Cytotoxic oplopane sesquiterpenoids from Arnoglossum atriplicifolium

Abstract Pale Indian plantain (Arnoglossum atriplicifolium (L.) H. Rob.) is a plant with traditional medicinal usage among the Cherokee Native American tribe for treating cancer. Two oplopane sesquiterpenoids were isolated from an extract of A. atriplicifolium from Western North Carolina. The compounds were isolated by bioassay-guided fractionation using an MCF-7 breast tumour cell line assay. The known compound (1S,6R,7R,8R)-1-acetoxy-6,7-diangeloxy-8,10-epoxy-2-oxo-oplopa-3,14Z,11,12-dien-13-al (1) had an EC50 value of 9.0 μM against MCF-7 cells, while the new compound (1S,3R,6R,7R,8R,11S)-1-acetoxy-6,7-diangeloxy-8,10,11,13-bisepoxyoplopan-2-one (2) had an EC50 value of 96 μM. The compounds were characterised by 1D and 2D NMR spectroscopy and by comparison with literature values in the case for 1. Based on NOESY analysis, a correction of the relative configuration for 1 is presented. The presence of these compounds may help to explain the folk remedy usage of this plant as an anticancer agent.


Introduction
The southern Appalachian mountains of the united States are the most biodiverse temperate region in the world and are home to many plant species with folk remedy usage for various ailments. We performed a targeted screening of plant extracts in a tumour cell assay in an effort to identify plant extracts as potential adjuvants for chemotherapy. Plants were selected for screening based on reports of folk remedy usage for cancer among early settlers or indigenous peoples in Western North Carolina. In this screening, an extract of Arnoglossum atriplicifolium (L.) H. Rob. (pale Indian plantain) was found to be moderately cytotoxic. The name A. atriplicifolium has superseded the previous name for this species, Cacalia atriplicifolia L. In James Mooney's The Sacred Formulas of the Cherokees, the plant is reported to be well regarded as a cancer remedy, among other uses (Mooney 1891). Compounds previously reported from this plant include oplopane sesquiterpenoids (Bohlmann et al. 1984) and long chain esters of dammarenediol (Spencer 1981). A. atriplicifolium has also been examined as a source of natural rubber (Cornish & Williams 2006). In the present report, bioassay-guided fractionation of the crude extract afforded one known (1) and one new (2) oplopane sesquiterpenoid. We describe the isolation, structure determination and cytotoxicity testing results for 1 and 2.

