(6R,9S)-6′-(4″-Hydroxybenzoil)-Roseoside, a new megastigmane derivative, isolated from Ouratea polyantha Engl. (Ocnhaceae) and its effect over hepatic glucose-6-phosphatase

Abstract

A new megastigmane derivative, (6R,9S)-6′-(4″-hydroxybenzoil)-roseoside 1 and two known compounds, the biflavone agathisflavone 2 and the carboxylic acid derivative 4-hydroxibenzoic acid 3 were isolated and purified from leaves and stems of Ouratea polyantha Engl. Agathisflavone was isolated in a single high-speed counter-current chromatography run, meanwhile the megastigmane and the carboxylic acid derivatives were purified in two steps, a combination of high-speed countercurrent chromatography and analytical column chromatography. All structures were elucidated on the basis of spectral evidences and comparisons with literatures. Compound 1 was characterized by [α]_D²⁰, UV-Vis, IR, MS, ¹H-NMR, ¹³C-NMR, HMQC, HMBC, COSY and NOESY. Compounds 1 and 2 were tested as hepatic glucose-6-phosphatase inhibitors.

1. Introduction

Diabetes mellitus is the commonest endocrine disorder [1] and it is characterized by hyperglycaemia resulting from insufficient insulin production and/or insulin resistance [2]. In Venezuela It is commonly treated with herbal extracts. Although the utility of plant materials is well known in folk medicines based on the empirical knowledge, it needs to be scientifically confirmed because there are many risks on plant material ingestion by decoction, infusion, beverage, etc. [3].

The genus Ouratea belongs to the Ochnaceae family. It is widely represented in tropical areas and comprises approximately 35 genera and 600 species [4]. Several studies of Ouratea species describe the isolation of chloroisoflavonoids [5], biflavonoids [6,7], terpenes like megastigmane [8], kauranes [9], depside [10], among other metabolites. Rodriguez et al. [11] reported that MeOH extract from the aerial parts of Ouratea polyantha Engl. (Op) had a strong inhibition on glucose-6-phosphatase, (G-6-Pasa) EC 3.1.3.9.

The enzyme G-6-Pase is a multi-component system constituted by a catalytic subunit and three transporters: T1 for glucose-6-phosphate (G-6-P), T2 for phosphate/pyrophosphate, and T3 for glucose [12] and both, enzyme and its transporters, are potential targets for anti-diabetic therapy [13–14]. Fractions from methanol extract of O. polyantha showed inhibitory effects on hepatic microsomal G-6-Pase. Therefore, an effective separation and isolation technique is required for the identification of the bioactive fractions.

Countercurrent chromatography (CCC) is a technique used to separate mixtures into their individual components that was first developed in late 1970s and refined to High Speed Countercurrent chromatography (HSCCC) in 1980s when it overshadowed other chromatography methods with its superior capacity to achieve rapid and efficient separation. This chromatographic system is now employed in a wide range of applications, most notably for extracting bioactive compounds from medicinal plants [15].

In the present paper, we report the isolation and structural elucidation of three compounds from O. polyantha aerial parts (leaves and stems) that inhibit rat liver microsomal G-6-Pase.

2. Experimental

2.1. General

Phlorizin, G-6-P and histones II-A were acquired from Sigma-Aldrich (Milwaukee, USA). All other chemicals used were analytical grade.

High-speed counter-current chromatography were performed using two chromatograph, one was an Ito Multi-Layer Coil Separator-Extractor (P.C. Inc. Potomac, MD, USA) with a single column of 325 mL and 1.6 mm internal diameter with a β between 0.5 and 0.85, and the other one was a CCC-1000 High-speed Counter-Current Chromatograph (Pharma-Tech Research, Baltimore, MD, USA) equipped with three coils connected in series (inner diameter of tubing 1.6 mm) with a β between 0.5 and 0.75. The total capacity of the column is 325 mL.

The NMR was carried out in a JEOL spectrometer model Eclipse with an application camp of 270 MHz for ¹H and 67.5 MHz for ¹³C, and Bruker spectrometer with an application camp of 500 MHz for ¹H and 125 MHz for ¹³C.

2.2. Plant material

Aerial parts (leaves and stems) of O. polyantha Engl. were collected in the Sipapo River: near to Cerro Pelota, southern Laja de Garza, between Autana and Guayapola Rivers at the Amazon forest, Amazon State, Venezuela in 1992, and identified by Dr. Anibal Castillo from the Biology School, Science Faculty, Venezuela Central University. A voucher specimen (N° 3308AC) was deposited in the Venezuela National Herbarium, (VEN).

