New sesquineolignan glycoside isomers from the aerial parts of Leonurus japonicus and their absolute configurations

2.1 General experimental procedures

A Rudolph Autopol-I automatic polarimeter (Rudolph Research Analytical, USA) was applied to measure optical rotations. IR spectra were recorded using a Cary 600 FT-IR microscope (Agilent Technologies Inc., CA, USA). ECD spectra were detected on a Chirascan-plus Circular Dichroism spectrometer (Applied Photophysics Ltd., Leatherhead, England). A Bruker TIMS-TOF-MS instrument (Bruker Corporation, Billerica, MA, USA) was applied to acquire HR-ESI-MS data. NMR data were obtained using a Bruker-AVIIIHD-600 spectrometer (Bruker Corporation, Billerica, MA, USA) with the solvent peaks as the references. Column chromatography was performed using silica gel (200–300 mesh, Yantai Institute of Chemical Technology, Yantai, China), MCI gel CHP 20P (40–60 μm, Mitsubishi Chemical, Co., Japan), and Toyopearl HW-40F (TOSOH, Tokyo, Japan). TLC was performed using GF₂₅₄ silica gel plates (Anhui Liangchen Silicon Source Material Co. Ltd., Anhui, China). HPLC separation was carried out on an Agilent 1100 instrument equipped with a Zorbax SB-C₁₈ (9.4 × 250 mm², 5 μm). The hydrolysis reactions were carried out with snailase (Beijing BioDee Biotechnology Co., Ltd., Beijing, China). Curcumin was used as a positive control and purchased from Sichuan Weikeqi Biological Technology Co., Ltd. (Sichuan, China). NO levels were detected by nitric oxide assay kit from Beyotime Institute of Biotechnology (Shanghai, China). Rabbit anti-phospho-ERK1/2 (Thr202/Tyr204) and anti-ERK1/2 antibodies were purchased from Cell Signaling Technology (CST; Danvers, Massachusetts, US). Rabbit anti-phospho-NF-κB (Ser536) and anti-NF-κB antibodies were obtained from Chengdu Zen Bioscience Co., Ltd. (Chengdu, China). Rabbit anti-iNOS was bought from Beyotime Institute of Biotechnology (Shanghai, China). Rabbit anti-β-actin antibody was gained from GeneTex Inc. (Irvine, California, USA). Peroxidase-conjugated AffiniPure goat anti-rabbit immunoglobulin G (IgG; [H+L]) was purchased from ZSGB-BIO Commerce Store (Beijing, China). Shanghai EpiZyme Biotechnology Co., Ltd. (Shanghai, China) supplied Omni-ECL TMFemtoLiqht chemiluminescence Kit.

2.2 Plant material

The aerial parts of L. japonicus were purchased from Sichuan Neautus Traditional Chinese Medicine Co., Ltd. in 2019 and identified by Prof. Jihai Gao at Chengdu University of Traditional Chinese Medicine. A voucher specimen (LJ-20190517) has been deposited at the school of pharmacy, Chengdu university of Traditional Chinese Medicine, China.

2.3 Extraction and isolation

The aerial parts of L. japonicus (100 kg) were refluxed three times with 95 % EtOH (3 h each time) to afford an EtOH extract, which was evaporated to give a residue (5.5 kg). The residue was suspended in water containing 1 % hydrochloric acid and partitioned with EtOAc. The acid water-soluble fraction (AWSF, 660 g) was subjected to a silica gel column using a gradient elution of ethyl acetate–methanol–ammonia (100:7:5–0:7:5) to afford 11 fractions (F₁–F₁₁). Eluting with a step gradient of 10–100 % MeOH in H₂O, F₅ was separated using flash chromatography over MCI gel to obtain six subfractions (F_5-1–F_5-6). F_5-5 was fractioned via silica gel column chromatography over ethyl acetate–methanol–ammonia (100:5:5–0:5:5) to afford nine subfractions F_5-5-1–F_5-5-9. F_5-5-4 was further purified using Toyopearl HW-40F column chromatography (50 % MeOH in H₂O), followed by reserved-phase semi-preparative HPLC (50 % MeOH in H₂O) to afford 1 (2.3 mg, t_min = 102.4 min), 2 (1.3 mg, t_min = 90.2 min), and 3 (1.9 mg, t_min = 112.8 min).

