Bupleurum chinense Roots: a Bioactivity-Guided Approach toward Saponin-Type NF-κB Inhibitors

Abstract

The roots of Bupleurum chinense have a long history in traditional medicine to treat infectious diseases and inflammatory disorders. Two major compounds, saikosaponins A and D, were reported to exert potent anti-inflammatory activity by inhibiting NF-κB. In the present study, we isolated new saikosaponin analogues from the roots of B. chinese interfering with NF-κB activity in vitro. The methanol-soluble fraction of the dichloromethane extract of Radix Bupleuri was subjected to activity-guided isolation yielding 18 compounds, including triterpenoids and polyacetylenes. Their structures were determined by spectroscopic methods as saikogenin D (1), prosaikogenin D (2), saikosaponins B₂ (3), W (4), B₁ (5), Y (6), D (7), A (8), E (9), B₄ (10), B₃ (11), and T (12), saikodiyne A (13), D (14), E (15) and F (16), falcarindiol (17), and 1-linoleoyl-sn-glycero-3-phosphorylcholine (18). Among them, 4, 15, and 16 are new compounds, whereas 6, previously described as a semi-synthetic compound, is isolated from a natural source for the first time, and 13–17 are the first reports of polyacetylenes from this plant. Nine saponins/triterpenoids were tested for inhibition of NF-κB signaling in a cell-based NF-κB-dependent luciferase reporter gene model in vitro. Five of them (1, 2, 4, 6, and 8) showed strong (> 50%, at 30 µM) NF-κB inhibition, but also varying degrees of cytotoxicity, with compounds 1 and 4 (showing no significant cytotoxicity) presenting IC₅₀ values of 14.0 µM and 14.1 µM in the cell-based assay, respectively.

Key words

Bupleurum chinense - Apiaceae - NF-κB - saponin - polyacetylene - lysophosphatidylcholine

Introduction

The transcription factor NF-κB plays a key role in the control of genes involved in the innate and adaptive immune response. Its aberrant regulation is implicated in the pathogenesis of several disease states, including inflammation and autoimmune disorders [1]. NF-κB signaling is involved in cancer development and progression [2]. Furthermore, it is responsible for certain kinds of chemoresistance [3]. In consistence with its role as a central mediator in inflammatory responses, NF-κB is implicated in the aging process [4] and the development of various metabolic diseases [5], with beneficial effects observed upon its inhibition [6]. Therefore, the NF-κB signaling pathway appears as a promising drug target [7], which is also susceptible to the bioactivity of numerous small molecules from plant origin [8], [9], [10], [11].

The roots of Bupleurum chinense DC (Apiaceae) have a long history in traditional medicine and have been used in China for the treatment of inflammatory disorders and infectious diseases [12], [13]. The main constituents of this plant are triterpenoid saponins (“saikosaponins”), which were reported to have in vitro and in vivo anti-inflammatory, immunomodulatory [14], and hepatoprotective [15] activities. One of the major constituents, saikosaponin D, and its epimer, saikosaponin A, were reported to inhibit NF-κB activation [16].

However, there are no further studies that address the NF-κB inhibitory activity of additional saikosaponin analogues from this plant. We report here the isolation of 12 saikosaponins/saikogenins, of which nine have been tested for their potential to inhibit NF-κB transactivation in a cell-based in vitro luciferase reporter gene assay.

Results and Discussion

The dichloromethane (DCM) extract of B. chinense roots inhibited NF-κB-driven gene expression in a cell-based in vitro luciferase reporter assay at a concentration of 10 µg/mL by 83%. The MeOH soluble part of the DCM extract was separated by column chromatography (CC) on MCI^® gel. Aliquots of the collected fractions were combined based on the TLC profiles into 10 pooled fractions that were tested for NF-κB inhibition. Activity was detected in pooled fractions 5 – 8, with the highest activity in pooled fraction 6 (91% inhibition, 10 µg/mL), and a moderate activity in pooled fractions 5 (47% inhibition at 10 µg/mL) and 7 (52% inhibition at 10 µg/mL) (Fig. 1S, Supporting Information). From MCI fractions belonging to the pooled fractions 4, 5 and 6, 18 compounds were isolated ([Fig. 1]), including one sapogenin (1), saponins (2–12), polyines (13–17), and a phospholipid (18). Their structures were determined by 1D and 2D NMR (COSY, HSQC, HMBC) and mass spectrometry (HR-ESI-MS), and by comparison of the spectral data with literature values. Data regarding the inhibition of the NF-κB transactivation activity by pooled fractions 1 – 10 are shown in Fig. 1S of the supporting information. Also spectroscopic and spectrometric data for the identification of known compounds (1–3, 5, 7–14, and 17–18) are provided in the supporting information.

Fig. 1 Structures of compounds 1–18 isolated from B. chinense.

