Rubescins A, B and C: New Havanensin Type Limonoids from Root Bark of Trichilia rubescens (Meliaceae)

Abstract

Three new limonoids, rubescins A–C (1–3), and three known compounds including, havanensin type limonoid TS3 (4), β-sitosterol, and stigmasterol were isolated from the root bark of Trichilia rubescens. Their structures were elucidated by means of extensive spectroscopic analyses, particularly one dimensional (1D)- and 2D-NMR techniques in conjunction with mass spectrometry. Rubescins A–C (1–3) and limonoid TS3 (4) were evaluated for their protective effects against oxidative stress induced in HC-04 cells by H₂O₂. Compound (1) showed strong inhibitory effects on lactase dehydrogenase (LDH) leakage, being as active (IC₅₀ value of 0.0026 µM) as the positive control quercetin (IC₅₀ value of 0.0030 µM).

Plants belonging to the family Meliaceae, order Rutales, are well known sources of limonoids.¹⁾ This class of secondary metabolites has been a focus in natural products research for its structural diversity and potential biological significance.^1,2)

The tropical genus Trichilia, one of the 50 genera of the Meliaceae family,¹⁾ is comprised of approximately 90 species of trees that are widely distributed in tropical and subtropical regions of South America, continental Africa, and Madagascar¹⁾ with representative species being used in indigenous medicine.¹⁾ γ-Lactones, terpenoids, cycloartanes and limonoids have been reported as constituents of this genus.²⁾ Limonoids, the most characteristic secondary metabolites of the order Rutales, are tetranortriterpenoids derived from tirucallane (20S) or euphane (20R) triterpenoids with a 4,4,8-trimethyl-17-furanylsteroidal skeleton.²⁾ They exhibit a wide range of insecticidal,²⁾ antiviral,²⁾ insect antifeedant,²⁾ antibacterial,²⁾ antifungal,³⁾ anticancer,³⁾ antimalarial,⁴⁾ and cytotoxic⁵⁾ activities.

Trichilia rubescens OLIV. is a tree, growing mainly in tropical areas of Africa, and is used in Cameroonian folk medicine for the treatment of a variety of ailments including jaundice, fever, gonorrhea, malaria, and to induce labor in pregnant women.⁶⁾ Previous chemical investigations of the leaves of this species have led to the isolation of a series of havanensin-type limonoids denoted TS1, TS2, TS3, and trichirubines A and B which have been found to increase chloride conductance in epithelial cells and to exhibit antimalarial activity.^7,8)

As part of our ongoing research program on the Meliaceae family,⁹⁾ three new havanensin type limonoids named rubescins A–C (1–3) and three known compounds, were isolated from the root bark of Trichilia rubescens. Their structures were elucidated by spectroscopic methods.

In this report, we describe the isolation and structural elucidation of these limonoids, as well as their protective effects against oxidative stress induced in HC-04 cells by H₂O₂.

Results and Discussion

The air-dried and ground root bark of Trichilia rubescens was extracted at room temperature with a mixture of CH₂Cl₂–CH₃OH (1 : 1). Filtration and vacuum concentration of the resulting solution led to a dark greenish extract (yield: 15.5%). A bioassay-guided chromatographic fractionation over silica gel of the above extract, followed by purification using silica gel column chromatography resulted in the isolation of three new limonoids, rubescins A–C (1–3) and known compounds including, havanensin type limonoid TS3 (4),⁷⁾ β-sitosterol and stigmasterol as described in Experimental. The known compounds were identified by comparison of their spectroscopic data (¹H-, ¹³C-NMR and mass spectra) with those reported in literature.⁷⁾

Rubescin A 1, mp 234–236°C, [α]_D²⁶ − 41.5° (c=0.1, CHCl₃), was obtained as white crystals from a mixture of hexane–EtOAc (17 : 3). It reacted positively, both to Liebermann-Burchard and Erhlich tests suggesting its limonoidic nature. This was confirmed by the molecular formula determined as C₂₆H₂₈O₄ from its high-resolution electronic impact mass spectrum (HR-EI-MS) which showed the molecular ion peak [M]⁺ at m/z 404.1069 (Calcd for C₂₆H₂₈O₄, 404.1082) accounting for 13° of unsaturation. The infrared (IR) spectrum of compound 1 showed vibration bands at 1676 cm⁻¹ and 1647 cm⁻¹ indicative of the presence of two α,β-unsaturated carbonyl moieties. In accordance with the above molecular formula, 26 carbon signals were observed in the ¹³C broad band proton decoupled spectrum (Table 1). Its heteronuclear single quantum coherence (HSQC) spectrum revealed the presence of four tertiary methyl groups (δ_C 26.1, 24.7, 24.4, 18.2); ten olefinic carbons due to a β-substituted furan ring, two α,β-unsaturated carbonyl groups and an isolated double bond; two ketocarbonyl groups (δ_C 200.3, 193.0); four sp³ methylenes (one oxygenated); two sp³ methines and four sp³ quaternary carbons. These data suggested that 1 was a tetranorterpenoid similar to havanensin type limonoid TS3.⁷⁾

