Six new tigliane diterpenoids with anti-inflammatory activity from Euphorbia kansuensis

1 Introduction

In recent years, diterpenoids derived from the genus Euphorbia have attracted the attention of pharmacologists and natural products chemists because of their extensive biological properties and structural diversity (Vasas and Hohmann, 2014). Up to now, over 1000 diterpenoids covering >30 skeletal types have been discovered from Euphorbia plants (Shi et al., 2008; Vasas and Hohmann, 2014; Xu et al., 2021). Among these diterpenoids, tiglianes possessing a 5/7/6/3 tetracyclic ring system are a class of diterpenoids with potential values in natural products drug discovery. Prostratin is a natural diterpenoids of this type under clinical trials as an anti-HIV agent capable of activating latent viral reservoirs and inhibiting HIV-1 infections through kinase C (PKC)-dependent NF-κB activation (Miana et al., 2015).

Euphorbia kansuensis is a traditional Chinese medicinal plant often used for the treatments of pyretolysis, cholagogue, and apocenosis (Wang et al., 2008). Previous studies showed the presence of lathyrane diterpenoids and ursane triterpenoids in this plant (Wang et al., 2008; Zhang et al., 2013; Zhang et al., 2012), and our early chemical investigation also found the occurrence of ingenane and lathyrane diterpenoids (Yan et al., 2019; Yan et al., 2020). Among them, two lathyrane diterpenoids exhibited marked NO inhibitory activity (Yan et al., 2021). In our continuing efforts toward novel NO inhibitor from this plants, six new tigliane diterpenoids and 12 known diterpenoids were isolated from the EtOAc fraction of E. kansuensis. All the compounds were tested for NO production inhibitory effects in LPS-induced RAW264.7 cells, and compounds 1–3 and 6 showed marked activity with IC₅₀ values in the range of 4.8–11.3 μM. Herein, we reported the isolation, structural identification, and NO inhibitory activity of these diterpenoids.

2 Material and methods

3 Results and discussion

The roots of E. kansuensis were extracted with 95% EtOH to give a crude extract, which was suspended in H₂O and partitioned with EtOAc. The EtOAc extract was separated by various column chromatographic methods to afford compounds 1–18 as shown in Fig. 1.

Close

Fig. 1

Structures of compounds 1–18 and two hydrolyzed analogues 1c and 4a.

Export to PPT

Compound 1 was obtained as a yellow oil. Its molecular formula C₃₀H₃₆O₉ was determined by the HRESMS at m/z 585.2344 [M + HCOO]⁻ (calcd for C₃₁H₃₇O₁₁⁻, 585.2341) and 539.2263 [M − H]⁻ (calcd for C₃₀H₃₅O₉⁻, 539.2287). Its IR spectrum showed the presence of hydroxyl (3397 cm⁻¹) and carbonyl (1705 cm⁻¹) groups. The ¹H NMR data of 1 showed signals for five methyls [δ_H 0.90 (3H, d, J = 6.4 Hz), 1.22 (3H, s), 1.27 (3H, s), 1.77 (3H, br s), and 2.11 (3H, s)], an oxygenated methene [δ_H 3.99 (1H, d, J = 12.9 Hz) and 4.05 (1H, d, J = 12.9 Hz)], an oxygenated methine [δ_H 5.49 (1H, d, J = 10.3 Hz)], an aldehyde group [δ_H 9.63 (1H, d, J = 7.8 Hz)], and six olefinic protons (δ_H 5.69–7.59). The ¹³C NMR data and DEPT experiments resolved 30 carbon resonances attributable to a ketone carbonyl (δ_C 208.9), an 8-oxo-2,4,6-octatrienoyl (δ_C 125.6, 134.2, 136.9, 138.4, 142.6, 149.2, 166.1, and 193.1), an ester carbonyl (δ_C 173.8), two trisubstituted double bonds, five methyls, two sp³ methenes (one oxygenated at δ_C 68.0), five sp³ methines (one oxygenated at δ_C 77.6), three oxygenated sp³ tertiary carbons, and a characteristic quaternary carbon (δ_C 25.8). The abovementioned NMR data showed high similarity to those of a known tigliane phorbol ester, 12-O-(2′E,4′E)-6-oxohexa-2′,4′-dienoylphorbol-13-acetate (Wagh et al., 2017), except for the replacement of the 6-oxo-2,4-hexadienoyl side chain in the known compound by an 8-oxo-2,4,6-octatrienoyl side chain in 1. This long chain fatty acid moiety, C-1′−C-8′, was tied to C-12 of the phorbol backbone in 1 by HMBC correlation from H-12 to the ester carbony carbon (δ_C 166.1, C-1′) (Fig. 2). Further 2D NMR data interpretation confirmed the planar structure of 1.