Leolignoside A (1): white powder; ${[\mathrm{\alpha }]}_{\mathrm{D}}^{20}$ −4.8 (c 0.04, MeOH); UV (MeCN) λ_max (log ε) 193 (4.4), 281 (3.0) nm; IR (ATR) ν_max 3296, 2938, 2856, 1634, 1567, 1514, 1463, 1025, 771, 722 cm⁻¹; (+)-HR-ESI-MS m/z 797.2995 [M+Na]⁺ (calcd. for C₃₉H₅₀O₁₆Na, 797.2997); ¹H (CDCl₃, 600 MHz) and ¹³C NMR (CDCl₃, 150 MHz) data, see Table 1.

Leolignoside B (2): white powder; ${[\mathrm{\alpha }]}_{\mathrm{D}}^{20}$ −9.5 (c 0.04, MeOH); UV (MeCN) λ_max (log ε) 202 (3.9), 280 (2.7) nm; IR (ATR) ν_max 3355, 2925, 2854, 1666, 1595, 1517, 1457, 1075, 820, 742, cm⁻¹; (+)-HR-ESI-MS m/z 797.2995 [M+Na]⁺ (calcd. for C₃₉H₅₀O₁₆Na, 797.2997); ¹H (CDCl₃, 600 MHz) and ¹³C NMR (CDCl₃, 150 MHz) data, see Table 1.

Leolignoside C (3): white powder; ${[\mathrm{\alpha }]}_{\mathrm{D}}^{20}$ −4.8 (c 0.04, MeOH); UV (MeCN) λ_max (log ε) 199 (4.3), 281 (3.0) nm; IR (ATR) ν_max 3297, 2919, 2852, 1637, 1557, 1516, 1461, 1074, 763, 720 cm⁻¹; (+)-HR-ESI-MS m/z 797.2995 [M+Na]⁺ (calcd. for C₃₉H₅₀O₁₆Na, 797.2997); ¹H (CDCl₃, 600 MHz) and ¹³C NMR (CDCl₃, 150 MHz) data, see Table 1.

2.4 Enzymatic hydrolysis

Compounds 1 (1.9 mg), 2 (1.0 mg), and 3 (1.5 mg) in H₂O (4 mL) were hydrolyzed with snailase (15 mg) at 37 °C for 48 h, respectively. After that, EtOAc was applied to extract aglycones three times (3 × 16 mL). Then, the aglycones 1a (1.1 mg), 2a (0.6 mg), and 3a (0.8 mg) were purified using HPLC (50 % MeCN in H₂O), and their structures were determined using ¹H NMR (Table 2) and HR-ESI-MS. Compound 1a: ${[\mathrm{\alpha }]}_{\mathrm{D}}^{20}$ −16.0 (c 0.03, MeOH); CD (MeCN) 209 (Δε + 4.50), 239 (Δε + 0.41), 279 (Δε + 0.41) nm; (+)-HR-ESI-MS m/z 635.2461 [M+Na]⁺ (calcd. for C₃₃H₄₀O₁₁Na, 635.2468). 2a: ${[\mathrm{\alpha }]}_{\mathrm{D}}^{20}$  − 3.6 (c 0.03, MeOH); CD (MeCN) 217 (Δε + 0.14), 233 (Δε − 0.48), 282 (Δε + 0.10) nm; (+)-HR-ESI-MS m/z 635.2463 [M+Na]⁺ (calcd. for C₃₃H₄₀O₁₁Na, 635.2468). 3a: ${[\mathrm{\alpha }]}_{\mathrm{D}}^{20}$ +11.5 (c 0.03, MeOH); CD (MeCN) 216 (Δε + 0.23), 232 (Δε − 0.22), 289 (Δε + 0.09) nm; (+)-HR-ESI-MS m/z 635.2464 [M+Na]⁺ (calcd. for C₃₃H₄₀O₁₁Na, 635.2468).