Compound 6 was obtained as a white amorphous powder. Its HR-ESI-MS displayed a quasimolecular ion peak at 779.4567 [M + H]⁺, which, together with ¹³C NMR data provided a molecular formula of C₄₂H₆₆O₁₃. The ¹H NMR spectrum showed the presence of a cis-olefinic group, two anomeric protons, and six tertiary methyl groups. Through interpretation of the 2D NMR data, all proton and carbon resonance were assigned. HMBC correlation was observed between one of the olefinic protons (δ 5.94, H-11) and a tertiary carbon (δ 85.6, C-13), which in turn showed HMBC correlation with a methylene group (δ 76.3, CH₂-28), suggesting an ether linkage between them. This indicated a triterpenoid aglycone similar to that of saikosaponins A (8) or D (7) [17]. The NMR resonances belonging to the rings A and B of compound 6 were almost superimposable to that of compounds 7 and 8. The major difference was the existence in compound 6 of a ketone group (δ 214.8, C-16) instead of a hydroxylated methine. This was supported by the HMBC correlations of this ketone carbon with H-15, H-22, and H-28, as well as lower field shifts for C-14 (Δδ _C − 6.6 Hz), C-15 (Δδ _C − 9.9 Hz), C-17 (Δδ _C − 11.3 Hz), and C-18 (Δδ _C − 4.2 Hz). Therefore, the aglycone part of compound 6 was determined as 13,28-epoxy-3β,23-dihydroxy-olean-11-en-16-one [18]. The remaining 12 carbon signals, together with two anomeric protons at δ 4.53 (d, J = 7.7 Hz) and 4.38 (d, J = 7.7 Hz), suggested the existence of two sugar units. Further analysis of the resonance data revealed a β-glucopyranose and a β-fucopyranose. As only D-configuration of both sugars have been reported in saikosaponins, we tentatively assigned the D-configuration based on biosynthetic considerations. HMBC correlations between signals at δ 4.53 (H-1″) and δ 85.2 (C-3′) suggested the connection of β-glucopyranose to C-3′ of the β-fucopyranose moiety. The connection of the sugar chain to C-3 of the aglycone was deduced from the HMBC correlation between signals at δ 4.38 (H-1′) and δ 83.0 (C-3). The sugar sequence was supported by fragment ions at m/z 617 [M+H-Glc]⁺, 471 [M+H-Glc-Fuc]⁺, and 453 [M+H-Glc-Fuc-H₂O]⁺. Therefore, the structure of compound 6 is established as 13,28-epoxy-23-hydroxy-olean-11-en-16-one-3β-yl O-β-D-glucopyranosyl(1 → 3)-β-D-fucopyranoside. Compound 6 was once reported as semi-synthetic derivative from saikosaponin D, but never as a natural product [18]. It is now assigned the trivial name saikosaponin Y.

Compound 4 was obtained as a white amorphous powder. A molecular formula of C₄₂H₆₈O₁₂ was derived from its HR-ESI-MS and ¹³C NMR data. The UV spectrum showed strong absorption maxima at 245 (sh), 251, and 260 (sh) nm, indicating the existence of an heteroannular diene [19]. This was further supported by the ¹H NMR signals of cis-olefinic protons at δ 6.45 and 5.58, and ¹³C NMR signals at δ 137.2, 132.7, 127.1, and 126.7. Through interpretation of the 2D NMR data, all proton and carbon resonances were assigned. The ¹³C NMR data of the aglycone part were nearly superimposable to that of saikosaponin B₂ (3) [17], except for the upper field shift for C-4 (Δδ _C − 3.7 ppm), and downfield shifts for C-3 (Δδ _C 7.2 ppm), C-5 (Δδ _C 8.5 ppm), C-24 (Δδ _C 3.7 ppm), and C-23 (Δδ _C 36.9 ppm), which revealed the absence of a 23-hydroxyl group in compound 4 [17]. At the same time, upper field shifts were observed in the ¹H NMR for H-3 (Δδ _H − 0.46 ppm) and H-5 (Δδ _H − 0.42 ppm), owing to the releasing of steric hindrance from 23-hydroxyl group. When compared with saikosaponin E (9), these atoms, together with other atoms belonging to the ring A of the pentacyclic skeleton of compound 4, shared nearly identical chemical shifts with their counterpart in compound 9 [17]. Therefore, the aglycone of compound 4 was determined to be 16-epi-saikogenin C [19]. The remaining 12 carbon signals, together with two anomeric protons at δ 4.55 (d, J = 7.7 Hz) and 4.32 (d, J = 7.7 Hz), could be readily assigned to the sugar moiety. As in the case of compound 6, through further analysis of the NMR data, the existence of a β-glucopyranose and a β-fucopyranose was determined, and the configuration of both of them was also tentatively assigned as D for biogenetic reasons. Their sequence was elucidated by HMBC correlations at δ 4.55 (H-1″)/δ 84.9 (C-3′) and δ 3.10 (H-3′)/δ 105.6 (C-1″), which suggest the connection of β-glucopyranose to C-3′ of the β-fucopyranose moiety. HMBC correlations at δ 3.62 (H-3)/δ 105.6 (C-1′) and δ 4.32 (H-1′)/δ 83.1 (C-3) supported the connection of the sugar chain to C-3 of the aglycone. This linking pattern was supported by mass fragment peaks at m/z 747 [M+H-H₂O]⁺, 585 [M+H-H₂O-Glc]⁺, and 439 [M+H-H₂O-Glc-Fuc]⁺. Therefore, the structure of compound 4 is established as 16α,28-dihydroxyoleana-11,13(18)-diene-3β-yl O-β-D-glucopyranosyl-(1 → 3)-β-D-fucopyranoside. To our knowledge, it is reported here for the first time and is given the trivial name saikosaponin W.

The remaining triterpenoids were identified by comparison of their spectral data with literature values as saikogenin D (1) [17], prosaikogenin D (2) [20], saikosaponins B₂ (3) [17], B₁ (5) [17], D (7) [17], A (8) [17], E (9) [17], B₄ (10) [17], B₃ (11) [17], and T (12) [21].

