Identification of a Novel Retinoid X Receptor Agonist from Boenninghausenia albiflora var. japonica

Abstract

Retinoid X receptors (RXRs) are nuclear receptors involved in various crucial biological processes, such as gene regulation, metabolism, and cell differentiation. They predominantly function as heterodimers with other nuclear receptors and modulate gene expression in response to ligand binding. Additionally, they act as therapeutic targets for different conditions, such as cancer and metabolic disorders. Although synthetic RXR agonists, such as bexarotene, are used in clinical settings, they exert adverse side effects. In this study, to explore insights into potential natural RXR agonists as alternatives to existing drugs, we isolated 14 coumarins, including 1 new compound from Boenninghausenia albiflora var. japonica (Rutaceae). Among them, daphnoretin methyl ether (14), a known biscoumarin, was found to exhibit subtype-nonspecific RXR agonist activity.

Introduction

Retinoid X receptors (RXRs) are nuclear receptors that play crucial roles as transcription factors for gene regulation. They are activated by 9-cis retinoic acid and are involved in various biological processes, including cell differentiation, apoptosis, and metabolism.^1–3) RXR has three isoforms, RXRα, RXRβ, and RXRγ, each encoded by distinct genes and exhibiting unique tissue distributions and functions.^4,5) RXRs primarily function as heterodimers, partnering with other nuclear receptors such as retinoic acid, peroxisome proliferator-activated, and liver X receptors. These heterodimers bind to specific DNA sequences, known as the “response elements,” in the promoters of target genes, modulating their expression in response to ligand binding. Interestingly, RXRs also form homodimers or independently bind to DNA.^4,6) Interactions of RXRs with other nuclear receptors enhance the specificity of gene regulation, making them key players in cellular signaling pathways and influencing several physiological and developmental processes. Additionally, RXRs are involved in metabolic regulation and are implicated in various diseases and conditions, such as cancer, metabolic syndrome, and neurological disorders, making them important therapeutic targets.⁷⁾ In clinical practice, bexarotene (BEX), a synthetic RXR agonist, is primarily used to treat cutaneous T-cell lymphoma.⁸⁾ By selectively activating RXRs, BEX influences the expression of genes related to cell growth, differentiation, and apoptosis, thereby controlling malignant cell proliferation. However, BEX also causes serious side effects, such as hypertriglyceridemia, hypothyroidism, hepatomegaly, and anemia, limiting its widespread application.⁹⁾

Our group focuses on identifying natural RXR agonists as potential alternatives to existing drugs.^10–13) As part of our ongoing research, we explored Boenninghausenia albiflora (Hook.) Rchb. ex Meisn. var. japonica (Nakai ex Makino et Nemoto) Suzuki (Rutaceae), a plant native to East Asia traditionally used in Japanese and Chinese herbal medicine for its anti-inflammatory and analgesic effects. This plant is rich in coumarin derivatives exhibiting antitumor and antiviral activities.^14,15) In this study, we isolated a new compound, boenninghausenal (1), and 13 known coumarin derivatives, including daphnoretin methyl ether (14). We also examined the RXR agonist activities of the isolated coumarin derivatives, and 14 was found to exhibit pan-RXR agonist activity.

Results and Discussion

Compounds 1–14 were isolated from the acetone extracts of the aerial parts of B. albiflora var. japonica by repeated column chromatography (CC) or recrystallization. The structures of known compounds 2–14 were determined by comparing their NMR data with the literature data. The compounds were identified as chalepensin (2),^16,17) rutamarin (3),¹⁷⁾ chalepin (4),¹⁸⁾ clausindin (5),¹⁷⁾ minumicrolin (6),^19,20) murrangatin (7),^19,20) murralongin (8),²¹⁾ xanthyletin (9),²²⁾ luvangetin (10),²³⁾ bergapten (11),¹⁷⁾ xanthotoxin (12),¹⁷⁾ matsukaze-lactone (13),²⁴⁾ and daphnoretin methyl ether (14)²⁵⁾ (Fig. 1).