2.5 Cell culture

RAW 264.7 macrophages were purchased from the Shanghai Cell Bank of the Chinese Academy of Sciences and cultured in DMEM containing 10% FBS, 100 U/mL penicillin, and 100 µg/mL streptomycin at 37 °C in a humidified atmosphere with 5 % CO₂.

2.6 Anti-inflammatory activity assay

The doses of the tested compounds were chosen according to a cell viability assay in normal RAW 264.7 cells using CCK-8 kits. Then, in the anti-inflammatory assay, RAW 264.7 cells were seeded in a 96-well plate (5 × 10⁴ cells per well) and incubated for 24 h. Subsequently, the cells were treated with 1 μg/mL LPS for 24 h with or without various concentrations of tested compounds (curcumin was used as a positive control) (Huang et al., 2022, Li et al., 2021). Then, the supernatants were collected and assayed for NO production with an NO kit using Griess reagent. Absorbance was measured using a microplate reader at a wavelength of 570 nm. NO levels in samples were determined based on a standard curve.

2.7 Western blot analysis

RAW 264.7 cells were pretreated with compound 3a (12.5, 25, and 50 μM) for 1 h and then stimulated with 1 μg/mL LPS. To obtain total proteins, the cells were lysed in RIPA buffer containing proteinase inhibitors and crushed with an ultrasonic cell pulverizer. Equal quantities of protein (30 μg per lane) were separated using SDS-PAGE and transferred to PVDF membranes, which were blocked in 5% non-fat milk for 2 h at room temperature and incubated with primary antibodies against ERK, p-ERK, NF-κB, p-NF-κB, iNOS, and β-actin overnight at 4 °C. After three washes with TBST, membranes were incubated with HRP-conjugated goat anti-rabbit secondary antibodies for 1 h at room temperature and washed again. Protein bands were visualized by enhanced chemiluminescence. Western blot images were captured using a Tanon 5200 chemiluminescence imaging system.

2.8 Statistical analysis

Data are presented as mean ± standard deviation (SD). All experiments were conducted at least three times. One-way analysis of variance was conducted for comparisons among multiple groups, and the least significant difference test was used for comparisons between two groups. Figures were prepared using GraphPad Prism version 5.0 (GraphPad Software, Inc., San Diego, CA, US). In all comparisons, P<0.05 was considered statistically significant.

3.1 Inhibitory effect of AWSF on NO production in RAW 264.7 cells

To exclude the possibility that the cytotoxicity of AWSF may contribute to its anti-inflammatory effect, the viability of RAW 264.7 cells treated with various AWSF concentrations was evaluated using the CCK-8 assay. As shown in Fig. 2A, AWSF exhibited no cytotoxic effect at concentrations of 6.25, 12.5, 25, 50, and 100 μg/mL. Therefore, it is rational to apply these concentrations in the subsequent experiments.

Close

Fig. 2

Inhibitory effect of AWSF on LPS-induced NO production in RAW 264.7 cells. (A) Effect of AWSF on the survival rate of normal macrophages. (B) Effect of AWSF on NO release from macrophages. Data are shown as mean ± SD of three independent experiments. ^##P < 0.01 vs. untreated control; ^**P < 0.01 vs. model group.

Export to PPT

Because overproduction of NO is a typical marker for inflammation in activated macrophages, Griess reagent kits were applied to assess extracellular NO concentrations. As presented in Fig. 2B, NO production in RAW 264.7 macrophages was markedly increased by stimulation with LPS, and AWSF could significantly decrease the NO content in a dose-dependent manner with an IC₅₀ value of 14.59 μg/mL.

