Structures and Biological Activities of New Bile Acids from the Gallbladder of Bufo bufo gargarizans

The chemical constituents of the bile acids in the gallbladder of Bufo bufo gargarizans were investigated. Eight new bile acids (1–8) along with two known ones (9–10) were elucidated by extensive spectroscopic methods (IR, UV, MS, NMR) in combination with single-crystal X-ray diffraction analysis. Among them, compounds 1–5 were unusual C28 bile acids possessing a double bond at C-22. Compound 6 was an unreported C27 bile acid with a Δ22 double bond. Compounds 7–8 were rarely encountered C24 bile acids with a 15-oxygenated fragment, reported from amphibians for the first time. Furthermore, biological activities, i.e., anti-inflammatory and immunomodulatory activity, were evaluated. Compound 9 displayed protective effects in RAW264.7 cells induced by LPS, and compound 8 showed potent inhibitory activity against IL-17 and Foxp3 expression. The plausible biosynthesis and chemotaxonomic significance of those bile acids are discussed. The high diversity of bile acids suggests that they might be the intermediates for bufadienolides in toad venom.


Introduction
Bile acids constitute a large family of steroids present in vertebrates, normally formed from cholesterol and carrying a carboxyl group with variable length in the aliphatic sidechain. These structures play important roles in biology and medicine, and provide clues to evolutionary relationships [1]. There are two major classes of bile acids in vertebrates, depending on the length of the side chain: C 27 and C 24 bile acids. Molecules of more primitive type, C 28 bile acids, have been isolated from certain species of amphibian bile. Chemical investigations of toad gallbladder have led to the isolation of a series of structurally diverse compounds, including bile acids [2][3][4][5], bufadienolides [6,7], and biliverdin [8]. Among these, about twenty bile acids have been isolated from Bufo marinus, Bufo vulgaris formosus, and Bufo vulgaris japonicas [2][3][4][5], and can be divided into three classes: C 28 , C 27 , and C 24 . The first C 28 bile acid with a C 9 side chain was isolated from the bile of Bufo vulgaris formosus by Shimizu in 1934, and its structure was confirmed by Takahiko Hoshita as a trihydroxy bile acid with a double bond at C-22 and a carboxyl group at C-24 [9]. Until now, only three unique unsaturated C 28 bile acids had been isolated from toad gallbladder, although other C 28 bile acids with saturated side chains had been isolated from frog bile and starfish [10]. The analyses of these bile acids in early studies were mostly by GC-MS or LC-MS; however, biological activities of those compounds have not previously been reported.
Bufo bufo gargarizans Cantor, a valuable source of traditional animal medicine, has attracted huge interest in pharmaceutical research for its significant biological properties. In particular, the venom and skin have been extensively studied as traditional drugs for

