Five undescribed plant-derived bisphenols from Artemisia capillaris aerial parts: Structure elucidation, anti-hepatoma activities and plausible biogenetic pathway

2.1 General experimental procedures

Nuclear magnetic resonance (NMR) spectra were recorded on a Bruker DPX-400 (¹H NMR, 400 MHz; ¹³C NMR, 100 MHz) spectrometer, with tetramethylsilane (TMS) as the internal standard. Unless otherwise specified, chemical shifts (δ) were quoted in ppm with reference to the residual solvent signal (DMSO‑d₆: δ_H 2.50 ppm, δ_C 39.52 ppm). Positive- or negative-ion HR-ESI-TOF-MS data were measured using a Bruker microTOF-QII mass spectrometer (Bruker, Karlsruhe, Germany). Analytical HPLC was carried out on a Shimadzu LC-16 instrument, and UV detection was carried out using a YMC-Pack ODS-A C18 (4.6 × 250 mm, 5 μm) column. All samples were purified by preparative HPLC using a Shimadzu HPLC system equipped with a UV detector and a Cosmosil 5C₁₈-MS-II column (20 × 250 mm, 5 μm, Nacalai Tesque, Inc., Nijo Karasuma, Japan). Silica gel (100–200, 200–300, and 300–400 mesh, Qingdao Marine Chemical Factory, Qingdao, China), ODS (50 μm, ODS-A-HG, YMC Co. ltd., Japan), and Sephadex LH-20 (GE Healthcare Bio-Sciences AB, Sweden) were used for open column chromatography (CC) to separate samples, and thin-layer chromatography (TLC) analysis were carried out using precoated silica gel GF₂₅₄ plates. Solvents were of analytical grade (Guanghua Chemical Co., ltd. Guangzhou, P.R. China) for open CC and HPLC grade (Merck, Germany) for HPLC analysis.

2.2 Plant material

The dried aerial parts of A. capillaris Thunb. were gathered in Ganzhou, Jiangxi Province, China, in August 2019 and authenticated by Dr. Xiaobin Zeng of Shenzhen People’s Hospital. A voucher specimen (No. 20190805) has been deposited in the Center Lab of Longhua Branch, Shenzhen People’s Hospital, Second Clinical Medical College of Jinan University, Shenzhen, China.

2.3 Extraction and isolation

The air-dried aerial parts of A. capillaris (7.0 kg) were powered and extracted four times (5 days each) using 80 % EtOH at room temperature. A dark-brown crude extract (1.4 kg, 20 %) was afforded following concentration under reduced pressure. The extract was suspended in water (2 L) and partitioned successively with petroleum ether, ethyl acetate, and n-BuOH. The ethyl acetate fraction (114 g) was separated by ODS column chromatography, eluted with a MeOH − H₂O (5–100 %) gradient, and yielded ten fractions. Various column chromatographic separations of the 10 fractions afforded compounds 1–35. The n-BuOH fraction (200 g) was first separated by silina gel column chromatography, eluted with a CH₂Cl₂-MeOH (5–100 %) gradient, and yielded six fractions. Various column chromatographic separations of FB. 3 afforded compounds 36–38. Fig. 2 shows the extraction and isolation procedures for compounds from the aerial parts of A. capillaris. The purity of these compounds was determined as to be more than 95 % by HPLC.

Close

Fig. 2

Extraction and isolation procedure of compounds from A. capillaris aerial parts.

Export to PPT

2.4 Anti-Hepatoma activity assay in vitro

Huh 7 cells and HepG 2 cells were maintained in DMEM containing 10 % FBS (foetal bovine serum) and cultured at 37 °C (5 % CO₂, 95 % relative humidity). A cytotoxicity assay was performed, according to the CCK8 method, using 96-well microplates. Briefly, 200 μL of adherent cells were seeded into each well of the 96-well cell culture plates and allowed to adhere for 24 h before drug addiction, with an initial density of 5 × 10³ cells each well. Each tumor cell line was exposed to the test compounds at different concentrations for six times within 48 h.