3.2 AWSF decreased the protein expression of p-ERK and iNOS in LPS-activated RAW 264.7 cells

NO, a pro-inflammatory mediator, is synthesized excessively by iNOS when macrophages are stimulated by LPS (Feng et al., 2021, Xiang et al., 2018). Thus, we speculated that iNOS might be involved in the effect of AWSF on NO release. In addition, numerous studies have reported that ERK, a member of the MAPK family, is critical for activating NF-κB, which could subsequently activate iNOS expression (Peng et al., 2015, Zou et al., 2018). Considering that AWSF is a mixture, it is acceptable to preliminarily explore whether AWSF decreases NO overproduction by suppressing ERK and iNOS protein expression. As shown in Fig. 3, AWSF suppressed LPS-activated protein expression of p-ERK and iNOS in a dose-dependent manner, indicating that AWSF may exert anti-inflammatory activity by inhibiting p-ERK and iNOS expression.

Close

Fig. 3

Effect of AWSF on the expression of inflammatory-related proteins. (A) Expression levels of p-ERK1/2 (Thr202/Tyr204) and iNOS. (B) p-ERK/ERK ratio. (C) iNOS/β-actin ratio. Data are shown as mean ± SD. ^##P < 0.01 vs. untreated control; *P < 0.05 and ^**P < 0.01 vs. model group (n = 3).

Export to PPT

3.3 Structure elucidation of the compounds isolated from AWSF

Compound 1 was obtained as a white powder. HR-ESI-MS analysis of 1 gave an [M+Na]⁺ at m/z 797.2995 revealing a molecular formula of C₃₉H₅₀O₁₆ (calcd. for C₃₉H₅₀O₁₆Na, 797.2997). The IR spectrum of 1 displayed strong signals at 1634, 1567, 1514, and 3296 cm⁻¹, supporting the presence of aromatic rings and hydroxy groups. To provide more detailed structural information, the ¹H NMR spectrum (Table 1) displayed typical signals that corresponded to two 1,3,4-trisubstituted aromatic rings [δ_H 6.90 (1H, d, J=1.8 Hz), 6.90 (1H, d, J=8.4 Hz), 6.82 (1H, dd, J=8.4, 1.8 Hz), 6.81 (1H, d, J=8.4 Hz), 6.76 (1H, d, J=1.8 Hz), and 6.66 (1H, dd, J=8.4, 1.8 Hz)], a 1,3,4,5-tetrasubstituted aromatic ring [δ_H 6.56 (2H, br.s)], two exchangeable aromatic hydroxy protons (δ_H 5.62 and 5.56), four aromatic methoxy groups (δ_H 3.91, 3.84, and 3.83), multiple oxygenated methylene and methine groups, two non-oxygenated methine groups, and a methyl group. The ¹³C NMR data (Table 1) in combination with DEPT and HSQC spectra exhibited carbon signals corresponding to the above protons, including 19 methine signals [eight aromatic (δ_C 120.2, 119.1, 114.4, 114.2, 109.2, 108.7, 103.4, 103.4), nine oxygenated (δ_C 104.1, 86.1, 85.9, 85.7, 81.2, 76.6, 75.3, 74.0, 70.5), and two non-oxygenated (δ_C 54.6, 54.2)], five oxygenated methylene signals (δ_C 72.2, 71.7, 69.5, 65.3, 62.6), and five methyl signals [a methyl group (δ_C 15.6) and four methoxy groups (δ_C 56.4, 56.4, 56.1, 56.1)]. In addition, the ¹³C NMR and DEPT data indicated the presence of 10 aromatic quaternary carbons attributed to three phenyl units. From the above data, it can be seen that 1 contains three C6-C3 moieties substituted with four aromatic methoxy groups, two phenolic hydroxy groups, and a glycosyl group. Moreover, according to the coupling constant of the anomeric proton (J=7.8 Hz) and the ¹³C chemical shifts of the sugar moiety, it is inferred that there is a β-glucopyranosyl (Xiong et al., 2013). Therefore, compound 1 is a sesquineolignan β-glucopyranoside.

