Biotransformation of neuro-inflammation inhibitor kellerin using Angelica sinensis (Oliv.) Diels callus

Abstract

Kellerin, a sesquiterpene coumarin derivative, has been identified as a major constituent of Ferula sinkiangensis K. M. Shen. It has been proved to be a potential natural therapeutic agent for Alzheimer's disease because of its inhibition of inflammatory cytokines nitric oxide (NO), tumor necrosis factor-α (IL-6) (TNF-α), cyclooxygenase-2 (COX-2), interleukin-6 (IL-6) and interleukin-1β (IL-1β) in over-activated BV2 mouse microglial cells. Because of the multi chiral centers and the chemical instability of the sesquiterpene, coumarin, it is rather difficult to obtain this bioactive natural product by synthesis. Thus, biotransformation of kellerin was carried out to afford more novel derivatives using the callus of Angelica sinensis (Oliv.) Diels, which has an abundance of biosynthetic enzymes of phenylpropanoids. As a result, 14 products were obtained and identified, including four new sesquiterpene coumarin derivatives: 14′-hydroxy-(3′S,4′R,5′S,8′R,9′S,10′R)-kellerin (1), 5′,6′-ene-14′-hydroxy-(3′S,4′R,8′R,9′S,10′R)-kellerin (2), 5′,6′-ene-(3′R,8′R,9′S,10′R)-ferukrin (3), and 14′-hydroxy-(3′S,4′R,5′S,8′R,9′S,10′R)-deacetylkellerin (4), together with 10 other known compounds. Their structures were elucidated using comprehensive spectroscopic techniques and their possible biosynthetic pathways were proposed on the basis of the structural analyses. Furthermore, their anti-neuroinflammatory activities were assessed in BV2 cells by monitoring lipopolysaccharide-induced NO production, and the structure–activity relationships were discussed.

---

Introduction

Sesquiterpene coumarins are a small family of natural product, which are constructed of a coumarin moiety and a sesquiterpene fragment through an ether linkage. They are usually found in some plants of Umbelliferae, Compositae and Rutaceae families.¹ Sesquiterpene coumarins have been reported to be characteristic of the chemical composition of the Ferula plant genus, and are responsible for antiviral, antibacterial, antileishmanial, antitumor and anti-neuroinflammatory activities.^1–3 In previous research, a sesquiterpene coumarin, kellerin, a major constituent of Ferula sinkiangensis K. M. Shen, has been proved to be a potential natural therapeutic agent for Alzheimer's disease (AD) by inhibition of the inflammatory cytokines nitric oxide (NO), tumor necrosis factor-α (TNF-α), cyclooxygenase-2 (COX-2), interleukin-6 (IL-6) and interleukin-1β (IL-1β) in over-activated BV2 mouse microglial cells.⁴ However, it is difficult to synthesize and modify the bioactive natural product by chemical methods because of its multiple chiral centers and the chemical instabilities of the skeleton. Therefore, in this research, biotransformation was used in anticipation of producing more derivatives of kellerin more conveniently.

Biotransformation produces chemical transformation in enzymes, plant cells or microorganisms' systems,⁵ and it has become an economically and ecologically competitive technology for modifying chemical compounds, especially for bioactive natural compounds with complex structures.^6,7 The plant cultured cells can modify natural products with the advantages of stereo selectivity, mild conditions and diverse products when compared with chemical synthesis.^8–10 Angelica sinensis (Oliv.) Diels (A. sinensis) is a traditional medicinal plant, rich in phenylpropanoids, phthalides, terpenoids, essential oils and aromatic compounds.^11,12 Its abundant biosynthetic enzymes of phenylpropanoids and terpenoids have attracted the attention of researchers. Preliminary results obtained using the callus of A. sinensis showed that it was able to transform kellerin successfully. Therefore, it was chosen for use in producing derivatives of kellerin.

As far as is known, this is the first time the biotransformation of the natural sesquiterpene, coumarin, has been performed using A. sinensis callus. As a result, 14 products (1–14) were isolated and identified from the suspension cultures of A. sinensis callus. Next, the possible biosynthetic pathways of compounds 1–8 from kellerin were proposed. Furthermore, the primary transformation rules of sesquiterpene, coumarin, skeleton in A. sinensis callus, which fill the gaps not only in the study of the biotransformation of the sesquiterpene, coumarin but also on the application of A. sinensis callus were also considered. Finally, the neuro-inflammation inhibitory activities of kellerin and its transformed products were evaluated using NO assays in lipopolysaccharide (LPS)-induced BV2 microglial cells.¹³

Results and discussion

Identification of transformed products

A large number of assays were carried out to optimize the induction of A. sinensis callus and subculture conditions by varying the composition of callus induction media, plant organs and callus type, and large-scale tissue culture of A. sinensis callus was obtained.¹⁴ Kellerin (855 mg) was distributed into A. sinensis callus and they were co-incubated for one week under the optimal subculture conditions. After incubation, mixtures were separated into the cultures and medium, then they were further extracted and purified using repeated chromatographic and recrystallization methods.

As a result, 14 products were obtained, including four new sesquiterpene coumarin derivatives (1–4) together with 10 known ones (5–14). Their structures were identified as: 14′-hydroxy-(3′S,4′R,5′S,8′R,9′S,10′R)-kellerin (1), 5′,6′-ene-14′-hydroxy-(3′S,4′R,8′R,9′S,10′R)-kellerin (2), 5′,6′-ene-(3′R,8′R,9′S,10′R)-ferukrin (3), 14′-hydroxy-(3′S,4′R,5′S,8′R,9′S,10′R)-deacetylkellerin (4), (3′S,5′S,8′R,9′S,10′R)-deacetylkellerin (5),¹⁵ ferukrin (6),¹⁶ gummosin (7),¹⁷ 7-hydroxycoumarin (8),¹⁸ ferulic acid (9),¹⁹ bis(2-ethylhexyl)phthalate (10),²⁰ 3-heptyl-3-hydroxy-2,4(1H, 3H)-quinolinediones (11),²¹ 7-methoxy-1-methyl-9H-pyrido[3,4-b]indole (12),²² methyl butyrate (13),²³ (Z)-3-methyl-4-oxobut-2-enoic acid (14)²⁴ (Fig. 1). Among them, compounds 1–8 were derivatives of kellerin, whereas compounds 11, 12 and 14 were produced by the culture medium stimulated by the substrate kellerin.^11,12

Fig. 1 Structures of compounds 1–14.