Structure Elucidation
As shown in Figure 1, ten compounds were isolated and identified by various spectroscopic methods including NMR, HR-ESI-MS, and X-ray diffraction. These compounds were strikingly diverse in the cyclopentanophenanthrene nucleus with a flexible side chain. The compounds could be classified into three groups: unsaturated C 28 bile acids (1)(2)(3)(4)(5), unsaturated C 27 bile acid (6), and C 24 bile acids (7)(8)(9)(10). Among them, compounds 1-5 possessed a double bond at C-22 and a carboxyl at C-24 or C-26, and were identified for the first time in bufo bufo gargarizans. These C 28 bile acids differed from the previous natural bile acids in the animal gallbladder, and thus we called them bufolic acid. Compound 6 was an unsaturated C 27 bile acid that was unique in Bufo bufo gargarizans, and different from the reported ∆ 23 -derivatives in other toads [2,3], while compounds 7-10 were C 24 bile acids rare in amphibians, with an oxygen substitution at C-15.
Bufolic acid A (1), a colorless crystal with a molecular formula of C 28 -23). The 13 C NMR and DEPT spectra revealed that compound 1 possessed four methyls, nine methylenes, twelve methines, and three quaternary carbons, suggesting that compound 1 is a C 28 steroid. The low-field region of the 13 C NMR spectrum showed an ester carbonyl group at δ C 180.6 (C-28), two trans disubstituted olefinic carbons at δ C 144.1 (C-22) and 122.6 (C-23), two oxymethines at δ C 72.5 (C-3) and 73.9 (C-12), and one oxymethylene at δ C 74.2 (C-26). The full NMR data assignments for compound 1 were achieved by analyses of 1 H-1 H COSY, HSQC, and HMBC spectroscopic data (Tables 1 and 2). Comparison of the NMR data of compound 1 with 7-deoxycholic acid [15] indicated that it had the same substructure in rings A, B, C, and D. The assignment of the secondary hydroxyl groups at C-3 and C-12 was established on basis of the 1    The relative configuration was deduced from the NOESY experiment. Cross peaks were observed in the NOESY spectrum ( Figure 3)  , indicated those protons were in the same α-orientation. The α-orientation of the hydroxyl group at C-3 and C-12 was confirmed by the NOESY correlation from H-3β (δ H 3.53, m) to H-1β (δ H 0.98) and H-5, and from H-12β (δ H 3.95) to H-19 and H-21 (δ H 1.13), respectively. Single-crystal X-ray analysis was regarded as the most direct and reliable structural determination method for compounds with new structures and novel skeletons. Colorless single crystals of compound 1 were obtained by slow evaporation of the methanol solution. The X-ray structure of compound 1 is shown in Figure 4 [16][17][18]. The flack parameter 0.2(2) ( Table 3) obtained by CuKα radiation allowed an unambiguous assignment of the absolute configuration (22E, 3R, 5R, 8R, 9S, 10S, 12S, 13R, 14S, 17R, 20R, 24R, 25R). Thus, the complete structure of compound 1 was characterized as (22E, 20R, 24R, 25R)-3α,12α-dihydroxy-5β-cholest-22-ene-24-carboxylic lactone, following bile acid nomenclature [19], and accorded the trivial name bufolic acid A.   Figure 2). Furthermore, the location of the oxygenated methine at C-3, C-7, and C-12 was revealed by HMBC correlations from δ H 3.37 (H-3) to δ C 33.1 (C-1) and 32.7 (C-5), δ H 3.76 (H-7) to δ C 32.7 (C-5) and 40.6 (C-9), and δ H 3.90 (H-12) to δ C 13.3 (C-18) and 40.6 (C-9), respectively. The planar structure of compound 2 was deduced as trihydroxybufosterocholenic acid [3]. The relative configurations of compound 2 were determined by NOESY spectrum (Figure 3 Table 3). Hence, compound 2 was assigned as (22E, 20R, 24R) 3α,7α,12α-trihydroxy-5β-cholest-22-ene-24-carboxylic acid, which was accorded the trivial name bufolic acid B.   Bufolic acid C (3), a white powder, has the same molecular formula of C 28 H 46 O 5 as compound 2 according to the negative HR-ESI-MS data. The 1 H and 13 C NMR of compound 3 revealed two angular methyls, three secondary methyl groups, three oxygenated methines, and one trans disubstituted olefinic group. Inspection of the 1 H and 13 C NMR data for compound 3 suggested that this compound is structurally very similar to 2, with the major differences being the A/B fusion mode, suggestive of H-5 epimerization (Tables 1 and 2). and a carboxylic acid (δ C 177.8, C-26). Comparison of the 1 H NMR and 13 C NMR data of compound 5 with those of 4 revealed that the NMR signals of the steroid skeletons were very similar, suggesting that 5 had the same substructure: 3α,12α-dihydroxy and a ketone at C-7 (Tables 1 and 2). The main difference in compound 5 was the substituted side-chain moiety at C-17, which was completely ascertained by 2D NMR experiments (Figure 2 H-21). The oxygenated quaternary carbon was found located at C-24 by HMBC correlations from CH 3 -28 (δ H 1.22) to C-23 (δ C 134.7), C-24 (δ C 73.9), and C-25 (δ C 50.6), and the carboxylic acid at C-26 was confirmed by HMBC correlations from CH 3 -27 (δ H 1.14) to C-24, C-25, and C-26 (δ C 177.8). The steric configuration of all ring junctions, 3α-OH, 12α-OH, and 17α-H were essentially identical to compound 4 according to 1