Table 1. ¹³C-NMR Data of Rubescins A–C (1–3) in CDCl₃^a)

Position	1	2	3
	δCb)	δCb)	δCb)
1	200.3 (s)	201.7 (s)	200.8 (s)
2	127.2 (d)	131.2 (d)	129.7 (d)
3	149.5 (d)	150.9 (d)	150.3 (d)
4	46.2 (s)	42.4 (s)	42.8 (s)
5	135.1 (s)	46.1 (d)	58.0 (d)
6	148.8 (s)	73.6 (d)	76.1 (d)
7	193.0 (s)	75.3 (d)	204.0 (s)
8	51.8 (s)	47.4 (s)	54.6 (s)
9	41.7 (d)	139.1 (s)	49.1 (d)
10	47.8 (s)	47.8 (s)	46.2 (s)
11	21.0 (t)	124.8 (s)	71.4 (d)
12	37.5 (t)	37.9 (t)	45.1 (t)
13	47.2 (s)	46.9 (s)	45.6 (s)
14	147.4 (s)	158.2 (s)	149.3 (s)
15	128.8 (d)	121.0 (d)	127.1 (d)
16	35.1 (t)	34.9 (t)	34.8 (t)
17	51.2 (d)	49.5 (d)	52.2 (d)
18	24.7 (q)	18.8 (q)	24.2 (q)
19	26.1 (q)	24.6 (q)	15.4 (q)
20	124.5 (s)	124.6 (s)	123.9 (s)
21	139.6 (d)	139.5 (d)	139.7 (d)
22	110.9 (d)	111.0 (d)	110.8 (d)
23	142.6 (d)	142.6 (d)	142.7 (d)
28	82.4 (t)	79.3 (t)	79.8 (t)
29	24.4 (q)	31.0 (q)	19.4 (q)
30	18.2 (q)	20.2 (q)	27.7 (q)
1′			169.7 (s)
2′			21.7 (q)

a) ¹³C-NMR spectra were recorded at 125 MHz. b) Chemical shifts are expressed in δ (ppm) downfield from TMS.