Compound 1 was obtained as a yellow oil (methanol; MeOH), [α]²⁰_D + 29.6 (c 1.0, MeOH), high-resolution electrospray ionisation mass spectroscopy (HR-ESI-MS) showed a pseudo molecular ion peak at m/z 481.2196 [M + Na]⁺ (calcd 481.2197 for C₂₆H₃₄O₇Na), which indicated the molecular formula of 1 was C₂₆H₃₄O₇. The proton-nuclear magnetic resonance (¹H-NMR) and ¹³C-NMR spectra of compound 1 closely matched those of kellerin except for the absence of signals for a tertiary methyl group and the presence of resonance for a hydroxymethyl group in 1.⁴ The ¹H-NMR spectrum of 1 exhibited typical signals for 7-O-substituted coumarin at δ_H 7.90 (1H, d, J = 9.5 Hz, H-4), 7.57 (1H, d, J = 8.0 Hz, H-5), 6.97 (1H, d, J = 8.0, 2.0 Hz, H-6), 6.99 (1H, d, J = 2.0 Hz, H-8) and δ_H 6.25 (1H, d, J = 9.5 Hz, H-3). However, the existence of an oxygenated methylene group and a hydroxymethyl unit can also be predicted by the signals at δ_H 4.29 (1H, d, J = 10.9, 2.3 Hz, H-11′a), 4.22 (1H, d, J = 10.9, 3.0 Hz, H-11′b), 3.82 (1H, d, J = 11.5 Hz, H-14′a), 3.42 (1H, d, J = 11.5 Hz, H-14′b). Furthermore, four methyl groups were suggested according to the signals observed at δ_H 1.01 (3H, s, Me-14′), 1.29 (3H, s, Me-12′), 1.34 (3H, s, Me-15′) and δ_H 1.72 (3H, s, Me-17′). Then 26 carbon signals were displayed in the ¹³C-NMR spectrum of 1, which indicated a typical sesquiterpene coumarin skeleton including 9 carbons for umbelliferone, 15 which were ascribed to a sesquiterpene moiety and 2 for an acetyl group. Most importantly, the spectral data of one-dimensional (1D) NMR (Table 1) suggested 1 to be a kellerin derivative.

Table 1 NMR data of compounds 1–4 (¹H: 600 MHz, ¹³C: 150 MHz in methanol-d₄)

Position	1		2		3		4
	C	H	C	H	C	H	C	H
2	163.8		163.8		163.8		163.8
3	113.5	6.25 (1H, d, J = 9.5 Hz)	113.4	6.25 (1H, d, J = 9.5 Hz)	113.4	6.25 (1H, d, J = 9.5 Hz)	113.1	6.24 (1H, d, J = 9.5 Hz)
4	145.7	7.90 (1H, d, J = 9.5 Hz)	145.6	7.89 (1H, d, J = 9.5 Hz)	145.6	7.90 (1H, d, J = 9.5 Hz)	145.7	7.88 (1H, d, J = 9.5 Hz)
5	130.7	7.57 (1H, d, J = 8.0 Hz)	130.8	7.57 (1H, d, J = 8.0 Hz)	130.8	7.56 (1H, d, J = 8.0 Hz)	130.4	7.53 (1H, d, J = 8.0 Hz)
6	114.3	6.99 (1H, dd, J = 8.0, 2.0 Hz)	114.4	6.99 (1H, dd, J = 8.0, 2.0 Hz)	114.4	6.93 (1H, dd, J = 8.0, 2.0 Hz)	114.1	6.95 (1H, dd, J = 8.0, 2.0 Hz)
7	163.6		163.7		163.7		163.6	6.96 (1H, d, J = 2.0 Hz)
8	102.0	6.97 (1H, d, J = 2.0 Hz)	102.1	6.97 (1H, d, J = 2.0 Hz)	102.1	6.99 (1H, d, J = 2.0 Hz)	102.2
9	157.4		157.5		157.5		157.4
10	114.4		114.5		114.5		114.4
1′	31.6	1.80 (1H, d, J = 12.8 Hz)	31.7	1.80 (1H, d, J = 12.8 Hz)	31.7	1.80 (1H, d, J = 12.8 Hz)	31.6	1.80 (1H, d, J = 12.8 Hz)
		1.13 (1H, t, J = 3.2 Hz)		1.13 (1H, t, J = 3.2 Hz)		1.13 (1H, t, J = 3.2 Hz)		1.13 (1H, t, J = 3.2 Hz)
2′	23.7	2.05 (1H, m)/1.57 (1H, m)	23.9	2.05 (1H, m)/1.57 (1H, m)	23.9	2.05 (1H, m)/1.57 (1H, m)	23.7	2.05 (1H, m)/1.57 (1H, m)
3′	75.8	5.07 (1H, t, J = 2.7 Hz)	75.8	5.33 (1H, m)	79.7	3.15 (1H, m)	71.5	3.35 (1H, t, J = 2.7 Hz)
4′	39.0		39.0		39.0		41.5
5′	45.7	2.16 (1H, m)	144.4		148.3		37.1	2.16 (1H, m)
6′	20.2	1.75 (1H, m), 1.45 (1H, m)	121.8	5.27 (1H, m)	130.0	5.83 (1H, m)	20.2	1.75 (1H, m), 1.45 (1H, m)
7′	40.7	1.76 (2H, m)	45.1	1.76 (2H, m)	45.1	1.76 (2H, m)	40.7	1.76 (2H, m)
8′	74.1		74.1		74.1		74.1
9′	59.0	1.55 (1H, m)	59.5	1.55 (1H, m)	59.5	1.55 (1H, m)	59.0	1.55 (1H, m)
10′	43.7		43.6		43.6		43.7
11′	68.9	4.29 (1H, dd, J = 10.9, 2.3 Hz)	69.0	4.34 (1H, dd, J = 10.9, 2.3 Hz)	69.0	4.34 (1H, dd, J = 10.9, 2.3 Hz)	69.4	4.29 (1H, dd, J = 10.9, 2.3 Hz)
		4.22 (1H, dd, J = 10.9, 3.0 Hz)		4.28 (1H, dd, J = 10.9, 3.0 Hz)		4.28 (1H, dd, J = 10.9, 3.0 Hz)		4.23 (1H, dd, J = 10.9, 3.0 Hz)
12′	32.0	1.29 (3H, s)	31.0	1.24 (3H, s)	29.7	1.29 (3H, s)	32.0	1.29 (3H, s)
13′	23.0	1.01 (3H, s)	23.0	1.06 (3H, s)	16.0	1.07 (3H, s)	17.8	0.75 (3H, s)
14′	65.0	3.82 (1H, d, J = 11.5 Hz)	64.7	3.64 (1H, d, J = 11.5 Hz)	27.1	1.17 (3H, s)	77.4	3.64 (1H, d, J = 11.5 Hz)
		3.42 (1H, d, J = 11.5 Hz)		3.42 (1H, d, J = 11.5 Hz)				3.59 (1H, d, J = 11.5 Hz)
15′	25.5	1.34 (3H, s)	25.6	1.32 (3H, s)	31.0	1.32 (3H, s)	25.5	1.34 (3H, s)
16′	172.6		172.7
17′	19.2	1.72 (3H, s)	19.4	1.77 (3H, s)