Nine of these compounds (1, 2, 3, 4, 6, 8, 10, 11, and 12) present in bioactive pooled fractions 5 and 6 were tested at 30 µM for their NF-κB inhibitory activity in vitro ([Fig. 2 a]). At 30 µM, compound 6 markedly reduced fluorescence of the vital dye (cell tracker green [CTG]), indicating cytotoxicity and therefore a false positive result of NF-κB inhibition in the cell-based luciferase assay. Compound 8 (saikosaponin A) showed a strong decrease of NF-κB-dependent luciferase gene expression in vitro. However, it also showed a significant reduction of CTG fluorescence at 30 µM, indicating cytotoxicity of this compound. Compounds 10–12 showed no NF-κB inhibition and did not show any cytotoxicity. Interestingly, compounds 1–4 showed NF-κB inhibitory activities, but no or low cytotoxicity ([Fig. 2]). For compounds 1 and 4, which show strong NF-κB inhibition at 30 µM, IC₅₀ values were determined. Compound 1 possesses an IC₅₀ value of 14.0 µM and compound 4 an IC₅₀ value of 14.4 µM in the NF-κB-dependent luciferase reporter gene assay in vitro ([Fig. 3]).

Fig. 2  a NF-κB inhibition by selected compounds isolated from B. chinense detected in an in vitro cell-based luciferase assay. Luciferase activity (relative units to the internal control CTG fluorescence) are presented as mean ± SD (n = 3, Dunnettʼs test, *p < 0.05, ***p < 0.001, ns: not significant) and are normalized to the solvent control DMSO (0.1%). b Cells were stained with CTG 1 h before treatment of cells with the indicated compounds for 4 h. Relative fluorescence units were measured directly before measuring the NF-κB-dependent luciferase activity and are displayed here normalized to the solvent control DMSO (0.1%). Data are expressed as mean ± SD (n = 3, unpaired t-test, *p < 0.05, ***p < 0.001, ns: not significant).

Fig. 3 Concentration-response curves of NF-κB inhibition by B. chinese compound 1 and compound 4.

Some preliminary structure-activity relationships can be deduced regarding the aglycone moiety. Both of the highly cytotoxic saponins (6 and 8, type “b” and “c” [Fig. 1]) possess an ether bridge between C-13 and C-28 of the aglycone part. The three NF-κB inactive saponins (10–12, type “d” [Fig. 1]), which are artifacts from the isolation process [22], have a methoxy group at C-11 in common. The NF-κB active and less cytotoxic compounds (1–4, type “a” [Fig. 1]) possess a heteroannular 11,13(18)-diene system [22].

Compounds 1–3 showed a mild decrease in CTG fluorescence, indicating low cytotoxicity at 30 µM (without reaching significance for 1). Their activities in terms of inhibiting NF-κB transactivation seem to correlate inversely with the number of sugars attached at C-3 of the aglycone. The aglycone saikogenin D (1) was the strongest inhibitor (91% inhibition) with no significant cytotoxicity. Compound 2, identified as prosaikogenin D, with one sugar (fucose) attached, was less active (75% inhibition). Compound 3, identified as saikosaponin B2, and possessing one more sugar (glucose), was even less active (20% inhibition). Compound 4 (saikosaponin W), the C-23-deoxyanalogue of compound 3 (saikosaponin B2), was one of the most active, decreasing NF-κB activation by 76.9 ± 1.4%, with no indication of cytotoxicity at 30 µM. These data suggest, in the case of type “a” saponin/triterpenoid, that NF-κB inhibition is favorably influenced by a shorter (or no) sugar chain at aglycone C-3 position, and no C-23 hydroxyl group.

Compound 8, which is one of the two major saponins of Radix Bupleuri, can be transformed into its prosapogenin and genuine sapogenin after oral administration. Its incubation with intestinal flora led also mainly to genuine sapogenin. But when incubated in gastric juice, type “a” saponins turned out to be the major products [23]. Actually, Chinese pharmacopoeia includes vinegar processing (being fried in vinegar) of Radix Bupleuri. Due to the heated acidic condition, this process can lead to a sharp increase of the content of type “a” and concomitant decrease of type “c” saponins [24]. The NF-κB inhibition observed for compounds 1, 2, and 4 (type “a”) lend support to this processing method.

Besides one saponin (6), fractions belonging to pooled fraction 6 also contained a small amount of compound 18, obtained as a colorless gum, whose strong affinity to both normal and reversed stationary phases indicated amphiphilic properties. Its 1D and 2D NMR spectra suggested a 1,3-substituted glycerol, an esterified linoleic acid, and a choline moiety [25]. The remaining phosphate unit was identified by positive ESI-MS/MS, which showed a quasi-molecular ion (m/z 520) and a specific fragment at m/z 184 characteristic of a protonated phosphorylcholine head group [26]. Therefore, compound 18 was identified as a lysophosphatidylcholine (Lyso-PC) [25]. This is the first time that a lysophosphatidylcholine is isolated from the genus Bupleurum. Compound 18 showed moderate NF-κB inhibition (44%) at 30 µM.

Finally, four C15 polyacetylenes (13–16), and one C17 polyacetylene (17) were isolated in very small amounts. Their structures were determined through interpretation of NMR and MS spectra as well as by comparison with reference compounds. This is the first time polyacetylenes were isolated from B. chinense.

Compounds 13–16 were isolated as amorphous solids. Compound 13 was identified as saikodiyne A [27]. Present ¹³C NMR experiments showed a different shift for C-4 than previously reported. The corrected value is outlined in the supporting information. Compound 14 was identified as a reduced derivative of saikodiyne B [27]. Compound 17 was identified as falcarindiol, a C17 polyacetylene [28].

