 ± 0.59% at 2 h, and 93.60 ± 0.41% at 6 h, respectively. And the removal capacity of DPPH at 2 h was more visible than other times, which was consistent with the content of calycosin at this time. For series B, similar to that of series A, the DPPHs of the metabolites before 6 h were＞90%. The levels of DPPH were 91.29 ± 0.56% at 0 h, 94.93 ± 0.27% at 2 h, and 91.60 ± 0.25% at 6 h, respectively, and the peak level was appeared at 2 h. This was consistent with the growth trend of formononetin. Moreover, for series C, the metabolites at 0 h, 2 h, 6 h and 12 h showed beneficial DPPH scavenging effects with 90.76 ± 0.67%, 91.64 ± 0.41%, 92.13 ± 0.37% and 94.39 ± 0.20%, respectively. Meanwhile, the DPPH at 12 h was more than that of others times. Significantly, this result was closely related to the formation of genistein.

DPPH scavenging rate (%), NO production and T-AOC of the metabolites of A: calycosin-7-O-glucoside, B: ononin and C: genistin (n = 3, data are presented as mean ± SD; * p < 0.05, ** p < 0.01, *** p < 0.001 vs LPS group).

Anti-inflammatory and antioxidant activities of calycosin-7-O-glucoside, ononin and genistin at different transformation times were evaluated based on LPS-induced RAW264.7 cells. As shown in Fig. 5, the level of NO was significantly increased (p < 0.01), while T-AOC was obviously decreased compared with control group (p < 0.01), indicating the success of the model. For series A, the levels of NO at 0 h, 2 h, 6 h and 12 h were obviously decreased (p < 0.05), and the level of NO at 2 h was lower than that of other times. In addition, the levels of T-AOC at different times were also altered. Among them, the levels at 0 h, 2 h and 18 h were significantly increased (p < 0.05), and T-AOC content at 2 h was higher than that of the other two times. For series B, the metabolites of ononin at 0 h, 2 h and 6 h could significantly inhibit the production of NO in LPS-induced RAW264.7 cells (p < 0.05), and this inhibitory effect at 2 h was more pronounced (p < 0.01). For T-AOC, the metabolites of ononin at all times could all significantly elevate the level of T-AOC (p < 0.05), and the growth effect at 2 h was more violent than that of other times (p < 0.01). In series C, on the one hand, the metabolites of genistin could only significantly decrease the levels of NO at 6 h and 12 h (p < 0.05), and this suppression ratio at 12 h was more obvious. On the other hand, the metabolites of genistin contributed to the level of T-AOC at 0 h, 2 h, 12 h, 18 h and 24 h (p < 0.05), and the beneficial effect was optimized at 12 h (p < 0.05). Overall, the metabolites of calycosin-7-O-glucoside, ononin and genistin at different transformation times exhibit different antioxidant and anti-inflammatory activities. Among them, the metabolites of calycosin-7-O-glucoside and ononin at 2 h could both significantly inhibited NO secretion, and increased the level of T-AOC. While the metabolites of genistin at 12 h could only perform the same function.

Coincidentally, this phenomenon was consistent with the contents of calycosin at 2 h, formononetin at 2 h and genistein at 12 h, which were transformed from calycosin-7-O-glucoside, ononin and genistin. Speculating that the contents of glycosides were directly correlated with biological activities of biotransformation systems, and the results of DPPH experiments also confirmed the above speculations. However, the low biological activities at other times could not be attributed to the incomplete transformation, but rather to the low levels of other metabolites.

3.4

3.4 Network pharmacology analysis of the metabolites and the targets of inflammation and oxidation

3.4.1

3.4.1 Analysis of “compounds‑targets‑disease” network

249 biological targets with probability＞0 of the 18 metabolites (10 of series A, 10 of series B and 11 of series C) were obtained from the SwissTargetPrediction database. Based on the GeneCards database, 233 targets about “inflammation” and 242 about “oxidation” related to the 18 compounds were obtained. The “compounds-targets-disease” network is constructed in Fig. 6A. The yellow nodes represented the metabolic components, green nodes represented the targets, and red nodes represented the diseases (inflammation and oxidation).

A: “compounds-targets-disease” network analysis; B: GO enrichment and KEGG pathway analysis.

3.4.2

3.4.2 GO enrichment and KEGG pathway analysis

The GO enrichment analysis yielded 676 biological processes (BPs), 98 cellular components (CCs) and 191 molecular functions (MFs). Based on the p value arranged incrementally, the top 10 projects with lower P value of BP (p ≤ 1.73 × 10^-15), CC (p ≤ 2.15 × 10^-7) and MF (p ≤ 1.33 × 10^-11) for targets of 18 metabolites against inflammation and oxidation were selected for functional analysis, respectively (Fig. 6B). The results showed that the BPs enrichment items comprised protein phosphorylation, protein autophosphorylation, peptidyl-serine phosphorylation, response to xenobiotic stimulus, peptidyl-tyrosine phosphorylation, etc.; the CCs enrichment items contained cyclin-dependent protein kinase holoenzyme complex, cytosol, plasma membrane, membrane, neuronal cell body, etc.; the MFs enrichment items included protein kinase activity, transmembrane receptor protein tyrosine kinase activity, ATP binding, protein serine/threonine kinase activity, protein tyrosine kinase activity and so on. Furthermore, a total of 149 pathways were obtained from the KEGG pathway enrichment analysis, and the top 20 pathways were screened according to the p value (p ≤ 3.35 × 10^-9) for functional analysis (Fig. 6B). 18 metabolites exerted their anti-inflammatory and antioxidant activities through the following enriched pathways: prostate cancer, pathways in cancer, progesterone-mediated oocyte maturation, nitrogen metabolism, cellular senescence, cell cycle, endocrine resistance, PI3K-Akt signaling pathway, non-small cell lung cancer, EGFR tyrosine kinase inhibitor resistance and so on.