Fig. 1. Structures of Compounds 1–14

Boenninghausenal (1) was isolated as an amorphous, colorless solid. High-resolution electrospray ionization mass spectrometry (HRESIMS) spectrum in positive ion mode revealed a protonated molecule at m/z 259.0960 [M + H]⁺, consistent with the molecular formula C₁₅H₁₄O₄ (calcd. for C₁₅H₁₅O₄: 259.0965). Its IR spectrum exhibited absorption peaks corresponding to carbonyl groups (ν_max = 1723 and 1668 cm^–1), and the ¹H-NMR spectrum showed resonances corresponding to two equivalent methyl groups [δ_H 1.49 (6H, s, H₃-4′,5′)], one vinyl group [δ_H 5.11 (1H, d, J = 17.4 Hz, H_a-3′), 5.13 (1H, d, J = 10.5 Hz, H_b-3′), and 6.14 (1H, dd, J = 10.5, 17.4 Hz, H-2′)], one formyl group [δ_H 9.90 (1H, s, H-6′)], one chelated hydroxy group [δ_H 11.35 (1H, s, 7-OH)], and three singlets of sp² methine protons [δ_H 6.83 (1H, s, H-8), 7.55 (1H, s, H-4), and 7.71 (1H, s, H-5)]. The ¹³C-NMR and DEPT135 spectra revealed 15 carbon signals, including a singlet corresponding to two equivalent methyl groups [δ_C 26.0 (C-4′,5′)], an sp³ non-protonated carbon [δ_C 40.5 (C-1′)], six sp² non-protonated carbons [δ_C 113.1 (C-10), 118.1 (C-6), 133.3 (C-3), 158.6 (C-9), 159.0 (C-2), and 163.6 (C-7)], an sp² methylene group [δ_C 112.7 (C-3′)], four sp² methine groups [δ_C 104.2 (C-8), 134.3 (C-5), 136.7 (C-4), and 144.9 (C-2′)], and a formyl carbon [δ_C 194.9 (C-6′)]. The presence of the 1,1-dimethylallyl group was confirmed by heteronuclear multiple bond connectivity (HMBC) correlations between H₂-3′/C-1′, C-2′ and H₃-4′(5′)/C-1′, C-2′. Moreover, HMBC correlations of H-5/C-7, C-9, C-6′, H-8/C-6, C-10, H-6′/C-7, and 7-OH/C-6, C-8 suggested the presence of a 1,2,4,5-tetrasubstituted benzene ring with formyl and a hydroxy group positioned ortho to each other (Fig. 2). HMBC correlations of H-4/C-1′, C-2, C-9, H-5/C-4, and the molecular formula inferred from HRESIMS revealed the structure of a 3,6,7-trisubstituted coumarin with a 1,1-dimethylallyl group at the C-3 position. Consequently, compound 1 was identified as 3-(1,1-dimethylallyl)-6-formyl-7-hydroxycoumarin.

Fig. 2. HMBC Correlations in Compound 1

Given that several phenolic compounds demonstrated RXR agonist activity in our previous studies,^10–13) we assessed the RXR agonist activities of compounds 1–14 by luciferase reporter assay (Fig. 3a). Among these, only compound 14 exhibited RXRα agonist activity. Moreover, RXRα agonistic activity of 14 was dose-dependent, with a 50% effective concentration (EC₅₀) of 6.18 μM (95% confidence interval [CI]: 4.40–8.93 μM) and efficacy rate of approximately 95.8% relative to those of BEX (EC₅₀: 3.70 nM; 95% CI: 2.82–4.84 nM) in the luciferase reporter assay (Fig. 3b). Furthermore, the RXRα agonist activity of 14 was completely inhibited by UVI3003, an RXR antagonist, in a manner similar to the inhibition of the RXRα agonist activity of BEX by UVI3003 (Fig. 3c). To further assess the selectivity of 14 for different RXR isoforms, agonist activities for RXRβ and RXRγ were evaluated using the luciferase reporter gene assay (Fig. 3d and 3e). Notably, 14 exhibited agonist activities at both RXRβ and RXRγ, acting as a pan-RXR agonist.