Compounds 15 and 16 are both reported for the first time. Compound 15 was obtained as a white amorphous solid. The HR-ESI-MS, ¹³C NMR, and HSQC data indicated a molecular formula of C₁₅H₂₂O₂. As in the case of compound 13 and 14, ¹³C NMR spectrum of 15 suggested the presence of four acetylene carbon signals, at δ 84.2, 75.3, 66.2, and 64.1 ppm. Comparison of ¹³C NMR spectra of 15 with those of 13 and 14 suggested the existence of a five-membered aliphatic chain, whose connection to the alkene C-10 was supported by the HMBC correlations at δ 4.08 (H-10)/δ 37.9 (C-11), δ 1.48 (H-11)/δ 72.6 (C-10), δ 1.48 (H-11)/δ 150.7 (C-9). The missing of one of the two pairs of alkene protons in 15 at lower field coincided with the appearing of two vicinal methylene signals at δ 2.42, δ 1.73 (CH₂-2 and CH₂-3). At the same time, up-field shifts were observed for their vicinal hydroxymethylene protons (HO-CH₂-). Taking the solvent effect into consideration, the NMR data of 15 are highly superimposable to that of virol C, except for the signals of the longer aliphatic chain in virol C [29], [30]. By comparing their optical rotation, the configuration of 15 was determined as 10S [29]. Therefore, the structure of compound 15 was elucidated as (10S,8E)-pentadeca-8-en-4,6-diyne-1,10-diol. We assigned the trivial name saikodiyne E to compound 15.

Compound 16 was obtained as a white amorphous solid. The HR-ESI-MS, ¹³C NMR, and HSQC data indicated a molecular formula of C₁₅H₂₂O₂. The ¹³C NMR spectrum of 16 suggested the presence of four acetylene carbon signals, at δ 85.1, 79.4, 75.1, and 69.0 ppm. Like compounds 13 and 14, compound 16 possessed two olefinic bonds. One of the major differences revealed by the HSQC spectrum was the presence of a lower field hydroxymethine proton (δ 5.19, d, J = 8.3 Hz, H-8) in 16, while in both 13 and 14, a higher field hydroxymethine proton (δ 4.21, p, J = 5.5 Hz, H-10) was observed. The splitting pattern of the hydroxymethine proton at δ 5.19, together with the HMBC correlations at δ 5.47 (H-9)/δ 28.5 (C-11), δ 2.13 (H-11)/δ 134.1 (C-10), δ 2.13 (H-11)/δ 129.8 (C-9) revealed the shifting of the double bond from Δ ⁸ in 13 and 14 to Δ ⁹ in 16. The NMR data of 16 are highly superimposable to that of another C17 analogue, (2Z,8S,9Z)-heptadeca-2,9-dien-4,6-diyne-1,8-diol [31]. By comparison of their optical rotations, the configuration of 16 was determined as 8S [31]. Therefore, the structure of compound 16 was elucidated as (2Z,8S,9Z)-pentadeca-2,9-dien-4,6-diyne-1,8-diol. Compound 16 was named saikodiyne F.

Due to their limited amounts, compounds 13–17 have not been tested for their bioactivity.

In summary, this study examined nine triterpenoids (1–4, 6, 8, and 10–12) for their potential to inhibit NF-κB transactivation activity in vitro in a cell-based NF-κB-driven luciferase reporter gene model in HEK293 cells. Six of these compounds showed NF-κB inhibition. However, four of them appeared also to be cytotoxic, especially the type “b” and “c” saponins 6 and 8. Compound 4, a new saponin belonging to type “a” presented the best NF-κB inhibitory activity in relation to its cytotoxicity. Several polyacetylenes, including two new congeners, have been isolated in trace amounts for the first time from B. chinense. These findings enhance the understanding of this plant as an anti-inflammatory herbal medicine.

Materials and Methods

General experimental procedures

All solvents were obtained from VWR Chemicals and Carl Roth. TLC was performed on Silica gel 60_F254 and Silica gel 60 RP-18_F254s plates (Merck): detection with a mixture of 1% vanillin and 10% sulfuric acid in dehydrated ethanol. Open CC was carried out with MCI CH-P 20P resin (Mitsubishi Chemical), RP-18 silica gel (25 – 40 µm, Fuji silica), Sephadex LH-20 (GE Healthcare), and silica gel (15 – 40 µm, Merck) as stationary phases. Semi-preparative HPLC experiments were performed with L-6200A intelligent pump (at a flow rate of 3 mL/min) and L-4500 DAD detector (Merck-Hitachi) equipped with LicroCART 10 × 250 mm column packed with LiChrospher 100 RP-18 (particle size: 10 µm). Analytical HPLC was conducted on an Agilent 1100 system using a Zorbax SB-C18 Narrow Bore (3.5 µm) 2.1 × 150 mm column (Agilent). LC-ESI-MS analyses were carried out using the same column on a Dionex Ultimate 3000 UHPLC coupled with a Thermo LTQ XL linear ion trap mass spectrometer equipped with an H-ESI II probe in the positive and negative mode. Accurate masses were measured on a LC/FTMS system consisting of an Exactive Orbitrap mass spectrometer (ThermoFisher) coupled to a ThermoFisher Accela U-HPLC system. ¹H, ¹³C, and 2D NMR spectra (COSY, HSQC, and HMBC) were recorded in CD₃OD at 25 °C on a Unity 600 (Varian), and an Avance 700 spectrometer (Bruker). Chemical shifts are expressed in δ (ppm) with the solvent peak used as reference.

Plant material

Roots of B. chinense were obtained from two sources. Material-I, used for screening, was purchased from Plantasia (Batch No. 100093; Oberndorf, Austria). Material-II, used for preparative isolation, was obtained from the Bavarian State Research Center for Agriculture (LfL-Bayern, Freising-Weihenstephan). Voucher specimens (No.: IPW_Bup-chin_2008 and IPW_Bup-chin_2010) are kept at the Institute of Pharmaceutical Sciences, Department of Pharmacognosy, University of Graz.

Extraction and isolation

For screening, 10 g of Material-I were pulverized and extracted by accelerated solvent extraction with DCM and MeOH: preheat 1 min, static 5 min, flush 150%, purge 60 sec, cycles 3, pressure 68.9 bar, temperature 44 °C for DCM and 68 °C for MeOH.