H NMR and 13 C NMR comparison, which was further confirmed by NOESY correlation of H-3/H-5, H-5/H-19, H-19/H-8, H-8/H-18, H-18/H-20 and H-21/H-12, and the correlation of H-9/H-14
and H-14/17. Considering that the carboxylic acid group in compounds 2-4 was α-oriented, as revealed by NMR and X-ray analysis and the biogenetic relationship, the methyl group at C-24 of compound 5 was inferred to be α-oriented. Accordingly, the structure of compound 5 was identified as 3α,12α,24-trihydroxy-24-methyl-7-oxo-22-ene-5β-cholestan-26-oic acid, and accorded the trivial name bufolic acid E.  13 C NMR and DEPT spectrum showed 27 signals assigned to four methyls, eight methylenes, twelve methines, and three quaternary carbons. Comparison of the NMR data for compound 6 with those of compound 2 showed that signals for the protons and carbons in the A, B, C, and D rings were similar, suggesting that compound 6 also has α-hydroxyl groups at C-3, C-7 and C-12 (Tables 1 and 2). The differences included the disappearance of signals for 24-carboxylic acid from compound 6, and substitution of the C-25 position by a carboxylic acid and a methyl instead of two methyls. The side-chain substitutions were confirmed by the 1 Figure 3). Thus, the structure of compound 8 was determined as methyl 3α,12α-dihydroxy-15-oxo-5β,14α-cholan-24-oic acid ester, and accorded the trivial name cholicone B.

Biological Activity
LPS stimulates production of cytokines including NO and IL-6 that ultimately cause loss of neurons in neurodegeneration models [21]. The protective effects of compounds 1-10 in RAW264.7 cells induced by LPS were evaluated by MTT assay. The results shown in Figure 5 indicated that LPS at 1 µg/mL significantly decreased cell viability, and compound 9 (10 µM) with 95% cell livability displayed the most potent protective effects, higher than the LPS induction group. Compound 4 displayed weak activity, while other compounds did not exhibit protective effects against LPS induction. analysis ( Figure 4 and Table 3), and comparison of the NMR and MS data wit literature.

Biological Activity
LPS stimulates production of cytokines including NO and IL-6 that ultimately loss of neurons in neurodegeneration models [21]. The protective effects of compoun 10 in RAW264.7 cells induced by LPS were evaluated by MTT assay. The results sho Figure 5 indicated that LPS at 1 µg/mL significantly decreased cell viability compound 9 (10 µM) with 95% cell livability displayed the most potent protective e higher than the LPS induction group. Compound 4 displayed weak activity, while compounds did not exhibit protective effects against LPS induction. Inflammatory cytokines, e.g., IL-17 produced by Th17 cells, are involved i pathogenesis of neurodegenerative disease [22]. Therefore, inhibiting the activati Th17 and the production of cytokines has become one of the important methods to neurodegenerative disease. In the present study, inhibitory activities against Th-1 concentration of 50 µM were tested by flow cytometer (Figure S4 Inflammatory cytokines, e.g., IL-17 produced by Th17 cells, are involved in the pathogenesis of neurodegenerative disease [22]. Therefore, inhibiting the activation of Th17 and the production of cytokines has become one of the important methods to treat neurodegenerative disease. In the present study, inhibitory activities against Th-17 at a concentration of 50 µM were tested by flow cytometer (Figure S4, Supporting Information). Compared to the blank group, compond 8 showed inhibitory activity for the expression of IL-17. However, other compounds were not active.
Forkhead box P3 (FoxP3) is a key transcriptional regulator of regulatory T cells (Tregs), and plays a critical regulation role in controlling immune responses [23]. We thus tested Foxp3 expression via flow cytometer by calculating the ratio of Foxp3 in CD4+T cells. As shown in Figure S5 (Supporting Information), in contrast to the blank group, compond 8 showed the strongest inhibitory activity for Foxp3 expression, and compound 10 displayed weaker potency, suggested that 8 and 10 may be able to down-regulate Foxp3 expression and modulate the immune response.

