Auto-oxidation of Ent-beyer-15-en-19-al isolated from the essential oil of the heartwood of Erythroxylum monogynum Roxb.: formation of 15,16-epoxy-ent-beyeran-19-oic acid and other products

Chemical investigation of the essential oil obtained from the heartwood of Erythroxylum monogynum Roxb. yielded three beyerene type diterpenoids ent-beyer-15-ene (1), ent-beyer-15-en-19-ol (erythroxylol A) (2) and ent-beyer-15-en-19-al (3). Ent-beyer-15-en-19-al (3) was found to be unstable at room temperature, giving rise to hitherto unknown 15,16-epoxy-ent-beyeran-19-oic acid (4). This conversion involves the auto-oxidation of a C-4 axial aldehyde group of an ent-beyer-15-ene diterpenoid with the concurrent epoxidation of the C-15 double bond. This is the first report of the auto-oxidation of an aldehyde group to a carboxylic acid group with the concurrent epoxidation of a double bond in the same compound. Further investigation of this observation under controlled conditions resulted in the isolation and identification of ent-beyer-15-en-19-oic acid (5), two new epoxy hydroperoxides, 15,16-epoxy-19-nor-ent-beyeran-4α-hydroperoxide (6a), 15,16-epoxy-18-nor-ent-beyeran-4β-hydroperoxide (6b), and two new hydroperoxides, ent-beyer-19-nor-15-en-4α-hydroperoxide (7), ent-beyer-18-nor-15-en-4β-hydroperoxide (8) and ent-beyer-18-nor-15-en-4β-ol (9). Identification of these compounds was carried out by the extensive usage of spectroscopic data including 1D and 2D NMR. The acid 5 and the alcohol 9 have been reported previously as natural products from Elaeoselinum asclepium and Erythroxylum monogynum. The mechanistic basis of this auto-oxidation reaction is discussed.


Introduction
Erythroxylum monogynum Roxb. (Erythroxylaceae), is a small evergreen tree indigenous to Sri Lanka and India. The oil obtained by the distillation of its heartwood has been used traditionally as a wood preservative [1] and in the local perfumery industry. Many diterpenoids belonging to the kaurane and beyerane groups have been reported from this plant [2]. In continuing our search for industrially useful essential oils in Sri Lanka, we initiated an investigation of the heartwood of E. monogynum. The essential oil obtained by the initial hydrodistillation of the heartwood contained a mixture of monoterpenoids along with three high molecular weight compounds, one of which was identified as ent-beyer-15-ene (1) by Gas chromatography-mass spectrometry (GC-MS) and confirmed by 1 H and 13 C NMR data. With the view to identifying the remaining two high molecular weight compounds, steam distillation of the heartwood of E. monogynum has been carried out initially at 40 psi for 24 h (Stage 1) exhaustively to remove the monoterpene fraction and then at 70 psi for additional 12 h (Stage 2) to obtain mainly the high molecular weight fraction. The essential oil obtained in the Stage 2 showed the presence of three compounds on GC-MS analysis, one of which was identified as ent-beyer-15-ene (1). The remaining two compounds were identified as erythroxylol A (2) and an unsaturated diterpene aldehyde, ent-beyer-15-en- 19-al (3) by the analysis of their 1 H and 13 C NMR data. This is the first report of 3 from this plant as well as the complete assignment of its 1 H and 13 C NMR data supported by 2D NMR spectroscopic data. Compound 3 has been reported previously as an unstable oil from the timber of Erythroxylum zambesicum. Its structure had been established by its conversion to the corresponding alcohol (erythroxylol A, 2), and the observation of an aldehyde group, three tertiary methyl groups and a cis double bond in its 1 H NMR spectrum [3]. Prior to the report by Ansell [3] it had been reported as a semisynthetic product obtained from the oxidation of erythroxylol A (2) and to be a low melting solid (m. p. 63-65 °C), stable at 0 °C under nitrogen. The structure had been established by reducing it to ent-beyer-15-ene via its ethylene thioacetal [4]. Compound 3 has also been reported from Viguiera grammatoglossa [5] and Myriocephalus stuartii [6].