For large-scale isolation, 900 g of Material-II was pulverized and subsequently percolated with DCM (10 L). After evaporation to dryness (< 40 °C), the DCM extract (36 g) was then partitioned twice between n-hexane and 90% MeOH (1.5 : 1, v/v), yielding 24.7 g of a defatted, MeOH soluble part, which was subsequently fractionated by CC (Ø: 5.5 cm, L: 55 cm), using MCI CHP-20P resin as the stationary phase and an MeOH-water gradient (40% to 100%, v/v) as the mobile phase, to afford 101 fractions (Fr. 1 to Fr. 101, 700 mL each). After the fractions had been evaporated to dryness, they were transferred into small vials, and reconstituted to 10 mL. For bioassay, 30 µL were taken from each vial, and combined according to their TLC profiles to form 10 pooled fractions, tested for NF-κB inhibition.

MCI fractions belonging to the NF-κB inhibitory pooled fractions 5 and 6 were subjected to Sephadex LH-20 (35% MeOH in water) chromatography, in order to separate polyacetylenes from dominant saponins. The saponin containing subfractions were then subjected to RP-18 (Ø: 4 cm, L: 45 cm; eluted with 60 – 70% MeOH water solution, v/v) and silica gel (Ø: 3 cm, L: 30 – 50 cm; eluted with n-hexane-ethylacetate, 3 : 1 or 2.2 : 1, v/v) open CC, and finally semi-preparative RP-18 HPLC (35%, 40%, 45%, or 50% MeCN) to afford a sapogenin and eleven saponins, respectively: 1 (2.4 mg) [17], 2 (2.5 mg) [20], 3 (1.5 mg) [17], 4 (2.4 mg), 5 (0.6 mg) [17], 6 (1.3 mg) [18], 7 (5.7 mg) [17], 8 (4.7 mg) [17], 9 (4.4 mg) [17], 10 (20.7 mg) [17], 11 (6.2 mg) [17], and 12 (8.5 mg) [21]. Fractions belonging to pooled fraction 4, which also contained polyacetylenes, were processed in the same way. The polyacetylene enriched subfractions were purified directly by semi-preparative HPLC (isocratic 50% or 69% MeCN) to afford five polyacetylenes, respectively: 13 (0.8 mg) [27], 14 (0.8 mg) [27], 15 (0.9 mg), 16 (0.5 mg). and 17 (0.9 mg) [32]. The isolation of 18 (2.4 mg) was achieved through RP-18 SPE (80% MeOH) and semi-preparative RP-18 HPLC (60% MeCN) [25]. Among them, saponins 1–5 and 7–12 were isolated from the combined fractions 37, 38, 39, and 40 (460 mg) belonging to the fifth pooled fraction; polyacetylenes 13–15 were isolated from the combined fractions 33, 34, 35, and 36 (490 mg) belonging to the fourth and fifth pooled fractions; polyacetylene 16 was isolated from the combined fractions 31 and 32 (663 mg) belonging to the fourth pooled fraction; polyacetylene 17 was isolated from the combined fractions 41, 42, 43, and 44 (370 mg) belonging to the fifth pooled fraction; saponin 6 and lysophosphatidylcholine 18 were isolated from the combined fractions 47 and 48 (210 mg) belonging to the sixth pooled fraction. According to HPLC profiles and NMR spectra, all compounds tested for bioactivity are of at least 95% purity.

16α,28-dihydroxyoleana-11,13(18)-diene-3β-yl O-β-D-glucopyranosyl-(1 → 3)-β-D-fucopyranoside (4 , saikosaponin W): white amorphous powder; [α]_D ²⁰ − 29.0 (c 0.03, MeOH); UV (MeCN/H₂O) λ _max: 245 (sh), 251, and 260 (sh) nm; ¹H and ¹³C NMR (150 MHz, CD₃OD) data, see [Table 1]; HR-ESI-MS m/z 763.4641 [M – H]⁻ (calcd for C₄₂H₆₈O₁₂: 763.4633).

Table 1  ¹H and ¹³C NMR spectroscopic data (in CD₃OD) of compounds 4 and 6 ^a, J in Hz.