---

In the heteronuclear multiple bond correlation (HMBC) experiment (Fig. 2), the long range correlations from δ_H 4.25 (2H, CH₂-11′) to δ_C 163.6 (C-7), 74.1 (C-8′), 59.0 (C-9′) and δ_C 43.7 (C-10′) were observed to construct the C-11′–O–C-7 bridge between sesquiterpene and the coumarin moieties. The cross peaks originating from δ_H 1.34 (Me-15′) to δ_C 45.7 (C-5′), 31.6 (C-1′), 59.0 (C-9′) fixed Me-15′ at C-10′. Furthermore, the HMBC correlations from the proton signals at δ_H 3.82 (1H, H-14′a), 3.42 (1H, H-14′b), 1.01 (Me-13′) to the carbon resonance at δ_C 75.8 (C-3′), 39.0 (C-4′) located the acetyl group at C-3′. Furthermore, the remaining methyl group Me-12′ was assigned to C-8′ because of the HMBC correlation between δ_H 1.29 (3H, s, Me-12′) and δ_C 40.7 (C-7′), 59.0 (C-9′), 74.1 (C-8′). Consequently, the planar structure of compound 1 was determined by 1D and two-dimensional NMR. Allowing for the biosynthesis pathway, the absolute configuration of C-3, C-5, C-8, C-9 and C-10 in compound 1 was identified to be same as for the substrate 3′S,5′S,8′R,9′S,10′R-kellerin. The relative configuration of the chiral carbon atom C-4′ was unambiguously determined by nuclear Overhauser spectroscopy (NOESY) (Fig. 3) correlations from Me-13′ to H-3′, H-5′, H-11′, and from Me-15′ to CH₂-14′ and Me-12′.

Fig. 2 The key HMBC () correlations of compounds 1–4.

Fig. 3 Key NOESY () corrections of compound 1.

The absolute configuration was further established by comparison of experimental and calculated electronic circular dichroism (ECD) spectra using time dependent density functional theory (TDDFT).²⁵ A conformational search using the Discovery Studio 3.0 program using a systematic method revealed preferred conformers within a 20 kcal mol⁻¹ energy threshold. These conformers were subjected to geometrical optimization and energy calculation with the Becke, three-parameter, Lee–Yang–Parr (B3LYP) function and 6-31G (d) combined with calculation of vibrational modes to confirm these minima. No imaginary frequencies were found. Calculation of the ECD spectra for these conformers were performed using TDDFT at the B3LYP/6-311+G (2d, p) level. The weighted ECD spectrum in MeOH is shown in Fig. 4A. The calculated ECD spectrum of (3′S,4′R,5′S,8′R,9′S,10′R)-1 closely matched the experimental data, in particular there was a positive Cotton effect (CE) in the 200–220 nm region, and a negative CE in the region of 300–350 nm. Differences between the calculated and experimental spectra were apparently because of minor differences between the calculated and solution conformers of this flexible molecule. Most importantly, the structure of compound 1 was finally determined to be 14′-hydroxy-(3′S,4′R,5′S,8′R,9′S,10′R)-kellerin.

Fig. 4 Comparison of calculated and experimental ECD spectra. (A) Comparison of calculated and experimental ECD spectra of compound 1. (B) Comparison of calculated and experimental ECD spectra of compound 4.

Compound 2 was obtained as yellow solid (MeOH), [α]²⁰_D − 24.0 (c 1.0, MeOH), and had a molecular formula of C₂₆H₃₂O₇, as established by an HR-ESI-MS quasi-molecular ion peak at m/z 479.2054 [M + Na]⁺ (calcd 479.2040 for C₂₆H₃₂O₇Na). The ¹H-NMR spectrum of 2 also exhibited typical signals of 7-O-substituted coumarin at δ_H 7.89 (1H, d, J = 9.5 Hz, H-4), 7.57 (1H, d, J = 8.0 Hz, H-5), 6.97 (1H, d, J = 8.0, 2.0 Hz, H-6), 6.99 (1H, d, J = 2.0 Hz, H-8) and δ_H 6.25 (1H, d, J = 9.5 Hz, H-3). However, the proton signals at δ_H 4.34 (1H, d, J = 10.9, 2.3 Hz, H-11′a) and 4.28 (1H, d, J = 10.9, 3.0 Hz, H-11′b) for an oxygenated methylene group, δ_H 3.64 (1H, d, J = 11.5 Hz, H-14′a) and δ_H 3.42 (1H, d, J = 11.5 Hz, H-14′b) for a hydroxymethyl unit, respectively, were found. It also displayed signals for four methyl groups at δ_H 1.06 (3H, s, Me-13′), 1.24 (3H, s, Me-12′), 1.32 (3H, s, Me-15′) and δ_H 1.77 (3H, s, Me-17′). Then 26 carbon signals were displayed in the ¹³C-NMR spectrum, which indicated a sesquiterpene coumarin skeleton including umbelliferone with 9 carbons and a sesquiterpene moiety with 15 carbons. The 1D NMR data of 2 revealed similar structural features to those found in 1, with the major differences being the presence of a double bond at C-5′ and C-6′ in 2, which was proved by resonances at δ_H 5.27 (1H, m, H-6′) and δ_C 144.4 (C-5′), 121.8 (C-6′). Most importantly, the spectral data of 1D NMR (Table 1) suggested that 2 was a kellerin analog.

In Fig. 2, the HMBC correlations from δ_H 4.31 (2H, CH₂-11′) to δ_C 163.7 (C-7), 74.1 (C-8′), 59.5 (C-9′) and δ_C 43.6 (C-10′) were verified by the presence of the C-11′–O–C-7 bridge between sesquiterpene and coumarin moieties. The long range interactions of δ_H 1.32 (Me-15′) and δ_C 144.4 (C-5′), 31.7 (C-1′) fixed Me-15′ at C-10′. Furthermore, correlations from the proton signals at δ_H 3.64 (1H, H-14′a), 3.42 (1H, H-14′b), 1.06 (Me-13′) to the carbon resonances at δ_C 75.8 (C-3′), 39.0 (C-4′) proved that a hydroxymethyl group and Me-13′ were connected at C-4′. Furthermore, the remaining methyl group Me-12′ was unambiguously assigned to C-8′ because of the HMBC correlations between the proton signals at δ_H 1.24 (Me-12′) and the carbon resonances at δ_C 45.1 (C-7′), 59.5 (C-9′) and 74.1 (C-8′). According to the biosynthesis pathway, it can be verified that the hydroxymethyl group was α-oriented, and the absolute configuration of 2 was identified to be the same as the substrate kellerin, that is 3′S, 4′R, 8′R, 9′S, 10′R. Most importantly, the structure of compound 2 was finally determined to be 5′,6′-ene-14′-hydroxy-(3′S,4′R,8′R,9′S,10′R)-kellerin.