3.4.3

3.4.3 Analysis of PPI core network

As shown in Fig. 7, 111, 201 and 162 targets of the metabolites in series A, B and C were respectively screened by drawing a Venn diagram intersected with the genes related to inflammation and oxidation. These intersection genes in the Venn diagram were inputted into the String database to construct the PPI network. The greater the degree, the more the connected edges, and the more important the target protein corresponding to this node in the network. Then the key target nodes were selected with DC, EC, BC and CC higher than the corresponding average values.

The metabolite target PPI network (Series A: the metabolites of calycosin-7-O-glucoside, Series B: the metabolites of ononin, Series C: the metabolites of genistin).

For series A (the metabolites of calycosin-7-O-glucoside), the network comprised 111 nodes and 87 edges and the average node degree was 1.57. A total of 9 key targets whose DC ≥ 5, EC ≥ 0.011, BC ≥ 226 and CC ≥ 0.04 were used to construct the interaction network with the metabolites of series A. The color and the size of the key target nodes represented the reaction degree. Consequently, the key targets were AKR1C3, CDK5, EGFR, ESR1, CDK1, CDK6, CYP19A1, PARP1 and IGFBP3. For series B (the metabolites of ononin), the network comprised 201 nodes and 452 edges and the average node degree was 4.23. A total of 15 key targets whose DC ≥ 12, EC ≥ 0.049, BC ≥ 360 and CC ≥ 0.07 were used to construct the interaction network with the metabolites. Consequently, the key targets were SRC, PIK3CA, CDK1, EGFR, MAPK14, LCK, CCND1, ABL1, MAP2K1, ESR1, HCK, SNCA, PDPK1, GSK3B and RAF1. For series C (the metabolites of genistin), the network comprised 162 nodes and 288 edges and the average node degree was 3.56. A total of 10 key targets whose DC ≥ 10, EC ≥ 0.055, BC ≥ 291 and CC ≥ 0.08 were used to construct the interaction network with the metabolites. Consequently, the key targets were HSP90AA1, SRC, PIK3CA, CDK1, EGFR, ESR1, RAF1, PDPK1, SNCA and DNM1.

Interestingly, different metabolites from three series could treat inflammation and oxidative stress by different targets. For series A, the activities of calycosin against inflammation and oxidative stress was closely related to the targets of ESR1, EGFR and CYP19A1, while it had an unremarkable number of targets. Conversely, 3′,4′,7-trihydroxyisoflavone and genistein, which possessed multiple phenolic hydroxyl groups, could present their activities through 8 key targets of CDK5, EGFR, ESR1, CDK1, CDK6, CYP19A1, PARP1 and IGFBP3. For series B, ononin was closely related to 3 targets of ABL1, MAPK14 and SRC, and this was the same as formononetin (EGFR, ESR1 and RAF1). Dihydroformononetin, whose polarity was stronger than that of formononetin, was strongly associated with 12 key targets, including SRC, PIK3R1, PIK3CA, CDK1, LCK, CCND1, ABL1, MAP2K1, ESR1, HCK, GSK3B and RAF1. In series C, genistein was closely related to CDK1, EGFR, ESR1 and SNCA. O-desmethylangolensin was connected to 8 targets, including HSP90AA1, PIK3CA, CDK1, EGFR, ESR1, RAF1, PDPK1 and SNCA. These results illustrated that metabolites of natural compounds seems to be more partial advantages in treating diseases, and the polarity and the functional groups of metabolites could also enhance their activities. Combined with activity experiments in this study, the above results also indirectly confirmed that the activities of metabolites should be also related to their levels in three series.

4 Discussion

Macromolecular compounds, especially glucosides, were difficult to be absorbed into the circulation system, and their bioavailability can only be manifested based on the substances after being degraded. In this process, intestinal microorganisms play of importance role in the degradation of these compounds. In this study, we first determined the contents of 6 isoflavones in RAs from 12 regions, the results showed that these contents showed differences in RAs of different origins and different species. Next, 3 isoflavone glycosides in RA, including calycosin-7-O-glucoside, ononin and genistin, were transformed by human feces in vitro. Their levels were drastically reduced, while the levels of their aglycones were increased with transformation. This result also confirmed the fact that glucosides can be converted into their aglycones. Certainly, 4 other metabolites, including 3′,4′,7-trihydroxyisoflavone, pratensein, dihydrogenistein and daidzein, were also identified by UHPLC-HRMS and ¹H NMR, and the levels of them at different times were not stereotyped. This means the existence of other metabolic pathways of the 3 isoflavone glycosides. Other metabolites which could not be detected by HPLC were visualized by UHPLC-HRMS. A total of 18 metabolites were identified and a metabolic pathway network was constructed based on them. The network indicated that deglycosylation, hydrogenation, B-ring removal, C-ring cleavage, hydroxylation or dihydroxylation, and O-methylation or O-demethylation were the major metabolic reactions of the isoflavone glycoside in RA by human feces. The above experiments will be essential for the interpretation of metabolic profiles of calycosin-7-O-glucoside, ononin and genistin.