3.1 Structural elucidation of five new bisphenols (13, 15, 25, 29, 31)

Compound 13 was isolated as white amorphous powder. Its molecular formula was determined as C₁₄H₁₂O₅ by the HR-ESI-TOF-MS at m/z 283.0653 [M + Na]⁺. The UV maximum absorptions at 220, 261, and 297 nm indicated the existence of aromatic moieties. In the ¹H NMR spectrum of 13, characteristic signals corresponding to a 1,2,4,5-tetrasubstituted benzene ring [δ_H 6.52 (s, H-3) and 7.31 (s, H-6)], a 1,4-disubstituted benzene ring [δ_H 6.91 (2H, d, J = 8.2 Hz, H-2′/H-6′) and 6.63 (2H, d, J = 8.2 Hz, H-3′/H-5′)], and methylene [δ_H 4.10 (2H, s, H₂-7′)] were observed. The ¹³C NMR spectrum of 13 revealed the presence of 14 carbon signals, which were assigned based on the DEPT-135 spectrum to a carbonyl group [δ_C 168.3 (C-7)], eight aromatic carbons [δ_C 120.0 (C-1), 135.8 (C-2), 118.2 (C-3), 148.9 (C-4), 142.8 (C-5), 118.1 (C-6), 132.0 (C-1′), 129.6 (C-2′/C-6′), 115.0 (C-3′/C-5′), and 155.3 (C-4′)], and a methylene carbon [δ_C 37.1 (C-7′)]. Based on the comprehensive analysis of ¹H–¹H COSY, HSQC, and HMBC spectra, the ¹H and ¹³C NMR spectral data of 13 were assigned (Table 1).

Interpretation of the ¹H–¹H COSY correlations (Fig. 3) led to the assignment of two spin-coupling systems (H-2′ to H-3′ and H-5′ to H-6′) in 13. In the HMBC spectrum, correlations between H-3 and C-1/C-5/C-7′, as well as between H-6 and C-2/C-4/C-7, allowed the establishment of a 1,2,4,5-tetrasubstituted benzene ring moiety (a), in which C-4 and C-5 were oxygenated due to its obvious downfield shift (δ_C 148.9 and 142.8). Meanwhile, HMBC correlations between H-2′/H-6′ and C-4′ and between H-3′/H-5′ and C-1′, in which C-4′ was oxygenated due to its obvious downfield shift (δ_C 155.3), indicated the existence of a p-hydroxybenzene unit (b). Furthermore, the HMBC correlations between H-3/H-2′/H-6′ and C-7′ as well as between H₂-7′ and C-1/C-3/C-2′/C-6′ implied that two fragments (a and b) were connected by methylene (C-2 − C-7′−C-1′ bond). Moreover, the remaining carbonyl and hydroxyl groups could be assigned to a carboxyl moiety based on molecular formula information, and this moiety was attached to the C-1 position on the basis of the HMBC correlation between H-6 and C-7. Hence, the planar structure of 13 was confirmed (Fig. 3) by the analysis, and compound 13 was identified as 2-C-p-hydroxybenzylprotocatechuic acid, which was renamed capillarisenol A.

Close

Fig. 3

Key ¹H–¹H COSY, HMBC and NOESY correlations of five new bisphenols (13, 15, 25, 29, 31) and one new aromatic compound 32.

Export to PPT

Compound 15 was obtained as an off-white amorphous powder. Its molecular formula was determined to be C₁₅H₁₄O₅ according to the HR-ESI-TOF-MS data at m/z 274.2749 [M + H]⁺. The UV spectrum of 15 showed the absorption maxima at 222, 259, and 296 nm. In the ¹H and ¹³C NMR spectra of 15, a 1,2,4,5-tetrasubstituted aromatic fragment, a 1,4-disubstituted aromatic fragment, a carboxyl group, a methylene, and a methoxyl group were observed. Extensive analysis of the 1D- and 2D NMR data (Table 1) of 15 indicated that its skeletal structure was identical to that of 13, and the difference between them is that the hydroxyl group (δ_C 148.9, C-4) in 13 was replaced by a methoxyl group (δ_C 150.4, C-4; δ_C 55.5, C-8) in 15, which was supported by the HMBC correlations from H₃-8 (s, δ_H 3.75) to C-4 (Fig. 2). Accordingly, compound 15 was determined to be 2-C-p-hydroxybenzylisovanillic acid, which was renamed capillarisenol B.