A rigorous survey of the literature revealed that 1 and (−)-(7R,7′R,7″R,8S,8′S,8″S)-4′,4″-dihydroxy-3,3′,3″,5-tetramethoxy-7,9′:7′,9-diepoxy-4,8″-oxy-8,8′-sesquineolignan-7″,9″-diol (Xiong et al., 2011) have a very similar sesquineolignan skeleton, both of which are linked by medioresinol and guaiacylglycerol. The main difference is that 1 has one more β-glucopyranosyl unit and an ethoxy group than the compound reported in the literature. Detailed comparison of the ¹³C NMR data between the two compounds showed that C-7″ and C-9″ resonances in 1 were shifted by Δδ_C+8.7 and +8.8 ppm, respectively, indicating that the β-glucopyranosyl and ethoxy groups are attached at the C-7″ and C-9″ positions. The inference of the planar structure of 1 was confirmed by 2D NMR including ¹H–¹H COSY, HSQC, and HMBC (Fig. 4). In the HMBC spectrum, H-7 shared correlations with C-9, C-8′, and C-9′, while H-7′ displayed correlations with C-8, C-9, and C-9′, in combination with the COSY correlations of H-7/H-8/H₂-9, H-7′/H-8′/H₂-9′, and H-8/H-8′, confirming the existence of the tetrahydrofuran ring in 1. Two symmetric methoxy groups are connected to the C-3 and C-5 positions based on the HMBC correlations of H-2/6 with C-1, C-3/5, C-4, and C-7, as well as OMe-3/5 with C-3/5. A methoxy group and a hydroxy group were determined to be located at the C-3′ and C-4′ positions, respectively, relying on the correlations from H-5′ to C-1′, C-3′, and C-6′; from H-6′ to C-2′ and C-4′; from OMe-3′ to C-3′; from OH-4′ to C-3′, C-4′, and C-5′; and from H-8′ to C-1′, C-2′, and C-3′. Similarly, it can be determined that a methoxy group and a hydroxy group were attached to C-3″ and C-4″, respectively. Furthermore, according to the HMBC correlations of H-7″ with C-1″, C-2″, C-6″, C-8″, and C-9″, together with the COSY correlations of H-7″/H-8″/H₂-9″, it is inferred that 1 contains a guaiacylglycerol fragment. The ethoxyl group at C-7″ was established by the HMBC correlation of H-7″ with C-10″, while the β-glucopyranosyl group was connected to C-9″ based on the HMBC correlation of H-1′″ with C-9″. Although no HMBC signal for the ether bond between H-8″ and C-4 was observed, the connection of 8″-O-4 can be inferred from the molecular formula and the chemical shifts of H-8″ (δ_H 4.36), C-8″ (δ_C 85.7), and C-4 (δ_C 135.5). Therefore, the planar structure of 1 was determined as 4′,4″-dihydroxy-3,3′,3″,5-tetramethoxy-7″-ethoxy-7,9′:7′,9-diepoxy-4,8″-oxy-8,8′-sesquineolignan-9″-O-β-glucopyranoside.

Close

Fig. 4

Key HMBC and ¹H–¹H COSY correlations of compounds 1–3.