Compound 3 was found to be unstable at room temperature converting into more polar compounds. Of these the major compound was isolated and identified as 15,16-epoxy-ent-beyeran-19-oic acid (4). According to published data, this is the first report of the auto-oxidation of an aldehyde group to a carboxylic acid group with the concurrent epoxidation of a double bond in the same compound. Prompted by this observation we setup an experiment to study this auto-oxidation process. In this experiment when 3 was exposed to air as a solution in cyclohexane at room temperature (28-30 °C) it underwent auto-oxidation to give two major products and five minor products. Of the two major products, the more polar one was found to be 4 while the other major compound was identified as ent-beyer-15-en-19oic acid (5). Further investigation of the reaction mixture enabled us to isolate and identify the five minor compounds produced during this auto-oxidation process as, 15,16-epoxy-19-nor-ent-beyeran-4α-hydroperoxide (6a), 15,16-epoxy-18-nor-ent-beyeran-4β-hydroperoxide (6b), ent-beyer-19-nor-15-en-4α-hydroperoxide (7), ent-beyer-18-nor-15-en-4β-hydroperoxide (8), and ent-beyer-18nor-15-en-4β-ol (9). Herein we report the occurrence of 3 in E. monogynum along with its complete assignment of 1 H and 13 C NMR data and the structure elucidation of compounds 4, 5, 6a, 6b, 7-9 ( Fig. 1) utilizing their spectroscopic data. The plausible pathway of the formation of these compounds during the auto-oxidation of 3 is also discussed.

Results and discussion
Heartwood of E. monogynum was subjected to steam distillation in two stages. The oil obtained in stage 1 was found to contain a mixture of monoterpenoids along with three high molecular weight compounds by GC-MS analysis (see Additional file 2). The oil obtained in stage 2 was shown to consist of mainly the above three major compounds by GC-MS analysis. The oil obtained in stage 2 was subjected to silica gel column chromatography to isolate these three compounds.
The remaining two compounds isolated from the essential oil were identified as ent-beyerene (1) by its GC-MS data [7] and comparison with reported 13 C NMR data [8] (Additional file 2: Table S1) and ent-beyer-15-en-19-ol (erythroxylol A) (2) by the analysis of its 1 H and 13 C NMR data and comparison with reported 13 C NMR data [8] (Additional file 2: Table S1). These two compounds have been reported previously from the timber of E. monogynum [4].
Thin layer chromatographic (TLC) analysis of a solution of compound 3 in cyclohexane showed that the compound was unstable when exposed to air and decomposed to a number of compounds, all of which were more polar than the parent compound. It was observed that the decomposition of 3 was prevented by the addition of butylated hydroxy toluene (BHT) to the solution supporting the view that the decomposition was an auto-oxidation occurring through a free radical mechanism. Compound 3 was found to be stable when stored at 0 °C under nitrogen in the absence of a solvent.
Periodic TLC analysis of a solution of 3 in cyclohexane indicated a changing pattern of spots characteristic of a radical reaction, which stabilized after 14 days to give a pattern of six spots, two of which were present in larger amounts than the others. Compound 4, the more polar one of the two major products was obtained as a viscous liquid, analyzed for C 20 H 30 O 3 by a combination of highresolution electro-spray ionization mass spectrometry (HRESIMS) and 13 C NMR spectroscopy. The 1 H and 13 C NMR spectra (Tables 1 and 2) together with HSQC data of 4 indicated that its structure was very similar to that of 3. In comparing the 13 C NMR spectrum of 4 with that of 3, the oxidation of the aldehyde group to a carboxylic acid group is clearly indicated by the appearance of a signal at δ C 183.9 and the loss of the signal at δ C 205.9. This structural change is also reflected in the 1 H NMR spectrum, where the signal due to aldehyde proton at δ H 9.75 in 3 is absent in 4. Epoxidation of the double bond is indicated by the loss of two olefinic CH groups [δ H 5.68 d, (J = 5.6 Hz), 5.46 d, (J = 5.6, Hz); δ C 134.5, 136.8] and the appearance of two new oxygenated methines [δ H 3.43 d, (J = 