Compound 25 was isolated as yellow oil. Its molecular formula was established as C₁₇H₂₀O₂ based on its negative HR-ESI-TOF-MS data (m/z 255.1310 [M - H]^–). The UV maximum absorptions at 204, 225, and 279 nm supported the presence of aromatic moieties. The ¹H NMR spectrum of DMSO‑d₆ indicated a 1,2,4,5-tetrasubstituted benzene ring due to the presence of two aromatic proton singlets at δ 6.67 (1H, s) and 6.77 (1H, s) and a 1,4-disubstituted benzene ring due to aromatic protons at δ 6.86 (2H, d, J = 8.4 Hz) and 6.64 (2H, d, J = 8.4 Hz), one two-proton singlet at δ 3.73 (2H, s), and one three-proton singlet at δ 2.04 (3H, s) assignable to the methylene and methyl groups likely attached to the aromatic ring, respectively, as well as two overlapping three-proton doublets at δ 1.00 (6H, d), assignable to two methyl groups likely attached to the methine group. Furthermore, the 17 carbons were resolved as 14 carbon signals in the ¹³C NMR, confirming the presence of structurally symmetric subunits in the compound. Based on 1D- and 2D NMR experiments, the ¹H- and ¹³C NMR data of 25 were assigned and listed (Table 3).

The ¹H–¹H COSY correlations led to the establishment of three spin-coupling systems in 25, as shown in blue bold lines in Fig. 3. In the HMBC spectrum, key correlations between H-3 and C-1/C-10, between H-6 and C-2/C-4/C-7, between H-7 and C-6, between H₃-8/H₃-9 and C-5, and between H₃-10 and C-1/C-3 established a carvacrol motif (a). In addition, the HMBC correlations between H-2′/H-6′ and C-4′ and between H-3′/H-5′ and C-1′ suggested the existence of an p-hydroxybenzene unit (b), in which C-4′ was oxygenated due to its obvious downfield shift (δ 155.2). Furthermore, the HMBC correlations between H-3/H-6′ and C-7′ as well as between H₂-7′ and C-2′/C-6′ implied that two substructures (a and b) were also connected through the C-4 − C-7′−C-1′ bond. Consequently, compound 25 was identified as 4-C-p-hydroxybenzylcarvacrol, which was renamed capillarisenol C.

The molecular formula of compound 29 was determined to be C₁₇H₂₀O₂ based on HR-ESI-TOF-MS analysis (m/z 255.1309 [M - H]^–), which was indicative of eight degrees of unsaturation. The UV spectrum of 29 showed the absorption maxima at 204, 226, and 279 nm. The 1D- and 2D NMR data (Table 3) of 29 resembled those of 25, and one of the major differences was the replacement of a methyl group at C-2 in 25 by an isopropyl motif (δ_C 26.1, C-7; δ_C 22.6, C-8/C-9) at the same position in 29, as supported by the HMBC correlations from H₃-8/H₃-9 (δ_H 1.11) to C-2 and from H-3 (δ_H 6.84) to C-7 as well as a COSY correlation of H-7 (δ_H 3.11) to H₃-8/H₃-9. The other major difference was the replacement of an isopropyl motif at C-5 in 29 by a methyl group (δ_C 18.9, C-10) at the same position in 29, which was confirmed by the HMBC correlations from H₃-10 (δ_H 2.02) to C-4 and C-6. The planar structure of 29 was further confirmed with the aid of 2D NMR spectroscopic analysis (Fig. 3). Thus, compound 29 was assigned to be 4-C-p-hydroxybenzylthymol, which was renamed capillarisenol D.