The ¹H-NMR spectrum of compound 1 (Table 2), coupled with its ¹H–¹H correlation spectroscopy (COSY) and HSQC spectra, displayed a set of signals at δ_H 7.39 (1H, t, J=1.6 Hz, H-23)/δ_C 142.6; 7.28 (1H, t, J=1.6 Hz, H-21)/δ_C 139.6 and 6.34 (1H, d, J=1.6 Hz, H-22)/δ_C 110.9 supporting the presence of a furan ring moiety. An α,β-unsaturated carbonyl moiety in ring A was indicated by a pair of a doublets at δ_H 6.76 (d, J=10.0 Hz, H-3) and δ_H 5.92 (d, J=10.0 Hz, H-2) and carbon resonances at δ_C 149.5, 127.2 and 200.3 assigned to C-3, C-2 and C-1,⁷⁾ respectively. A proton of an isolated double bond was evidenced by the broad singlet at δ_H 5.85 (1H, br s, H-15)/δ_C 128.8 which showed an heteronuclear multiple bond correlation (HMBC) correlations with carbon C-16 at δ_C 35.1. The presence of two diastereotopic oxymethylene protons attributed to protons H-28 was inferred from an AB spin system of two doublets at δ_H 4.44 (1H, d, J=9.0 Hz, H-28a)/δ_C 82.4 and δ_H 4.14 (1H, d, J=9.0 Hz, H-28b)/δ_C 82.4. Many other proton signals were also observed, including four methyl as singlets between δ_H 1.47 and 0.72. In the HMBC spectrum (Fig. 1), cross-peaks between H-29 (δ_H 1.56) and C-3 (δ_C 149.5), C-4 (δ_C 46.2), C-5 (δ_C 135.1), C-28 (δ_C 82.4); H-2 (δ_H 6.76) and C-4 (δ_C 46.2), C-10 (δ_C 47.8) and H-19 (δ_H 1.47) and C-1 (δ_C 200.3), C-5 (δ_C 135.1), C-9 (δ_C 41.7) permitted the assignment of one ketonic carbonyl at C-1 and a double bond between C-2 and C-3. The existence of the second α,β-unsaturated carbonyl moiety in ring B was revealed by HMBC correlations observed between H-30 (δ_H 1.46) and C-7 (δ_C 193.0), C-9 (δ_C 41.7), C-14 (147.4) and between H-28 (a, b) (δ_H 4.44, 4.14) and C-4 (δ_C 46.2), C-6 (δ_C 148.8) suggesting that the second conjugated ketonic carbonyl group was attached to C-7. The isolated double bond between C-14 and C-15 was deduced from HMBC correlations between H-18 (δ_H 0.72) and C-12 (δ_C 37.5), C-13 (δ_C 47.2), C-14 (δ_C 147.4), C-17 (δ_C 51.2); and between H-16 (δ_H 2.52) and C-15 (δ_C 128.8), C-14 (δ_C 147.4), C-17 (δ_C 51.2). The existence of an ether bridge between C-6 and C-28, forming a dihydrofuran ring, was confirmed by HMBC correlations originating from the oxymethylene protons H-28 (a, b) (δ_H 4.44, 4.14) to the carbons C-4 (δ_C 46.2) and C-6 (δ_C 148.8). The relative stereochemistry of compound 1 was deduced from biogenetic considerations, which give the β-orientation of proton H-17 as in the limonoid precursors which are euphane or tirrucalane.¹⁰⁾ Accordingly, the structure of compound 1, named rubescin A, was assigned as depicted.

Table 2. ¹H-NMR (Mult. J Hz) Data of Rubescins A–C (1–3) in CDCl₃^a)

Position	1	2	3
	δHd)	δHd)	δHd)
1
2	5.92 (1H, d, 10.0)	5.93 (1H, d, 10.0)	5.90 (1H, d, 10.0)
3	6.76 (1H, d, 10.0)	6.99 (1H, d, 10.0)	6.99 (1H, d, 10.0)
4
5		3.00 (1H, d, 12.0)	2.19 (1H, d, 15.0)
6		4.44 (1H, dd, 3.4,12.0)	5.04 (1H, d, 15.0)
7		4.17 (1H, d, 3.4)
8
9	2.56 (1H, td, 7.9, 1.9)		2.35 (3H, s)
10
11a	2.40 (1H, tq, 1.9, 7,5)	6.86 (1H, dd, 2.5, 7.0)	6.36 (1H, dt, 7.0, 2.0)
11b	1.60 (1H, tq, 1.9, 7,5)
12a	1.91 (1H, qd, 7.5, 2.3)	2.30 (1H, dd, 7.0,15.0)	2.20 (1H, dd, 7.0, 15.0)
12b	1.76 (1H, qd, 7.5, 2.3)	2.06 (1H, br d,c) 15.0)	1.88 (1H, d, 15.0)
13
14
15	5.85 (1H, br sb))	5.67 (1H, br s)	5.72 (1H, br s)
16a	2.52 (2H, m)	2.54 (2H, m)	2.61 (1H, m)
16b			2.51 (1H, m)
17	2.97 (1H, dd, 8.0, 10.0)	2.97 (1H, q, 8,5)	2.88 (1H, q, 8.0)
18	0.72 (3H, s)	0.67 (3H, s)	0.77 (3H, s)
19	1.47 (3H, s)	1.43 (3H, s)	1.50 (3H, s)
20
21	7.28 (1H, t, 1.6)	7.29 (1H, t, 1.5)	7.27 (1H, br sb))
22	6.34 (1H, d, 1.6)	6.32 (1H, d, 1.5)	6.32 (1H, br sb))
23	7.39 (1H, d, 1.6)	7.39 (1H, d, 1.5)	7.39 (1H, t, 1.5)
28a	4.44 (1H, d, 9.0)	3.82 (1H, d, 7.5)	3.87 (1H, d, 7.0)
28b	4.14 (1H, d, 9.0)	3.67 (1H, d, 7.5)	3.71 (1H, d, 7.0)
29	1.56 (3H, s)	1.33 (3H, s)	1.40 (3H, s)
30	1.46 (3H, s)	1.36 (3H, s)	1.81 (3H, s)
1′
2′			2.09 (3H, s)
7-OH		2.03 (1H, s)

a) ¹H-NMR spectra were recorded at 500 MHz. b) Broad singlet. c) Broad doublet. d) Chemical shifts are expressed in δ (ppm) downfield from TMS. The couplings in all spin systems were obtained by simulation analysis using the Spinworks program.