Compound 3, a white powder, [α]²⁰_D − 23.7 (c 1.0, MeOH), had a molecular formula of C₂₄H₂₈O₅ based on a quasi-molecular ion peak at m/z 419.2 [M + Na]⁺ (calcd 419.2 for C₂₄H₂₈O₅Na) in ESI-MS that required 10 indices of hydrogen deficiency. The ¹H-NMR and ¹³C-NMR data of 3 were superimposable with those of kellerin, with the exception of slight changes such as the absence of acetyl group signals and the appearance of resonances of olefin. The ¹H-NMR spectrum of 3 showed the signals for a typical 7-O-substituted coumarin at δ_H 7.90 (1H, d, J = 9.5 Hz, H-4), 7.56 (1H, d, J = 8.0 Hz, H-5), 6.93 (1H, d, J = 8.0, 2.0 Hz, H-6), 6.55 (1H, d, J = 2.0 Hz, H-8) and δ_H 6.25 (1H, d, J = 9.5 Hz, H-3). However, the resonances of olefin at δ_H 5.83 (1H, m, H-6′) in the downfield and the proton signals at δ_H 4.34 (1H, d, J = 10.9, 2.3 Hz, H-11′a) and 4.28 (1H, d, J = 10.9, 3.0 Hz, H-11′b) for an oxygenated methylene group can be observed. Furthermore, according to the signals observed at δ_H 1.07 (3H, s, Me-14′), 1.17 (3H, s, Me-13′), 1.29 (3H, s, Me-12′) and δ_H 1.32 (3H, s, Me-15′) showed four methyl groups. Then 24 carbon signals were exhibited in the ¹³C-NMR spectrum, which indicated a typical sesquiterpene coumarin skeleton including 9 carbons for umbelliferone, and 15 carbons which could be ascribed to a sesquiterpene moiety. So the spectra data of 1D NMR (Table 1) suggested 3 to be a sesquiterpene coumarin.

In the HMBC spectrum (Fig. 2), the correlations between δ_H 4.31 (2H, CH₂-11′) and δ_C 163.6 (C-7), 74.1 (C-8′), 59.0 (C-9′), 43.7 (C-10′) indicated the presence of the C-11′–O–C-7 bridge between the sesquiterpene and coumarin moieties. The HMBC correlations from δ_H 1.32 (Me-15′) to δ_C 148.3 (C-5′) and δ_C 31.7 (C-1′) were observed, and this suggested that Me-15′ was connected to C-10′. Furthermore, a hydroxyl group was located at C-3′, the presence of which was supported by the HMBC correlations from the proton signals at δ_H 1.17 (Me-13′), 1.06 (Me-14′) to the carbon resonance at δ_C 79.7 (C-3′), 39.0 (C-4′), 148.3 (C-5′). Me-12′ was assigned to C-8′ because of the HMBC correlations from δ_H 1.29 (Me-12′) to δ_C 45.1 (C-7′), 59.5 (C-9′) and δ_C 74.1 (C-8′). However, the absolute configuration of C-3′, the chiral center of compound 3, was elucidated by comparing the data of H-3′ and C-3′ obtained from the ¹H-NMR and ¹³C-NMR experiments. With deuterated methanol (CD₃OD) as the deuterated reagent, the proton signals at δ_H 3.50–3.30 (H-3′) and the carbon resonance at δ_C 77.0–71.0 (C-3′) indicated that the hydroxyl group was β-oriented (3′S), whereas at δ_H 3.25–3.10 (H-3′) the carbon resonance at δ_C 80.0–79.0 (C-3′) was α-oriented (3′R).⁴ Thus, the orientation of the hydroxyl group was α-oriented with δ_H 3.15 (H-3′) and δ_C 79.7 (C-3′). The absolute configuration of 3 was identified to be 3′S, 8′R, 9′S, 10′R. Most importantly, the structure of compound 3 was finally determined to be 5′,6′-ene-(3′S,8′R,9′S,10′R)-ferukrin.

Compound 4 was isolated as a yellow oil, [α]²⁰_D − 26.2 (c 1.13, MeOH), HR-ESI-MS showed a quasi-molecular ion peak at m/z 439.2092 [M + Na]⁺ (calcd 439.2091 for C₂₄H₃₂O₆Na), which indicated that the molecular formula was C₂₄H₃₂O₆. The ¹H-NMR spectrum of 4 exhibited resonances for a typical 7-O-substituted coumarin at δ_H 7.88 (1H, d, J = 9.5 Hz, H-4), 7.53 (1H, d, J = 8.0 Hz, H-5), 6.95 (1H, d, J = 8.0, 2.0 Hz, H-6), 6.96 (1H, d, J = 2.0 Hz, H-8) and 6.24 (1H, d, J = 9.5 Hz, H-3). However, an oxygenated methylene group at δ_H 4.29 (1H, d, J = 10.9, 2.3 Hz, H-11′a), 4.23 (1H, d, J = 10.9, 3.0 Hz, H-11′b), a hydroxymethyl unit at δ_H 3.64 (1H, d, J = 11.5 Hz, H-14′a), 3.59 (1H, d, J = 11.5 Hz, H-14′b) appeared in the ¹H-NMR. Furthermore, three methyl groups were displayed at δ_H 0.75 (3H, s, Me-13′), 1.29 (3H, s, Me-12′) and δ_H 1.34 (3H, s, Me-15′). Then 9 carbons for umbelliferone and 15 carbons ascribed to a sesquiterpene moiety indicated a typical sesquiterpene coumarin skeleton (Table 1).

The C-11′–O–C-7 bridge between the sesquiterpene and coumarin moieties was supported by the long range HMBC correlations from δ_H 4.26 (2H, CH₂-11′) to δ_C 163.6 (C-7), 74.1 (C-8′), 59.0 (C-9′) and δ_C 43.7 (C-10′) in Fig. 2. The correlations between δ_H 1.34 (Me-15′) and δ_C 37.1 (C-5′), 31.6 (C-1′) revealed a tertiary methyl Me-15′ was fixed at C-10′. Then the HMBC interactions from the proton signals at δ_H 3.64 (1H, H-14′a), 3.59 (1H, H-14′b), 0.75 (Me-13′) to the carbon resonances at δ_C 71.5 (C-3′), 41.5 (C-4′), 37.1 (C-5′) were proved to be from the location of C-3′ with a β-oriented hydroxyl group. Furthermore, the residual index Me-12′ was assigned to C-8′ because of the HMBC correlations between the proton signals at δ_H 1.29 (3H, s, Me-12′) and the carbon resonances at δ_C 40.7 (C-7′), 59.0 (C-9′) and 74.1 (C-8′). Most importantly, the structural elucidation of compound 4, which was straightforward, was determined to be 14′-hydroxy-(3′S,4′R,5′S,8′R,9′S,10′R)-deacetylkellerin.