2.3. Animals

Male Sprague–Dawley rats of 180–220 g were used after an overnight fast period.

2.4. Purification of Microsomes and Glucose-6-phosphatase Assay

The microsomal fraction was purified following the method described Marcucci et al. [16], in brief: the rat livers were homogenized in 3 volumes of 0.32 M sucrose, 3 mM MgCl₂, centrifuged at 20000g for 20 min. at 4° C, the pellet was discarded and the supernatant was centrifuged at 105000g for 1 h at 4° C, and the pellet constituted the microsomal fraction. The microsomal fraction was resuspended in 0.25 mM sucrose, 1 mM MgCl₂, 5 mM HEPES (pH 6.5) to give a final protein concentration of 20 mg/mL and frozen at −80°C until use. Protein concentration was estimated using the Lowry method [17] modified by Markwell et al. [18].

G-6-Pase assays were performed by the method described by Burchell et al. [19] with intact and disrupted (histone treated) microsomes. In brief, the G-6-Pase hydrolyze glucose-6-phosphate (G-6-P) and produce glucose an inorganic phosphate ion, the last one was converted to the blue complex with ammonium heptamolypdate in acidic medium and measured at 820 nm using a Novaspec II spectrophotometer (Pharmacia).

2.5. Metabolites Isolation

Dried and powdered vegetal material, leaves and stems (944.04 g), was macerate and extracted with MeOH at room temperature to produce methanol extract, MeOH_rt. After that, the residual material was extracted with MeOH using a soxhlet extractor to obtain the hot methanolic extract (MeOH_Δ). Both extracts were concentrated in vacuo yielding 174.82 g (18.52%) and 8.14 g (0.86%) for MeOH_rt and MeOH_Δ, respectively. They were tested as G-6-Pase inhibitors (Table 1).

Table 1.

Effects of O. polyantha MeOH extracts, fractions and metabolites on the hepatic microsomal glucose-6-phosphatase.

	ppm	n	Intact microsomes (%)	Disupted microsomes (%)
Control		4	---	---
MeOHrt	10.0	3	67.5	13.6
MeOHΔ	10.0	3	38.6	6.7
CH2Cl2a	10.0	3	41.5	12.8
AcOEta	10.0	3	55.6	7.5
OpA	10.0	4	90.6	94.3
OpB	10.0	4	67.9	42.8
OpC	10.0	4	87.6	71.2
OpD	10.0	4	92.5	45.2
Agathisflavone	2.5	3	63.6	63.1
(6R,9S)-Roseosideb	20.0	3	58.5	12.2c
(6R,9S)-6′-(4″-hydroxybenzoil)-roseoside	100.0	8	13.7	2.2c
Phlorizin	20.0	3	27.8	0

^adata from Rodríguez et al., [11];

^bdata from Avila et al., [24];

^cenzyme activation.

Twenty grams of the MeOH_rt extract were suspended in a mixture of methanol-water (1:1; v/v) yielding two fractions, soluble (Op-1) and insoluble (Op-2). Op-1 was concentrated in vacuo, while Op-2 was air-dried. Both fractions were treated with acetone separately, yielding four new fractions named OpA (5.19 g, 25.95%) and OpB (1.36 g, 6.80%) from Op-1 and OpC (5.96 g, 29.80%) and OpD (1.72 g, 8.60%) from Op-2. OpA and OpC were acetone non-soluble while OpB and OpD were acetone soluble.

Fractions OpA, OpB and OpC were fractionated using a CCC-1000 HSCCC from Pharma-Tech Research (Unplublished results), and OpD was analyzed using an HSCCC instrument from P.C. Inc.

OpB and OpD fractions had a common biflavonoid aglycone as the major compound determined by the iodine reagent. In order to obtain this biflavonoid as pure compound the sample solution was prepared by dissolving the crude fraction, OpD, in a solution composed of the upper and lower phases (1:1, v/v) of the solvent system used in HSCCC separation. The selected solvent system (CHCl₃-MeOH-H₂O, 4:3.5:2, v/v/v) was thoroughly equilibrated in a separation funnel by repeated vigorous shaking at room temperature. The two phases were separated shortly before use. The separation was initiated by filling the entire column with the stationary phase (upper phase) and this was immediately followed by sample inyection dissolved in a mixture of stationary and mobile phases. The mobile phase was then eluted through the column at 2 mL/min while the column was rotated at 800 rpm in the combined head to tail elution mode.

