Highly oxidized and rearranged schinortriterpenoids with neuroprotective activity from the stems and leaves of Schisandra chinensis

1 Introduction

Schisandra chinensis is a deciduous woody liana famous for its importance as a raw material for medicine, food, and chemicals (Kim et al., 2015; Li et al., 2022; Xue et al., 2015). Northeast China, Korea peninsula, and far east Russia to Japan is the region for the growing of high-quality S. chinensis fruit (SCF) (Li et al., 2022). It has long been recognized as a medicinal and edible crop with multiple biological activities (Li et al., 2022). Modern pharmacology has shown that SCF extracts has anti-oxidant (Park et al., 2020; Zhang et al., 2013), anti-tumor (Jeong, et al., 2017), anti-depressants (Yan et al., 2016), and anti-inflammatory (Agnieszka et al., 2018; Guo et al., 2008) activities. Additionally, it was considered a nutritious and functional ingredient in yogurt, jams, and wine (Mocan, et al., 2016). The fruits were nutritious and effectively used, but the stems and leaves S. chinensis (SCSL) are treated as waste as agricultural by-products. Noteworthily, the leaves of S. chinensis have been described as a sustainable source of antioxidant and antibacterial compounds (Mocan, et al., 2014). Similarly, SCSL was used in folklore as a herbal remedy for frostbite, asthma, and influenza in China (Shao, et al., 1995). Nowadays, SCSL have received increasing attention due to their richness in bioactive chemicals. SCSL are rich in lignans (Shi et al., 2014), schinortriterpenoids (SNTs) (Shi et al., 2011; Shi et al., 2014), and phenylpropanoids (Yang et al., 2021). SNTs are a highly oxidized and rearranged triterpenoids mainly found in Schisandraceae plants, and they have exhibited multiple pharmacological activities, such as neuroprotective (He et al., 2020; Wang et al., 2021; Yang et al., 2022), antiviral (Cheng et al., 2010; Song et al., 2013), immunosuppressive (Song et al., 2018), and cytotoxic activities (Nguyen et al., 2021; Xiao et al., 2009). Therefore, SNTs, as important physiologically active components in SCSL, are preferred ingredients for the development of SCSL bioactivities.

So far, there have been many studies on the chemical composition and bioactivity of SCF (Venkanna et al., 2014; Xue et al., 2015; Zhang et al., 2013), but there are few studies on the systematic isolation and characterization of the chemical composition of SCSL. The isolation and purification of SNTs is a challenging task because SNTs molecules have a large amount of quaternary carbon, resulting in insufficient two-dimensional NMR signals. Although 55 SNTs have been isolated and identified from SCSL (Shi et al., 2011; Shi et al., 2014; Wang et al., 2013; Yang et al., 2022), the number of SNTs found in SCSL is still lower than expected because SCSL contains structurally diverse SNTs. Therefore, it is important to investigate the phytochemical and pharmacological activities of SCSL in depth. The purpose of this study was to obtain chemical and biological information on SCSL and to provide a deeper understanding of SCSL as a recoverable resource. As a result, 18 undescribed SNTs and 6 analogues were obtained (Fig. 1). Herein, we reported the separation, characterization, and neuroprotective activity of SNTs 1–24. In addition, we analyzed the interactions between growth factor receptor-bound protein 2 (GRB2) and compounds 2, 6, and 7 by molecular docking.

Close

Fig. 1

The structures of 1–24.

Export to PPT

2 Materials and methods

3 Results and discussion

Schinoxlactone A (1), a colorless needle crystal, possessed C₂₉H₃₆O₁₀ molecular formula based on HRESIMS at m/z 567.2195 ([M + Na]⁺, calcd. for C₂₉H₃₆O₁₀Na, 567.2206) revealed that 12 indices of hydrogen deficiency. The ¹H NMR spectroscopic data (Table 1) exhibited the existence of a secondary methyl (δ_H 1.09, Me-21), three tertiary methyls (δ_H 1.97, Me-27; 1.12, Me-18; 1.08, Me-29), one oxygenated methylene (δ_H 3.43, 3.37, H₂-30), three oxygenated methines (δ_H 4.01, H-1; 4.13, H-2; 3.82, H-15), two olefinic protons (δ_H 7.56, H-24; 5.74, H-22). The ¹³C NMR (Table 2) and HSQC spectra exhibited four methyls, six methylenes include one oxygenated, nine methines include two olefinic and three oxygenated, ten quaternary carbons include two olefinic, two ester carbonyl, and five oxygenated, indicating 1 was a highly oxygenated 14(13 → 12):16(17 → 13)-diabeoschiartane SNT.