Fig. 3. Effects of Coumarins Isolated from B. albiflora var. japonica on RXRs

(a) RXRα agonist activities of compounds 1–14 at 20 μM determined by luciferase reporter gene assay. ** indicates p < 0.01 versus DMSO. (b) Dose-dependent effects of 14 (0.2, 0.5, 1, 2, 5, 10, and 20 μM; open circles) and BEX (0.2, 0.5, 1, 2, 5, 10, 20, and 100 nM; filled circles) on RXRα determined by luciferase reporter gene assay. (c) Effects of the RXR antagonist, UVI3003 (10 μM), on the RXRα agonist activities of 14 (10 μM) and BEX (100 nM). ** indicates p < 0.01 versus DMSO and ^## indicates p < 0.01 versus UVI3003 (–) groups, respectively. (d and e) Dose-dependent effects of 14 (0.2, 0.5, 1, 2, 5, 10, 20, and 50 μM; open circles) and BEX (0.01, 0.1, 0.2, 10, 100, and 1000 nM; filled circles) on RXRβ (d) and RXRγ (e) determined by luciferase reporter gene assay.

In this study, we evaluated the RXRα agonist activity of 13 additional coumarin derivatives in addition to 14; however, none of these compounds exhibited agonist activity. Furthermore, to the best of our knowledge, no previous reports have described coumarin derivatives with RXRs agonist activity. Given this context, we hypothesized that the distinctive structure of 14, characterized by the dimerization of two coumarin scaffolds via an ether bond, plays a significant role in its binding to the RXR protein. To investigate the molecular basis of the binding mode and affinity of 14 with RXRα protein, receptor–ligand docking simulations were performed using the AutoDock Vina 1.2.5 program.²⁶⁾ The docking pose and predicted interactions that achieved the highest score in the docking simulation are presented in Fig. 4. Hydrogen bonding between ARG316 and the RXR ligand is crucial for receptor binding.^12,27,28) Moreover, our previous studies revealed that 6OHA, an RXR agonist we identified, also formed hydrogen bonds with ASN306 and ALA327, in addition to ARG316.¹³⁾ In this study, 14 was found to form hydrogen bonds between its methoxy groups at the 6′ and 7′ positions with ALA327 and ARG316, respectively, while the carbonyl group at the 2′ position formed a hydrogen bond with ASN306. Additionally, CH···O interactions were observed between LEU309 and LEU325 and the methoxy groups at the 7′ and 6′ positions, respectively. Various π interactions, including π–donor hydrogen bonds, σ–π, amide–π, and π–alkyl interactions, were also observed. Therefore, multiple intermolecular interactions, particularly hydrogen bonds, contribute to the affinity of 14 for the receptor. Based on the docking pose, it was confirmed that the twisting of the two coumarin moieties at the ether bond forms a structure that fits the ligand-binding domain. However, the docking score obtained for 14 was –6.80, which is higher than those of BEX (–8.97) and 6OHA (–8.49), suggesting that 14 has a lower affinity for the receptor compared to existing agonists. This lower affinity is thought to be consistent with the higher EC₅₀ value of 14 compared to BEX and 6OHA.

Fig. 4 Docked Model of Compound 14 with the RXRα Ligand-Binding Domain (PDB ID: 5MKU)

Dashed lines indicate the non-bonded interactions, including conventional hydrogen bonds, CH···O interactions, π–donor hydrogen bonds, σ–π, amide–π, and π–alkyl interactions in the colors shown in the legend. (a) Interactions between 14 and RXRα LBD. Amino acid residues and helices (H3, H5, H7, and H10/H11) involved in these interactions are also shown. For clarity, π–alkyl interactions have been omitted from this figure. (b) Two-dimensional (2D) diagram showing all interactions between 14 and RXRα LBD.