The fractions were monitored and combined by TLC similarity yielding 16 collective fractions (Fig. 2). From OpD-Fr₅ (57.00 mg) a pure biflavonoid, agathisflavone was isolated. OpD-Fr₉ was purified on a silice gel column chromatography using the organic phase of the solvent system CH₂Cl₂-n-BuOH-MeOH-H₂O (4:0.1:1.5:2, v/v) as eluent to afford 4.91 mg of (6R,9S)-6′-(4″-hydroxybenzoil)-roseoside, while OpD-Fr_10–11 was purified on a silice gel column chromatography using the mixture CH₂Cl₂-MeOH (94:6, v/v) as an eluent to obtain 2.27 mg of 4-hydroxybenzoic acid. In brief, three pure compounds were identified: agathisflavone, (6R,9S)-6′-(4″-hydroxybenzoil)-roseoside and 4-hydroxybenzoic acid and their structure are shown in Fig. 2.

Figure 2.

Structures of compounds 1, 2 and 4.

2.5.1. (6R,9S)-6′-(4″-hydroxybenzoil)-roseoside

Amorphus solid, Rf value: 0.68 (AcOEt-MeOH-H₂O; 100:13.5:10; v/v/v), UV-Vis (MeOH) λ_max: 229, 298 nm; IR (films) υ_max: 3410.5, 2970.9, 2929.7, 1692.0, 1652.6, 1589.5, 1500.7, 1442.7, 1374.5, 1277.5, 1163.9, 1067.6, 854.0 cm⁻¹; ¹H-NMR (500 MHz, CD₃OD) and ¹³C-NMR (125 MHz, CD₃OD), given in Table 2; TSI-MS m/z (rel. Int.); 506.88 (100), 488.88 (10), 282.96 (97), 264.96 (75), 207.00 (92), 189.00 (34), 149.04 (52), 138.96 (75), 120.96 (45), scheme 1.

Table 2.

¹H (500 MHz) and ¹³C-NMR (125 MHz) Chemical Shifts of (6R,9S)-6′-(4″-hydroxybenzoil)-Roseoside.

H/C	δH (ppm)	δC (ppm)	DEPTa	HMQC	HMBC	COSY	NOESY
1		40.93	C		H-11; H-12
2	2.44	49.28	CH2	H-2	H-11; H-12	H′-2	H′-2;H-11;H-7
	2.15			H′-2		H-2	H-2;H-11;H-12
3		199.93	C
4	5.85	125.70	CH	H-4	H-10	H-13	H-13
5		165.53	C		H-10
6		78.46	C		H-10;H-11;H-12;H-13
7	5.76	132.44	CH	H-7			H-2; H-12
8	5.69	132.25	CH	H-8	H-10	H-9	H-2; H-10
9	4.39	73.49	CH	H-9	H-10; H-1′	H-8; H-10	H-10; H-2′
10	1.25	20.78	CH3	H-10	H-8	H-9	H-8; H-9
11	0.94	23.31	CH3	H-11	C-1; C-2; C-6; C-12		H-2; H′-2
12	1.00	22.03	CH3	H-12	C-1; C-2; C-6; C-11		H′-2; H-7
13	1.91	18.21	CH3	H-13	C-4; C-5; C-6	H-4	H-4
1′	4.29	99.79	CH	H-1′	C-9; H-10	H-2′	H-5′
2′	3.22	73.49	CH	H-2′	C-4′	H-1′; H-3′	H-9; H-10
3′	3.29–3.32	76.81	CH	H-3′		H-2′	H-5′;H-3″;H-5″
4′	3.36	70.59	CH	H-4′	H-2′	H-5′
5′	3.40–3.47	74.18	CH	H-5′	H′-6′	H-6′; H′-6′	H-3′
6′	4.61	63.53	CH2	H-6′	C-5′; C7″	H′-6′	H′-6′;H-5′
	4.37			H′-6′		H-5′; H-6′	H-6′
1″		120.73	C		H-3″; H-5″
2″	7.87	114.88	CH	H-2″	C-7″	H-3″	H-3″
3″	6.81	131.42	CH	H-3″	C-1′	H-2″	H-3′;H-2″
4″		162.22	C
5″	6.81	131.42	CH	H-5″	C-7″	H-6″	H-3′;H-6″
6″	7.87	114.88	CH	H-6″	C-1′	H-5″	H-5″
7″		166.58	C		H-2″; H-6′; H-6″