Export to PPT

The relative configuration of 1 was determined through a comprehensive analysis of coupling constants, chemical shifts, and NOESY data. First, the ¹H NMR and ¹³C NMR shifts for the bis-tetrahydrofuran ring in 1 were analyzed, revealing that both H-7 and H-7′ were at 4.6–5.1 ppm, H-8 and H-8′ at 3.0–3.2 ppm, H-9a and H-9′a at 4.2–4.4 ppm, H-9b and H-9′b at 3.8–4.0 ppm, C-7 and C-7′ at 82–87 ppm, C-8 and C-8′ at 54–56 ppm, and C-9 and C-9′ at 71–74 ppm. According to empirical rules (Xiong et al., 2011, Shi, 2009, Greger and Hofer, 1980), it can be deduced that the bis-tetrahydrofuran lignan fragment of 1 is of the eq/eq type, and H-7/H-8 and H-7′/H-8′ are in trans orientation. These relative configurations can further be confirmed by Δδ_{H–9a, 9b} (0.35 ppm) and Δδ_{H–9′a, 9′b} (0.39 ppm) (Shao et al., 2018), both of which were in the range of 0.3–0.4 ppm. The above inference is also consistent with the NOESY data of 1, in which H-2/6 showed correlations to H-8, while H-2′/H-6′ shared correlations with H-8′. Regarding the guaiacylglycerol fragment, the relative configuration was determined by the coupling constant of H-7″ and H-8″ (J_7″,8″). According to the empirical rules for 8,4′-oxolignans, J_7″,8″ can be used for the determination of their relative configurations and is influenced by the tested solvent (Xiong et al., 2011, Gan et al., 2008). Therefore, the guaiacylglycerol unit of 1 was inferred to be an erythro configuration based on J_7″,8″ = 3.6 Hz (CDCl₃).

To further determine the absolute configuration, compound 1 was enzymatically hydrolyzed to obtain aglycone 1a (Table 2) and a glucose. The glucose was identified as D-glucose using TLC and ${[\mathrm{\alpha }]}_{\mathrm{D}}^{20}$ compared with a D-glucose reference standard (Xiong et al., 2013, Husdon and Dale, 1916). Because the cotton peaks in ECD spectra are widely used to determine the absolute configurations of lignans, the ECD data of 1 were measured. Numerous studies have shown that in 8,4′-oxyneolignans, the Cotton effect at 230–240 nm can be used to determine the configuration of C-8 in the aryl glycerol units (positive for 8S and negative for 8R) (Xiong et al., 2011, Gan et al., 2008, Huo et al., 2008, Greca et al., 1994), while in the eq/eq-type 7,9′:7′,9-diepoxylignans, the Cotton effect at around 280 nm can be validated for the configuration assignment of the bis-tetrahydrofuran ring (positive for 7S,7′S,8R,8′R and negative for 7R,7′R,8S,8′S) (Shi, 2009, Hofer and Schölm, 1981, Hosokawa et al., 2004). However, in the sesquineolignans with a bis-tetrahydrofuran ring and an aryl glycerol unit, these empirical rules could not be applied to determine the absolute configuration of the aryl glycerol unit but could be used only for configuration determination of the bis-tetrahydrofuran lignan fragment. This is because the bis-tetrahydrofuran lignan fragment could also generate an obvious Cotton effect at 230–245 nm in addition to the Cotton effect at around 280 nm (Hofer and Schölm, 1981; Hosokawa et al., 2004; Muhit et al., 2016; Li et al., 2008,2014), which means that the Cotton effect at 230–245 nm of such sesquineolignans is caused by a comprehensive effect of the aryl glycerol unit and the bis-tetrahydrofuran lignan fragment. In this study, to begin with, a positive Cotton effect at 279 nm (Δε + 0.41) indicated that the eq/eq-type bis-tetrahydrofuran lignan fragment in 1a had a 7S,7′S,8R,8′R configuration. Then, the absolute configuration of the guaiacyl glycerol unit in 1a was determined through ECD calculations, using the time-dependent density functional theory method. Based on the erythro relative configuration of the guaiacyl glycerol unit, the absolute configuration of C-7″ and C-8″ might be 7″S,8″R or 7″R,8″S. Thus, compound 1a could be inferred as (7S,7′S,7″S,8R,8′R,8″R)-1a or (7S,7′S,7″R,8R,8′R,8″S)-1a. ECD calculations showed that the experimental ECD spectrum of 1a matched the calculated ECD spectrum of (7S,7′S,7″S,8R,8′R,8″R)-1a (Fig. 5). Consequently, the structure of 1 was determined as (–)-(7S,7′S,7″S,8R,8′R,8″R)-4′,4″-dihydroxy-3,3′,3″,5-tetramethoxy-7″-ethoxy-7,9′:7′,9-diepoxy-4,8″-oxy-8,8′-sesquineolignan-9″-O-β-D-glucopyranoside (named leolignoside A).