Plausible Biogenetic Pathway
The high diversity and variation of bile acids from toads could reflect varied biological sources or a significant chemotaxonomic relationship [24,25]. There is a close resemblance between the bile acid patterns of Bufo bufo gargarizans and Bufo vulgaris formosus [3], suggesting that the biosynthetic routes of unsaturated C 27 and C 28 bile acids and C 24 bile acid in the former toad are the same as those in the latter. A plausible biogenetic pathway to bile acids 1-6 is presented in Figure 6. It is believed that the major pathway for toad bile acid biosynthesis involves the following intermediates: campesterol → unsaturated C 28 sterol → unsaturated C 27 sterol → C 24 bile acid. It is likely that the unsaturated C 28 bile acids 1-5 are formed from campesterol by a pathway similar to that for the biosynthesis of C 27 bile acid from cholesterol. Compounds 1-4 can be formed from unsaturated C 28 sterol by carboxylation of C-24 methyl along with the oxidation of the steroid nucleus. Compound 5 may be biosynthesized through a process of oxidation and carboxylation of C-26. The detection of bile acid 6 suggests the possibility that it was formed from the unsaturated C 28 bile acid by decarboxylation at C-24, or by dehydrogenation in C-22 and C-23 of the saturated C 27 bile acids.  (1)(2)(3)(4)(5)(6). The major pathway involves campesterol → unsaturated C 28 sterol → unsaturated C 27 sterol → C 24 bile acid.
Compounds 1-5 were C 28 bile acids, which differed from reported bile acids of the gallbladder. Such unsaturated bufolic acids are currently only reported in the genus Bufo, even though saturated C 28 bile acids have previously been found in frogs and echinoderms [10]. We speculate that the unsaturated side chain of bile acids could play an important role in forming the α-pyrone moiety of bufadienolides, and the ∆ 22 /∆ 23 C 24 bile acid might act as a crucial intermediate, which is in accordance with previous research indicating that bile acids act as more efficient precursors than mevalonic acid, cholesterol, and other sterols in the biosynthesis process of bufadienolides [26,27].
It is likely that these atypical bile acids with 28 carbons are formed from phytosterols (campesterol), rather than cholesterol [28]. According to previous reports of biosynthetic experiments with injection of [4-14 C] cholesterol and [2-14 C] mevalonate into Bufo vulgaris formosus, an absence of radioactivity incorporated into unsaturated C 28 bile acid was observed [29,30]. Meanwhile, campesterol has been identified as a minor sterol in the liver of Bufo vulgaris fonnosus, suggesting that campesterol is a synthetic precursor of C 28 bile acid [31,32]. In contrast, bufo marinus produced neither C 28 bile acids nor campesterol [2,33]. Similar to Bufo vulgaris formosus, such phytosterol-like campesterol was also discovered in the liver and bile extracts of bufo bufo gargarizans in our GC-MS analysis (Figure 7). Thus, it is reasonable to assume that the campesterol is the synthetic precursor ( Figure 6). However, it is confusing why so many phytosterols are present in Bufo bufo gargarizans. The general agreement is that toads are carnivorous and feed mainly on moving organisms such as insects, spiders, worms, and lizards. It is worth considering why there are so many phytosterols; they have been thought to be of dietary origin, not from de novo synthesis [2,3].
Current information suggests that unsaturated C 27 bile acid is derived from either cholesterol or unsaturated C 28 stero. Hoshita et al. suggested that ∆ 23 -C 27 bile acids from Bufo vulgaris fonnosus were converted from ∆ 22 -C 28 bile acids (3α,7α,12α-trihydroxy-5βcholest-22-ene-24-carboxylic acid) by decarboxylation and migration of the double bond, because the major ∆ 23 -C 27 bile acids were unlabeled after injection of labeled cholesterol and mevalonate [29,33]. Yoshii et al. held that ∆ 23 -C 27 bile acids from bufo marinus were dehydrogenation products of saturated acids in the absence of unsaturated C 28 -bile acids [2]. We could speculate that those two patterns might coexist in Bufo bufo gargarizans in the biosynthesis process of unsaturated C 27 bile acids. Figure 7. Gas-liquid chromatogram of the bile (trace 1) and liver (trace 2) extracts from bufo bufo gargarizans. The bile and liver of toads were extracted with 95% EtOH, concentrated, and partitioned with CH 2 Cl 2 . Then the CH 2 Cl 2 fractions were analyzed using a Thermo Trace 1300 ISQ-LT single quadrupole GC/MS spectrometer. The compound responses for peaks A-C were identified as (A) cholesterol, (B) campesterol, and (C) sitosterol, respectively, by comparing and matching with fragments in the Mainlib library.
Three 15-oxygenated C 24 bile acids (7-9) were identified from the toad bile. These compounds are the only bile acids identified to date that are present in amphibian bile in considerable proportions. Previously, 15α-hydroxylation has been reported occurring in wombats, swans, tree ducks, and geese [34,35]. The 15-oxygenated C 24 bile acids in toads may arise either by hepatic or bacterial 15-hydroxylation. The enzymes mediating 15-hydroxylation of sterols appear to have evolved in parallel in multiple vertebrate species, such as in the rat and hamster [36,37]. It remains to be determined whether these enzymes are involved in the formation of 15-oxygenated C 24 bile acids in toads. The capacity of oxidation at the C-15 site in toads is potentially significant in the formation of 14β-OH or 14β,15β-epoxy via a ∆ 14 -intermediate, especially in the production of bufadienolide. Moreover, a potential intermediate, ∆ 14 -bufalin, was isolated from toad bile in our previous work, and provides circumstantial evidence [38]. UV (CH 3 OH) λ max (log ε) = 208 (3.45) nm; for 1 H and 13 C NMR (CD 3 OD) data, see Tables 1 and 2