Close

Fig. 2

Key ¹H−¹H COSY (-) and HMBC (arrows) correlations of 1 and 6.

Export to PPT

The relative configuration of 1 was assigned to be the same as that of 12-O-(2′E,4′E)-6-oxohexa-2′,4′-dienoylphorbol-13-acetate based on their similar 1D NMR and NOESY data. As shown in Fig. 3, the strong NOESY correlation of H₃-17/H-11, H₃-17/H-8, and H-8/H-11 suggested that H-8, H-11, and CH₃-17 were on the same side and were arbitrarily designated as β-orientation, while the NOE correlations of H₃-18/H-12 and H₃-16/H-14 indicated that H-12, and H-14 were α-oriented. In particular, the E-geometry for Δ^2′ and Δ^6′ were determined by the coupling constants of H-2′/H-3′ (J = 15.4 Hz) and H-6′/H-7′ (J = 15.4 Hz), respectively. These assignments were supported by the NOE correlations of H-2′/H-4′ and H-6′/H-8′, respectively. The E-geometry for Δ^4′ was established by the strong NOE correlations of H-4′/H-6′ and H-5′/H-3′. As the chemical shifts of C-4 (δ_C 73.7), C-9 (δ_C 78.2), and C-10 (δ_C 56.1, δ_H 3.26) in 1 were very similar to those of the known compound (C-4: δ_C 73.7; C-9: δ_C 78.2; C-10: δ_C 56.2, δ_H 3.27), the orientations of 4-OH, 9-OH, and 10-OH in 1 were determined to be the same as those in the known compound.

Close

Fig. 3

Key NOE correlations (↔) of 1 and 6.

Export to PPT

ECD calculation was used to identify the absolute configuration of 1. As shown in Fig. 4, the experimental ECD curve of 1 showed three cotton effects around 205 (−), 229 (+), and 300 (−) nm, respectively, which matched well with those calculated for 1a, an isomer with a 4R,8S,9S,10S,11R,12R,13S,14R configuration, suggesting that 1 had the same absolute configuration as that of 1a. Moreover, the structure of 1 was further confirmed by chemical method. Briefly, alkaline hydrolysis of 1 generate a product 1c, which was identified as phorbol by its 1D NMR and ESIMS data (Neeman et al., 1979) (Fig. S53–55), followed by co-injection of 1c with an authentic phorbol sample on HPLC equipped with a chiral column (Fig. S3). Compound 1 was given the trivial name euphkanoid A.

Close

Fig. 4

The experimental and calculated ECD curves of 1.

Export to PPT

Compound 2 had the molecular formula C₃₂H₃₈O₉ determined by its HRESIMS and ¹³C NMR data. The 1D NMR data of 2 were similar to those of 1, with the major difference being the presence of additional signals attributable to one double bond in 2, suggesting that 2 had a 10-oxo-2,4,6,8-decatetraenoyl moiety at C-12 instead of an 8-oxo-2,4,6-octatrienoyl moiety. This was supported by detailed analysis of the HMBC and ¹H−¹H COSY correlations of 2 (Fig. S1). The relative configuration of 2 on the chiral centers was determined to be the same as that of 1 by comparing its 1D NMR and NOESY data with those of 1. In particular, the E-geometry for Δ^2′, Δ^4′, Δ^6′ and Δ^8′ were determined by the large coupling constants of H-2′/H-3′ (J = 15.2 Hz), H-4′/H-5′ (J = 14.5 Hz), H-6′/H-7′ (J = 14.5 Hz), and H-8′/H-9′ (J = 15.2 Hz), respectively. These assignments were further supported by the corresponding NOE correlations of H-2′/H-4′, H-3′/H-5′, H-6′/H-8′, and H-7′/H-9′ (Fig. S2). As the experimental ECD curve of 2 showed similar tendency to that of 1 (Fig. S4), the absolute configuration of 2 was assigned to be the same as that of 1, and compound 2 was given the trivial name euphkanoid B.