Compared with NMR data of schisdilactone I (Li et al., 2013), it implicated a similar substructure of rings A–C, the change only involved the dioxygen quaternary carbon C-17 (δ_C 107.3) in the side chains (Zeng et al., 2014), which can be deduced from the HMBC correlations of H₃-18, H-20/C-17 (Fig. 2). Rings E and D were implied by the ¹H−¹H COSY correlations of H₂-11/H-12, coupled with the HMBC correlations of H-11β/C-8, C-9, C-14; H-15/C-12, C-13; H₃-18/C-12, C-16. In addition, the 9,17:15,17-diepoxy moiety was confirmed by 2 remaining unsaturation, and the NMR data C-9 (δ_C 88.5), C-15 (δ_C 81.2), and C-17 (δ_C 107.3), together with single-crystal X-ray diffraction analysis. Finally, ring D was fused with ring C by C-9 and C8, which can be established by the HMBC correlation of H-19α/C-11; H-11β, and H-19β/C-8.

Close

Fig. 2

Key HMBC () and ¹H−¹H COSY () correlations of compounds 1, 3, 8, and 15.

Export to PPT

The relative configuration of 1 was derived from the NOESY spectrum (Figure S7). Biogenetically, H-5 has been identified as α-orientated (He et al., 2020; Wang et al., 2021). The correlations of H-5/H₂-30, H-8; H-1/H₃-29 revealed that H₂-30, H-8 were α-orientated, and H-1, H₃-29 were β-orientated. As well, the correlations of H-12/H-15, H₃-18 suggested they shared the same orientation. Noteworthily, there no coupling constant was observed for the adjacent signals H-1 (δ_H 4.01, s) and H-2 (δ_H 4.13, s) in the ¹H NMR spectrum, indicating the dihedral angle was close to 90°, so that H-2 should be α-orientation (Lei et al., 2009; Shi et al., 2014). Besides, the E geometry of Δ²² was determined through the correlations of H-24/H-20 (Wang et al., 2013). X-ray diffraction analysis is a common method for confirming the absolute configuration of natural products. Finally, its absolute configuration was further established as 1R,2R,4R,5S,8R,9S,10R,12R,13R,14R,15S,17S,20R based on X-ray diffraction analysis (Flack parameter of 0.02(2)) (Fig. 5).

Schinoxlactone B (2) was isolated as gum, and the molecular formula C₂₉H₃₆O₁₀ of 2 was assigned by HRESIMS m/z 567.2179 ([M + Na]⁺, calcd. for C₂₉H₃₆O₁₀Na 567.2206) and the same as 1. Analysis of the NMR data (Tables 1 and 2) of 2 suggested it was very similar to 1, except for that minor difference of δ_C-21 (|Δ| = 1.1 ppm) and the configuration of Δ²². In the ECD spectrum (Fig. 3), the characteristic negative Cotton effect at 275 nm was opposite to 1, which indicates the configuration of C-20 as S. Besides, the NOESY correlation of H-22 (δ_H 5.74) and H-24 (δ_H 7.56) (Figure S7), indicating that the Z configuration of Δ²² (Wang et al., 2021). Thus, 2 was identified as the isomer of 1.

Close

Fig. 3

Experimental ECD spectra of 1–24.

Export to PPT

Schinoxlactone C (3), gum, the molecular formula C₂₉H₃₆O₁₀ assigned based on HRESIMS at m/z 567.2201 ([M + Na]⁺, calcd. for C₂₉H₃₆O₁₀Na, 567.2206), ascertained 12 indices of hydrogen deficiency. The NMR data (Tables 1 and 2) suggested the structure of 3 was very similar to 1. The differences in the structures of 1 and 3 were as follows: (1) The C-2 and C-17 positions of 3 were substituted with hydroxyl and carbonyl groups, respectively; which can be ascertained by the key HMBC correlations of H-2α/C-3, C-10; H-2β/C-3, and H-16α, H₃-18, H₃-21/C-17(δ_C 214.5) (Fig. 2). (2) 3 has an oxa-bridged hemiketal moiety between C-9 and C-15, which can be demonstrated by the NMR data of C-9 (δ_C 81.4), C-15 (δ_C 108.3), and the requirement of one remaining unsaturation. The relative configuration of 3 was established based on the NOESY spectrum (Figure S7). In the NOESY spectrum, the correlation of H-22 (δ_H 5.58) and H-24 (δ_H 7.30), indicating that the Z configuration of Δ²² (Wang et al., 2021). The diagnostic Cotton effect at 275 nm (positive) and 308 nm (negative) advised the configuration of C-20R (Fig. 3) (Huang et al., 2007; Huang et al., 2007; Huang et al., 2008). Furthermore, the absolute configuration 1R,4R,5S,8R,9S,10R,12R,13R,14R,15S,20R of 3 was deduced by X-ray diffraction analysis (Flack parameter of 0.07 (11)) (Fig. 5).