Gut microbes can convert glycosides into aglycones by secreting β-D-glucosidase, which can improve the bioavailability of glycosides by cutting off their sugar ligands. Lactobacillus spp., Bifidobacterium spp. and Bacteroides spp. in the human intestine have been reported to be all contribute to the conversion of glycosides (Zhang et al., 2014; Murota et al., 2018; Peiroten et al., 2020 ). In this study, the transformation of the 3 isoflavone glycosides into their aglycones may be completed with the participation of gut microbes.

However, equol, as the reduction metabolite of daidzein produced in most literature (Shimada et al., 2010), was not found in our study, suggesting that the metabolism by intestinal bacterial is diverse between individuals due to different daily diet composition, the physiological environment in intestine and the state of host–microbial symbiosis. Whereas, 5-hydroxyequol was found as the metabolite of genistein, which has one more hydroxyl group than equol. Three enzymes include daidzein reductase (DZNR), dihydrodaidzein reductase (DHDR) and tetrahydrodaidzein reductase (THDR) produced by gut microbes (‘Hugonella massiliensis’ DSM 101782 and Senegalimassilia faecalis KGMB 04484) (Soukup et al., 2021), are considered to be associated with the reduction of daidzein, and they can convert daidzein into dihydrodaidzein, tetrahydrodaidzein and equol, respectively. In addition to dihydrodaidzein, O-desmethylangolensin, which was reported to be another metabolite of daidzein (Maruo et al., 2008; Soukup et al., 2021), was identified in our study. Likewise, dihydrogenistein and dihydroformononetin were also found in metabolic profiles of genistein and formononetin. The above metabolites proved that the metabolic behaviors of these aglycones were highly similar, and intestinal flora were closely linked to their metabolism. The related literature has reported that the partial intestinal bacteria such as HGH6 can convert daidzein and genistein into dihydrodaidzein and dihydrogenistein, respectively, which reduction of the double bond between C-2 and C-3 is specific for isoflavone (Hur et al., 2000). The another specifically hydrogenase, which named as flavone reductase (FLR), has been discovered from the gut bacterium, Flavonifractor plautii (ATCC 49531) (Yang et al., 2021). FLR can reduce the C2-C3 double bond of flavonoids.

Genistein and daidzein can be first metabolized to 6′-hydroxy-O-desmethylangolensin or O-desmethylangolensin through the fragmentation of the C-ring, and then re-metabolized to 2-(4-hydroxyphenyl)-propanoic acid by the loss of B-ring (Peiroten et al., 2020). Likewise, 2-(3,4-dihydroxyphenyl)-propionic acid might be transformed from 3′,4′,7-trihydroxyisoflavone, which has one more hydroxyl group at C-3′ than that of genistein and daidzein. Whereas, benzopyran-4,7-diol has been found as the metabolite of daidzein by human fecal under anoxic conditions in the literature (Matthies et al., 2009). However, the specific bacteria or enzymatic reaction involved in this metabolism are blurry, that deserves to be explored in further study.

The reactions of hydroxylation, O-methylation and O-demethylation are often driven by the involvement of gut microbes. The C-4′ hydroxyl group of genistein and daidzein can be substituted by a methoxyl group to form biochanin A and formononetin, respectively. The other researchers have found that Eubacterium limosum (ATCC 8486) from human intestine, possess the O-demethylation effect on biochanin A and formononetin (Hur and Rafii 2000). Meanwhile, the demethylation rate by E. limosum in vitro was less than that in vivo, suggesting the presence of other bacteria in intestine with similar functions. In our study, the transformation processes from calycosin and formononetin to genistein and pratensein were discovered for the first time. For the above-mentioned or other unmentioned bacteria involved in the metabolic reaction, it is necessary for further screening and research.

As we known, the aglycone is the bioactive form of the isoflavone glycoside. Calycosin, formononetin and genistein show more excellent anti-inflammatory and antioxidant activities than their prototypes. On the LPS-induced zebrafish models, formononetin showed higher anti‑inflammatory effects than ononin, and formononetin can also decrease the levels of triacylglycerols (TAGs), and regulate the glycosylphosphatidylinositol (GPI) -anchor biosynthesis, while ononin not (Luo et al., 2019). Genistin in fermented soymilk was converted into genistein, and showed higher potential for anti-inflammatory and treatment of inflammatory bowel diseases (Hiramatsu et al., 2020). In our study, calycosin, formononetin and genistein show the significant therapeutic effects of anti-inflammatory and antioxidant (p < 0.05) when their levels reach the maximum value. With the further metabolism of these aglycones, the therapeutic effect was gradually weakened. Whereas, anti-inflammatory and antioxidant activities are measured only in vitro in this study, thus the experiment in vivo needs to be supplemented in the further research. Additionally, although other metabolites, such as dihydrogenistein and pratensein, also have certain anti-inflammatory and antioxidant effects, they cannot reverse this trend of the therapeutic effect because of their limited contents. However, the roles of these metabolites against inflammation and oxidative stress in the whole transformation system are mysterious.