Fig. 1. Relevant COSY and HMBC Correlations of Compounds 1, 2 and 3

Rubescin B 2, mp 258–260°C, [α]_D²⁶ +39.6° (c=0.1, CHCl₃), was isolated as white needles from the mixture of hexane–EtOAc (4 : 1). It also reacted positively, both to Liebermann-Burchard and Erhlich tests suggesting its limonoidic nature. Its molecular formula, C₂₆H₃₀O₄ implying 12 sites of unsaturation, was established from its HR-EI-MS which gave a molecular ion peak [M]⁺ at m/z 406.2130 (Calcd for C₂₆H₃₀O₄, 406.2143), indicating that compound 2 has two hydrogen (one unsaturation less) atoms more than compound 1. Its IR absorption bands implied the presence of a free hydroxyl (3529 cm⁻¹) and an α,β-unsaturated ketone functionalities (1664 cm⁻¹). The ¹H-NMR (Table 2) and ¹³C-NMR data (Table 1) of 2 showed the characteristic limonoid β-substituted furan ring occurring at δ_H 7.39/δ_C 142.6, δ_H 7. 29/δ_C 139.5 and δ_H 6.32/δ_C 111.0, and an α,β-unsaturated ketone in ring A, indicated by a pair of doublets at δ_H 6.99 (d, J=10.0 Hz)/δ_C 150.9 and δ_H 5.93 (d, J=10.0 Hz)/δ_C 131.2 and δ_C 201.7. The comparison of the NMR spectra of compounds 1 and 2 showed a very close similarity except for two major differences. The first difference is the disappearance in the ¹³C-NMR spectrum of compound 2 of three sp² carbons signals due to the second α,β- unsaturated ketone moiety formed by C-5, C-6 and C-7 carbon atoms which were replaced by three sp³ carbons at δ_C 46.1, 73.6 and 75.3. The second difference is the appearance in ¹³C-NMR spectrum of the signals due to an additional trisubstituted double bond at δ_C 139.1 and δ_C 124.8, with an olefinic proton signal at δ_H 6.86 (1H, dd, J=2.5, 7.0 Hz). This double bond was assigned at C-9 and C-11 according to HMBC correlations between H-19, H-30 and the olefinic quaternary carbon at δ_C 139.1 (C-9). The relative configurations of protons H-5, H-6 and H-7 in compound 2 were established via nuclear Overhauser effect (NOE) experiments. Thus, the irradiation of H-6 proton signal at δ_H 4.44 led to the enhancement of the intensity of H-7 proton signal at δ_H 4.17, whereas no enhancement was observed for H-5 proton signal. This observation led to the conclusion that H-5 and H-6 are trans-oriented, while H-6 and H-7 are cis-oriented. These assumptions were also confirmed by the coupling constant values between H-5 and H-6 (³J_H5/H6=12.0 Hz) and between H-6 and H-7 (³J_H6/H7=3.6 Hz) which matched well with those of similar protons described in literature for trichirubine B,⁸⁾ another havanensin type limonoid. From the above spectroscopic data, structure 2 was assigned to this compound, and given the trivial name rubescin B.