Close

Fig. 5

Experimental and calculated ECD spectra of compounds 1a–3a.

Export to PPT

Compound 2 showed very similar spectroscopic data to those of 1, indicating that 2 is also a sesquineolignan derivative with a similar structure. HR-ESI-MS displayed that 2 had the same molecular formula as 1 (C₃₉H₅₀O₁₆), thus 2 was inferred to be an isomer of 1. Detailed comparison of the ¹H and ¹³C NMR data between 2 and 1 (Table 1) revealed that the most prominent difference is that C-8″ resonance in 2 was shifted by Δδ_C−2.4 ppm, while the coupling constant of H-7″ and H-8″ (J_7″,8″) in 2 was increased to 6.0 Hz. Thus, it can be assumed that 2 is the C-8″ epimer of 1. 2D NMR data analysis further verified that 2 had the same planar structure as 1 (Fig. 1). Moreover, based on ¹H and ¹³C NMR data of the bis-tetrahydrofuran ring, it was clear that the bis-tetrahydrofuran lignan fragment of 2 was also of the eq/eq type. In particular, Δδ_{H–9a, 9b} and Δδ_{H–9′a, 9′b} (0.3–0.4 ppm) (Shao et al., 2018), together with the NOESY correlations of H-2/H-6 with H-8 and H-2′/H-6′ with H-8′, confirmed it was of the eq/eq type. The relative configuration of the guaiacyl glycerol fragment was determined as threo according to the coupling constant of J_7″,8″ = 6.6 Hz (Xiong et al., 2011, Gan et al., 2008).

Using the same methods as described for 1, compound 2 was enzymatically hydrolyzed to obtain its aglycone 2a (Table 2) and a D-glucose. The ECD spectrum of 2a showed a positive Cotton effect at 282 nm, similar to that of 1a, indicating that the eq/eq-type bis-tetrahydrofuran lignan fragment of 2a was also a 7S,7′S,8R,8′R configuration. Then, compound 2a was assumed as (7S,7′S,7″S,8R,8′R,8″S)-2a or (7S,7′S,7″R,8R,8′R,8″R)-2a for ECD calculation according to the above threo configuration of the guaiacyl glycerol unit. As shown in Fig. 5, the calculated ECD spectrum of (7S,7′S,7″S,8R,8′R,8″S)-2a agreed with the experimental ECD spectrum of 2a. Therefore, compound 2 was determined to be (–)-(7S,7′S,7″S,8R,8′R,8″S)-4′,4″-dihydroxy-3,3′,3″,5-tetramethoxy-7″-ethoxy-7,9′:7′,9-diepoxy-4,8″-oxy-8,8′-sesquineolignan-9″-O-β-D-glucopyranoside (leolignoside B).