Conclusion

In this study, we aimed to explore insight into potential natural RXR agonists as alternatives to existing drugs, such as BEX, whose applications are limited by their side effects. Isolation and analysis of compounds from B. albiflora var. japonica revealed a new compound, boenninghausenal (1), and 13 known coumarin derivatives, including daphnoretin methyl ether (14). The structures of these compounds were assessed using various analytical techniques, including NMR spectroscopy, and their RXR agonist activities were evaluated by luciferase reporter assay. Notably, compound 14 exhibited RXRα agonist activity, with an EC₅₀ value of 6.18 μM, acting as a pan-RXR agonist targeting all three RXR isoforms (RXRα, RXRβ, and RXRγ). Docking simulations revealed that 14 interacted with RXRα via multiple intermolecular forces, including hydrogen bonds, CH···O, and various π interactions. Overall, our results revealed 14 as a promising natural RXR agonist that can be used as a potential RXR-mediated pathway-targeting therapeutic agent for cancer and various metabolic diseases.

Experimental

General Experimental Procedures

IR spectra were recorded using the FTIR-8400S spectrophotometer (Shimadzu, Kyoto, Japan). NMR spectra were acquired using the JNM-ECZ 400S spectrometer (JEOL, Tokyo, Japan) with tetramethylsilane as the internal standard. ESIMS data were obtained using the Agilent 6230 LC/TOF mass spectrometer (Agilent, Santa Clara, CA, U.S.A.). Silica gel AP-300 (Toyota Kako, Aichi, Japan), Sephadex LH-20 (GE Healthcare, Chicago, IL, U.S.A.), and Cosmosil 75C18-OPN (Nacalai Tesque, Kyoto, Japan) were used for CC. Silica gels 60 F254 and RP-18 F254S (Merck, Darmstadt, Germany) were used for TLC.

Minimum essential medium, penicillin, and streptomycin were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). Fetal bovine serum was purchased from Biowest (Nuaillé, France). BEX was purchased from Toronto Research Chemicals, Inc. (North York, ON, Canada). UVI3003 was purchased from Cayman Chemical (Ann Arbor, MI, U.S.A.). All test compounds were dissolved in dimethyl sulfoxide and stored at −20°C until use. Human embryonic kidney 293(HEK293) cells were provided by RIKEN BioResource Center (Tsukuba, Japan).

Plant Material

The aerial parts of B. albiflora var. japonica were collected in Gujo city (Gifu, Japan) in September 2022. Taxonomic identification was performed by Ken-ichi Nakashima. A voucher specimen (AGU20220901) was deposited at Aichi Gakuin University (Nagoya, Japan).

Extraction and Isolation

Dried aerial parts of B. albiflora var. japonica (1.75 kg) were extracted with acetone (7.5 L × 3) at room temperature (25–30°C) and evaporated under reduced pressure to obtain the acetone extract (71.3 g). The acetone extract was subjected to silica gel CC using a stepwise gradient of CHCl : MeOH (1 : 0, 20 : 1, and 10 : 1, v/v) as the eluent. The fractions were pooled by TLC, yielding 5 fractions (Frs. A–E).

Fraction A (23.1 g) was further fractionated by silica gel CC using a stepwise gradient of n-hexane:acetone (5 : 1, 3 : 1, and 1 : 1, v/v), yielding 4 fractions (Frs. A1–4). Fr. A2 was further fractionated by silica gel CC using n-hexane:ethyl acetate (100 : 1, 20 : 1, 10 : 1, and 8 : 1, v/v) to yield 10 combined subfractions (Frs. A2-1–10). Fr. A2-9 was identified as pure compound 9 (38.4 mg). Frs. A2-1, A2-2, A2-4, and A2-6 were subjected to Sephadex LH-20 CC and eluted with MeOH to yield compounds 5 (101 mg), 2 (10.3 mg), 1 (9.9 mg), and 3 (232 mg), respectively. Fr. A3 was further fractionated by silica gel CC using a stepwise gradient of n-hexane : ethyl acetate (50 : 1, 20 : 1, 10 : 1, 8 : 1, and 5 : 1, v/v) to yield 5 combined fractions (Frs. A3-1–5). Frs. A3-2 and A3-4 were recrystallized in MeOH to yield compounds 11 (32.9 mg) and 12 (181 mg), respectively.