^aat 270 MHz

2.5.2. Agathisflavone (6→8″-biapigenin)

Yellow solid, Rf value: 0.93 (AcOEt-MeOH-H₂O; 100:13.5:10; v/v/v), UV-Vis (MeOH) λ_max: 229, 298 nm; ¹H-NMR (500 MHz, CD₃OD): δ 13.30 (1H, bs, 5-OH), 13.04 (1H, bs, 5″-OH), 8.01 (2H, d, J=8.75 Hz, H-2′, H-6′), 7.58 (2H, d, J=8.75 Hz, H-2′, H-6′), 6.97 (2H, d, J=8.75 Hz, H-3′, H-5′), 6.77 (2H, d, J=8.75 Hz, H-3′, H-5′), 6.88 (1H, s, H-3), 6.79 (1H, s, H-3″), 6.70 (1H, s, H-8), 6.38 (1H, s, H-6″); ¹³C-NMR (125 MHz, CD₃OD), δ 182.51 (C-4″), 182.30 (C-4), 164.16 (C-2″), 163.99 (C-2), 163.06 (C-7″), 162.95 (C-7), 161.65 (C-4‴), 161.48 (C-4′, C-5), 161.06 (C-5″), 157.14 (C-9″), 155.23 (C-9), 129.02 (C-2‴, C-6‴), 128.50 (C-2′, C-6′), 121.83 (C-1′), 121.66 (C-1‴), 116.45 (C-3′, C-5′), 116.33 (C-3‴, C-5‴), 104.11 (C-10), 103.95 (C-10″), 103.68 (C-3″), 103.31 (C-3), 103.05 (C-6″), 99.54 (C-8), 99.00 (C-6), 93.85 (C-8″).

3. Results and discussion

(6R,9S)-6′-(4″-hydroxybenzoil)-roseoside 1 was isolated as an amorphous solid ([α]_D²⁰ – 57). The UV spectrum of 1 (λ_max 229 nm) and the IR absorption (ν_max 1652.6 cm⁻¹) indicated the presence of an α,β-unsaturated ketone. The molecular formula of 1 was established as C₂₆H₃₄O₁₀ from the TSI-MS [m/z 507.88 (M)⁺, 529.40 (M+Na)⁺ and 1035.40 (2M-H+Na)⁺] and NMR data. The mass fragmentation of the ion peak at 506.88 gave fragments for water loss at 488.88 (rel. Int., 10), sugar fragmentations at 282.96 (97) and 264.96 (75), alkene fragmentations at 207.00 (92) and 149.04 (52), ester fragmentation at 120.96 (45) and ester fragmentation throught McLafferty rearrangement (Figure 3). The IR spectrum showed the presence of hydroxyl (broad band 3410.5 cm⁻¹), a conjugated ester (1692.0, 1277.5, 1067.6 cm⁻¹) and a trans olefin (1589.5 cm⁻¹). The ¹H and ¹³C-NMR spectra (Table 2), which were assigned by 2D experiments (COSY, HMQC, HMBC and NOESY) of 1 showed the presence of a 4-hydroxybenzoil and β-glucopyranosyl units and an aglycone moiety.

Figure 3.

Proposed fragmentation mechanism of (6R,9S)-6′-(4″-hydroxybenzoil)-roseoside.

The ¹H-NMR spectrum showed two doblet signals at δ 7.87 (2H, d, J=8.75, H-2″ and H-6″) and 6.82 (2H, d, J=8.75, H-3″ and H-5″), attributed to the AA′BB′ system in a 1,4- substituted benzene ring, assigned to the 4-hydroxybenzoil. In the HMBC spectrum, the proton signal at δ 7.87 showed a cross peak with the carbon signal at δ 166.58 (C-7″) which exhibited a long range coupling with the proton signal at δ 4.61 (1H, bs, H-6′) assignable to one proton of methylene of sugar unit, thus the location of 4-hydroxybenzoil group in the pyranosyl moiety was established at C-6′ (Fig. 4). In the ¹H-NMR spectrum, a signal at δ 4.29 (1H, d, J=7.8 Hz, H-1′) with a large coupling constant (7.8 Hz) suggested an anomeric proton with β-configuration. Furthermore, the presence of the COSY cross peak between H-6′ and the signal at δ 3.40–3.47 (1H, m, H-5′) and NOESY correlation between H-5′ and the signals at δ 4.29 (1H, d, J=7.8 Hz, H-1′) and δ 3.29–3.32 (1H, m, H-3′) demonstrated the presence of glucopyranosyl moiety (Figs. 5 and 6).

Figure 4.

COSY (bold bonds) and HMBC (→) correlation of (6R,9S)-6′-(4″-hydroxybenzoil)-roseoside.