Schinoxlactone D (4) was obtained as gum. HRESIMS at m/z 567.2188 ([M + Na]⁺, calcd. for C₂₉H₃₆O₁₀Na 567.2206) implies that 4 has the same molecular formula as 3. Analysis of the NMR data (Tables 1 and 2) indicated 3 and 4 were similar in structure, except for the E configuration of Δ²², which was demonstrated by the NOESY correlation of H-20/H-24 (Figure S7) and the two singlets at δ_H-22 = 5.51 and δ_H-24 = 7.81 (Wang et al., 2013). In the ECD spectra (Fig. 3), C-20R configuration was deduced by intense Cotton effect at 275 nm (positive) and 308 nm (negative) the same as 3. Therefore, 4 was the Δ²² trans isomer of 3.

Schinoxlactone E (5) was isolated as gum. HRESIMS at m/z 583.2137 ([M + Na]⁺, calcd. for C₂₉H₃₆O₁₁Na, 583.2155) advised the molecular formula C₂₉H₃₆O₁₁, along with 12 indices of hydrogen deficiency. Comparing the NMR data (Tables 1 and 2) of 5 and 3 revealed the structure of 5 was a 7/5/5-fused carbocyclic core SNT and very similar to 3, and their differences only involved the hydrogen at the C-2 position of 5 was substituted by a hydroxyl group, which can be demonstrated the ¹H−¹H COSY correlation of H-1/H-2, as well as the HMBC correlation of H-2 (δ_H 4.21)/C-10 (Figure S6). The relative configuration of 5 was assigned based on the NOESY spectrum (Figure S7), and comparison with the NMR data of 3. In the NOESY spectrum, the correlation of H-22 (δ_H 5.16) and H-24 (δ_H 7.27) interpreted the existence of the Z configuration of Δ²² (Wang et al., 2021). In the ECD spectrum (Fig. 3), the characteristic cotton effect of 5 was close to 3 and 4, corresponding to the configuration of C-20R. Finally, its absolute configuration was further confirmed as 1R,2R,4R,5S,8R,9S,10R,12R,13R,14R,15S,20R due to calculated and experimental ECD (Fig. 4), and biosynthetic pathways.

Close

Fig. 4

Calculated ECD spectra of 5 and 6.

Export to PPT

Schinoxlactone F (6) was obtained as gum, HRESIMS at m/z 583.2148 ([M + Na]⁺, calcd. for C₂₉H₃₆O₁₁Na 583.2155) led to the molecular formula C₂₉H₃₆O₁₁, which indicated 12 indices of hydrogen deficiency. The similarity of the NMR data (Tables 1 and 2) of 6 and schisarisanlactone A, indicated 6 was a 14(13 → 12):16(17 → 13)-diabeoschiartane SNT and structurally similar to schisarisanlactone A (Lo et al., 2013), except for the chemical shifts of C-30 and C-2. C-30 and C-2 were confirmed as oxygenated methylene and oxygenated methine, supported by the ¹H−¹H COSY correlations of H-1/H₂-2, along with the HMBC correlation of H-2 (δ_H 4.08)/C-10, H₃-29/C-30 (δ_C 69.1) (Figure S6). The relative stereochemistry of 6 was elucidated by the NOESY spectrum (Figure S7) and the data of schisarisanlactone A. The correlation between H-5α/H₂-30; H-1/H₃-29 assigned H₂-30 as α-oriented, and H-1, H₃-29 as β-oriented. Also, the correlations of H-12/H-8, H₃-18 revealed that they shared the same orientation. The adjacent signals H-1 (δ_H 4.05, s) and H-2 (δ_H 4.08, s) interpreted H-2 was α-oriented the same as 1 (Lei et al., 2009; Shi et al., 2014). In the NOESY spectrum (Figure S7), the correlation of H-22 (δ_H 5.36) and H-24 (δ_H 7.31), indicating that the Z configuration of Δ²² (Wang et al., 2013). The S configuration of C-20 was deduced from the diagnostic Cotton effect at 275 nm (negative) and 308 nm (positive) (Fig. 3) (Huang et al., 2007; Huang et al., 2007; Huang et al., 2008). Finally, the structure of 6 was defined by calculated and experimental ECD (Fig. 4), and biosynthetic pathways.