Rubescin C 3, mp 250–252°C, [α]_D²⁶ −74.1° (c=0.1, CHCl₃), was isolated as white needles from hexane–EtOAc mixture (7 : 3). It also reacts positively, both to Liebermann-Burchard and Erhlich tests suggesting that compound 3, like compounds 1 and 2, is also a limonoid, having a molecular formula of C₂₈H₃₂O₆ as determined from HR-EI-MS, which showed the molecular ion peak [M]⁺at m/z 464.2175 (Calcd for C₂₈H₃₂O₆, 464.2197), accounting for 13° of unsaturation. Its IR spectrum displayed absorption bands at 1730 cm⁻¹, 1671 cm⁻¹and 1597 cm⁻¹ corresponding to ketone carbonyl, α,β-unsaturated carbonyl and ester carbonyl functionalities, respectively. The comparison of the ¹³C-NMR (Table 1) and ¹H-NMR (Table 2) spectral data of compound 3 with those of compound 1 revealed close similarity. The only difference between the two compounds was the absence of signals corresponding to C-5 and C-6 double bond in the ¹³C-NMR spectrum of compound 3, which were replaced by two sp³ signals resonating at δ_C 58.0 for C-5 and δ_C 76.1 for C-6. Additional signals were also observed in the ¹H-NMR spectrum of compound 3 at δ_H 2.09 (3H, s) and were attributed to a methyl of an acetyl moiety. The cross peaks observed in the HMBC spectrum between proton at δ_H 6.36 (dt) and carbons at δ_C 45.1 (C-12), δ_C 54.6 (C-8), δ_C 45.6 (C-13), δ_C 46.2 (C-10) and δ_C 169.7 (C-1′) led to the conclusion that the acetyl group was located at C-11. The β-orientation assigned to this acetyl group at C-11 position was deduced from the value of the coupling constants ³J_H11/H12a=7.0 Hz between H-11 and H-12a indicating their trans relationship and confirmed by comparison of this value with that of trichirubine A (³J_H11/H12a=8.0 Hz),⁸⁾ a limonoid in which the C ring substitution patent is similar to that of compound 3. From the above spectroscopic data, the structure of compound 3, to which the trivial name, rubescin C has been given, was therefore established as shown.

Oxidative stress is thought to be involved in the pathophysiology of malaria and the development of anemia induced by malaria.¹¹⁾ Thus, Trichilia rubescens, which is used in Cameroonian folk medicine for the treatment of malaria, might contain antioxidative compounds. That is the reason why, in addition to known havanensin limonoids TS3, the three new derivatives, whose structures have been described above, were evaluated for their capacity to protect HC-04 cells against oxidative stress induced by H₂O₂.

Viability of the cells remained unchanged when cells were treated with compounds 1, 3 and 4 for 24 h at concentrations of 1, 4, 20 and 50 µg/mL (Table 3). However, compound 2 and quercetin significantly (p<0.05) induced cell proliferation, the former at the tested concentrations and the later at the highest tested concentration compared with untreated control cells.

Table 3. Effect of Compounds 1–4 and Quercetin on HC-04 Cells Viability

Compounds	Concentrations (µg/mL)
	0	1	4	20	50
1	100.00±0.23	100.00±0.65	100.00±0.36	100.00±0.39	96.50±0.36
2	100.00±0.30	129.76±1.20a)	130.24±0.45a)	134.63±0.85a)	135.12±1.20a)
3	100.00±0.35	100.00±0.93	100.00±0.69	98.53±1.26	97.68±0.64
4	100.00±0.21	100.00±0.84	100.00±1.64	105.85±0.75	105.85±0.95
Quercetin	100.00±1.20	100.00±1.25	100.00±0.65	108.53±0.56	222.92±0.85a)

Values represent the mean±S.D. of samples in triplicate. a) Significantly different (p<0.05) compared to untreated control.

Lactase dehydrogenase (LDH) activity in the culture medium of HC-04 cells was significantly (p<0.05) increased when cells were treated with 1 mmol/L H₂O₂ for 24 h compared with untreated control cells. The enzyme activity in the medium was significantly (p<0.05) decreased in dependent-concentration manner following pre-treatment with compound 1 compared with the H₂O₂ intoxicated non-treated group while compounds 2 and 3 exhibited strong inhibition of LDH leakage into culture medium only at 1 and 4 µg/mL. The half inhibition concentration (IC₅₀) value (Table 4), concentration of compound required to inhibit LDH leakage by 50% from the intoxicated HC-04 cells of compound 1 was found to be 0.0026 µM.

Table 4. Effects (IC₅₀ Values) of the Compounds on LDH Release from HC-04 Cells Exposed to H₂O₂

	Compounds
	1	2	3	4	Quercetin
IC50 (µM)	0.0026±0.0001	ND	ND	ND	0.0030 ±0.001

Values represent the mean±S.D. of samples in triplicate. IC₅₀: Half inhibition concentration. ND: Not determined because the activity of the compound was not concentration-dependent at the tested concentrations. LDH: Lactate dehydrogenase.