By comparison of experimental and calculated ECD spectra, the absolute configuration of compound 4 was further established and is shown in Fig. 4B. The calculated ECD spectrum of (3′S,4′R,5′S,8′R,9′S,10′R)-4 closely matched the experimental data, in particular a negative CE in the 200–220 nm and 270–350 nm regions, and a positive CE in the region of 220–270 nm. Differences between calculated and experimental spectra were apparent because of minor differences between the calculated and solution conformers of this flexible molecule. Thus, the absolute configuration of 4 was established as 3′S, 4′R, 5′S, 8′R, 9′S, 10′R.

Proposed biotransformation pathway of products

The possible metabolic pathways of the substrate are outlined in Fig. 5. Firstly, kellerin was transformed into compound 1 by hydroxylation at Me-14′. Then the linkage between the coumarin and sesquiterpene moieties were broken to yield 8, 7-hydroxycoumarin. Compound 5 was produced from kellerin through deacetylation. Then 5 was dehydrated at C-3′ to produce the intermediate 5a. Compound 6 was produced by the addition of a double bond at C-2′, 3′ in 5a. After that, compound 5 was further dehydrated at C-8′, and C-12′ to give compound 7. In addition, compound 5 was hydroxylated at Me-14′ to give compound 4. Compound 6 was converted to generate compound 3 via dehydrogenation of C-5′, and C-6′. The formation of compound 2 was also involved in the dehydrogenation at C-5′, C-6′ of 1.

Fig. 5 Possible biosynthesis pathways of compounds 1–8 from kellerin in Angelica sinensis (Oliv.) Diels callus.

Furthermore, compounds 11, 12, 14 were produced using a culture medium stimulated by kellerin, as shown in Fig. 6. Aniline and 2-heptylmalonic acid were condensed to generate the intermediate 11a, which was transformed to compound 11 using addition and oxidation. Harmane was converted into compound 12 using hydroxylation and methylation reactions. Compound 14 was derived from 5-hydroxy-4-methylfuran-2(5H)-one using a cycloreversion reaction and oxidation process. Compounds 9, 10 and 13 were derived from A. sinensis callus.

Fig. 6 Possible biosynthesis pathways of compounds 11–12, 14 in Angelica sinensis (Oliv.) Diels callus stimulated by kellerin.

It is the first time that the biotransformation of sesquiterpene coumarins using the A. sinensis callus system has been explored. Next, using the characteristics of the transformed products, a brief summary of the method using A. sinensis callus to obtain sesquiterpene coumarins will be given. Firstly, the sesquiterpene moiety of the substrate was susceptible to being modified, which involved hydroxylation, dehydrogenation, dehydration and reduction reactions. Secondly, the ether linkage (C-11′–O–C-7 bridge) between sesquiterpene and the coumarin moieties was readily fractured and this gave coumarin. Thirdly, the products, produced by the culture medium stimulated by the substrate, were usually obtained from oxidation, reduction and methylation processes.

Anti-neuroinflammatory effects of transformed products in LPS-induced BV2 microglial cells

The anti-inflammatory activities of the identified products were evaluated using a NO assay in LPS-induced BV2 microglial cells. However, the cytotoxic activities were assayed to avoid possible effects of reduced viability on NO release (Fig. 1S; ESI†). The n-butanol (n-BuOH) extract (BE) showed anti-inflammatory activities with an IC₅₀ value (the concentration of an inhibitor where the response (or binding) is reduced by half) of 19.53 μg mL⁻¹ (Table 2). The ethyl acetate (EtOAc) extract (EE) could reduce the production of NO significantly at a concentration of 100 μg mL⁻¹. Compounds 1, 2, 3, 6, 7 and 9 exhibited stronger inhibitory activities on NO production than that of the positive control minocycline. Out of these compounds, 1 and 7 exhibited significant inhibitory effects with an IC₅₀ value of 7.55, at a concentration of 24.47 μM. Compounds 2, 3, 6, 9 could reduce the production of NO significantly with IC₅₀ values ranging from 3.72 μM to 35.46 μM (Table 2). Compounds 5, 8, 13 exhibited activities with IC₅₀ values ranging from 47.34 μM to 74.31 μM (Fig. 7A and B and Table 2). In addition, compounds 4, 11 and 14 showed a trend of inhibition on NO production. It is worth noting that all the samples tested, except for compound 10 did not exhibit obvious cytotoxicities at their effective concentrations (Fig. 1S; ESI†).

Table 2 Effects of extracts and transformed products from the suspension cultures on NO production by LPS-activated microglia cells (mean ± SEM)^d

Sample name	IC50^a	Sample name	IC50^a
a IC50 (μg mL^−1 for extracts and μM for compounds).b Compounds 1, 7, 10 showed cytotoxicity at 100 μM.c Positive control.d SEM = standard error of the mean.
EE	>100	Compound 8	74.31 ± 4.50
BE	19.53 ± 3.54	Compound 9	24.64 ± 4.33
Compound 1^b	7.55 ± 1.97	Compound 10^b	61.52 ± 3.51
Compound 2	21.62 ± 3.30	Compound 11	>100
Compound 3	24.89 ± 2.67	Compound 12	>100
Compound 4	>100	Compound 13	64.41 ± 3.98
Compound 5	47.34 ± 3.77	Compound 14	>100
Compound 6	35.46 ± 2.74	Substrate	3.72 ± 2.02
Compound 7^b	24.47 ± 3.71	minocycline^c	35.82 ± 3.60

---

Fig. 7 Anti-neuroinflammatory activities of transformed products and substrate (Sub.) on LPS-induced NO production in BV2 microglial cells (A: inhibitory effect of the transformed products 1–7 and positive control on LPS-induced NO production in BV2 microglial cells; B: inhibitory effect of the transformed products 8–14, substrate and positive control on LPS-induced NO production in BV2 microglial cells. Each bar represents the means ± standard error (SE) of three independent experiments. Significance: *P < 0.001 compared to LPS groups. ^#P < 0.001 compared to control groups. Con: control, NO: nitric oxide, LPS: lipopolysaccharide, MINO: minocycline).