	4		6
C	δ H	δ C	δ H	δ C
a Measured at 600 MHz (for 1H) and 150 MHz (for 13C).
1	1.88 (1H) 1.02 (1H)	39.4	1.87 (1H) 0.94 (1H)	39.2
2	1.95 (1H) 1.75 (1H)	27.0	1.96 (1H) 1.80 (1H, d, J = 14.2)	26.1
3	3.18 (1H, dd, J = 11.7,4.5)	90.6	3.63 (H-3)	83.0
4		40.3		44.1
5	0.85 (1H)	56.7	1.21 (1H)	48.0
6	1.64 (1H) 1.46 (1H)	19.4	1.54 (2H)	18.1
7	1.38 (2H)	33.4	1.50 (1H) 1.14 (1H)	32.0
8		41.9		43.0
9	1.99 (1H, br s)	54.8	1.91 (1H)	53.8
10		37.5		37.1
11	5.58 (1H, dd, J = 10.7,1.2)	127.1	6.04 (1H, d, J = 10.4)	134.6
12	6.45 (1H, dd, J = 10.8,2.9)	126.7	5.53 (1H, dd, J = 10.3,3.1)	129.6
13		137.2		85.6
14		42.3		50.9
15	1.93 (1H) 1.41 (1H)	32.0	2.79 (1H, d, J = 14.5) 1.80 (1H, d, J = 14.2)	45.4
16	4.03 (1H, t, J = 3.2)	69.1		214.8
17		45.4		57.4
18		132.7	2.28 (1H, dd, J = 14.2,3.3)	56.3
19	2.48 (1H, d, J = 14.7) 1.69 (1H, d, J = 14.8)	39.4	1.55 (1H) 1.43 (1H)	40.0
20		33.2		32.4
21	1.56 (1H) 1.30 (1H)	35.8	1.51 (1H) 1.24 (1H)	36.6
22	2.01 (1H) 1.61 (1H)	24.5	2.11 (1H) 1.26 (1H)	25.1
23	1.06 (3H, s)	28.2	3.67 (1H) 3.29 (1H)	64.7
24	0.84 (3H, s)	16.5	0.71 (3H, s)	12.7
25	0.92 (3H, s)	18.6	0.98 (3H, s)	18.8
26	0.73 (3H, s)	17.5	1.21 (3H, s)	20.1
27	1.23 (3H, s)	22.1	1.02 (3H, s)	20.7
28	3.74 (1H, d, J = 11.7) 3.26 (1H)	65.1	3.90 (d, J = 7.9) 3.44 (d, J = 8.0)	76.3
29	0.85 (3H, s)	25.3	0.93 (3H, s)	33.8
30	0.97 (3H, s)	32.9	0.90 (3H, s)	23.5
Fuc
1′	4.32 (1H, d, J = 7.7)	106.6	4.38 (1H, d, J = 7.7)	105.7
2′	3.67 (1H)	71.9	3.65 (1H)	71.9
3′	3.60 (1H, dd, J = 9.7,3.3)	84.9	3.60 (1H)	85.2
4′	3.85 (1H)	72.4	3.85 (1H)	72.3
5′	3.64 (1H)	71.2	3.66 (1H)	71.3
6′	1.26 (3H, d, J = 6.4);	16.9	1.27 (3H, d, J = 6.4);	16.9
Glc
1″	4.55 (1H, d, J = 7.7)	105.6	4.53 (d, J = 7.7)	105.7
2″	3.28 (1H)	75.4	3.29 (1H)	75.4
3″	3.37 (1H)	77.7	3.36 (1H)	77.7
4″	3.32 (1H)	71.3	3.33 (1H)	71.2
5″	3.28 (1H)	77.9	3.28 (1H)	77.9
6″	3.84 (1H) 3.67 (1H)	62.5	3.84 (1H, dd, J = 12.0,2.0) 3.68 (1H)	62.4

13,28-epoxy-23-hydroxy-olean-11-en-16-one-3β-yl O-β-D-glucopyranosyl(1 → 3)-β-D-fucopyranoside (6 , saikosaponin Y): white amorphous powder; ¹H and ¹³C NMR (150 MHz, CD₃OD) data, see [Table 1]; HR-ESI-MS m/z 777.4437 [M – H]⁻ (calcd 777.4425 for C₄₂H₆₅O₁₃).

(10S,8E)-pentadeca-8-en-4,6-diyne-1,10-diol (15 , saikodiyne E): white amorphous solid; [α]_D ²⁰ + 19.1 (c 0.05, MeOH). UV (MeCN/H₂O) λ _max: 236, 264, 278, 295, 314 nm; ¹H and ¹³C NMR, see [Table 2]; HR-ESI-MS m/z 293.1759 [M + CH₃COO]⁻ (calcd 293.1753 for C₁₇H₂₅O₄).

Table 2  ¹H and ¹³C NMR spectroscopic data (in CD₃OD) of compounds 15 and 16 ^a, J in Hz.

	15		16
C	δ H	δ C	δ H	δ C
a Except for 13C NMR of 16 (measured at 175 MHz), all spectra were measured at 600 MHz (for 1H) and 150 MHz (for 13C).
1	3.63 (2H, t, J = 6.3)	61.4	4.30 (2H, dd, J = 6.4,1.6)	61.1
2	1.73 (2H, p, J = 6.7)	32.3	6.23 (1H, dt, J = 11.1,6.4)	147.8
3	2.42 (2H, t, J = 7.0)	16.6	5.63 (1H, d, J = 11.1)	109.1
4		84.2		75.1
5		66.2		79.4
6		75.3		69.0
7		74.1		85.1
8	5.71 (1H, d, J = 15.8)	109.0	5.19 (1H, d, J = 8.3)	59.0
9	6.24 (1H, dd, J = 15.9,5.8)	150.7	5.47 (1H, ddt, J = 10.1,8.3,1.6)	129.8
10	4.08 (1H, qd, J = 6.2,1.6)	72.6	5.56 (1H, dtd, J = 10.9,7.6,1.2)	134.1
11	1.49 (2H, q, J = 7.1)	37.9	2.13 (2H, q, J = 7.1)	28.5
12	1.41 (1H) 1.32 (1H)	26.2	1.41 (2H, p, J = 7.2)	30.1
13	1.30 (2H)	32.9	1.32 (2H)	32.5
14	1.33 (2H)	23.7	1.33 (2H)	23.6
15	0.91 (3H, t, J = 6.9)	14.4	0.91 (3H, t, J = 7.0)	14.4

(2Z,8S,9Z)-pentadeca-2,9-dien-4,6-diyne-1,8-diol (16 , saikodiyne F): white amorphous solid; [α]_D ²⁰ + 35.1 (c 0.03, MeOH). UV (MeCN/H₂O) λ_max : 216, 243, 256, 270, 286 nm; ¹H and ¹³C NMR, see [Table 2]; HR-ESI-MS m/z 291.1602 [M + CH₃COO]⁻ (calcd 291.1596 for C₁₇H₂₃O₄).