A major contributor to oxidative damage is H₂O₂, which is converted to HO· by superoxide dismutase that leaks from the mitochondria.¹²⁾ In the literature, oxidative stress is induced within cells through treatment with exogenous hydrogen peroxide.^12–14) This oxidant is known to cause significant molecular damage within cells, leading in particular to the peroxidation of lipids, which subsequently rises to the production of a range of reactive oxygenated species (ROS) causing further damage to DNA, proteins and lipids.¹⁵⁾

Prerequisite for proper evaluation of the antioxidative properties of individual molecules or mixtures is the exact determination of cytotoxicity associated with prolonged incubation of the cells. The effects of the tested compounds on cell survival were determined by the reduction of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium salt (MTS), a measure of mitochondrial function.¹⁶⁾ Compounds 1, 3 and 4 did not affect HC-04 cells proliferation and therefore may be safe at tested concentrations. Compound 2 may induce cells proliferation. Regarding the standard quercetin, it seemed to precipitate in the culture medium at 20 and 50 µg/mL and leads to the increase of absorbance.

Cellular lesion results in membrane damage and leakage of some cytosolic enzymes including LDH.¹⁷⁾ In our experiments, intoxication of HC-04 cells with H₂O₂ significantly (p<0.05) raised the level of LDH to about 59% in H₂O₂ control compared to normal control. The leakage of the enzyme in the incubation medium indicates that cellular membranes were damaged. As shown by the activity of quercetin, antioxidant standard compound,^13,18) it is observed that apart from compound 4, the other tested compounds, particularly compound 1 interestingly at lower concentrations, significantly (p<0.05), reduced LDH leakage in media, suggesting protective effects against membrane oxidative damage in HC-04 cells exposed to H₂O₂. Conclusively, the active compounds, including rubescin A–C (1–3), identified in this study may be useful for fighting against oxidative stress and related health conditions. Further studies will be necessary to determine, using lower concentration, the IC₅₀ values of the others compounds 2, 3 and 4.

Experimental

General Procedures

Optical rotations were recorded on a Perkin-Elmer Model 2000 polarimeter in CHCl₃ solutions. Melting points were determined on a Buchii melting point apparatus and are uncorrected. IR spectra were recorded on Bruker Fourier transform/infrared (ATR) spectrophotometer. ¹H- and ¹³C-NMR spectra were run on a Bruker AV NMR instrument equipped with 5 mm ¹H and ¹³C probes operating at 500 and 125 MHz, respectively with tetramethylsilane (TMS) as internal standard. Mass spectra were obtained on a Varian Mass Spectrometer. Silica gels (Merck, 230–400, 70–230 mesh) were used for flash and column chromatography. Thin layer chromatography (TLC) analyses were performed on Kiesel gel 60F254 precoated alumina sheets (0.2 mm layer thickness). Spots were visualized under UV lamp (254, 365 nm) or by heating after spraying with 10% H₂SO₄ reagent. Different mixtures of hexane, EtOAc, CH₂Cl₂ and MeOH were used as eluent solvents.

Plant Material

The root bark of Trichilia rubescens were collected in October 2011 at Mendong (Eloumden) locality situated in the Central Region of the Republic of Cameroon. The botanical identification of the plant was made by Mr. Nana plant taxonomist at the National Herbarium of Cameroon, where a voucher specimen was deposited under N° 38705/SRF Cam.

Extraction and Isolation

The air dried and powdered root bark (1 kg) of Trichilia rubescens was extracted by maceration at room temperature for 24 h with a mixture of CH₂Cl₂–MeOH (1 : 1) to yield, after filtration and removal of the solvent under vacuum, 155 g of a dark greenish extract.

One hundred and fourty four grams of this extract was subjected to flash silica gel chromatography with hexane–EtOAc mixtures (4 : 1, 3 : 1), EtOAc–MeOH (1 : 20) and MeOH to afford four fractions (F_1, F₂, F₃ and F₄) monitored by TLC.

Fraction F₂ (20 g) was subjected to silica gel column chromatography using a gradient mixture of hexane–EtOAc (20 : 1 to 5 : 1) to yield ninety fractions of 125 mL each. Fractions of similar compounds were combined, based on their TLC profiles to give five series (S₁ (1–29), S₂ (30–41), S₃ (46–66), S₄ (69–72) and S₅ (77–90)).