The molecular formula of compound 31 was determined to be C₂₅H₂₂O₉ by its HR-ESI-TOF-MS at m/z 465.1051 [M - H]^–. The UV absorption at 215 and 267 nm implied the existence of an aromatic ring in 31. The ¹³C NMR spectrum of 31 revealed that this compound contained 25 carbons, which consisted of two methylene, one methine, two olefinic, eighteen aromatic, and two carbonyl carbons. In the ¹H NMR spectrum, two double doublets at δ_H 2.87 (dd, J = 8.7, 14.4 Hz) and 2.98 (dd, J = 3.4, 14.4 Hz) were attributed to the geminal protons of the benzylic methylene, which were coupled with a double doublet at δ_H 4.98 (dd, J = 3.4, 8.7 Hz), assignable to the oxymethine proton. The characteristic ABX-type aromatic proton signals at δ_H 6.49 (dd, J = 1.1, 8.1 Hz), 6.62 (d, J = 8.1 Hz), and 6.67 (d, J = 1.1 Hz) suggested the presence of a 1,3,4-trisubstituted aromatic ring. These data indicated that compound 31 had a similar structure to rosmarinic acid (compound 17) (Petersen and Simmonds, 2003). In comparison with the ¹H NMR spectrum of rosmarinic acid, two doublets at δ_H 6.11 and 7.81 (J = 15.7 Hz) and two singlets at δ_H 6.60 and 7.10 were observed in the spectrum of 31 instead of the characteristic ABX-type aromatic proton signals in that of rosmarinic acid. Based on this finding, a 1,2,4,5-tetrasubstituted aromatic ring of trans-caffeoyl moiety was present in the molecule of 31 in place of the 1,3,4-trisubstituted one of rosmarinic acid, which was also supported by λ^max at 288 and 340 nm in the UV spectrum as well as by ¹³C NMR data (Table 2). In addition, a pair ortho-coupling protons were observed at δ_H 6.65 and 6.87, accompanied by diphenyl methylene protons at δ_H 3.81. Therefore, it was suggested that a 4-hydroxybenzylic moiety was attached to C-6 of rosmarinic acid, which was further supported by the presence of signals at δ_C 36.4, 115.2 (2C), 129.2 (2C), 131.1, and 155.4 in ¹³C NMR spectrum and by the HMBC correlations of H-7′’ to C-1/C-5 and H-5 to C-7′’ (Fig. 2). Compound 31 showed a specific rotation of + 76.7° (in methanol), which was the same behaviour as that of rosmarinic acid ([α]_D²⁵: + 70.8°). Therefore, it was proposed that the stereochemistry at C-8′ was R configuration. Thus, compound 31 was deduced to be 6-C-p-hydroxybenzylrosmarinic acid, which was renamed capillarisenol E.

3.2 Structural elucidation of another new aromatic compound 32

Compound 32 was obtained as yellow oil, with the molecular formula C₁₉H₂₂O₇ based on its HR-ESI-TOF-MS at m/z 385.1203 [M + Na]⁺. The UV absorption at 215 and 267 nm of 32 indicated the presence of an aromatic ring. The ¹H NMR spectrum of 32 exhibited six aromatic [δ_H 7.20 (2H, s, H-2 and H-6), 7.37 (2H, d, J = 8.5 Hz, H-2′ and H-6′), 6.92 (2H, d, J = 8.5 Hz, H-3′ and H-5′)], one methine [δ_H 4.88 (1H, m, H-7′)], one methylene [δ_H 4.25 (2H, m, H-8′)], and four methoxyl [δ_H 3.81 (6H, s, H₃-8 and H₃-10), 3.74 (3H, s, H₃-9′), 3.73 (3H, s, H₃-9)] proton signals. The ¹³C NMR and DEPT-135 spectra showed 19 carbon resonances shared by twelve aromatic, four methoxyl, one carbonyl, one methine, and one methylene carbons. The ¹H- and ¹³C NMR spectral data of 32 are listed in Table 4, and the signal assignments were determined by the HSQC technique. The ¹H–¹H COSY correlations led to the establishment of three spin-coupling systems in 32 (blue bold lines in Fig. 3). Combined with ¹H NMR, ¹³C NMR, and HRESIMS, one hydroxyl group and one carbonyl group were found in compound 32. The hydroxyl group was attached to C-7′, which was confirmed by the chemical shifts of the methine carbon of C-7′ (69.8) as well as the HMBC correlation of H-2′/C-7′. One carbonyl group was located at C-1, which was confirmed by the HMBC correlation of H-2/C-7. Four methoxyl groups were attached to C-3, C-4, C-5, and C-4′ based on the HMBC correlations of H-8/C-3, H-9/C-4, H-10/C-5, and H-9′/C-4′, respectively. To further determine the absolute configuration at C-7′, the molecular optical rotation value of compound 32 were compared with that of R-suspensaside and S-suspensaside (Guo, et al., 2007). Compound 32 showed a specific rotation of −4.6° (methanol), which was nearly equal to that of S-suspensaside ([α]²⁵_D = -4.7°). This method was successfully applied previously for the configuration confirmation (Ge, et al., 2018; Ma, et al., 2016; Xiao, et al., 2022). Therefore, it was proposed that the stereochemistry at C-7′ was S configuration. Based on the analysis, the structure of 32 was thus determined to be (7′S)-2-hydroxy-2-(4-methoxyphenyl)ethyl3,4,5-trimethoxy benzoate.