NF-κB transactivation activity

NF-κB activity was evaluated in HEK293/NF-κB-luc cells (HEK293 cells stably transfected with a NF-κB-responsive luciferase reporter gene Panomics, RC0014) as previously described [33], [34]. Cells were loaded with CTG CMFDA (C2925; Invitrogen), a fluorescent probe that is retained inside living cells [11], [35], and 4 × 10⁴ cells/well were seeded in 96-well plates. After 24-h incubation, cells were pretreated as indicated for 30 min and activated with TNF-α (2 ng/mL, Sigma; T-6676) for 4 h. Cells were then lysed with luciferase lysis buffer (Promega; E1531), the luminescence of the firefly luciferase and the fluorescence of stained cells with CTG (also: CMFDA, 5-chloromethylfluorescein diacetate) were quantified with a GeniosPro plate reader (Tecan). The luciferase signal resulting from the NF-κB reporter activation was normalized by the CTG-derived fluorescence to account for potential differences in cell number. The impact of cytotoxicity on cell numbers was evaluated by comparing fluorescence of the cells treated by the solvent vehicle with that of cells treated with the indicated samples. Parthenolide (Sigma-Aldrich; P0667 ≥ 98% HPLC), a known NF-κB inhibitor, was used as positive control (10 µM). Results were normalized to the solvent control DMSO (0.1%).

Statistical analysis and calculation of IC₅₀ values

Statistical analysis and nonlinear regression were performed using Prism software (version 4.03; GraphPad Software Inc). To calculate the IC₅₀ values, data were curve fitted and non-linear transformed using a sigmoidal dose response with variable slope. Results of bioassay data are expressed as the mean ± standard deviation (SD) of at least three independent experiments performed in quadruplicate. Results of IC₅₀ values are expressed with 95% confidence interval (CI). Dunnettʼs test was used for statistical analyses (*p < 0.05, **p < 0.01, ***p < 0.001, ns: not significant).

Conflict of Interest

The authors declare no conflict of interest.

Acknowledgements

The authors appreciate the professional assistance by Dr. Karin Ardjomand-Wölkart, Dr. Nadine Kretschmer, Dr. Marlene Monschein, Ms. Scarlet Hummelbrunner, Ms. Elvira Knauder, and Ms. Andrea Fleck of the University of Graz. The authors gratefully acknowledge the Funding provided by the Austrian Science Fund (FWF): S 10705, and S 10704 (NFN “Drugs from Nature Targeting Inflammation”) and P25971-B23 (“Improved cholesterol efflux by natural products”). We also gratefully acknowledge the financial support from University of Graz (IncentForsch: Pharm.Wiss. Prämie FWF). This paper is dedicated to Professor Dr. Max Wichtl in recognition of his outstanding contribution to pharmacognosy research.

^* These authors contributed equally to this work.