Series S₁, mainly low polar compounds rich extract, was subjected to silica gel column chromatography with a mixture of hexane–EtOAc as eluent to yield β-sitosterol (20 mg) and stigmasterol (15 mg) using hexane–EtOAc (19 : 1). S₂ crystalizes to afford, after filtration, compound 1 (30 mg). S₄ was equally subjected to silica gel column chromatography to give compound 2 (40 mg) using hexane–EtOAc (4 : 1). S₅ was respectively chromatographed on a silica gel column eluted successively with a gradient mixtures of hexane–EtOAc to yield compound 3 (55 mg) using hexane–EtOAc (7 : 3) and compound 4 (150 mg) using hexane–EtOAc (3 : 2).

Rubescin A (1): White crystals. mp 234–236°C. [α]_D²⁶ −41.5° (c=0.1, CHCl₃). IR (KBr) cm⁻¹: 3495, 1660. ¹H-NMR (see Table 2) and ¹³C-NMR (see Table 1). EI-MS m/z: 404 [M]⁺ (75), 389 (100), 95 (70), 43 (60), 91(53). HR-EI-MS m/z 404.1069 [M] ⁺ (Calcd for C₂₆H₂₈O₄, 404.1082).

Rubescin B (2): White needles. mp 258–260°C. [α]_D²⁶ +39.6° (c=0.1, CHCl₃). IR (KBr) cm⁻¹: 3529, 1664. ¹H-NMR (see Table 2) and ¹³C-NMR data (see Table 1). EI-MS m/z: 406 [M]⁺ (66), 95 (60), 281 (48), 43 (34), 81 (28). HR-EI-MS m/z: 406.2130 [M] ⁺ (Calcd for C₂₆H₃₀O₄, 406.2143).

Rubescin C (3): White needles. mp 250–252°C. [α]_D²⁶ −74.1° (c=0.1, CHCl₃). IR (KBr) cm⁻¹: 1730, 1671, 1597. ¹H-NMR (see Table 2) and ¹³C-NMR data (see Table 1). Electrospray ionization-time-of-flight (ESI-TOF)-MS m/z [M+H]⁺ 465.2, [M+Na]⁺ 487.5. HR-EI-MS m/z: 464.2175 [M]⁺ (Calcd for C₂₈H₃₂O₆, 464.2197).

Cells Culture and Treatment

Hepatocyte line HC-04 established from normal human liver cells¹⁹⁾ and characterized,²⁰⁾ were obtained from Dr. Urs A. Boelsterli. Cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Hyclone, Fisher Scientific, PA, U.S.A.) supplemented with 10% heat-inactivated fetal bovine serum, penicillin–streptomycin cocktail purchased from Gibco in a humidified incubator with 5% CO₂ at 37°C. Cells were grown to 80% confluence before pretreatment. Isolated compounds and quercetin the reference antioxidant compound were dissolved in dimethyl sulfoxide (DMSO) before adding to culture medium.

Cell Viability Assay

Cell viability was quantified using the MTS assay (CellTiter 96 Aqueous Cell Proliferation Assay, Promega, Madison, WI, U.S.A.) kit following the manufacturer’s instructions. For this assay, 25000 cells in 100 µL complete-DMEM were seeded/well in triplicate in a 96-wells plate and incubated at 37°C in 95% air and 5% CO₂ for 24 h. Then, groups of cells were treated with Serum Free-DMEM, 0.1% DMSO, 1 mmol/L H₂O₂, tested compounds at concentrations 1, 4, 20 and 50 µg/mL and incubated for more than 24 h. At the end of this second incubation time, 20 µL of MTS reagent was added in the dark to each well and incubated for 1 h. After, the reaction was stopped by adding 25 µL of the 10% sodium dodecyl sulfate (SDS) in the dark. A 1/5 dilution with double distilled water was done in a new 96-wells plate and the absorbance was recorded at 490 nm using microplate reader for viability percentages calculation.

Determination of Cell Integrity (LDH Leakage Assay)

Cell integrity was determined by monitoring LDH leakage by commercial assay kit (Sigma) following the manufacturer’s instructions. Cells were plated at a concentration of 2×10⁵ in 1 mL DMEM per well in 24-wells plate and cultured at 37°C in 95% air and 5% CO₂ for 24 h. Quercetin was used as reference antioxidant compound and cells were treated with test compounds at concentrations 1, 4, 20 and 50 µg/mL and incubated again for 24 h and followed by 1 mmol/L H₂O₂ intoxication and another 24h incubation period. After the last incubation, 360 µL of the incubation medium was pipetted and put into eppendorf tubes containing 40 µL of 10% Triton X-100. Then, the remaining medium was aspirated and cells lysed in 1 mL 10% Triton X-100. The activity of LDH in the incubation solution (medium) and in the detergent extract (lysate) of cells was determined in 96-wells plate by recording absorbance using microplate reader at 490 nm and 690 nm. The results are expressed as the percentage of the total LDH content of the well which appeared in the medium (LDH medium): percent release=100×LDH medium/(LDH medium+LDH lysate).