Schinoxlactone G (7) was isolated as gum, HRESIMS at m/z 583.2141 ([M + Na]⁺, calcd. for C₂₉H₃₆O₁₁Na 583.2155) advised the molecular formula C₂₉H₃₆O₁₁ the same as 6. The NMR data (Tables 1 and 2) of 7 was very similar to 6 except for that minor difference of δ_C-21, (|Δ| = 1.2 ppm), which can be further demonstrated by the opposite characteristic Cotton effect at 275 nm and 308 nm in the ECD spectrum (Fig. 3). Thus, 7 was identified as the C-20 epimer of 6.

Schinoxlactone H (8), colorless needle crystal, the molecular formula C₂₉H₄₀O₁₀ was implied by HRESIMS at m/z 571.2501 ([M + Na]⁺, calcd for C₂₉H₄₀O₁₀Na 571.2519), corresponding to 10 indices of hydrogen deficiency. The NMR data (Tables 2 and 3) revealed that 8 was an 18(13 → 14)-abeoschiartanes SNT, and particularly similar to wuweizidilactone F (Huang et al., 2008) except that the presence of the hydroxy-substituted at C-2 and C-15, which was demonstrated by the ¹H−¹H COSY correlations of H-1/H-2 (δ_H 4.07); H-15 (δ_H 3.86)/H₂-16, together with the HMBC correlations of H-1/C-2 (δ_C 74.1); H₃-18/C-15 (δ_C 79.9) (Fig. 2). The relative configuration of 8 was deduced by the NOESY spectrum (Figure S7). The correlation of H-5α/H₂-30 and H-1/H₃-29 assigned H₂-30 as α-oriented, and H-1, H₃-29 as β-oriented. Moreover, the correlation of H₃-18/H-8, H-15 exhibited they were β-oriented, and the correlations of H₃-21/H-22 suggested that they shared the same orientation. Also, the α-oriented of H-2 was established by the adjacent signals H-1 (δ_H 4.09, s) and H-2 (δ_H 4.07, s) the same as 1 (Lei et al., 2009; Shi et al., 2014). The presence of C-23S and α,β-unsaturated-γ-lactone moiety led to a negative Cotton effect at 210 nm of the experimental ECD (Shi et al., 2014). Furthermore, the absolute configuration of 8 was demonstrated by X-ray diffraction analysis (Flack parameter of −0.09(7)) (Fig. 5).

Close

Fig. 5

The X-ray ORTEP drawing of 1, 3 and 8.

Export to PPT

Schinoxlactone I (9), white amorphous powder, its HRESIMS data (m/z 571.2480 [M + Na]⁺, calcd. for 571.2519) intimated the molecular formula C₂₉H₄₀O₁₀ the same as 8. The NMR data (Tables 2 and 3) indicate that 9 and 8 have the same planar structure. The conformation of 9 was determined by the NOESY spectrum (Figure S7), compared with 8, as well as biosynthetic considerations. The HMBC correlations (Figure S6) of H-5α/H₂-30; H₃-18/H-8, H-15, H-22 and H-1/H₃-29 assigned H₂-30 as α-oriented, and H-1, H-8, H-15, H-22, Me-18, Me-29 as β-oriented. Moreover, there were no correlations between H-22/H₃-21, H-17 assigned H-17, Me-21 as α-oriented. Also, the α-oriented of H-2 was confirmed by the adjacent signals H-1 (δ_H 4.09, s) and H-2 (δ_H 4.07, s) the same as 1 (Lei et al., 2009; Shi et al., 2014). In the ECD spectrum, the S configuration of C-23 was deduced by a characteristic negative Cotton effect at 210 nm the same as 8 (Fig. 3) (Shi et al., 2014). Thus, 9 was identified as the C-20R epimer of 8.

Schinoxlactone J (10) was obtained as gum, HRESIMS at m/z 571.2497 ([M + Na]⁺, calcd for C₂₉H₄₀O₁₀Na 571.2519) ascertained the molecular formula C₂₉H₄₀O₁₀ the same as 8. The NMR data (Tables 2 and 3) 8 and 10 intimated their structure was very similar, except that the presence of the hydroxy-substituted at C-7 and the absence of the hydroxy-substituted at C-30 in 10, which was confirmed by the HMBC correlations of H-5/C-7 (δ_C 73.7); H₃-30 (δ_H 1.26)/C-4, C-29, coupled with the ¹H−¹H COSY correlations of H₂-6/H-7 (δ_H 4.14)/H-8. The relative configuration of 10 was confirmed based on the NOESY spectrum (Figure S7), and comparison with the NMR data of 8. The correlations of H-5α/H-7 assigned H-7 as α-oriented. The similarity diagnostic ECD spectra ascertained the absolute configuration C-23 was S for 10 and 8. Thus, the structure of 10 was defined.