3.0 Hz), 3.04 d, (J = 3.0 Hz); δ C 55.9, 60.1). The tertiary methyl group at δ H 1.25 assigned to 18-H 3 [9] showed HMBC correlations with C-3 (δ C 37.7), C-4 (δ C 43.7), C-5 (δ C 56.9) and the carboxyl carbonyl carbon (δ C 183.9), placing the methyl and carboxylic acid groups at C-4. The tertiary methyl group at δ H 1.02 showed HMBC correlations with the oxygenated methine carbon at δ C 60.1 (C-16) and methylene carbon at δ C 46.6 (C-14) placing it at C-13. The remaining tertiary methyl group at δ H 0.84 showed HMBC correlations with the methine carbons at δ C 56.9 (C-5), and 55.7 (C-9) placing it at C-10. The oxygenated methine proton at δ H 3.04 (d, J = 3.0 Hz) assigned to H-16 showed HMBC correlation with the quaternary carbon at δ C 38.9 (C-13) and the remaining oxygenated methine proton at δ Compound 5, the less polar major product was obtained as a colorless viscous liquid. The 1 H and 13 C NMR spectroscopic data together with HSQC and DEPT135 of 5 (Additional file 2: Table S2 and Figures S16-S21) showed very close resemblance of its structure to those of 3 and 4. The presence of a carboxylic acid C = O was indicated by the signal at δ C 183.8 in its 13 C NMR spectrum while the presence of the two olefinic protons as in the case of 3 were clearly evident by the presence of the signals at δ H 5.73 d (J = 5.7 Hz); δ C 134.8 (CH-15) and δ H 5.45 d (J = 5.7 Hz); δ C 136.5 (CH-16). Thus 5 was identified as ent-beyer-15-en-19-oic acid by the comparison of its 13 C NMR data with reported data for ent-beyer-15-en-19-oic acid (Additional file 2: Table S2), which has been isolated from Elaeoselinum asclepium [10]. As the reported 13 C NMR data of this compound was not supported by 2D NMR (HSQC and HMBC) spectroscopic data and its 1 H NMR data was not available, we assigned the 1 H and 13 C NMR data of 5 with the help of HSQC and HMBC correlations (Additional file 2: Table S2).
The most polar minor auto-oxidation product (6) obtained as a colorless viscous liquid was determined to be a mixture of two isomers 6a and 6b epimeric at C-4 in 1:2 ratio from the following evidence. GC-MS analysis showed two peaks in the GC, both of which showed the same [M] + m/z 306.4; 1 H NMR spectrum of 6 showed two doublets at δ H 3.40 (J = 3.0 Hz) and δ H 3.47 (J = 3.0 Hz) in 2:1 ratio and 6 tertiary methyl groups and 13 C NMR spectrum showed 37 carbon signals. 1 H and 13 C NMR spectroscopic data (Tables 1 and 2 of which two overlapped at δ C 21.5 in the 13 C NMR spectrum. These data suggested that these two isomers lack a carbon atom from each of these two isomers and presence of 15 (16) epoxide as in the case of compound 4. 13 C NMR spectrum of 6 did not show either CHO or CO 2 H carbon signals but showed two signals at δ C 84.7 and 84.0 for oxygenated quaternary carbons indicating that the 18 (or 19) C has been lost from the beyerane skeleton and a hydroperoxide (-OOH) group has been attached to 4-C, suggesting the possibility of these two being C-4 epimers of each other. Analysis of HMBC correlations (Fig. 4) permitted the unambiguous assignment of NMR signals of each of the epimers (Tables 1 and 2). Differentiation of epimers 6a and 6b and the relative configuration at C-4 in 6b was achieved by a series of Selective NOESY Gradient experiments (Fig. 5). Irradiation of the signal at δ H 0.92 (20-H 3 ) in a Selective NOESY Gradient experiment caused enhancement of the methyl signal at δ H 1.13 (19-H 3 ) and the oxygenated methine proton at δ H 3.40 (15-H) suggesting that these two methyl groups and the oxygenated methine proton belong to the epimer 6b and both methyl groups are on the same side of the molecule confirming α-orientation of the C-4 methyl group. Further it is evident that the H-15 is also α-orientated confirming the formation of epoxide from the β-face (exo epoxide) as in the case of 4. Irradiation of the methyl signal at δ H 1.13 (19- Methine protons of the epoxide ring at δ H 3.03 (H-16) and δ H 3.40 (H-15) showed HMBC correlations to quaternary carbons δ C 39.0 (C-13) and δ C 44.2 (C-8) respectively while both of them showed correlations to δ C 46.7 (C-14). The remaining key HMBC correlations useful for the establishment of connectivity in the molecule are shown in Fig. 4.