Acknowledgements

One of the authors T.T.A., acknowledges the Canadian Commonwealth Exchange Program-Africa for their support through four month (August–December 2009) scholarship at the University of Manitoba. Likewise, N.F.N. thanks the Fulbright Program for the award which allowed him to visit and carry out experiments in the laboratory of José E. Manautou at the School of Pharmacy, Connecticut, U.S.A.

References

• 1)  Gouvea C. F., Dornelas M. C., Rodriguez A. P. M., Ann. Bot. (Lond.), 101, 39–48 (2008).
• 2)  Champagne D. E., Koul O., Isman M. B., Scudder G. G. E., Towers G. H. N., Phytochem., 31, 377–394 (1992).
• 3)  Mulholland D. A., Parel B., Coombes P., Curr. Org. Chem., 4, 1011–1054 (2000).
• 4)  Mohamad K., Hirasawa Y., Litaudon M., Awang K., Hadi A. H. A., Takeya K., Ekasari W., Widyawaruyanti A., Zaini N. C., Morita H., Bioorg. Med. Chem., 17, 727–730 (2009).
• 5)  Wong C. P., Shimada M., Nagakura Y., Nugroho A. E., Hirasawa Y., Kaneda T., Awang K., Hadi A. H. A., Mohamad K., Shiro M., Morita H., Chem. Pharm. Bull., 59, 407–411 (2011).
• 6)  Louppe D., “Bois d’œuvre 1,” Prota, France, 2008, p. 785.
• 7)  DeCarvalho A. C. V., Ndi C. P., Tsopmo A., Tane P., Ayafor J., Conolly J. D., Teem J. L., Mol. Med., 8, 75–87 (2002).
• 8)  Sabrina K., Martin M. T., Grellier P., Kasenene J., Sevenet T., Antimicrob. Agents Chemother., 48, 3196–3199 (2004).
• 9)  Vardamides J. C., Sielinou T. V., Akone H. S., Nkengfack A. E., Abegaz B. M., Nat. Prod. Commun., 5, 1535–1538 (2010).
• 10)  Corrado T., “Bioactive Compounds from Natural Sources: Isolation, Characterisation and Biological Properties,” Chap. 13, ed. by Nakatani M., London, 2001, pp. 529–550.
• 11)  Kulkarni A. G., Suryakar A. N., Sadeshmunth A. S., Rathi D. B., Indian J. Clin. Biochem., 18, 136–149 (2003).
• 12)  Xu D., Zhou C. Neural., Neural. Regener. Res., 5, 1391–1395 (2010).
• 13)  Subramaniam S. R., Ellis E. M., Toxicol. Appl. Pharmacol., 250, 130–136 (2011).
• 14)  Back M. J., Jun M., Jeong W. S., Int. J. Mol. Sci., 13, 2314–2330 (2012).
• 15)  Esterbauer H. C., J. Clin. Nutr., 57, 779S–786S (1993).
• 16)  Meng X., Reeves W. B., Am. J. Physiol. Renal Physiol., 278, 83–90 (2000).
• 17)  Rang H. P., Dale M. M., Ritter J. M., “Pharmacology,” 4th ed., Churchill Livingstone, New York, Philadelphia, Sydney, Toronto, 1999.
• 18)  Alía M., Ramos S., Mateos R., Granado-Serrano A. B., Bravo L., Goya L., Toxicol. Appl. Pharmacol., 212, 110–118 (2006).
• 19)  Sattabongkot J., Yimamnuaychoke N., Leelaudomlipi S., Rasameesoraj M., Jenwithisuk R., Coleman R. E., Udomsangpetch R., Cui L., Brewer T. G., Am. J. Trop. Med. Hyg., 74, 1390–1401 (2006).
• 20)  Lim P. L. K., Tan W., Latchoumycandane C., Mok W. C., Khoo Y. M., Lee H. S., Sattabongkot J., Beerheide W., Lim S. G., Tan T. M. C., Boelsterli U. A., Toxicol. In Vitro, 21, 1390–1401 (2007).