Fraction B (8.84 g) was recrystallized in MeOH to yield compound 13 (935 mg). Following the isolation of 13, Fr. B filtrate was subjected to Sephadex LH-20 CC and eluted with MeOH, yielding 6 fractions (Frs. B1–6). Fr. B3 was further fractionated by silica gel CC using a stepwise gradient of n-hexane:acetone (5 : 1, 3 : 1, and 1 : 1, v/v), resulting in 7 combined subfractions (Frs. B3-1–7). Frs. B3-2 and B3-6 yielded pure compounds 10 (364 mg) and 8 (22.3 mg), respectively. Fr. B3-4 was recrystallized in dichloromethane to yield compound 4 (17.0 mg). Fr. B5 was fractionated by silica gel CC using a stepwise gradient of n-hexane:ethyl acetate (4 : 1, 2 : 1, and 1 : 1, v/v), yielding 4 combined subfractions (Frs. B5-1–4). Fr. B5-2 was identified as pure compound 10 (23.1 mg). Fr. B5-3 was purified using the Sep-Pak C₁₈ cartridge and eluted with MeOH/H₂O (2 :3) to obtain compounds 13 (24.0 mg) and 14 (14.8 mg).

Fraction C (5.58 g) was subjected to Sephadex LH-20 CC and eluted with MeOH to yield 4 fractions (Frs. C1–4). Fr. C4 was further fractionated via silica gel CC using a stepwise gradient of n-hexane:ethyl acetate (4 : 1, 2 : 1, and 1 : 1, v/v), resulting in 3 combined subfractions. The second subfraction was repeatedly purified using the Sep-Pak C₁₈ cartridge and eluted with MeCN/H₂O (1 : 1) and MeOH/H₂O (1 : 3) to yield compounds 6 (1.28 g) and 7 (47.8 mg).

Boenninghausenal (1)

¹H-NMR (CDCl₃) δ: 1.49 (6H, s, H₃-4′ and H₃-5′), 5.11 (1H, d, J = 17.4 Hz, H_a-3′), 5.13 (1H, d, J = 10.5 Hz), 6.14 (1H, dd, J = 10.5, 17.4 Hz, H-2′), 6.83 (1H, s, H-8), 7.55 (1H, s, H-4), 7.71 (1H, s, H-5), 9.90 (1H, s, H-6′), 11.35 (1H, s, 7-OH); ¹³C-NMR (CDCl₃) δ: 26.0 (2C, q, C-4′, 5′), 40.5 (s, C-1′), 104.2 (d, C-8), 112.7 (t, C-3′) 113.1 (s, C-10), 118.1 (s, C-6), 133.3 (s, C-3), 134.3 (d, C-5), 136.7 (d, C-4), 144.9 (d, C-2′), 158.6 (s, C-9), 159.0 (s, C-2), 163.6 (s, C-7), 194.9 (d, C-6′); IR (KBr) cm^–1: 3441, 1723, 1668, 1634, 1173, 980; UV λ_max (MeOH) nm (log ε): 234 (4.29), 260 (4.44), 311 (4.08), 339 (4.11), 400 (3.54); HRESIMS (positive) m/z 259.0960 [M + H⁺] (Calcd for C₁₅H₁₅O₄⁺: 259.0965).