Figure 5.

NOESY correlation of (6R,9S)-6′-(4″-hydroxybenzoil)-roseoside.

The position of the glucosydic linkage was deducted from the ¹³C-NMR, chemical shift of C-9 (δ_C 73.49), and this was confirmed by long correlation between H-1′ and C-9. In addition to the signals of the 4-hydroxybenzoil and glucopyranosyl groups, the remaining proton signals, corresponding to 13 carbons, were assignable to the aglycone, which include two trans-coupled olefinic protons at δ 5.76 (1H, d, J=15.6 Hz, H-7) and δ 5.69 (1H, dd, J=15.6, 6.8 Hz, H-8), and a olefinic proton at δ 5.85 (1H, s, H-4) located on α-position of an α,β-unsaturated ketone and four methyl groups at 0.94, 1.00, 1.25 and 1,91. The deshielded methyl signal at δ 1.91 (3H, d, J=0.95 Hz, H-13) and the olefinic proton H-4 indicated that the methyl is directly attached to an olefinic carbon, this was supported by the HMBC correlation between δ_H 1.91 (H-13) and δ_C 125.70 (C-4). The COSY and HMBC experiments showed a cross peak between a doublet at δ 1.25 (3H, d, J=6.4 Hz, H-10) an a oxygenated methine at δ 4.39 (1H, dd, J=14.9, 6.2 Hz, H-9). In addition, gem-dimethyl groups at δ 0.94 (3H, s, H-11) and δ 1.00 (3H, s, H-12) were observed. Geminally coupled signals at δ 2.44 (1H, d, J=16.8 Hz, H-2a) and δ 2.15 (1H, d, J=16.8 Hz, H-2b) suggested that a carbonyl group was linked to this methylene. Finally, the absolute configuration of C-6 proved to be R by the negative value of its optical rotation as determine Yamano and Ito in the total synthesis of four stereoisomers of roseoside [20].

The stereochemistry of C-9 was determined by comparison with the chemical data for roseoside [20]. Therefore, the absolute configuration of the C-9 (δ 73.49) in 1 was assigned as S. On the basis of this data, 1 turned out to be (6R,9S)-6′-(4″-hydroxybenzoil)-roseoside. To the best of our knowledge, this is the first report of the isolation and characterization of (6R,9S)-6′-(4″-hydroxybenzoil)-roseoside from nature.

Agathisflavone (C₃₀H₁₈O₁₀) was isolated as a yellow amorphous solid and exhibited a (M-1)⁺ molecular ion peak at m/z 537.2813. In the ¹³C-NMR displayed 30 carbons including two carbonyl carbons at δ 182.3 and 182.5, 16 quaternary sp² carbons with eight linked to an oxygen atom and 12 tertiary sp² carbons. The ¹H-NMR spectrum showed two singlets at δ 13.32 and 13.06 attributed to hydroxyl groups, two sets of A₂B₂-type doublets, one set at δ 8.01 and 6.97 and the other at δ 7.58 and 6.77 attributed to two AA′BB′ systems in two para-substituted aromatic rings, and four singlets at δ 6.88, 6.79, 6.38 and 6.37 attributed to four hydrogen attached to sp² carbon atoms. The complete assignment of the proton and carbon was made by the comparison with literature data [21, 22]. 4-Hydroxybenzoic acid was isolated as a white solid and was identified on the basis of its ¹H-NMR and ¹³C-NMR data and a comparative TLC with an authentic sample.

As shown in table 1, (6R,9S)-6′-(4′-hydroxybenzoil)-roseoside showed a statistically significant (p<0.02)inhibition of 13.7% of the G-6-Pase in intact microsomes without affecting the enzyme activity of the disrupted system, the effect which is lower in comparison to 27.8% of inhibition of the G-6-Pase in intact microsomes obtained with phlorizin, a known inhibitor of T1 transporter [23]. These results allowed to suggest that 1 inhibits the G-6-P transporter (T1) without affecting the catalytic subunit or phosphate/pyrophosphate transporter (T2) of the G-6-Pase system. The (6R,9S)-roseoside, isolated by Avila et al. [24], exerted an inhibition of 58.5% of the enzyme in intact microsomes, suggesting that 4-hydroxybenzoil moiety exerts a negative effect in the biological activity.

Figure 1.

Purification of agathisflavone, (6R,9S)-6′-(4″-hydroxybenzoil)-roseoside and 4-hydroxybenzoic acid from OpD fraction of O. polyantha Engl.