Schinoxlactone K (11) was obtained as gum, the molecular formula C₂₉H₄₀O₉ deduced by HRESIMS at m/z 555.2543 ([M + Na]⁺, calcd. for C₂₉H₄₀O₉Na, 555.2570) implied 10 indices of hydrogen deficiency. The NMR spectrum (Tables 2 and 3) exhibited 11 was an 18(13 → 14)-abeoschiartanes SNT and possessed six rings. 11 and 19 have the same molecular formula, and a comparison of the NMR data can be inferred that 11 was structurally similar to 19. As shown in Figure S6, the ¹H–¹H COSY correlation of H-1/H-2, along with the HMBC correlations of H-2 (δ_H 4.07)/C-10; H₃-30 (δ_H 1.29)/C-4, C-5, suggested that C-2 was oxygenated methine and C-30 was methyl, respectively. The relative stereochemistry of 11 was elucidated by the NOESY spectrum (Figure S7), and a comparison with the NMR data of 19. The correlations of H-5α/H₂-30; H-1/H₃-29; and H₃-18/H-8, H-15 assigned H₂-30 as α-oriented, and H-1, H-8, H-15, Me-18, Me-29 as β-oriented. The configuration of H-2α was also demonstrated by H-1 (δ_H 3.97, s) and H-2 (δ_H 4.07, s) the same as 1 (Lei et al., 2009; Shi et al., 2014). The similarity ECD spectra deduced to the same absolute configuration of C-23S for 11 and 19. Combining the biosynthetic pathway and the above information, the structure of 11 was defined.

Schinoxlactone L (12) was isolated as gum, and the molecular formula C₃₁H₄₂O₁₀ was demonstrated by HRESIMS at m/z 597.2642 ([M + Na]⁺, calcd. for C₃₁H₄₂O₁₀Na, 597.2676), advised 11 indices of hydrogen deficiency. The NMR spectrum (Tables 2 and 3) of 12 exhibited 31 carbon resonances, indicating 12 was an 18(13 → 14)-abeoschiartanes SNT and the presence of an acetoxy group (δ_H 2.09; δ_C 21.4, 171.0). Comparison of the NMR data of 12 with kadnanolactone H (Gao et al., 2008) leads to the assumption that 12 was structurally similar to kadnanolactone H, with the difference being the substituents on C-2, C-15, and C-30. As shown in Figure S6, the ¹H–¹H COSY correlation of H-1/H-2, combined with the HMBC correlations of H-2 (δ_H 4.06)/C-10; H₃-30 (δ_H 1.29)/C-4, C-5, it was presumed that C-2 was oxygenated methine and C-30 was methyl, respectively. Furthermore, the acetoxy group attached to C-15 was supported by the HMBC correlations of H-15 (δ_H 4.86), H₃-31 (δ_H 2.09)/C-32 (δ_C 171.0). Its relative stereochemistry was identical to the NOESY correlations of kadnanolactone H. Drawing on the experience of 1, H-2 was also assigned to the α-orientation (Lei et al., 2009; Shi et al., 2014). The diagnostic negative Cotton effect at 210 nm ascertained C-23S of 12 the same as 11 and 19. Combining the biosynthetic pathway and the above information, the structure of 12 was defined.

Schinoxlactone M (13) was isolated as a colorless oil. Its molecular formula was confirmed as C₂₈H₃₆O₁₀ by HRESIMS at m/z 555.2186 ([M + Na]⁺, calcd. for C₂₈H₃₆O₁₀Na, 555.2206), suggested 11 indices of hydrogen deficiency. 13 can be confirmed as an 18-norschiartane SNT by the 28 carbon resonances in the NMR spectrum (Tables 2 and 3) and possessed seven rings. It was structurally similar to wuweizidilactone O (Shi et al., 2014), the difference was the substituents on C-12 and C-19. The HMBC correlations of H-19 (δ_H 3.37)/C-5, C-8; H-11/C-19 (δ_C 76.7); H-12 (δ_H 3.66)/C-9, C-14, combined with the ¹H–¹H COSY correlations of H₂-11/H-12 (Figure S6), it was presumed that C-19 and C-12 were attached with a hydroxyl group, respectively. According to NOESY spectral analysis, the relative stereochemistry of 13 can be determined the same as wuweizidilactone O. The correlations of H-5α/H₃-30; H-1/H₃-29; H-8/H-7, H-19; H-12/H-17 assigned H-8 and H-19 have the same orientation, and H-12 and H-17 have the same orientation (Figure S7). In the ECD spectrum, the C-23S was deduced from a characteristic negative Cotton effect at 210 nm the same as wuweizidilactone O (Fig. 3). So, the structure of 13 was confirmed.