For network pharmacology. Firstly, the relevant targets were obtained from the SwissTargetPrediction and GeneCards databases, and the “compounds-targets-disease” network was constructed. Next, these targets were analyzed by GO enrichment and KEGG pathway analysis. Several of these pathways, including nitrogen metabolism, cellular senescence, cell cycle, PI3K-Akt signaling pathway and EGFR tyrosine kinase inhibitor, resistance, attracted our attention. Nitrogen metabolism suggested that these metabolites can exert anti-inflammatory effects by affecting the production of NO, which was one of the inflammatory factors. Cellular senescence and cell cycle proved that these metabolites contributed to the recovery of cell damage caused by inflammation and oxidative stress. PI3K-Akt pathway is an intracellular signal transduction pathway that responds to extracellular signals and promotes metabolism, proliferation, cell survival, and angiogenesis. The inhibition of PI3K-Akt pathway can enhance the expression of nitric oxide synthase, which can induce inflammation and oxidative stress (Mayer and Arteaga 2016). EGFR tyrosine kinase inhibitor can inhibit abnormal cell proliferation and exert anti-inflammatory and antioxidant activities by inhibiting the overexpression of EGFR (Rayego-Mateos et al., 2018).

Especially, in the PPI core network analysis, the key targets of EGFR, ESR1 and CDK1 were screened out in three transformation series. EGFR (epidermal growth factor receptor) is a member of human epidermal growth factor receptor (HER) family and a membrane tyrosine kinase receptor expressed in the kidney. Activation of the EGFR signaling pathway is linked to the regulation of several cellular responses, including cell proliferation, inflammatory processes, and extracellular matrix regulation. The decreased expression of EGFR exerts anti-inflammatory activity in inflammatory diseases such as asthma and enteritis. (Rayego-Mateos et al., 2018). ESR1 (estrogen receptor 1) and ESR2 genes encode estrogen receptor, which is related to the estrogenic effect of isoflavones. However, ESR1 is more involved in the cell differentiation process and apoptosis whereas ESR2 is more competent in spermiogenesis regulation (Krela-Kazmierczak et al., 2019). CDK1 (cyclin-dependent kinase 1) play an important role in the cell cycle of eukaryotic cells, controlling DNA replication and segregation, transcriptional programs, resumption of meiosis and cell morphogenesis (Bielak-Zmijewska et al., 2010). These targets may be the common mechanism of the three series against inflammation and oxidative stress.

Additionally, by comparing the number of edges on the metabolic components connecting the targets, 3′,4′,7-trihydroxyisoflavone and genistein with each 8 edges might play a more important role in the anti-oxidative and anti-inflammatory effects of series A. Subsequently, dihydroformononetin (12 edges) and O-desmethylangolensin (8 edges) were the key metabolites of series B and C, respectively. Consequently, the anti-inflammatory and antioxidant mechanisms about the key targets and the comparison of the activities of these key metabolites with their prototype components are worthy of further study.

5 Conclusion

In this study, 3 isoflavone glucosides from RA were transformed at different times by human feces in vitro. A total of 10 metabolites, including 3 prototype glucosides (calycosin-7-O-glucoside, ononin and genistin), 3 corresponding aglycones (calycosin, formononetin and genistein), and 4 other metabolites identified by UHPLC-HRMS and ¹H-NMR (3′,4′,7-trihydroxyisoflavone, pratensein, dihydrogenistein and daidzein), were quantified by HPLC, and the metabolic process of pratensein was discovered for the first time. Furthermore, 8 additional metabolites were characterized by UHPLC-HRMS. The metabolic network of 18 metabolites was constructed, and the metabolic reactions of these metabolites were deglycosylation, hydrogenation, B-ring removal, C-ring cleavage, hydroxylation or dihydroxylation, and O-methylation or O-demethylation. In the previous studies, the metabolites of calycosin-7-O-glucoside, ononin and genistin in RA were only qualitatively identified, while lacked of quantitative research. In this study, we elucidated their metabolic process and trend of content changes through both qualitative and quantitative research. This improved the problem of previous research that only focused on qualitative identification, while lacked of quantitative research. At the same time, 4 metabolites were separated and further characterized through multiple methods. The above results provide guidance to explain the transformation of flavonoid glycoside components in the intestine.