• References

• 1 Baldwin AS. The transcription factor NF-κB and human disease. J Clin Invest 2001; 107: 3-6
• 2 Karin M. Nuclear factor-κB in cancer development and progression. Nature 2006; 441: 431-436
• 3 Wang CY, Cusack JC, Liu R, Baldwin AS. Control of inducible chemoresistance: enhanced anti-tumor therapy through increased apoptosis by inhibition of NF-κB. Nat Med 1999; 5: 412-417
• 4 Salminen A, Hyttinen JMT, Kaarniranta K. AMP-activated protein kinase inhibits NF-κB signaling and inflammation: impact on healthspan and lifespan. J Mol Med 2011; 89: 667-676
• 5 Tornatore L, Thotakura AK, Bennett J, Moretti M, Franzoso G. The nuclear factor kappa B signaling pathway: integrating metabolism with inflammation. Trends Cell Biol 2012; 22: 557-566
• 6 Gareus R, Kotsaki E, Xanthoulea S, van der Made I, Gijbels MJJ, Kardakaris R, Polykratis A, Kollias G, de Winther MPJ, Pasparakis M. Endothelial cell-specific NF-κB inhibition protects mice from atherosclerosis. Cell Metab 2008; 8: 372-383
• 7 Gilmore TD, Garbati MR. Inhibition of NF-κB signaling as a strategy in disease therapy. Curr Top Microbiol Immunol 2011; 349: 245-263
• 8 Bremner P, Heinrich M. Natural products as targeted modulators of the nuclear factor-kappaB pathway. J Pharm Pharmacol 2002; 54: 453-472
• 9 Salminen A, Lehtonen M, Suuronen T, Kaarniranta K, Huuskonen J. Terpenoids: natural inhibitors of NF-kappaB signaling with anti-inflammatory and anticancer potential. Cell Mol Life Sci 2008; 65: 2979-2999
• 10 Fakhrudin N, Waltenberger B, Cabaravdic M, Atanasov AG, Malainer C, Schachner D, Heiss EH, Liu R, Noha SM, Grzywacz AM, Mihaly-Bison J, Awad EM, Schuster D, Breuss JM, Rollinger JM, Bochkov V, Stuppner H, Dirsch VM. Identification of plumericin as a potent new inhibitor of the NF-κB pathway with anti-inflammatory activity in vitro and in vivo . Br J Pharmacol 2014; 171: 1676-1686
• 11 Tran TVA, Malainer C, Schwaiger S, Atanasov AG, Heiss EH, Dirsch VM, Stuppner H. NF-κB inhibitors from Eurycoma longifolia . J Nat Prod 2014; 77: 483-488
• 12 Zhao GP, Dai S, Chen RS. eds. Dictionary of Traditional Chinese Medicine. 2nd ed.. ed. Shanghai: Shanghai Scientific and Technical Publishers; 2006
• 13 Ashour ML, Wink M. Genus Bupleurum: a review of its phytochemistry, pharmacology and modes of action. J Pharm Pharmacol 2011; 63: 305-321
• 14 Benito PB, Martinez MJA, Sen AMS, Gomez AS, Matellano LF, Contreras SS, Lanza AMD. In vivo and in vitro antiinflammatory activity of saikosaponins. Life Sci 1998; 63: 1147-1156
• 15 Lu XL, He SX, Ren MD, Wang YL, Zhang YX, Liu EQ. Chemopreventive effect of saikosaponin-d on diethylinitrosamine-induced hepatocarcinogenesis: involvement of CCAAT/enhancer binding protein β and cyclooxygenase-2. Mol Med Rep 2012; 5: 637-644
• 16 Lu CN, Yuan ZG, Zhang XL, Yan R, Zhao YQ, Liao M, Chen JX. Saikosaponin a and its epimer saikosaponin d exhibit anti-inflammatory activity by suppressing activation of NF-κB signaling pathway. Int Immunopharmacol 2012; 14: 121-126
• 17 Ishii H, Nakamura M, Seo S, Tori K, Tozyo T, Yoshimura Y. Isolation, characterization, and nuclear magnetic resonance spectra of new saponins from the roots of Bupleurum falcatum L. Chem Pharm Bull 1980; 28: 2367-2373
• 18 Ogihara Y, Nose M. Novel cleavage of the glycosidic bond of saponins in alcoholic alkali metal solution containing a trace of water. J Chem Soc Chem Commun 1986; 1417
• 19 Kobayashi Y, Takeda T, Ogihara Y. New triterpenoid glycosides from the leaves of Bupleurum rotundifolium L. Chem Pharm Bull 1981; 29: 2222-2229
• 20 Shimizu K, Amagaya S, Ogihara Y. New derivatives of saikosaponins. Chem Pharm Bull 1985; 33: 3349-3355
• 21 Liang H, Zhao Y, Qiu H, Huang J, Zhang R. New saikosaponin from Bupleurum chinese DC. Acta Pharm Sin 1998; 33: 37-41
• 22 Shimaoka A, Seo S, Minato H. Saponins isolated from Bupleurum falcatum L.; components of saikosaponin b. J Chem Soc Perkin 1 1975; 2043-2048
• 23 Shimizu K, Amagaya S, Ogihara Y. Structural transformation of saikosaponins by gastric juice and intestinal flora. J Pharmacobiodyn 1985; 8: 718-725
• 24 Xu L, Tian JX, Song R, Liu GQ, Tian Y, Zhang ZJ. LC-MS/MS determination and comparison of saikosaponin a,b2,c,d in crude and processed Radix Bupleuri by vinegar. J China Pharm Univ 2012; 43: 334-340
• 25 Min J, Lee YJ, Kim YA, Park HS, Han SY, Jhon GJ, Choi W. Lysophosphatidylcholine derived from deer antler extract suppresses hyphal transition in Candida albicans through MAP kinase pathway. Biochim Biophys Acta 2001; 1531: 77-89
• 26 Houjou T, Yamatani K, Nakanishi H, Imagawa M, Shimizu T, Taguchi R. Rapid and selective identification of molecular species in phosphatidylcholine and sphingomyelin by conditional neutral loss scanning and MS3. Rapid Commun Mass Spectrom 2004; 18: 3123-3130
• 27 Morita M, Nakajima K, Ikeya Y, Mitsuhashi H. Polyacetylenes from roots of Bupleurum falcatum . Phytochemistry 1991; 30: 1543-1545
• 28 OʼNeill T, Johnson JA, Webster D, Gray CA. The Canadian medicinal plant Heracleum maximum contains antimycobacterial diynes and furanocoumarins. J Ethnopharmacol 2013; 147: 232-237
• 29 Ohta T, Uwai K, Kikuchi R, Nozoe S, Oshima Y, Sasakl K, Yoshizaki F. Absolute stereochemistry of cicutoxin and related toxic polyacetylenic alcohols from Cicuta virosa . Tetrahedron 1999; 55: 12087-12098
• 30 Fiandanese V, Bottalico D, Cardellicchio C, Marchese G, Punzi A. Stereoselective total synthesis of (S)-Virol C and (S)-1-dehydroxyvirol A. Tetrahedron 2005; 61: 4551-4556
• 31 Estevez-Braun A, Estevez-Reyes R, Moujir LM, Ravelo AG, Gonzalez AG. Antibiotic activity and absolute configuration of 8S-heptadeca-2(Z), 9(Z)-diene-4,6-diyne-1,8-diol from Bupleurum salicifolium . J Nat Prod 1994; 57: 1178-1182
• 32 Blunder M, Liu X, Kunert O, Winkler NA, Schinkovitz A, Schmiderer C, Novak J, Bauer R. Polyacetylenes from Radix et Rhizoma Notopterygii incisi with an inhibitory effect on nitric oxide production in vitro . Planta Med 2014; 80: 415-418
• 33 Xie LW, Atanasov AG, Guo DA, Malainer C, Zhang JX, Zehl M, Guan SH, Heiss EH, Urban E, Dirsch VM, Kopp B. Activity-guided isolation of NF-κB inhibitors and PPARγ agonists from the root bark of Lycium chinense Miller. J Ethnopharmacol 2014; 152: 470-477
• 34 Vogl S, Picker P, Mihaly-Bison J, Fakhrudin N, Atanasov AG, Heiss EH, Wawrosch C, Reznicek G, Dirsch VM, Saukel J, Kopp B. Ethnopharmacological in vitro studies on Austriaʼs folk medicine – an unexplored lore in vitro anti-inflammatory activities of 71 Austrian traditional herbal drugs. J Ethnopharmacol 2013; 149: 750-771
• 35 Stern ST, Zolnik BS, McLeland CB, Clogston J, Zheng J, McNeil SE. Induction of autophagy in porcine kidney cells by quantum dots: A common cellular cesponse to nanomaterials?. Toxicol Sci 2008; 106: 140-152