The other known compounds were identified as protocatechuic acid (1) (Li, et al., 2022), protocatechuyl aldehyde (2) (Li, et al., 2005), 4-hydroxybenzoic acid (3) (Zhang, et al., 2021), 4-hydroxybenzaldehyde (4) (Zhang, et al., 2021), 3-hydroxy-4-methoxybenzoic acid (5) (Yang, et al., 2017), 3-hydroxy-4-methoxy-benzaldehyde (6) (Yang, et al., 2017), caffeic acid (7) (Ge, et al., 2018), 3,5-dimethoxy-4-hydroxybenzoic acid (8) (Feng, et al., 2011), cumaric acid (9) (Ge, et al., 2018), 3,4,5-trimethoxybenzoic acid (10) (Feng, et al., 2011), 3,4-dihydroxybenzoic acid ethyl ester (11) (Yin, et al., 2013), 3,4,α-trihydroxy-methyl phenylpropionate (12) (Gu, et al., 2007), amburoside A (14) (Sauvain, et al., 1999), trans-6-[(1E)-2-carboxyethenyl]-3-(3,4- dihydroxyphenyl)-2,3-dihydro-1,4-benzodioxin-2-carboxlic acid (16) (Wan, et al., 2008), rosmarinic acid (17) (Refaey, et al., 2021), 2-(3-hydroxy-4-methylphenyl) propanoic acid (18) (Austgulen, et al., 1987), caffeic acid ethyl ester (19) (Ge, et al., 2018), amburoside B (20) (Sauvain, et al., 1999), (+)-epi-pinoresinol (21) (Ge, et al., 2019), (+)-diasyringaresinol (22) (Chang, et al., 1998), (+)-lirioresinol A (23) (Liu, et al., 2013), 2-caffeoyloxy-3-[2-(4-hydroxybenzyl)-4,5-dihydroxy]phenylpropionic acid (24) (Kikuzaki and Nakatani, 1989), isosalvianolic acid C (26) (Liu, et al., 2014), cis-p-hydroxyl ethyl cinnamate (27) (Lu, et al., 2015), trans-p-hydroxyl ethyl cinnamate (28) (Lu, et al., 2015), sakuranetin (30) (Ferreira, et al., 2020), luteolin (33) (Ferreira, et al., 2020), apigenin (34) (Ferreira, et al., 2020), 8-C-p-hydroxybenzylluteolin (35) (Merghem, et al., 1995), 3,4,α-Trihydroxy- isobutylphenylpropionate (36) (Gu, et al., 2007), isobutyl protocatechuate (37) (Li, et al., 2022), and 5,6-dihydroxy-7,3′,4′-trimethoxyflavone (38) (Sato and Tamura, 2015).

3.3 Hypothetical biogenetic pathway for new bisphenols

The hypothetical biogenetic pathway for four new bisphenols (13, 15, 25, 29) is shown in Scheme 1. The reaction may be derived from the precursor of 4-hydroxybenzaldehyde (4), with a subsequent attack on protocatechuic acid (1), 3-hydroxy-4-methoxy benzoic acid (5), carvacrol, and thymol, respectively, through a Knoevenagel condensation reaction and eduction of the exocyclic double bond, then yielded capillarisenols A-D. Furthermore, 4-hydroxybenzaldehyde (4) and rosmarinic acid (17) could also follow the same hypothetical biogenetic pathway to generate compounds 24 and 31 (capillarisenol E). This biogenetic pathway plausibly elucidates the structures of new bisphenols.

Close

Scheme 1

Hypothetical biogenetic pathway for new bisphenols (13, 15, 25, 29, 31).

Export to PPT

3.4 Anti-hepatoma activities of all compounds (1–38)

The anti-hepatoma activities of all compounds are shown in Table 5. Of the five new bisphenols (capillarisenols A-E), only capillarisenol C (25) exhibited significant inhibition of Huh7 and HepG2 cells proliferation, with IC₅₀ values of 4.96 and 8.58 μM, respectively. While the first-line anti-HCC drug lenvatinib exhibited anti-hepatoma activities against Huh7 and HepG2 cells with IC₅₀ values of 29.85 and 30.17 μM, respectively. In addition, new compound 32 also showed significant inhibition of Huh7 and HepG2 cells proliferation, with IC₅₀ values of 40.27 and 16.56 μM, respectively. Luteolin (33) is a common flavonoid, and its IC₅₀ values for Huh7 and HepG2 cells inhibition were 38.14 and 21.07 μM. These results indicated that capillarisenol C is a potential anti-cancer compound, and its mechanism is under study.