Euphkanoid C (3) had the molecular formula of C₃₄H₄₀O₉ as determined by its HRESIMS data, C₂H₂ more than that of 2. The 1D NMR data of 3 bored a high resemblance to those of 2 except for the presence of additional signals attributable to a double bond, suggesting that 3 possessed a 12-oxo-2,4,6,8,10-dodecapentaenoyl moiety at C-12 instead of a 10-oxo-2,4,6,8-decatetraenoyl moiety in 2. Further analysis of its 2D NMR data secured the planar structure of 3 (Fig. S1). The relative configuration of 3 in the chiral carbons was assigned to be the same as that of 2 by comparison of their similar 1D NMR and NOESY data. In particular, the E-geometry for Δ^2′, Δ^4′, Δ^6′, Δ^8′, and Δ^10′ were determined by the large coupling constants of H-2′/H-3′ (J = 15.2 Hz), H-4′/H-5′ (J = 14.5 Hz), H-6′/H-7′ (J = 14.3 Hz), H-8′/H-9′ (J = 14.5 Hz), and H-10′/H-11′ (J = 15.2 Hz), respectively. The stereochemistry of 3 was assigned to be the same as that of 2 based on their similar ECD spectra, both of them showed two apparent Cotton effects around 204 (−) and 230 (+) nm (Fig. S4 in Supplementary Material).

Compound 4 was isolated as a colorless oil. The pseudomolecular ions in HRESIMS at m/z 675.4258 [M + Na]⁺ (calcd. for C₄₀H₆₀O₇Na⁺, 675.4231) allowed the molecular formula C₄₀H₆₀O₇ to be assigned to 4. The ¹H and ¹³C NMR data were very similar to those of a co-isolated compound, prostratin (7) (Gustafson et al., 1992), except for the presence of additional signals attributable to a linoleoyl group (Huh et al., 2010), suggesting that 4 was a O-linoleoyl derivative of 7. The linoleoyl group was further located at C-20 by the HMBC correlations from H_a-20 and H_b-20 to the additional ester carbonyl carbon (δ_C 173.5, C-1′, Fig. S1). This was also supported by the severely downfield-shifted H₂-20 signals in 4 with respect to those of 7 (δ_H 3.79 and 3.86 in 7, δ_H 4.42 and 4.47 in 4). The relative configuration on the phorbol backbone in 4 was assigned to be the same as that of 7 based on their similar 1D NMR data, while the structure of the linoleoyl group remained to be determined. Thus, alkaline hydrolysis of 4 in NaOMe was carried out to generate a mixture of 12-deoxyphorbol (4a) and methyl linoleate. The 12-deoxyphorbol isolated from the mixture was identified by comparing the experimental ¹H and ¹³C NMR data with those reported in the literature (Sosath, et al., 1988), and was further confirmed by 2D NMR data analysis, while the structure of the isolated methyl linoleate was confirmed by GC–MS (t_R = 6.87 min, m/z 294.1 [M]⁺) and 1D NMR data (Marwah et al., 2007) (Fig. S63 − 65). Thus, the structure of 4 was established and was given a trivial name euphkanoid D.

Compound 5 had a molecular formula C₄₀H₅₈O₇, as determined by its HRESIMS and ¹³C NMR data. The 1D NMR data of 5 were very similar to those of 4 except for the presence of additional signals attributable to a cis double bond and absence of signals for two methenes, suggesting that 5 had a linolenoyl group at C-20 instead of linoleoyl group. The linolenoyl moiety was further identified by comparison of the 1D NMR data of 5 with those of a methyl linolenate reported in the literature (Chang et al., 2000). HMBC correlations from H_a-20 and H_b-20 to the ester carbonyl carbon (δ_C 173.5, C-1′) located the linolenoyl group at C-20 (Fig. S1). The structure of 5 was further confirmed by using the same method as described for 4, which afforded the same hydrolysis product 4a and a methyl linolenate. The structure of methyl linolenate was confirmed by comparing the observed GC–MS (t_R = 8.44 min, m/z 292.1 [M]⁺) and 1D NMR data with those reported in the literature (Chang et al., 2000) (Fig. S66 − 68). Compound 5 was given a trivial name euphkanoid E.