Subsequently, the anti-inflammatory and antioxidant activities of the metabolites at different times were evaluated by DPPH scavenging in vitro and LPS-stimulated RAW 264.7 cell. By comparing the results with HPLC quantitative analysis, it is shown that the anti-inflammatory and antioxidant activities of the metabolites were closely related to the formation of the aglycones. Meanwhile, network pharmacology was applied to reveal the relationship and mechanism between anti-inflammatory and antioxidant activities and all metabolites. These results also provide reasonable support for the transformation system. However, there were still few deficiencies in this study. The important role of gut microbes in the whole biotransformation experiment was mysterious, and the reasons for the activities difference of the metabolites at different times should not be ignored. These challenges will be revealed in further studies.

Acknowledgments

The work had been financially supported by Shandong Province Chinese herbal medicine and decoction piece standard research topic (2020-201); Yantai campus integration development project (NO. 2019XDRHXMPT18); Binzhou Medical University Scientific Research Fund for High-level Talents (BY2018KYQD11, 2019KYQD05 and 2019KYQD06); Young and Creative Team for Talent Introduction of Shandong Province (10073004); The Project of Shandong Provincial Natural Fund (ZR2021QH009, ZR2020MH371 and ZR2020MH372); Medical and Health Technology Development Plan at Shandong Province (202006031281).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

Adesso S., Russo R., Quaroni A., . Astragalus membranaceus extract attenuates inflammation and oxidative stress in intestinal epithelial cells via NF-kappaB activation and Nrf2 response. Int. J. Mol. Sci.. 2018;19
[CrossRef] [Google Scholar]
Alam M.A., Subhan N., Rahman M.M., . Effect of citrus flavonoids, naringin and naringenin, on metabolic syndrome and their mechanisms of action. Adv. Nutr.. 2014;5:404-417.
[CrossRef] [Google Scholar]
Bi Y., Bao H., Zhang C., . Quality control of Radix Astragali (The Root of Astragalus membranaceus var. mongholicus) along its value chains. Front. Pharmacol.. 2020;11:562376
[CrossRef] [Google Scholar]
Bielak-Zmijewska A., Sikora-Polaczek M., Nieznanski K., . Curcumin disrupts meiotic and mitotic divisions via spindle impairment and inhibition of CDK1 activity. Cell Prolif.. 2010;43:354-364.
[CrossRef] [Google Scholar]
Chen Z., Liu L., Gao C., . Astragali Radix (Huangqi): a promising edible immunomodulatory herbal medicine. J. Ethnopharmacol.. 2020;258:112895
[CrossRef] [Google Scholar]
Hiramatsu E.Y., de Avila A.R.A., Genova V.M., . Biotransformation processes in soymilk isoflavones to enhance anti-inflammatory potential in intestinal cellular model. J. Food Biochem.. 2020;44:e13149.
[Google Scholar]
Huang Y., Kwan K.K.L., Leung K.W., . The extracts and major compounds derived from Astragali Radix alter mitochondrial bioenergetics in cultured cardiomyocytes: comparison of various polar solvents and compounds. Int. J. Mol. Sci.. 2018;19
[CrossRef] [Google Scholar]
Hur H.G., Lay J.O. Jr., Beger R.D., . Isolation of human intestinal bacteria metabolizing the natural isoflavone glycosides daidzin and genistin. Arch. Microbiol.. 2000;174:422-428.
[CrossRef] [Google Scholar]
Hur H.-G., Rafii F., . Biotransformation of the isoflavonoids biochanin A, formononetin, and glycitein by Eubacterium limosum. FEMS Microbiol. Lett.. 2000;192:21-25.
[CrossRef] [Google Scholar]
Krela-Kazmierczak I., Skrzypczak-Zielinska M., Kaczmarek-Rys M., . ESR1 gene variants are predictive of osteoporosis in female patients with Crohn's disease. J. Clin. Med.. 2019;8
[CrossRef] [Google Scholar]
Liu X.Y., Fan M.L., Wang H.Y., . Metabolic profile and underlying improved bio-activity of Fructus aurantii immaturus by human intestinal bacteria. Food Funct.. 2017;8:2193-2201.
[CrossRef] [Google Scholar]
Luo L., Zhou J., Zhao H., . The anti-inflammatory effects of formononetin and ononin on lipopolysaccharide-induced zebrafish models based on lipidomics and targeted transcriptomics. Metabolomics. 2019;15:153.