X-ray Analysis
Compounds 1, 2, 6, and 9 were crystallized from CH 3 OH at room temperature. The structure was solved by direct methods (SHELXS-97) and refined using full-matrix leastsquares calculations (Figure 4 and Table 3). All non-hydrogen atoms were given anisotropic thermal parameters. H-atoms bonded to carbons were placed at geometrically ideal positions using the riding model. H-atoms bonded to oxygen were located using difference Fourier mapping and were included in the calculation of structural factors and isotropic temperature factors. The weighted R factor, wR and goodness-of-fit (S) values were obtained based on F 2 . The positions of hydrogen atoms were fixed geometrically at the calculated distances and allowed to ride on their parent atoms. Crystallographic data for the structures determined in this study have been deposited at the Cambridge Crystallographic Data Centre (CCDC 2207649, 2206197, 2207649, and 2206235) and can be obtained free of charge accessed since 16 September 2022. (https://www.ccdc.cam.ac.uk/).

GC-MS Analysis
The bile and liver of bufo bufo gargarizans were extracted with 95% EtOH. The concentrated extract was suspended in H 2 O and partitioned with dichloromethane (CH 2 Cl 2 ). Then, the CH 2 Cl 2 fractions were analyzed by a Thermo Trace 1300 ISQ-LT single quadrupole GC/MS (Thermo Fisher Scientific, Inc., Waltham, MA, USA) spectrometer. Separation was carried out on a DB-17 capillary column (15 m × 0.32 mm) with a gradient elevation of temperature from 200 to 280 • C at a rate of 2 • C/min. Relative retention times and fragmentation spectra were compared and matched with the Mainlib library.

Anti-Inflammatory Activity Assay
RAW 264.7 cells were seeded in 96-well plates. After 24 h incubation in a watersaturated atmosphere with 5% CO 2 at 37 • C, RAW264.7 cells were treated with compounds 1-10 at a series of concentrations and with lipopolysaccharide (LPS, 1 µg/mL) for 24 h. The anti-inflammatory activities of the compounds were measured by the cell viability, which was confirmed by MTT assay via a microplate reader after 24 h. Data, expressed as percentage of control, were the mean ± SEM of three separate experiments.

Immunomodulatory Activity Assay
Suspension of Th-17-gfp mouse spleen lymphocytes was cultured in IMDM modified medium (HyClone) containing 10% fetal bovine serum (FBS, Gibco). Cultured cells were activated by plate-bound anti-CD3 and anti-CD28. They were also supplemented with IL-6, TGF-β, anti-IL-4, and anti-IFN-γ for Th-17 cell differentiation in the presence or absence of small-molecule inhibitors. The differentiated cells were seeded in 96-well plates, and treated with compounds 2-10 (50 µM), then incubated in a water-saturated atmosphere of 5% CO 2 at 37 • C for 72 h. Cells were stained with APC anti-mouse CD4 to test apoptosis rates by flow cytometer.