Molecular Docking Studies

The crystallographic structure of RXRα (PDB ID: 5MKU) was downloaded and observed using BIOVIA Discovery Studio 2021 Visualizer, which is commonly used to draw various compounds.²⁹⁾ Geometrical optimization of 14 was conducted using the MMFF94s force field in CONFLEX 8. Polar hydrogen atoms were added, and grid box parameters were defined using MGL Tools 1.5.6. Fourteen amino acid residues in the ligand-binding domain (Ile268, Asn306, Leu309, Ile310, Phe313, Arg316, Ile324, Leu326, Val332, Val342, Ile345, Phe346, Val349, Cys432, and Leu436) were designated as flexible. Docking simulations were performed using AutoDock Vina 1.2.5 with default parameters, except for search exhaustiveness, which was set to 32, and the scoring function was set to Vinardo.^26,30) Docking simulations were performed 5 times with different random seeds. The best docking results for 14 were visualized using Discovery Studio 2021 Visualizer, and the intermolecular interactions were analyzed.

Cell Culture and Luciferase Reporter Gene Assay for RXRα

HEK293 cells were cultured in minimum essential medium supplemented with 10% (v/v) fetal bovine serum, 50 U/mL penicillin, and 50 μg/mL streptomycin. The cells were transfected by calcium phosphate co-precipitation as described below. pCMX-RXRα (30 ng) and CRBPII-tk-Luc (120 ng) were used together with the addition of the pCMX-β-gal expression vector (30 ng) and carrier DNA pUC18 (420 ng) to yield 600 ng of total DNA per well. After 8 h of transfection, the cells were thoroughly washed with fresh medium, and incubation was continued in the presence of the compounds at the indicated concentrations in the medium containing 10% (v/v) fetal bovine serum for 24 h. Then, luciferase and β-galactosidase activities of the cell lysate were analyzed using the Spark multimode microplate reader (TECAN, Männedorf, Switzerland). Luciferase activity was normalized to β-galactosidase activity (internal control) and expressed as the mean ± standard deviation.

Luciferase Reporter Gene Assay for RXRβ and RXRγ

Luciferase reporter gene assay for RXRβ and RXRγ was performed using the human RXRβ and RXRγ reporter assay kits (Indigo Biosciences, PA, U.S.A.), respectively, following the manufacturer’s protocol. Luciferase activity of the cell lysate was analyzed using the Spark multimode microplate reader.

Statistical Analysis

All experiments were performed in triplicate except for those shown in Figs. 3d and 3e, which were repeated twice. Statistical analyses were conducted using the GraphPad Prism v.9.0 software. Data are represented as the mean ± standard deviation. One-way ANOVA with Dunnett’s test and two-way ANOVA with Tukey’s test were used for multiple comparisons in Figs. 3a and 3c, respectively. Statistical significance was set at p < 0.05.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplemen-tary materials.