Results and discussion
High-resolution MS analysis of 1 gave a pseudomolecular ion of [M + H] + = 501.2122, corresponding to a molecular formula of C 27 H 32 O 9 . The 1 H NMR spectrum of 1 in CDCl 3 at 25 °C displayed several key signals. Notably, several signals in the 1 H NMR spectrum of 1 were broadened, which initially led to difficulties in assignment and determination of multiplicity for several protons. Oplopane sesquiterpenoids were reported from A. Atriplicifolium (Bohlmann et al. 1984), and the NMR experiments in that study were performed in C 6 D 6 at 70 °C. In our study, when NMR data for 1 were collected under those conditions, the structural assignment was simplified. Comparison of our data for 1 with that reported by Bohlmann et al. suggested that 1 was a known compound, although a complete NMR data assignment and clarification of configuration was not previously presented. 13 C NMR data were collected for 1, and a full 2D NMR assignment was performed to confirm the structural assignment.
The J value of 13 Hz for the coupling of H-4 and H-9 agreed with the presence of the trans ring fusion in the oplopane skeleton. NOESY correlations between H-4 and H-6 and between H-5 and H-9 suggested that the six-member ring prefers a chair-like conformation. The small (J = 3.0) coupling constants for H-1 and H-7 suggested cis relationships with their neighbouring protons. Key NOESY correlations were observed between H-1 and H-10a, as well as between H-10b and H-7. These correlations suggested that the epoxide methylene was in an equatorial position. Bohlmann et al. depicted the configuration of the epoxide group in a structural drawing of 1, but they did not comment on the epoxide configuration. In a synthetic study of oplopane sesquiterpenoids with C-8 epoxides, Piers and Gavai (1990) found that proton H-9 could couple to one of the epoxide protons (H-10a in this case) by W-type coupling if the epoxide methylene was in an axial orientation. This would have given rise to a doublet of doublets for H-10a, while both H-10a and b would have been doublets if the epoxide methylene was in an equatorial position. The observation of doublets for both H-10a and b and the NOESY data together implied that C-10 was in an equatorial position. The overlap of signals in the vinylic region of the 1 H NMR spectrum precluded NOESY analysis of the configuration of the C-3 double bond. In agreement with Bohlmann et al., the chemical shift of H-15 (δ H 1.94 in C 6 D 6 , δ H 2.08 in CDCl 3 ) suggested a Z configuration for the C-3 double bond, based on the comparison of our data with those of other oplopane sesquiterpenoids (Joseph-Nathan, Villagómez, Rojas-Gardida et al. 1989;Kikuchi & Suzuki 1992;Li et al. 2012). Due to the paucity of material isolated, the absolute configuration was not determined for 1, and the configuration depicted is suggested by biosynthetic analogy to previously reported oplopanes isolated from other members of the tribe Senecioneae. Thus, the structure of 1 is identified as (1S,6R,7R,8R)-1-acetoxy-6,7-diangeloxy-8,10-epoxy-2-oxo-oplopa-3,14Z,11,12-dien-13-al.
High-resolution MS analysis of 2 gave a pseudomolecular ion at [M + H] + = 505.2402, indicating a molecular formula of C 27 H 36 O 9 for 2. The 1 NMR spectrum for 2 had features similar to that of 1, with several key differences. Signals for two angeloxy groups were observed for 2, while signals for two sets of epoxidised methylene protons were also detected. A triplet methyl signal at δ H 1.08 implied the presence of an ethyl group in the structure of 2, while an aldehyde proton signal was not observed for 2. A methyl singlet was observed at δ H 1.08, which was not found for 1.
The 13 C NMR spectrum of 2 in CDCl 3 contained 27 signals, two of which were partially overlapped. Signals for a ketone carbon (δ C 211.9), three ester carbons (δ C 165.9, 166.2, and 170.0), four alkene carbons (δ C 127.0, 127.1, 140.0, and 140.3), three oxygenated carbons (δ C 70.7, 72.6, and 74.7) and four epoxide carbons (δ C 57.2, 39.4, 50.8, 56.6, and 50.4) were observed. Because of the similarities observed between 1 and 2, a structure was proposed for 2 that had an ethyl group attached at C-3 of the oplopane skeleton, an epoxide at C-11 and C-13, and no other double bonds on the oplopane skeleton. Correlations found by HMBC analysis of 2 were consistent with the proposed skeleton structure (Figure 1). Due to the small amount of material isolated, the absolute configuration of 2 was not determined, although the relative configuration has been elucidated. The coupling constants for the ring protons and NOESY analysis indicated that the ring system for 2 possessed the same relative configuration as that of 1. A NOESY correlation was observed between H-3 and H-5, showing that C-3 had R configuration. NOESY correlations between H-13 and H5, as well as between H-12 and H-6 supported the assignment of relative configuration of C-11. This assignment was in agreement with that found for oplopanes from Pittocaulon filare (Asteraceae, Senecioneae) by Arciniegas et al. in a recent study (2014). The absolute configuration was inferred by analogy with other oplopane sesquiterpenoids found from the Senecioneae tribe. Thus, 2 is a new compound identified as (1S,3R,6R,7R,8R,11S)-1-acetoxy-6,7-diangeloxy-8,10,11,13-bisepoxyoplopan-2-one.
Compounds 1 and 2 were tested for their cytotoxicity to MCF-7 cells. Compound 1 was modestly active, with an EC 50 value of 9.0 μM, while 2 was only weakly active, with an EC 50 of 96 μM against MCF-7 cells. Given the relatively low potency of the compounds in this study, it is difficult to correlate the activity of these compounds with the purported anticancer activity mentioned by Mooney. Further studies of the efficacy of A. atriplicifolium extract and 1 in a skin cancer model might shed light on the connection between these compounds and the traditional use. Since the aldehyde functional group in 1 appears to be important for its activity, it is likely that 1 is generally cytotoxic, and therefore is not likely to be a candidate for further development as an antitumour agent.

General
Optical rotations were recorded on a Jasco P-2000 series polarimeter operating at 589 nm. IR data were collected using a Perkin-Elmer Spectrum One FT-IR with an ATR sampling accessory. NMR spectra were recorded on a JEOL Eclipse + 300 with a 5 mm FG/ TH tunable probe (direct geometry) operating at 300 MHz for 1 H and 75.6 MHz for 13 C. High-resolution electrospray ionisation mass spectrometry (HR-ESI-MS) was performed using a LCT premier XE TOF mass spectrometer (Waters, Manifold, MA) equipped with an ESI interface and controlled by MassLynx V4.1 software. Silica gel and C 18 reverse phase silica chromatography packing material were purchased from SiliCycle. Microsorb silica gel and C 18 reverse phase HPLC columns (10 × 250 mm, 5 μm particle size, 100 Å pore size) were purchased from Agilent Technologies. Preparative HPLC purifications were performed on a Perkin-Elmer 300 Series HPLC or on a Dionex ultiMate 3000 HPLC. Normal phase flash chromatography was performed on a Teledyne Isco Combiflash Rf system with a RediSep Rf gold column (20 × 120 mm).

Plant material
Whole plant material for A. atriplicifolium was collected in Buncombe County, North Carolina, uSA, in August 2007. The plant was identified by Joshua A. Kelly. A voucher specimen of A. atriplicifolium (number JK2 BCI00027V) is deposited at the herbarium of the Bent Creek Germplasm Repository at the North Carolina Arboretum, Asheville, NC, uSA.

Cytotoxicity testing
The compounds were tested for cytotoxic activity against MCF-7 human breast tumour cells and FS-4 human foreskin fibroblast cells as described elsewhere (Flood 2010).