It appears that the remaining three methyl signals in the 1 H NMR spectrum of 4 responsible for tertiary methyl groups belong to the epimer 6a. The methyl protons at δ H 1.30 (18-H 3 ) in compound 6a showed HMBC correlations with the quaternary carbon at δ C 84.0 (C-4), methylene carbon at δ C 34.9 (C-3) and methine carbon at δ C 55.9 (C-5) placing this methyl at C-4. Methyl protons at δ H 1.03 (20-H 3 ) showed HMBC correlations with the methylene carbon at δ C 39.3 (C-1) and methine carbons at δ C 55.9 (C-5) and δ C 56.1 (C-9). The remaining methyl group at δ H 1.01 (17-H 3 ) showed HMBC correlations to quaternary carbon at δ C 39.1 (C-13), methine carbon at δ C 60.20 (C-16), methylene carbons at δ C 35.4 (C-12) and 46.9 (C-14). Methine protons of the epoxide ring at δ H 3.47 (15-H) and δ H 3.02 (16-H) showed HMBC correlations to quaternary carbons δ C 44.1 (C-8) and 39.1 (C-13) respectively while both of them showed correlations to δ C 46.9 (C-14). The remaining key HMBC correlations useful for the establishment of connectivity in the molecule are shown in Fig. 4. Although these two epimers were inseparable under normal phase chromatographic techniques, it was possible to separate them by reverse phase analytical TLC and the two compounds were subjected to HRESIMS to determine their molecular formulae. Compound 6a analyzed for C 19 3 , 305.21180). These data confirmed the structures of 6a and 6b as15,16-epoxy-19-nor-entbeyeran-4α-hydroperoxide and 15,16-epoxy-18-nor-entbeyeran-4β-hydroperoxide respectively. The 1 H and 13 C NMR spectra obtained for these two samples were found to be consistent with the assignments made for 6a and 6b based on the above analysis of the spectra of their mixture 6 (Additional file 2: Figures S32-S37).
Compound 7 obtained as a colorless viscous liquid, analyzed for C 19 H 30 O 2 by a combination of HRESIMS and 13 C NMR spectroscopy. 1 H and 13 C NMR spectroscopic data (Tables 1 and 2 (19-H 3 ) suggesting that these two methyls are in the same side of the molecule indicating the α-orientation of the C-4 methyl group. Irradiation of the methyl signal at δ H 1.12 (19-H 3 ) caused enhancement of signal at δ H 0.74 (20-H 3 ) further confirming the above suggestion (Additional file 2: Figures S50 and S51). Irradiation of the signal at δ H 0.85 (20-H 3 ) or δ H 1.28 (18-H 3 ) of 7 did not cause enhancements of either signals (Additional file 2: Figures S43 and S44) suggesting that these two methyl groups are not in the same face and hence suggested the β-orientation of methyl group at C-4 (18-H 3 ). Thus, the compounds 7 and 8 were identified as ent-beyer-19-nor-15-en-4α-hydroperoxide and ent-beyer-18-nor-15-en-4βhydroperoxide respectively.
Compound 9 was obtained as a white, amorphous solid. Comparison of 1 H and 13 C NMR spectroscopic data (Additional file 2: Table S3) together with HSQC and DEPT135 of compound 9 with those of 7/8 suggested that this is a C-4 alcohol with C-15 olefinic double bond. Although both ent-beyer-18 (19)-nor-15-en-α-and β-ols were known, their spectroscopic assignments were not supported by 2D NMR (HSQC and HMBC) spectroscopic data. Hence, we assigned the 1 H and 13 C NMR data of this compound with the help of HSQC and HMBC data. The HMBC correlations useful for the assignment of 1 H and 13 C NMR signals confirming the structure of compound 9 are shown in (Additional file 2: Table S3). Stereochemistry at C-4 has been established by carrying out Selective NOESY Gradient experiments. Irradiation of the signal at δ H 0.71 (20-H 3 ) of 9 in a Selective NOESY Gradient experiment caused enhancement of the methyl signal at δ H 1.14 (19-H 3 ) indicating that these two methyl groups are in the same side of the molecule confirming α-orientation of the C-4 methyl group. Irradiation of the methyl signal at δ H 1.14 (19-H 3 ) caused enhancement of signal at δ H 0.71 (20-H 3 ) further confirming the above α-orientation of the C-4 methyl group (Additional file 2: Figures S57 and S58). Thus, the compound 9 was identified as ent-beyer-18-nor-15-en-4β-ol, which has been isolated previously as a natural product from E. monogynum [11].