According to the biological results and reported structure–activity relationships (SAR) in previous research,⁴ the SAR of sesquiterpene coumarins were further enriched as shown in Fig. 8. Firstly, as for the anti-neuroinflammatory effects of bicyclic sesquiterpene coumarins, with a terminal double bond substituted at C-8′, the order is β-OH, acetoxyl group, carbonyl group, and α-OH (Fig. 8A). Secondly, the effects of sesquiterpene coumarins with the presence of methyl and hydroxyl groups at C-8′, depended on the substitution at C-3′. Following the sequential order of acetoxyl group, α-OH, β-OH, and carbonyl group at C-3′, neuroinflammation inhibitory activities decreased (Fig. 8B). Thirdly, the presence of the C-14′-methyl was favorable to anti-inflammatory activity, although the hydroxylation of C-14′ decreased the effect, for example, with compound 1 versus substrate (IC₅₀ of 7.55 μM versus 3.72 μM). Finally, compounds with a double group present at Δ^5′,6′, for example, compounds 1 and 2, showed a dramatic decrease of activity, which indicated that the saturated β-ring in the bicyclic skeleton of sesquiterpene coumarins was critical for bioactivity.

Fig. 8 SAR of bicyclic sesquiterpene coumarins on anti-neuroinflammation activity (A): bicyclic sesquiterpene coumarins with a terminal double bond substituted at C-8′; (B): bicyclic sesquiterpene coumarins with the presence of methyl and hydroxyl groups at C-8′.

The possible anti-inflammatory mechanism for kellerin derivatives was likely to be similar to that of kellerin. In previous work, the primary mechanism of kellerin was tested by using a quantitative real-time polymerase chain reaction (qRT-PCR). The results showed that kellerin could significantly inhibit the mRNA expression of inflammatory factors TNF-α, IL-6, IL-1β and COX-2 induced by LPS in BV2 microglial cells at concentrations of 1–10 μM.⁴ Therefore, the kellerin derivatives described in this paper might also exhibit their effect by inhibiting the release of inflammatory cytokines in over-activated microglials. Thus, further research will be done to investigate the anti-inflammatory mechanism of the kellerin derivatives identified in this research.

Experimental

General experimental procedures

NMR spectra were recorded on an AVANCE III HD 600 MHz instrument (Bruker, Switzerland), with trimethylsilane (TMS) as an internal standard. The chemical shifts were measured relative to TMS and expressed in δ (ppm) and the coupling constants (J) are reported in Hertz (Hz). A quadropole time-of-flight mass spectrometer (Q-TOF-MS): micrOTOF-Q mass spectrometer (Bruker), was used to collect HR-ESI-MS data, in the m/z (rel.%) mode. Optical rotations were detected using a modular circular polarimeter: MCP 200 Polarimeter (Anton Paar). Silica gel (SiO₂: 50–74 μm) for chromatography was obtained from the Qingdao Ocean Chemical Group Co. (China). Sephadex LH-20 was purchased from Pharmacia (Switzerland). Octadecylsilyl (ODS; 100–50 μm) chromatographic columns were purchased from YMC (Japan). High performance chromatography (HPLC) separation was carried out on a ODS-A C₁₈ column (5 μm, 250 mm × 20 mm) equipped with a SPD-20A ultraviolet detector (Shimadzu) and a LC-20AR series pumping system (Shimadzu). Dimethyl sulfoxide (DMSO), deuterated DMSO (DMSO-d₆) and CD₃OD were obtained from the Sigma-Aldrich Company (St. Louis, MO, USA). All reagents were chromatographic grade or analytical grade and were obtained from Tianjin DaMao Chemical Company (Tianjin, China).

Substrate and culture system

Kellerin was isolated from the chloroform extract of the gum resin of F. sinkiangensis and was authenticated by comparing its physical and spectroscopic data to reported values.⁴ Its purity was determined to be 99.8% by HPLC analysis.

Petiole I (the portion adjacent to the leaf) of A. sinensis was collected from the South China Botanical Garden (Guangzhou, China), as explant material. Huang et al. induced the A. sinensis callus according to the optimized composition of callus induction media, plant organs, and callus type reported in previous work.¹⁴ Tissue culture conditions for A. sinensis were optimized using leaves and petioles (types I and II) as explant sources and petiole I was found to be the best plant organ for callus induction.

Petiole I was placed in running tap water for 30 min to wash away dust and other particles, then surface sterilized for 1 min by soaking in 75% (v/v) ethanol, then completely sterilized for 8–10 min in a 0.1% (w/v) aqueous solution of mercuric chloride, and washed five times in sterilized water (H₂O). Then petiole I was cut into approximately 1.5 cm segments, and cultured on Murashige and Skoog medium in 100 mL conical flasks, supplemented with 5 mg L⁻¹ of α-naphthaleneacetic acid (NAA), 0.5 mg L⁻¹ of 6-benzyladenine (BA), 0.7 mg L⁻¹ of 2,4-dichlorophenoxy acetic acid (2,4-D), 30 g L⁻¹ of sucrose, and 7.5 g L⁻¹ of agar, for induction of primary calli in a growth room at 25 °C ± 1 °C under complete darkness.¹⁴ After that, the primary calli were sub-cultured in 250 mL Erlenmeyer flasks containing Gamborg's B-5 medium with 1.0 mg L⁻¹ 2,4-D, 0.2 mg L⁻¹ of BA and 30 g L⁻¹ of sucrose at 25 °C ± 1 °C under darkness every three weeks.

Culture conditions and biotransformation procedures

The fresh primary A. sinensis callus was cultured on the B-5 liquid medium [15 g L⁻¹ sucrose (w/v)] with NAA (1.0 mg L⁻¹), 2,4-D (0.5 mg L⁻¹), and kinetin (0.05 mg L⁻¹). The pH value of the medium was adjusted to 5.75. Callus cells (6 g) were inoculated into a 250 mL Erlenmeyer flask with 120 mL B-5 medium, and then left under complete darkness at 25 °C ± 1 °C on a rotary shaker at 110 rpm for 20 days.

Kellerin (855 mg) was dissolved in MeOH (46.0 mL) and distributed equally into 57 bottles of the suspension cultures of A. sinensis callus. The control groups had the same concentration of kellerin but the flasks only contained the culture medium without cells or by just adding the same volume of MeOH to suspension cultured cells without substrate. All of the flasks were incubated under the same conditions throughout the experiment for 7 days at 25 °C ± 1 °C in the dark.