Schinoxlactone N (14) was isolated as gum, HRESIMS at m/z 615.2758 ([M + Na]⁺, calcd. for C₃₁H₄₄O₁₁Na, 615.2781) implied the molecular formula C₃₁H₄₄O₁₁ and requiring 10 indices of hydrogen deficiency. The NMR spectrum (Tables 2 and 3) interpreted that 14 was a schiartanes SNT and the existence of an acetoxy group (δ_H 2.14; δ_C 21.7, 171.2). The data of 14 and 20 were similar, their difference was the substituent on C-15. It was presumed that 14C-15 has an acetoxy by the HMBC correlations of H-15 (δ_H 4.90), H₃-31 (δ_H 2.14)/C-32 (δ_C 171.2) (Figure S6). Comparing the NOESY spectra of 14 and 20, it was clear that they have the same relative configuration. In the ECD spectrum., the C-23S was deduced by a characteristic negative Cotton effect at 210 nm the same as 20 (Fig. 3). So, the structure of 14 was demonstrated.

Schinoxlactone O (15) was isolated as gum, and possessed a C₂₉H₃₆O₁₁ molecular formula based on HRESIMS at m/z 583.2137 ([M + Na]⁺, calcd. for C₂₉H₃₆O₁₁Na, 583.2155) implied that 12 indices of hydrogen deficiency. Comparing the NMR data (Table 4) of 15 with schisdilactone V (Yang et al., 2022) leads to the assumption that they were structurally similar. The ¹H–¹H COSY correlations of H-8/H₂-7/H₂-6/H-5, along with The HMBC correlation of H-8 (δ_H 3.37)/C-9 (Fig. 2), indicated that the C-8 hydroxyl group of schisdilactone V was reduced to produce 15. Its relative stereochemistry was due to the NOESY correlations (Figure S7) and the NMR data of kadnanolactone H. According to the NOESY spectrum, H-5/H₂-30, H-8, and H-22 (δ_H 5.34)/H-24 (δ_H 7.30) (Figure S7), implies that H-8, H₂-30 were α-oriented, and the Z configuration of Δ²² (Wang et al., 2013). In the ECD spectra (Fig. 3), the configuration of C-20S was implied by intense Cotton effect at 275 nm (negative) and 308 nm (positive) (Fig. 3) (Yang et al., 2022). Combined with biosynthetic considerations, the structure of 15 was defined.

Schinoxlactone P (16) was isolated as gum, and the molecular formula was demonstrated as C₂₉H₃₄O₁₂ by HRESIMS at m/z 597.1921 ([M + Na]⁺, calcd. for C₂₉H₃₄O₁₂Na, 597.1948), indicating 13 indices of hydrogen deficiency. The NMR spectral data (Table 4) suggested that 16 had a ternary oxygen ring (δ_H 3.67; δ_C 63.4, 60.4), and was similar to schisdilactone V (Yang et al., 2022). The ternary oxygen ring at C-7 and C-8 was demonstrated by the HMBC correlation of H-19β/C-8 (δ_C 60.4), combined with the ¹H–¹H COSY correlation of H-5/H₂-6/H-7 (δ_H 3.67) (Figure S6). According to NOESY spectral analysis (Figure S7) and compared to schisdilactone V NMR data, the relative stereochemistry of 16 can be established. The NOESY correlations of H-5/H₂-30, H-7, and H-22 (δ_H 5.34)/H-24 (δ_H 7.30) (Figure S7), implies that H-7, H₂-30 were α-oriented, and the Z configuration of Δ²² (Wang et al., 2013). The configuration of C-20S was confirmed by ECD spectrum the same as 15 (Fig. 3) (Yang et al., 2022). Combined with biosynthetic considerations, the structure of 16 was defined.

Schinoxlactone Q (17) was obtained as gum, and had a molecular formula C₂₉H₄₀O₁₀ demonstrated by HRESIMS at m/z 571.2490 ([M + Na]⁺, calcd. for C₂₉H₄₀O₁₀Na, 571.2519), which displayed 10 indices of hydrogen deficiency. The NMR data (Table 4) of 17 and 23 leads to the assumption they were structurally similar except for the unsaturation and the substituents of C-4 and C-30. C-14 was methylene in 17 established by the HMBC correlations (Figure S6) of H-14β (δ_H 2.50)/C-15, C-16. Furthermore, the ¹H–¹H COSY correlations of H₂-7 (δ_H 3.64, 3.55)/H₂-6/H-5, and the HMBC correlation of H-8 (δ_H 3.98)/C-15, combined with the remaining unsaturation was 0, which indicated the carbon bond C-7 and C-8 were oxidation and breakage. Besides, C-30 should be methylene demonstrated by the HMBC correlations of H₃-30 (δ_H 1.23)/C-4, C-5, C-29. According to the NOESY spectrum (Figure S7), H-5/H₃-30; H-7α/H-5, H-8; H₃-29/H-1, and the correlation H-20/H-24, implies that H-8, Me-30 were α-oriented, H-1, Me-29 were β-oriented, and the E configuration of Δ²² (Wang et al., 2013). The R configuration of C-20 was deduced by the characteristic ECD spectrum the same as 23 (Fig. 3). So, the structure of 17 was confirmed.