The susceptibility of 4-axial aldehyde groups in the diterpenes towards auto-oxidation giving rise to carboxylic acids and hydroperoxides via radical mechanisms has been previously reported [10,12,13]. Although 3 was known to be an unstable compound, there have been no previous reports on the products obtained from the auto-oxidation of 3. The formation of the epoxy compounds 4, 6a and 6b during the auto-oxidation of 3 can be rationalized by considering the steps involved in the auto-oxidation of aldehydes to carboxylic acids. Autooxidation of an aldehyde to the corresponding carboxylic acid is a facile reaction and takes place via a free radical mechanism [14,15]. Acyl peroxy radicals and per-acids are generated as intermediates during the reaction (Scheme 1). Both these species are capable of epoxidizing an olefinic double bond. Thus, both catalyzed and uncatalyzed processes for the epoxidation of an olefin coupled to the oxidation of an aldehyde to the corresponding carboxylic acid by molecular oxygen have been reported [16][17][18][19][20].
We envisage the epoxidation taking place via an intermolecular reaction between a C-19 peroxy group and the C-15,16 olefin group giving rise to the exo-epoxide because the approach of the two species to form an endoepoxide would be sterically hindered. Steric hindrance of the α-face of the molecule exerted by the axial 20-methyl group would also allow epoxidation to compete with the usual bimolecular termination reaction (Scheme 1) between the peroxy acid and the aldehyde to give two molecules of carboxylic acids. This interpretation is supported by the observation that the C-20,29 double bond of betulonaldehyde does not undergo epoxidation during the auto-oxidation of its unhindered C-28 formyl group [21] which can be approached without hindrance from the β-face of the molecule. It is interesting to note that betulonaldehyde on auto-oxidation gave in addition to betulonic acid, two epimeric C-17 hydroperoxides which would correspond to 7 and 8 in the current study. In a related process, 14-hydroxypimara-8,15-dien-19-oic acid has been isolated from the auto-oxidation of pimara-8(14),15-dien-19-al (which also has an axial aldehyde group) [12]. It has been proposed that the 14-hydroxy compound arises from the ring opening hydrolysis followed by dehydration of the corresponding 8(14)-epoxy carboxylic acid on the basis of chromatographic evidence, although such an epoxy carboxylic acid has not been isolated from the reaction mixture from the autooxidation reaction.
The formation of the epimeric epoxy hydroperoxides 6a and 6b, the epimeric hydroperoxides 7 and 8, and the alcohol 9 can be explained (Scheme 2) as arising from the reactions of the tertiary cycloalkyl radical at C-4 which can be formed by the decarbonylation of the acyl radical and decarboxylation of the acyloxy radical that are formed during the auto-oxidation process [12,15].
The ease of decarbonylation of tertiary acyl radicals is well known. The decarbonylation and decarboxylation reactions are further aided by the loss of steric strain when the sp 3 carbon (C-4) changes to a planar sp 2 carbon removing the 1,3-diaxial interaction of the radical on C-19 with the 20 α-methyl group. The 20 α-methyl group also directs the approach of molecular oxygen and the peracid group to preferentially approach the planar C-4 radical from the β-face by sterically hindering the αface approach. This results in the excess of 6b over 6a, as observed in the 1 H NMR spectrum of 6, (the mixture of 6a and 6b) isolated by column chromatography of the auto-oxidation reaction mixture. Further, of the two possible epimeric alcohols, only the β-alcohol 9 could be detected. However, we note that the preference for the β-face approach of molecular oxygen is not reflected in the relative isolated yields of the hydroperoxides 7 and 8, probably due to experimental losses of yield. Although 9 has been reported as a natural product isolated from the timber of E. monogynum [11], our results support the suggestion by Caputo [13] and Tanaka [12] that it was an artifact arising from the auto-oxidation of 3. The auto-oxidation of 3 is inhibited when it is found as a component of the essential oil of E. monogynum by the presence of other components such as 4, which can act as radical chain breakers. However, we note that no trace of 3 could be detected by GC analysis in a 1 year old sample of the essential oil which had been stored under ambient conditions.