Extraction and isolation of biotransformation products

After incubation, the mixtures were filtered to separate the cultures and the medium. Then the medium was concentrated to 200 mL by evaporation in vacuo and extracted using the same volume of EtOAc and n-BuOH successively for three times to obtain Fr. 1 (EE; 0.70 g) and Fr. 2 (n-BuOH extract; 0.98 g). The cultures were dried under 55 °C, immersed in MeOH for 12 h and extracted using ultrasonic mixing for 30 min to obtain the MeOH extract (5.30 g), which was then resuspended in water and partitioned using EtOAc or n-BuOH to give Fr. 3 (EE; 0.21 g) and Fr. 4 (n-BuOH extract) (1.12 g), respectively. For the control group, parallel procedures were employed.

The yield of kellerin derivatives was calculated according using the formula:

Yield of product = amount of the product (mg)/[total amount of the substrate (mg) − amount of the substrate (non-transformed)].

The EE (0.91 g) of cultures and medium (Fr. 1 and Fr. 3) was combined and then subjected to HPLC on an ODS column with an elution gradient of MeOH/H₂O (0 : 100–100 : 0) to get five subfractions (Fr. 1a–5a). Next, Fr. 1a was separated using HPLC with an ODS column (250 mm × 10 mm, flow rate 3 mL min⁻¹) eluted with acetonitrile (MeCN)/H₂O (37 : 63) to give compound 11 (2.0 mg, retention time (t_R) = 27 min) and compound 12 (1.3 mg, t_R = 43 min). Then Fr. 2a was purified using HPLC with an ODS column (250 mm × 10 mm, flow rate 3 mL min⁻¹) to yield compound 8 (1.9 mg, 0.30%, t_R = 40 min) and compound 9 (2.3 mg, t_R = 56 min) using MeOH/H₂O (35 : 65) as the eluting solvent. Fr. 3a was firstly fractionated using silica gel column chromatography (3 × 50 cm) with a gradient elution of petroleum ether/acetone (0 : 100–100 : 0) to get sub-fractions, 3a-1 and 3a-2. Then 3a-1 was purified using HPLC with an ODS column (250 mm × 10 mm, flow rate 3 mL min⁻¹) using MeOH/H₂O (85 : 15) as a mobile phase, and this provided compound 10 (1.5 mg, t_R = 43 min) and compound 13 (1.8 mg, t_R = 59 min). Compound 14 (1.3 mg, t_R = 25 min) was obtained from sub-fraction 3a-2 using HPLC separation with an ODS column (250 mm × 10 mm, flow rate 3 mL min⁻¹) using MeOH/H₂O (56 : 44) as eluting solvent. Compound 5 (0.8 mg, 0.13%, t_R = 45 min) and compound 6 (1.3 mg, 0.20%, t_R = 59 min) were obtained from Fr. 5a using HPLC with an ODS column (250 mm × 10 mm, flow rate 3 mL min⁻¹) eluted with MeCN/H₂O (40 : 60). Both the combinations (BE, 2.10 g) of Fr. 2 and Fr. 4 were subjected to HPLC on an ODS column gradient eluted with MeOH/H₂O (from 0 : 100 to 100 : 0) to get seven fractions (Fr. 1b–7b). Fr. 1b was isolated using HPLC with an ODS column (250 mm × 10 mm, flow rate 3 mL min⁻¹) with MeOH/H₂O (55 : 45) and then purified with Sephadex LH-20 chromatography eluted with MeOH to give compound 1 (2.1 mg, 0.33%) and compound 2 (1.0 mg, 0.16%). Then, compounds 3 (1.2 mg, 0.19%, t_R = 53 min) and 4 (1.5 mg, 0.24%, t_R = 73 min) were obtained from Fr. 3b using HPLC with an ODS column (250 mm × 10 mm, flow rate 3 mL min⁻¹) eluted with MeOH/H₂O (55 : 45). Fr. 4b was loaded on a silica gel column (3 × 50 cm) with gradient elution of dichloromethane/MeOH of increasing polarity to get sub-fraction 4b-1 and 4b-2. Then 4b-1 was purified using HPLC with an ODS column (250 mm × 10 mm, flow rate 3 mL min⁻¹) with MeCN/H₂O (43 : 57) to obtain compound 7 (1.5 mg, 0.24%, t_R = 19 min) and substrate (220.0 mg, t_R = 39 min).

Structure characterization of new compounds

14′-Hydroxy-(3′S,4′R,5′S,8′R,9′S,10′R)-kellerin (1). Yellow oil. ¹H-NMR (600 MHz, CD₃OD), see Table 1. ¹³C-NMR (150 MHz, CD₃OD), see Table 1. HR-ESI-MS [M + Na]⁺ m/z 481.2196 (calcd for C₂₆H₃₄NaO₇, 481.2197).

5′,6′-Ene-14′-hydroxy-(3′S,4′R,8′R,9′S,10′R)-kellerin (2). Yellow solid. ¹H-NMR (600 MHz, CD₃OD), see Table 1. ¹³C-NMR (150 MHz, CD₃OD), see Table 1. HR-ESI-MS [M + Na]⁺ m/z 479.2054 (calcd for C₂₆H₃₂NaO₇, 479.2040).

5′,6′-Ene-(3′R,8′R,9′S,10′R)-ferukrin (3). White powder. ¹H-NMR (600 MHz, CD₃OD), see Table 1. ¹³C-NMR (150 MHz, CD₃OD), see Table 1. ESI-MS [M + Na]⁺ m/z 419.2 (calcd for C₂₄H₂₈NaO₅, 419.2).

14′-Hydroxy-(3′S,4′R,5′S,8′R,9′S,10′R)-deacetylkellerin (4). Yellow oil. ¹H-NMR (600 MHz, CD₃OD), see Table 1. ¹³C-NMR (150 MHz, CD₃OD), see Table 1. HR-ESI-MS [M + Na]⁺ m/z 439.2092 (calcd for C₂₄H₃₂NaO₆, 439.2091).

Viability assay, and measurement of NO production

Cell viability was determined using the methylthiazolyldiphenyl-tetrazolium bromide (MTT) reduction assay.⁸ Briefly, cells were plated on 96-well microtiter plates and pretreated with EE, BE at concentrations of 100 μg mL⁻¹, 30 μg mL⁻¹, 10 μg mL⁻¹, 1 μg mL⁻¹, 0 μg mL⁻¹ and the tested compounds were used at concentrations of 100 μM, 30 μM, 10 μM, 1 μM, 0 μM in the presence of LPS (100 ng mL⁻¹) for 24 h. Compared with the experimental group, the cells without LPS were used as a control group. Next the cells were incubated with MTT (0.25 mg mL⁻¹) for 4 h at 37 °C. The supernatant of the cells was removed and the formazan crystals were prepared as stock solutions in DMSO. The optical densities were read on a plate reader (Bio-Tek, USA) at 490 nm.