[CrossRef] [Google Scholar]
Lv Y.W., Hu W., Wang Y.L., . Identification and determination of flavonoids in astragali radix by high performance liquid chromatography coupled with DAD and ESI-MS detection. Molecules. 2011;16:2293-2303.
[CrossRef] [Google Scholar]
Maruo T., Sakamoto M., Ito C., . Adlercreutzia equolifaciens gen. nov., sp. nov., an equol-producing bacterium isolated from human faeces, and emended description of the genus Eggerthella. Int. J. Syst. Evol. Microbiol.. 2008;58:1221-1227.
[CrossRef] [Google Scholar]
Matthies A., Blaut M., Braune A., . Isolation of a human intestinal bacterium capable of daidzein and genistein conversion. Appl. Environ. Microbiol.. 2009;75:1740-1744.
[CrossRef] [Google Scholar]
Mayer I.A., Arteaga C.L., . The PI3K/AKT pathway as a target for cancer treatment. Annu. Rev. Med.. 2016;67:11-28.
[CrossRef] [Google Scholar]
Murota K., Nakamura Y., Uehara M., . Flavonoid metabolism: the interaction of metabolites and gut microbiota. Biosci. Biotechnol. Biochem.. 2018;82:600-610.
[CrossRef] [Google Scholar]
Park H.Y., Kim M., Han J., . Stereospecific microbial production of isoflavanones from isoflavones and isoflavone glucosides. Appl. Microbiol. Biotechnol.. 2011;91:1173-1181.
[CrossRef] [Google Scholar]
A.P. Peiroten, Gaya, M. f. L. J, Application of recombinant lactic acid bacteria and bifidobacteria able to enrich soy beverage in dihydrodaidzein and dihydrogenistein, Food Res Int. vol. 134 (2020) p. 109257 Doi: 10.1016/j.foodres.2020.109257.
Peiroten A., Gaya P., Alvarez I., . Production of O-desmethylangolensin, tetrahydrodaidzein, 6'-hydroxy-O-desmethylangolensin and 2-(4-hydroxyphenyl)-propionic acid in fermented soy beverage by lactic acid bacteria and Bifidobacterium strains. Food Chem.. 2020;318:126521
[CrossRef] [Google Scholar]
Rayego-Mateos S., Rodrigues-Diez R., Morgado-Pascual J.L., . Role of Epidermal Growth Factor Receptor (EGFR) and its ligands in kidney inflammation and damage. Mediators Inflamm.. 2018;2018:8739473.
[CrossRef] [Google Scholar]
Shimada Y., Yasuda S., Takahashi M., . Cloning and expression of a novel NADP(H)-dependent daidzein reductase, an enzyme involved in the metabolism of daidzein, from equol-producing Lactococcus strain 20–92. Appl. Environ. Microbiol.. 2010;76:5892-5901.
[CrossRef] [Google Scholar]
Soukup S.T., Stoll D.A., Danylec N., . Metabolism of Daidzein and Genistein by gut bacteria of the class Coriobacteriia. Foods. 2021;10
[CrossRef] [Google Scholar]
H.f. Su, S. Shaker, Y. Kuang et al., Phytochemistry and cardiovascular protective effects of Huang‐Qi (Astragali Radix), Medicinal Research Reviews vol. 41 (2021) p. 1999-2038 Doi: 10.1002/med.21785.
Tang Z., Huang G., . Extraction, structure, and activity of polysaccharide from Radix astragali. Biomed. Pharmacother.. 2022;150:113015
[CrossRef] [Google Scholar]
Xiao W., Han L., Shi B., . Microwave-assisted extraction of flavonoids from Radix Astragali. Separ. Purif. Technol.. 2008;62:614-618.
[CrossRef] [Google Scholar]
Yang L., Han X., Xing F., . Total flavonoids of astragalus attenuates experimental autoimmune encephalomyelitis by suppressing the activation and inflammatory responses of microglia via JNK/AKT/NFkappaB signaling pathway. Phytomedicine. 2021;80:153385
[CrossRef] [Google Scholar]
Yang G., Hong S., Yang P., . Discovery of an ene-reductase for initiating flavone and flavonol catabolism in gut bacteria. Nat. Commun.. 2021;12:790.
[CrossRef] [Google Scholar]
Yang B., Yang N., Chen Y., . An integrated strategy for effective-component discovery of Astragali Radix in the treatment of lung cancer. Front. Pharmacol.. 2020;11:580978
[CrossRef] [Google Scholar]
Yanjing X., Xiumei Y., . Bioconversion of genistein to (-)-5-hydroxy-equol by a newly isolated cock intestinal anaerobic bacterium. J. Chinese Pharmaceut. Sci.. 2015;24
[CrossRef] [Google Scholar]
Zhang W., Jiang S., Qian D., . Analysis of interaction property of calycosin-7-O-beta-D-glucoside with human gut microbiota. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci.. 2014;963:16-23.
[CrossRef] [Google Scholar]