Schinoxlactone R (18), gum, HRESIMS at m/z 581.1979 ([M + Na]⁺, calcd. for C₂₉H₃₄O₁₁Na, 581.1999), corresponding to the molecular formula C₂₉H₃₄O₁₁, indicating 13 indices of hydrogen deficiency. A comparison of the data of 18 with lanciformin E (Shi et al., 2014) shows that they were structurally similar except for the unsaturation and the substituents of C-7, C-19, and C-30. The ¹H–¹H COSY correlations H-5/H₂-6/H-7 (δ_H 3.54), combined with the HMBC correlations of H₂-19 (δ_H 2.17)/C-5, C-8; H₃-29/C-30 (δ_C 67.5), inferred C-19 and C-30 were methylene, C-7 was methine (Figure S6). The HMBC correlations of H₂-19 (δ_H 2.17)/C-5, C-8; H₃-29/C-30 (δ_C 67.5), combined with the ¹H–¹H COSY correlations H-5/H₂-6/H-7 (δ_H 3.54) (Figure S6), inferred C-19 and C-30 were methylene, C-7 was methine. The HMBC correlations of H₂-19, H-11β, H-16β/C-8 (δ_C 68.3); H-16β/C-14 (δ_C 74.0), and combined with the remaining one unsaturation, it was assumed that C-8 and C-14 form a ternary oxygen ring. The relative configuration of 18 was compared to lanciformin E and demonstrated by the NOESY spectrum (Figure S7). According to the NOESY correlations of H-5/H₂-30, H-7; H-1/H₃-29, and H-22 (δ_H 5.27)/H-24 (δ_H 7.31), which implies that H-7, H₂-30 were α-oriented, H-1, Me-29 were β-oriented, and the Z configuration of Δ²² (Wang et al., 2013). The S configuration of C-20 was determined by the characteristic ECD spectrum the same as lanciformin E (Fig. 3) (Shi et al., 2014). Combined with biosynthetic considerations, the structure of 18 was defined.

Based on the NMR data and literature, other SNTs were identified to be 2β-hydroxy-micrandilactone C (19) (Wang et al., 2011), kadcoccilactone E (20) (Mi et al., 2020), shisdilactone R (21) (Wang et al., 2021), shisdilactone S (22) (Wang et al., 2021), shisdilactone N (23) (Wang et al., 2021), and shisdilactone M (24) (Wang et al., 2021).

Neurodegenerative diseases, including Alzheimer's disease, Huntington's disease, multiple sclerosis, amyotrophic lateral sclerosis, and Parkinson's disease, are becoming prevalent conditions in society. According to a new report by Alzheimer's Disease International, the number of patients is projected to be as high as 152 million in 2050 (Zhou et al., 2020). To date, neurodegenerative diseases remain an unsolved problem for modern therapies, with their high morbidity and mortality associated with neuronal death and synaptic loss in different regions of the brain (Golpich et al., 2016). PC12 cells are cloned from transplantable rat pheochromocytomas with both neurosecretory cell and neuronal traits, and studies have been conducted to use the PC12 cell line for drug screening, neuronal cell differentiation, ion channels, receptors, and transmitter secretion in a variety of studies (Su et al., 2013). In the treatment of neurodegenerative diseases, natural products derived from medicinal plants have attracted the attention of researchers due to their few side effects, low toxicity, and good compliance. Recently, Prof. Sun and Puno found that SNTs with different carbon skeletons exhibited good neuroprotective activity, and the authors called for more attention to the neuroprotective activity of SNTs (He et al., 2020; Wang et al., 2021). In this study, 24 SNTs of 7/5/5, 7/6/5, and 7/8 carbon skeleton were isolated from SCSL, and the neuroprotective effect of compounds 1–24 was evaluated using PC12 cells to discover lead compounds for the treatment of neurodegenerative diseases and evaluate their structure activity relationship. As shown in Table 5, compounds 2, 6, and 7 showed strong neurite outgrowth-promoting activity with 13.1 %, 12.0 %, and 12.2 % cell differentiation rate (positive group: 15.4 %), respectively. Compounds 1, 2, and 11 at a concentration of 25 μM also had a certain neuroprotective effect on CORT-induced PC12 cell injury, the cell viability was increased by 21.1 %, 19.5 %, and 24.4 % (positive group: 30.6 %), respectively (Fig. 6).