Determination of NO production was performed by measuring the accumulation of nitrite in the culture supernatant using Griess reagent, as previously described. Cells were seeded into 96-well microtiter plates and then stimulated with LPS (100 ng mL⁻¹) in the presence of EE, BE at concentrations of 100 μg mL⁻¹, 30 μg mL⁻¹, 10 μg mL⁻¹, 1 μg mL⁻¹, 0 μg mL⁻¹ and each tested sample was used at a concentration of 100 μM, 30 μM, 10 μM, 1 μM, 0 μM at 37 °C for 24 h. Compared with the experimental group, cells without LPS were used as a control group. A portion (50 μL) of the culture supernatant fluids was mixed with 50 μL of Griess reagent at room temperature. The absorbance was measured at 540 nm with a plate reader (Bio-Tek, USA) after fifteen minutes.

Conclusions

In this research, it is the first time sesquiterpene coumarins have been biotransformed using A. sinensis callus. It was discovered that the sesquiterpene moiety of the sesquiterpene coumarin skeleton is susceptible to being modified, which involved hydroxylation, dehydrogenation, dehydration and reduction reactions. However, the ether linkage bridged sesquiterpene and coumarin moieties readily fracture. Research is continuing to reveal more biotransformation regulars of the A. sinensis callus.

The anti-neuroinflammatory effect assay revealed that the sesquiterpene coumarin skeleton is a potential bioactive molecule for the development of neuroinflammatory inhibitors. Based on previous research, the SAR between sesquiterpene coumarin and neuroinflammation inhibitory activities was further deduced. The anti-inflammation activities are closely related to 3′-acetyl, 3′-α/β-OH and 14′-methyl moieties. It was then found that the IC₅₀ values seem to reduce following the order: 3′-acetyl < 3′-α-OH < 3′-β-OH < 3′-carbonyl in the skeleton with methyl and hydroxyl at 8′ whereas it was found that 3′-β-OH < 3′-acetyl < 3′-carbonyl < 3′-α-OH in the structure with a terminal double bond at 8′. The presence of 14′-methyl and a saturated β-ring may increase the inhibitory activities. The SAR mentioned previously may provide some useful guidance for future design of anti-neuroinflammatory agents with preferred physical–chemical properties.

Acknowledgements

The supporters of this work include the National Natural Science Foundation of China (U1403102, 81173531, 81473330, 81673323), the Shenyang Science and Technology Research Project (F15-199-1-26), the Research Project for Key Laboratory of Liaoning Educational committee (LZ2015067), the Natural Science Foundation of Liaoning Province (2015020732), the Innovation Team Project of Liaoning Province (LT2015027), the Fund for Long-term Training of Young Teachers in Shenyang Pharmaceutical University (ZCJJ2013409), Program for Liaoning Excellent Talents in University (LR2015022).

Notes and references

1. A. Gliszczyńska and P. E. Brodelius, Phytochem. Rev., 2012, 11, 77–96.
2. H. M. Gao and J. S. Hong, Trends Immunol., 2008, 29, 357–365.
3. M. T. Heneka and M. K. O'Banion, J. Neuroimmunol., 2007, 184, 69–91.
4. Y. C. Xing, N. Li, D. Zhou, G. Chen, K. Jiao, W. L. Wang, Y. Y. Si and Y. Hou, Planta Med., 2016 DOI:10.1055/s-0042-109271.
5. Y. Simeo and J. V. Sinisterra, Mini-Rev. Org. Chem., 2009, 6, 128–134.
6. S. Deng, B. J. Zhang, C. Y. Wang, Y. Tian, J. H. Yao, L. An, S. S. Huang, J. Y. Peng, K. X. Liu and X. C. Ma, Bioorg. Med. Chem. Lett., 2012, 22, 1615–1618.
7. R. N. Patel, Coord. Chem. Rev., 2008, 252, 659–701.
8. K. Ishihara, H. Hamada, T. Hirata and N. Nakajima, J. Mol. Catal. B: Enzym., 2003, 23, 145–170.
9. K. D. Deepak, S. K. Amit, K. B. S. Cyrus, O. R. Scott and P. O. John, Chem. Res. Toxicol., 2002, 15, 269–299.
10. Y. Simeo and J. V. Sinisterra, Mini-Rev. Org. Chem., 2009, 6, 128–134.
11. L. Z. Yi, Y. Z. Liang, H. Wu and D. L. Yuan, J. Chromatogr. A, 2009, 1216, 1999–2001.
12. J. P. Ma, Z. B. Guo, L. Jin and Y. D. Li, Chin. J. Nat. Med., 2015, 13, 0241–0249.
13. Y. Hou, C. F. Wu, J. Y. Yang, X. He, X. Y. Bi, L. Yu and T. Guo, Progr. Neuro. Psychopharmacol. Biol. Psychiatr., 2006, 30, 1523–1528.
14. B. Huang, L. J. Han, S. M. Li and C. Y. Yan, Pharmacogn. Mag., 2015, 13, 574–578.
15. Y. E. Perelson, V. I. Sheichenko and Y. E. Sklyar, Khim.-Farm. Zh., 1977, 11, 33–36.
16. M. E. Perelson, Y. E. Sklyar and N. V. Veselovskaya, Khim.-Farm. Zh., 1977, 11, 366–369.
17. J. A. Marco, J. F. Sanz and A. Yuste, Liebigs Ann. Chem., 1991, 9, 929–931.
18. J. L. Wang, L. N. Li and L. D. He, Nat. Prod. Res. Dev., 2007, 19, 427–429.
19. X. J. Yang, W. X. Huang and N. L. Wang, Chin. Herb. Med., 2005, 36, 1604–1607.
20. R. J. Chang, C. H. Wang, Q. Zeng, B. Guan, W. D. Zhang and H. Z. Jin, Arch. Pharmacal Res., 2013, 36, 1291–1301.
21. K. Stanislav, K. Antonín, P. Jiří and K. Janez, Heterocycles, 2002, 57, 1659–1682.
22. F. Insaf, B. Jalloul, Z. Mansour, B. G. Fatima and B. J. Hichem, J. Enzyme Inhib. Med. Chem., 2015, 30, 371–377.
23. Y. G. Jiao, F. Zhang and T. S. Dibble, J. Phys. Chem. A, 2015, 119, 7282–7292.
24. F. Robert and T. Kiyoshi, Chemosphere, 1999, 38, 973–978.
25. N. M. O'Boyle, A. Tenderholt and K. M. Langner, J. Comput. Chem., 2009, 29, 839–845.

---

Footnote

† Electronic supplementary information (ESI) available: The spectroscopic data of known compounds and cell viability results of transformed products. See DOI: 10.1039/c6ra22502k

---