Conclusions
The auto-oxidation of the aldehyde group of ent-beyer-15-en-19-al to a carboxylic acid group is a facile process and takes place both with and without the concurrent epoxidation of the 15,16-double bond. Our results suggest that 4-hydroxy-19-nor-ent-beyer-15-ene which has been reported previously as a natural product from E. monogynum may be an artefact arising from the autooxidation reaction. Steric hindrance exerted by the axial 20-methyl group plays a determining role in the product distribution of the auto-oxidation reaction. The usage of the essential oil from the heartwood of E. monogynum in perfumery will be limited by the instability in the presence of oxygen of ent-beyer-15-en-19-al, which is a major component of the oil.

General experimental procedures
Optical rotations were measured in CHCl 3 with a BioBase Automatic polarimeter BK-P2. IR spectra were recorded as Attenuated total reflections spectra on a Perkin Elmer Spectrum 2. 1D and 2D NMR spectra were recorded in CDCl 3 with a Bruker Ascend 400 spectrometer at 400 Mz for 1 H NMR and 100 MHz for 13 C NMR using residual CHCl 3 as the internal reference. High resolution MS were measured on a Q-Exactive (ThermoScientific) equipment, with H-ESI source. Thin layer chromatography was carried out on Merck analytical normal phase (G60, F 254 , 0.2 mm) and reverse phase (Silica gel 60 RP-18, 0.25 mm) plates. Preparative layer chromatography was carried out on Analtech normal phase plates (G60, 0.5 mm). All solvents used for chromatography were of AR grade from Sigma-Aldrich. Compounds were visualized by spraying with anisaldehyde-sulfuric acid reagent made by mixing anisaldehyde (0.5 ml) with glacial acetic acid (10 ml), followed by 85 ml of methanol and concentrated sulfuric acid (5 ml). Dry column chromatography was carried out using Analtech silica gel (35-75 micron, 150 A). Column fractions were analyzed by TLC and similar fractions were combined and evaporated under reduced pressure. GC-MS analysis was carried out using an Agilent 7890 A GC system equipped with 5975C inert XL MSD Triple-Axis Detector, and HP-5 MS fused silica capillary with a (5% Phenyl)-methylpolysiloxane stationary phase (30 m × 0.25 mm Id, x 0.25 µm film thickness) capillary column. Helium (99.999%) was used as carrier gas. Mass spectra were acquired in the EI mode at 70 eV within the range of 40.5 to 500 mass units.

Plant material
The heartwood of E. monogynum was collected from a tree found at a home garden in Weerawila, Sri Lanka (6˚14ʹ 56.9ʹʹ N, 81˚13ʹ 46.6ʹʹ E). The species was identified by Prof. D. S. A. Wijesundara, National Institute of Fundamental Studies, Hantane, Kandy, Sri Lanka (former Director of the Royal Botanic Gardens, Peradeniya and the National Herbarium of Sri Lanka). A voucher specimen (Voucher no.43-002-010) was deposited in the herbarium at the R & D division at LINK Natural Products, Sri Lanka.

Extraction of essential oil
The dried heartwood (46 kg) of E. monogynum was cut into 8 mm pieces and subjected to steam distillation in two stages. A steam pressure of 40 psi was used, and the flow rate of the condensate was 500-600 mL min −1 during the first stage (24 h) to obtain 76.8 g of a pale yellow, light oil. Distillation was continued for an additional 12 h (stage 2) at a steam pressure of 70 psi and a condensate flow rate of 800-900 mL min −1 to obtain 40.8 g of a thick oil. Both fractions were dried over anhydrous sodium sulfate and stored at − 4 °C in a refrigerator. GC-MS analysis of the oil obtained in stage 2, indicated the presence of three major compounds, of which two were identified as ent-beyer-15-ene (1) and erythroxylol A (2). GC-MS analysis of the oil obtained in stage 1 showed that it also contained these three compounds, in lower concentrations along with lower boiling monoterpenoids.

Ent-beyer-15-ene Diterpenoids from the essential oil of E. monogynum
The oil obtained in stage 2 of the above steam distillation (1.08 g) was subjected to dry column chromatography using gradient elution with hexane-dichloromethane and 70 fractions (10 mL) were collected. These were analyzed by TLC and fractions containing same compound were combined. Fraction 2 on evaporation under reduced pressure gave ent-beyer- 15- 13 C NMR data [7,8] (Additional  file 2: Table S3).