Close

Fig. 6

Neuroprotective activities of 1–24 (25 μM) against CORT-induced injury of PC12 Cells. DIM (12.5 μM) was used as positive control (Results are expressed as the mean ± SD (n = 3). ^***p < 0.001, ^**p < 0.05).

Export to PPT

To explore the possible mechanism of action of these compounds on neurite outgrowth. We searched for 103 targets in the Swiss Target Prediction database (David et al., 2014) with the most active compound 2 and retrieved 3586 targets in GeneCards (Barshir et al., 2021) with Neurite outgrowth as the keyword. The 50 target proteins shared by compound 2 and Neurite outgrowth were imported into the STRING 11.0 (Damian et al., 2019) database, and a PPI network analysis was performed. In the KEGG pathway in the model of analysis, we screened the neurotrophin signaling pathway. When the degree statistics were performed on all the targets, it was found that the first three target protein degree were more significant and were in the Neurotrophin signaling pathway. We then performed molecular docking screening on the first three target proteins and found that the docking binding energy of the compound and the protein GRB2 was much better than that of HRAS and MAPK1.

GRB2 protein contains one SH2 domain and two SH3 domains. Neurite outgrowth was markedly suppressed by a cell-permeable peptide inhibitor of PD 098059 and GRB2, a medication that can prevent MEK1 activation in various cell types (Drosopoulos et al., 1999). To elucidate the potential mechanism, compounds 2, 6, and 7 were docked with GRB2 (PDB ID: 7MPH) (Xiao et al., 2021). The hydroxyl-2, hydroxyl-14, hydroxyl-30, and carbonyl-3 of 2 shaped hydrogen bonds with the key amino acid residues LYS-69, ASN-143, ASP-104, and HIS-107, according to the docking results (Fig. 7D). The key amino acid residues SER-141, ASN-143, TRP-60, and PHE-61 form hydrogen bonds with the hydroxyl-2, carbonyl-3, and carbonyl-26 of 6 shaped hydrogen bonds (Fig. 7E). The hydroxyl-14, hydroxyl-2, carbonyl-3, and carbonyl-26 of 7 formed hydrogen bonds with LYS-109, SER-141, ASN-143, and TRP-60 (Fig. 7F).

Close

Fig. 7

(A) The Venn results of potential target genes of compound 2 therapy for Neurite outgrowth. (B) The PPI network map of 50 target genes (genes in red are associated with the Neurotrophin signaling pathway). (C) Count and list the top 30 genes of PPI network map. (D–F) Molecular binding mode of compounds 2, 6, and 7 with GRB2.

Export to PPT

4 Conclusions

In conclusion, eighteen (1–18) undescribed highly oxygenated and rearranged SNTs and six analogues (19–24), were isolated from SCSL and evaluated for their neuroprotective activity. Their diverse structures included 14(13 → 12):16(17 → 13)-diabeoschiartane (1–7), 18(13 → 14)-abeoschiartanes (8–12, 19), 18-norschiartane (13), schiartanes (14, 20), 16,17-secopreschisanartanes (15–17, 21–24), lancifoartane (18) skeletons. Compounds 1–7 feature a rare 7/5/5-fused carbocyclic core. Their structures were elucidated by HRESIMS, NMR, ECD, single-crystal X-ray diffraction, and biogenetic considerations. Compounds 2, 6, and 7 showed strong neurite outgrowth-promoting activity, and compounds 1, 2, and 7 exhibited moderate neuroprotective activities against CORT-induced PC12 cell injury. The structure–activity relationship of SNTs with 7/5/5 and 7/6/5 carbon skeletons was analyzed. For SNTs with a carbon backbone of 7/5/5, the against CORT-induced injury was enhanced when the carbonyl group at the C-17 position formed a 9,17:15,17-diepoxy moiety structural fragment, and for SNTs with a carbon backbone of 7/6/5, the against CORT-induced injury was enhanced when the hydroxyl group at the C-15 position was replaced by an acetoxy group. SNTs with a 5/5/7/5/5/5/6/5 thickened heptacyclic structure had better neurite outgrowth-promoting activity in the C-20S configuration than the C-20R configuration. PPI network and molecular docking analysis suggest that compounds 2, 6, and 7 can exert neuroprotective effects through the regulation of GRB2. The discovery of the neuroprotective activity SNTs could benefit the further utilization and development of SCSL. Therefore, SNTs, as important physiologically active components in SCSL, are preferred ingredients for the development of SCSL bioactivities.