References

• 1)  Leid M., Kastner P., Lyons R., Nakshatri H., Saunders M., Zacharewski T., Chen J.-Y., Staub A., Garnier J.-M., Mader S., Chambon P., Cell, 68, 377–395 (1992).
• 2)  Mangelsdorf D. J., Borgmeyer U., Heyman R. A., Zhou J. Y., Ong E. S., Oro A. E., Kakizuka A., Evans R. M., Genes Dev., 6, 329–344 (1992).
• 3)  Krężel W., Rühl R., de Lera A. R., Mol. Cell. Endocrinol., 491, 110436 (2019).
• 4)  Germain P., Chambon P., Eichele G., Evans R. M., Lazar M. A., Leid M., De Lera A. R., Lotan R., Mangelsdorf D. J., Gronemeyer H., Pharmacol. Rev., 58, 760–772 (2006).
• 5)  Krężel W., Kastner P., Chambon P., Neuroscience, 89, 1291–1300 (1999).
• 6)  Evans R. M., Mangelsdorf D. J., Cell, 157, 255–266 (2014).
• 7)  Sharma S., Shen T., Chitranshi N., Gupta V., Basavarajappa D., Sarkar S., Mirzaei M., You Y., Krezel W., Graham S. L., Gupta V., Mol. Neurobiol., 59, 2027–2050 (2022).
• 8)  Farol L. T., Hymes K. B., Expert Rev. Anticancer Ther., 4, 180–188 (2004).
• 9)  Scarisbrick J. J., Morris S., Azurdia R., Illidge T., Parry E., Graham-Brown R., Cowan R., Gallop-Evans E., Wachsmuth R., Eagle M., Wierzbicki A. S., Soran H., Whittaker S., Wain E. M., Br. J. Dermatol., 168, 192–200 (2013).
• 10)  Kotani H., Tanabe H., Mizukami H., Makishima M., Inoue M., J. Nat. Prod., 73, 1332–1336 (2010).
• 11)  Inoue M., Tanabe H., Nakashima K., Ishida Y., Kotani H., J. Nat. Prod., 77, 1670–1677 (2014).
• 12)  Nakashima K., Murakami T., Tanabe H., Inoue M., Biochim. Biophys. Acta, Gen. Subj., 1840, 3034–3041 (2014).
• 13)  Nakashima K., Yamaguchi E., Noritake C., Mitsugi Y., Goto M., Hirai T., Abe N., Sakai E., Oyama M., Itoh A., Inoue M., ACS Chem. Biol., 15, 1526–1534 (2020).
• 14)  Chaya N., Terauchi K., Yamagata Y., Kinjo J., Okabe H., Biol. Pharm. Bull., 27, 1312–1316 (2004).
• 15)  Nakano D., Ishitsuka K., Matsuda N., Kouguchi A., Tsuchihashi R., Okawa M., Okabe H., Tamura K., Kinjo J., J. Nat. Med., 71, 170–180 (2017).
• 16)  Elgamal M. H. A., Elewa N. H., Elkhrisy E. A. M., Duddeck H., Phytochemistry, 18, 139–143 (1979).
• 17)  Wu T. S., Shi L. S., Wang J. J., Iou S. C., Chang H. C., Chen Y. P., Kuo Y. H., Chang Y. L., Tenge C. M., J. Chin. Chem. Soc. (Taipei), 50, 171–178 (2003).
• 18)  Tong L., Xiong R., Jiang H., Jiang X., Sun H., Jiang H., Shen J., Heterocycles, 81, 1697–1702 (2010).
• 19)  Ito C., Furukawa H., J. Chem. Soc. Perkin 1, 1990, 2047–2055 (1990).
• 20)  Luo X., Qi S., Yin H., Xiao Z., Zhang S., Magn. Reson. Chem., 47, 1110–1114 (2009).
• 21)  Kinoshita T., Wu J. B., Ho F. C., Phytochemistry, 43, 125–128 (1996).
• 22)  Liu X. G., Zhang S. S., Jiang C. Y., Wu J. Q., Li Q., Wang H., Org. Lett., 17, 5404–5407 (2015).
• 23)  Patra A., Mitra A. K., Org. Magn. Reson., 17, 222–224 (1981).
• 24)  Basa S. C., Das S. P., Tripathy R. N., Elango V., Shamma M., Heterocycles, 22, 333–337 (1984).
• 25)  Reisch J., Novak I., Szendrei K., Minker E., Planta Med., 16, 372–376 (1968).
• 26)  Trott O., Olson A. J., J. Comput. Chem., 31, 455–461 (2010).
• 27)  Wang G. H., Jiang F. Q., Duan Y. H., Zeng Z. P., Chen F., Dai Y., Chen J. B., Liu J. X., Liu J., Zhou H., Chen H. F., Zeng J. Z., Su Y., Yao X. S., Zhang X. K., Cancer Res., 73, 307–318 (2013).
• 28)  Scheepstra M., Nieto L., Hirsch A. K., Fuchs S., Leysen S., Lam C. V., in het Panhuis L., van Boeckel C. A. A., Wienk H., Boelens R., Ottmann C., Milroy L.-G., Brunsveld L., Angew. Chem. Int. Ed., 53, 6443–6448 (2014).
• 29)  Scheepstra M., Andrei S. A., De Vries R. M. J. M., Meijer F. A., Ma J. N., Burstein E. S., Olsson R., Ottmann C., Milroy L. G., Brunsveld L., ACS Chem. Neurosci., 8, 2065–2077 (2017).
• 30)  Quiroga R., Villarreal M. A., PLOS ONE, 11, e0155183 (2016).