Appendix A

Supplementary material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.arabjc.2023.105005.

Appendix A

Supplementary material

The following are the Supplementary data to this article:

Supplementary data 1

Download

Introduction

Materials and methods

Chemicals and reagents
Quantitative analysis of 6 isoflavones in multi-batch RA by HPLC
Biotransformation in vitro of calycosin-7-O-glucoside, ononin and genistin by human feces
Separation and identification analysis of the unknown metabolites
Quantitative analysis of the identified metabolites by HPLC
Identification analysis of the other metabolites by LC-MS/MS
Antioxidant activity of the metabolites by DPPH radical scavenging assay in vitro
Anti-inflammatory and antioxidative activities of the metabolites based on LPS-stimulated RAW 264.7 cell
Network pharmacology analysis of the metabolites and the targets of inflammation and oxidation

Results

Quantitative analysis of 6 isoflavones in multi-batch RA by HPLC
Biotransformation results of calycosin-7-O-glucoside, ononin and genistin by human feces
The anti-inflammatory and antioxidative activity results of the metabolites
Network pharmacology analysis of the metabolites and the targets of inflammation and oxidation

Discussion

Conclusion

Fulltext Views
645

PDF downloads
333

View/Download PDF
Download Citations

BibTeX
RIS

Show Sections

[1] Adesso S., Russo R., Quaroni A., . Astragalus membranaceus extract attenuates inflammation and oxidative stress in intestinal epithelial cells via NF-kappaB activation and Nrf2 response. Int. J. Mol. Sci.. 2018;19
[CrossRef] [Google Scholar]

[2] Alam M.A., Subhan N., Rahman M.M., . Effect of citrus flavonoids, naringin and naringenin, on metabolic syndrome and their mechanisms of action. Adv. Nutr.. 2014;5:404-417.
[CrossRef] [Google Scholar]

[3] Bi Y., Bao H., Zhang C., . Quality control of Radix Astragali (The Root of Astragalus membranaceus var. mongholicus) along its value chains. Front. Pharmacol.. 2020;11:562376
[CrossRef] [Google Scholar]

[4] Bielak-Zmijewska A., Sikora-Polaczek M., Nieznanski K., . Curcumin disrupts meiotic and mitotic divisions via spindle impairment and inhibition of CDK1 activity. Cell Prolif.. 2010;43:354-364.
[CrossRef] [Google Scholar]

[5] Chen Z., Liu L., Gao C., . Astragali Radix (Huangqi): a promising edible immunomodulatory herbal medicine. J. Ethnopharmacol.. 2020;258:112895
[CrossRef] [Google Scholar]

[6] Hiramatsu E.Y., de Avila A.R.A., Genova V.M., . Biotransformation processes in soymilk isoflavones to enhance anti-inflammatory potential in intestinal cellular model. J. Food Biochem.. 2020;44:e13149.
[Google Scholar]

[7] Huang Y., Kwan K.K.L., Leung K.W., . The extracts and major compounds derived from Astragali Radix alter mitochondrial bioenergetics in cultured cardiomyocytes: comparison of various polar solvents and compounds. Int. J. Mol. Sci.. 2018;19
[CrossRef] [Google Scholar]

[8] Hur H.G., Lay J.O. Jr., Beger R.D., . Isolation of human intestinal bacteria metabolizing the natural isoflavone glycosides daidzin and genistin. Arch. Microbiol.. 2000;174:422-428.
[CrossRef] [Google Scholar]

[9] Hur H.-G., Rafii F., . Biotransformation of the isoflavonoids biochanin A, formononetin, and glycitein by Eubacterium limosum. FEMS Microbiol. Lett.. 2000;192:21-25.
[CrossRef] [Google Scholar]

[10] Krela-Kazmierczak I., Skrzypczak-Zielinska M., Kaczmarek-Rys M., . ESR1 gene variants are predictive of osteoporosis in female patients with Crohn's disease. J. Clin. Med.. 2019;8
[CrossRef] [Google Scholar]

[11] Liu X.Y., Fan M.L., Wang H.Y., . Metabolic profile and underlying improved bio-activity of Fructus aurantii immaturus by human intestinal bacteria. Food Funct.. 2017;8:2193-2201.
[CrossRef] [Google Scholar]

[12] Luo L., Zhou J., Zhao H., . The anti-inflammatory effects of formononetin and ononin on lipopolysaccharide-induced zebrafish models based on lipidomics and targeted transcriptomics. Metabolomics. 2019;15:153.
[CrossRef] [Google Scholar]

[13] Lv Y.W., Hu W., Wang Y.L., . Identification and determination of flavonoids in astragali radix by high performance liquid chromatography coupled with DAD and ESI-MS detection. Molecules. 2011;16:2293-2303.
[CrossRef] [Google Scholar]

[14] Maruo T., Sakamoto M., Ito C., . Adlercreutzia equolifaciens gen. nov., sp. nov., an equol-producing bacterium isolated from human faeces, and emended description of the genus Eggerthella. Int. J. Syst. Evol. Microbiol.. 2008;58:1221-1227.
[CrossRef] [Google Scholar]

[15] Matthies A., Blaut M., Braune A., . Isolation of a human intestinal bacterium capable of daidzein and genistein conversion. Appl. Environ. Microbiol.. 2009;75:1740-1744.
[CrossRef] [Google Scholar]

[16] Mayer I.A., Arteaga C.L., . The PI3K/AKT pathway as a target for cancer treatment. Annu. Rev. Med.. 2016;67:11-28.
[CrossRef] [Google Scholar]

[17] Murota K., Nakamura Y., Uehara M., . Flavonoid metabolism: the interaction of metabolites and gut microbiota. Biosci. Biotechnol. Biochem.. 2018;82:600-610.
[CrossRef] [Google Scholar]

[18] Park H.Y., Kim M., Han J., . Stereospecific microbial production of isoflavanones from isoflavones and isoflavone glucosides. Appl. Microbiol. Biotechnol.. 2011;91:1173-1181.
[CrossRef] [Google Scholar]

[19] A.P. Peiroten, Gaya, M. f. L. J, Application of recombinant lactic acid bacteria and bifidobacteria able to enrich soy beverage in dihydrodaidzein and dihydrogenistein, Food Res Int. vol. 134 (2020) p. 109257 Doi: 10.1016/j.foodres.2020.109257.