Compound 6 was obtained as a white powder. Its molecular formula was determined to be C₂₂H₃₀O₇ by the HRESIMS peak at m/z 405.1921 [M − H]⁻ (calcd for C₂₂H₂₉O₇⁻, 405.1919). The 1D NMR data of compound 6 resembled closely to those of a known tigliane diterpenoid, prostratin (7) (Gustafson et al., 1992), with the major difference being in the chemical shift values for protons and carbons around C-13 and C-16, suggesting that the acetoxyl was located at C-16 in 6 instead of C-13 in 7. This was confirmed by the HMBC correlations from two additional oxymethene protons (δ_H 4.10 and 4.24, H₂-16) to the ester carbonyl carbon (δ_C 173.5) (Fig. 2). The relative configuration of 6 was determined by analysis of its NOESY data. Firstly, the configuration of the chiral centers on A/B/C rings of 6 was readily assigned to be the same as that of 7 by comparing their 1D NMR data. In particular, the NOE correlations from H_a-16/H_b-16 to H-14 indicated that C-15 − C-16 was α-oriented (Fig. 3). As the absolute configuration of 7 has been confirmed by total synthesis and X-ray single crystal diffraction (Hirose et al., 2021). The absolute configuration of 6 on A/B/C rings could be determined to be the same as that of 7 based on their similar ECD curves, both of them showed two cotton effects around 227 (+) and 207 (−) nm, respectively (Fig. S69). Thus, the remaining C-15 could be assigned as S-configuration according to the strong NOE correlations from H_a-16/H_b-16 to H-14. Compound 6 was given the trivial name euphkanoid F.

The known compounds prostratin (7) (Gustafson et al., 1992), phorbol-13-acetate (8) (Yang et al., 2014), fischeroside A (9) (Pan et al., 2011), 12-deoxyphorbol-13,20-diacetate (10) (Wang et al., 2010), phorbol (1c) (Neema and Simmo, 1979), 12-deoxyphorbol (4a) (Sosath et al., 1988), 12-deoxyphorbol-13-hexadecanoate (11) (Ma et al., 1997), langduin A (12) (Ma et al., 1997), neostellerin (13) (Feng and Pei, 2001), abbeokutone (14) (Yang and Yue, 2008), atisane-3-oxo-16α,17-diol (15) (Ding et al., 1990), ent-atisane-3β,16α,17-triol (16) (Jia and Ding, 1991), ent-(3β,13S)-dihydroxyatis-16-en-14-one (17) (Wang et al., 2014), and helioscopinolide A (18) (Borghi et al., 1991) were identified by comparison of their NMR data with those reported.

All the isolates (1–18) and two hydrolysis products (1c and 4a) were screened for their inhibitory effects on NO production in LPS-induced RAW 264.7 macrophages. The results showed that compounds 4, 5, 1c, 4a, and 7–18 showed weak activity (IC₅₀ > 20 μM), while 1–3 and 6 showed marked inhibitory activity with IC₅₀ values in the range of 4.8–11.3 μM, better than or comparable to the positive control quercetin (IC₅₀ = 12.3 μM) (Table 3). Further cytotoxicity evaluations showed that 1–6 did not show any significant cytotoxicity against RAW 264.7 cells at concentrations up to 50 μM (Table S1), which confirmed that their NO production inhibitory effects were not related to the cytotoxicity but to their anti-inflammatory effects. Based on the biological data, the brief structure–activity relationships for these structurally diverse diterpenoids were discused. Generally, tiglianes with 5/7/6/3 carbon ring system showed greater activity than that of diterpenoids with 6/6/6 ring system (1–6 vs 14–18). In tigliane series, the conjugated olefinic aldehyde moiety were beneficial to the activity, as indicated by the increased activity of 1–3 vs 7 and 8. Besides, the presence of the bulky groups at C-20 decreased the activity as shown by 4 and 5 vs 9, 10 and 4a.

4 Conclusions

In the study, six new tigliane diterpenoids and 12 known diterpenoids were isolated from the roots of E. kansuensis. Their structures were determined by NMR data analysis, ECD calculation, and chemical methods. Compounds 1–3 were rare examples of tigliane diterpenoids featuring a conjugated olefinic aldehyde moiety at C-12. Up to now, only one example has been found in nature. Compounds 1–3 and 6 could inhibit NO release in LPS-induced RAW264.7 cells with IC₅₀ values in the range of 4.8–11.3 μM, better than or comparable to the positive control quercetin. To our best knowledge, tigliane diterpenoids mainly possess anti-HIV and anticancer activities, while their anti-inflammatory potential have rarely been investigated (Wang et al., 2015). Our bioactivity assay for the first time demonstrated that tigliane diterpenoids exhibited marked anti-inflammatory effect in vitro. These findings not only enriched the chemical diversity of Euphorbia diterpenoids, but also indicated that tigliane diterpenoids may also serve as potential tools for further anti-inflammation drug development.