Compound 3 showed similar spectroscopic data and the same molecular formula as 1 and 2, inferring that 3 may be a sesquineolignan glycoside analog with the same planar structure. This deduction was demonstrated through 2D NMR experiments. Further comparison of ¹H and ¹³C NMR data of 3 and 2 uncovered that the resonance signals of the guaiacyl glycerol β-D-glucopyranoside fragment in 3 and 2 were almost the same, including the chemical shifts and coupling constants. Thus, the relative configuration of the guaiacyl glycerol unit in 3 was also threo. Applying empirical rules (Shi, 2009, Greger and Hofer, 1980), the ¹H and ¹³C NMR data of the bis-tetrahydrofuran ring in 3 indicated that the bis-tetrahydrofuran lignan unit is not of the eq/eq type but of the eq/ax type. The relative configuration was further verified by Δδ_{H–9a, 9b} (0.27 ppm, in the range of 0.205–0.36 ppm) and Δδ_{H–9′a, 9′b} (0.52 ppm > 0.50 ppm) (Shao et al., 2018). After enzyme hydrolysis of 3, the aglycone 3a (Table 2) and a D-glucose were obtained. In the ECD spectrum of 3a, a positive Cotton effect at 289 nm suggested that the eq/ax-type bis-tetrahydrofuran lignan fragment of 3a has a 7R,7′S,8S,8′S configuration (Muhit et al., 2016, Li et al., 2008, Li et al., 2014). Then, two possible isomers, (7R,7′S,7″R,8S,8′S,8″R)-3a and (7R,7′S,7″S,8S,8′S,8″S)-3a, were applied to ECD calculations. The results showed that 7R,7′S,7″S,8S,8′S,8″S was the most probable configuration for 3a (Fig. 5). Accordingly, compound 3 was determined to be (–)-(7R,7′S,7″S,8S,8′S,8″S)-4′,4″-dihydroxy-3,3′,3″,5-tetramethoxy-7″-ethoxy-7,9′:7′,9-diepoxy-4,8″-oxy-8,8′-sesquineolignan-9″-O-β-D-glucopyranoside (leolignoside C).

3.4 Inhibitory effects of compounds 1a–3a on NO production in RAW 264.7 cells

Similar to AWSF, compounds 1a–3a were first evaluated for their cytotoxic effects in RAW 264.7 cells. As presented in Fig. 6A, these compounds all showed no cytotoxicity at concentrations of 12.5, 25, and 50 μM, which provided evidence for further anti-inflammatory assessments.

Close

Fig. 6

Effects of compounds 1a–3a on NO release from macrophages. (A) Effects of compounds 1a–3a on the survival rate of normal macrophages. (B) Effects of compounds 1a–3a on NO release from macrophages. Data are shown as mean ± SD of three independent experiments. ^##P < 0.01 vs. untreated control; ^**P < 0.01 vs. model group.

Export to PPT

The results shown in Fig. 6B reveal that sesquineolignans 1a, 2a, and 3a exerted inhibitory effects on NO release in LPS-injured RAW 264.7 cells. The IC₅₀ values of 2a and 3a was 22.67 and 33.58 μM, respectively. Curcumin was used as a positive control (IC₅₀ = 27.90 μM). Although compound 2a showed the most potent effect, its inhibitory effect on inflammatory-related protein expression was not investigated because the amount of sample available was low. Compound 3a was selected for the subsequent experiments.

3.5 Inhibitory effect of compound 3a on p-ERK and p-NF-κB expression in LPS-injured RAW 264.7 cells

Based on the preliminary exploration of AWSF, the inhibitory effect of compound 3a on protein expression was also examined. As shown in Fig. 7, the LPS-induced increase in p-ERK and p-NF-κB was reversed by 3a. However, unfortunately, iNOS expression was not investigated due to the limited amount of 3a. In summary, compound 3a exerted anti-inflammatory activity through the ERK/NF-κB signaling pathway.

Close

Fig. 7

Effect of compound 3a on the expression of inflammation-related proteins. (A) Expression levels of p-ERK1/2 (Thr202/Tyr204) and p-NF-κB (Ser536). (B) p-ERK/ERK ratio. (C) p-NF-κB/NF-κB ratio. Data are shown as mean ± SD. ^##P < 0.01 vs. untreated control; *P < 0.05 and ^**P < 0.01 vs. model group (n = 3).

Export to PPT