[20] Peiroten A., Gaya P., Alvarez I., . Production of O-desmethylangolensin, tetrahydrodaidzein, 6'-hydroxy-O-desmethylangolensin and 2-(4-hydroxyphenyl)-propionic acid in fermented soy beverage by lactic acid bacteria and Bifidobacterium strains. Food Chem.. 2020;318:126521
[CrossRef] [Google Scholar]

[21] Rayego-Mateos S., Rodrigues-Diez R., Morgado-Pascual J.L., . Role of Epidermal Growth Factor Receptor (EGFR) and its ligands in kidney inflammation and damage. Mediators Inflamm.. 2018;2018:8739473.
[CrossRef] [Google Scholar]

[22] Shimada Y., Yasuda S., Takahashi M., . Cloning and expression of a novel NADP(H)-dependent daidzein reductase, an enzyme involved in the metabolism of daidzein, from equol-producing Lactococcus strain 20–92. Appl. Environ. Microbiol.. 2010;76:5892-5901.
[CrossRef] [Google Scholar]

[23] Soukup S.T., Stoll D.A., Danylec N., . Metabolism of Daidzein and Genistein by gut bacteria of the class Coriobacteriia. Foods. 2021;10
[CrossRef] [Google Scholar]

[24] H.f. Su, S. Shaker, Y. Kuang et al., Phytochemistry and cardiovascular protective effects of Huang‐Qi (Astragali Radix), Medicinal Research Reviews vol. 41 (2021) p. 1999-2038 Doi: 10.1002/med.21785.

[25] Tang Z., Huang G., . Extraction, structure, and activity of polysaccharide from Radix astragali. Biomed. Pharmacother.. 2022;150:113015
[CrossRef] [Google Scholar]

[26] Xiao W., Han L., Shi B., . Microwave-assisted extraction of flavonoids from Radix Astragali. Separ. Purif. Technol.. 2008;62:614-618.
[CrossRef] [Google Scholar]

[27] Yang L., Han X., Xing F., . Total flavonoids of astragalus attenuates experimental autoimmune encephalomyelitis by suppressing the activation and inflammatory responses of microglia via JNK/AKT/NFkappaB signaling pathway. Phytomedicine. 2021;80:153385
[CrossRef] [Google Scholar]

[28] Yang G., Hong S., Yang P., . Discovery of an ene-reductase for initiating flavone and flavonol catabolism in gut bacteria. Nat. Commun.. 2021;12:790.
[CrossRef] [Google Scholar]

[29] Yang B., Yang N., Chen Y., . An integrated strategy for effective-component discovery of Astragali Radix in the treatment of lung cancer. Front. Pharmacol.. 2020;11:580978
[CrossRef] [Google Scholar]

[30] Yanjing X., Xiumei Y., . Bioconversion of genistein to (-)-5-hydroxy-equol by a newly isolated cock intestinal anaerobic bacterium. J. Chinese Pharmaceut. Sci.. 2015;24
[CrossRef] [Google Scholar]

[31] Zhang W., Jiang S., Qian D., . Analysis of interaction property of calycosin-7-O-beta-D-glucoside with human gut microbiota. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci.. 2014;963:16-23.
[CrossRef] [Google Scholar]

Study on the biotransformation and activities of three isoflavone glycosides from Radix Astragali by human feces in vitro

Abstract

Keywords

Radix Astragali

Isoflavone glycoside

Human feces

Anti-inflammatory

Antioxidant

Network pharmacology

1 Introduction

2 Materials and methods

2.1 Chemicals and reagents

2.2 Quantitative analysis of 6 isoflavones in multi-batch RA by HPLC

2.3 Biotransformation in vitro of calycosin-7-O-glucoside, ononin and genistin by human feces

2.4 Separation and identification analysis of the unknown metabolites

2.5 Quantitative analysis of the identified metabolites by HPLC

2.6 Identification analysis of the other metabolites by LC-MS/MS

2.7 Antioxidant activity of the metabolites by DPPH radical scavenging assay in vitro

2.8 Anti-inflammatory and antioxidative activities of the metabolites based on LPS-stimulated RAW 264.7 cell

2.9 Network pharmacology analysis of the metabolites and the targets of inflammation and oxidation

3 Results

3.1 Quantitative analysis of 6 isoflavones in multi-batch RA by HPLC

3.2 Biotransformation results of calycosin-7-O-glucoside, ononin and genistin by human feces

3.2.1 HPLC analysis results of the metabolites of calycosin-7-O-glucoside, ononin and genistin

3.2.2 Separation and identification of the unknown metabolites

3.2.3 The quantitative analysis of the identified metabolites by HPLC

3.2.4 Identification analysis of the other metabolites by LC-MS/MS

3.3 The anti-inflammatory and antioxidative activity results of the metabolites

3.4 Network pharmacology analysis of the metabolites and the targets of inflammation and oxidation

3.4.1 Analysis of “compounds‑targets‑disease” network

3.4.2 GO enrichment and KEGG pathway analysis

3.4.3 Analysis of PPI core network

4 Discussion

5 Conclusion

Acknowledgments

Declaration of Competing Interest

References

Appendix A

Supplementary material

Appendix A

Supplementary material

Supplementary data 1

Suggested read for related articles: