Cytotoxic tremulanes and 5,6-secotremulanes, four new sesquiterpenoids from a plant-associated fungus X1-2

Abstract Two new tremulanes and two new 5,6-secotremulanes, davotremulanes A-D 1–4, along with four known compounds 5–8, were isolated from the culture extract of X1-2, an unidentified plant-associated fungus, which was isolated from the endangered plant, Davidia involucrate Baill. in Shennongjia District. The structures of new compounds 1–4 were established on the basis of extensive spectroscopic analysis. Compounds 1–8 were evaluated for cytotoxic activity to four cancer cell lines, and compounds 1, 2 and 5 displayed selectively moderate activities to A549 cell line with IC50 at 15.3, 25.2, 35.2 μg/mL.


Introduction
Tremulanes, a family of sesquiterpenes, were firstly isolated from the Aspen rotting fungus Phellinus tremulae (Ayer & Cruz 1993) and were rare discovered in nature, which are isomeric with the lactarane skeleton sesquiterpenes. Up to now, only about 30 tremulane derivatives are found in species P. tremulae (Ayer & Cruz 1993), Phellinus igniarius (Liu et al. 2007;Yin et al. 2014) and basidiomycete concybe siliginea Wu et al. 2010), Huperzia serrata (Ying et al. 2013). Tochtermann W. (Tochtermann et al. 1999) and Ashfeld BL. (Ashfeld & Martin 2005, 2006 synthesised some tremulane derivatives because of their unique skeleton structures. As part of the program to study the chemical diversity investigation of the plant-associated fungi from the endangered plants in Shennongjia District, we found that the extract of a plant-associated fungus X1-2 from the leaves of Davidia involucrate Baill showed selectively cytotoxtic activity to A549 cells in vitro and complexly chemical structure diversity of the secondary metabolites through HPLC-DAD analysis. At the same time, the chemical constituents of D. involucrate Baill were rarely involved, and only two papers referred to the caffeoyl and phenolic derivatives by Wu et al. 2008). This paper described the isolation, structural elucidation and preliminary bioassays (cytotoxic and antimicrobial activities) of eight secondary metabolites 1-8 in vitro.
In compound 1, there were five chiral carbons at C-3 (δ C 37.4), C-5 (δ C 72.4), C-6 (δ C 38.3), C-7 (δ C 39.7), C-9 (δ C 43.8), and the relative configuration of those carbons were set up by NoeSY spectrum. In the NoeSY spectrum, the cross peaks of H-5/H-15, H-13/H-5 and H-7, H-3/H-5 implied that the five protons of H-3, H-5, H-13, H-7 and H-15 were at the same side of both the five-membered and seven-membered rings (see Figure S9). The CD spectrum of compound 1 displayed the strong positive cotton effect at 219 nm and a weak negative cotton effect at 200 nm. According to the π-π* CD octant rule for olefins, the carboxyl group (C-12) and oxymethene group (C-11) lied in the positive contribution region, corresponding to the strong positive cotton effect at 219 nm in the CD spectrum. The methyl group (C-13) lied in the negative contribution region that was explained the weak negative cotton effect at 200 nm in the CD spectrum. Thus, the absolute configuration was established as 3S, 5R, 6R, 7R, 9R. Finally, the structure of compound 1 was fully established, a new compound, named as davotremulane A.
Davotremulane B (2) was obtained as colourless oil, with the molecular formula of C 15 H 22 o 4, established by the Hr-eSI-MS analysis. The 1D 13 C NMr spectrum of compound 2 displayed marked similarities to davotremulane A (1), except the olefinic carbons and carbonyl group at δ C 164.7, 123.3 and 175.7 in davotremulane B (2) instead of those carbons at δ C 139.0, 125.2 and 180.1 in davotremulane A (1). The above 1D 13 C NMr information of compound 2 hinted that the location of double bond and the free carboxyl group may be different from those carbons in davotremulane A (1). Further analysis of the 1 H-1 H CoSY, HSQC and HMBC NMr data verified the structure of compound 2 as shown in Figure 1. The Noe correlations from H-13 to H-5 and H-7, from H-1 to H-7 and H-15 proved the relative configuration of C-1 (δ C 37.7), C-5 (δ C 71.6), C-6 (δ C 39.1), C-7 (δ C 40.8) and C-9 (δ C 41.8), as shown in Figure S9. Based on the above spectral evidence, the planar structure of compound 2 was determined. The CD spectrum of compound 2 shared the similarity to the CD of compound 1. According to the π-π* CD octant rule for olefins, the absolute configuration of compound 2 was established as 1R, 5R, 6R, 7R, 9R. Thus, compound 2 was named as davotremulane B.
Compounds 3 and 4 were obtained as an inseparable mixture in the ratio of approximately 3:1 from the intensity of 1 H NMr resonances. The 1 H and 13 C NMr spectra of these two compounds showed two sets of resonances, respectively, and could be assigned clearly to each compound because their unequal concentrations resulted in an obvious difference of resonance intensities in the 1 H and 13 C NMr spectra. Comparing the 1D 13 C NMr data of davotremulane C (3) and D (4) to Conocenolide A and B (Liu et al. 2007), the two pairs of compounds shared remarkable similarities in 13 C NMr data, except the methyl group at C-14 (δ C 28.6, δ H 1.10) in conocenolide A, C-14 (δ C 28.5, δ H 1.10) in Conocenolide disappeared in those of 3 and 4. Instead, two additional oxymethylene signals δ C 70.6 for C-14 (δ H 3.49) in davotremulane C (3) and δ C 70.5 for C-14 (δ H 3.50) and in davotremulane D (4) were observed in the 1D NMr spectra of 3 and 4 (see Table S2). At the same time, the five carbon resonances at δ C 42.7 (C-8), 43.2 (C-9), 41.0 (C-10), 70.6 (C-14) and 22.7 (C-15) in compound 3 and another five carbon resonances at δ C 43.0 (C-8), 43.2 (C-9), 41.1 (C-10), 70.5 (C-14) and 22.8 (C-15) in compound 4 displayed significantly different from those carbons of conocenolide A and B. Detailed analysis the HMBC spectrum, the cross peaks of H-8/C-10, C-14, C-15; H-10/C-8, C-14, C-15; H-14 and H-15/C-8, C-9, C-10; H-14/C-15 and H-15/C-14 disclosed the existence of a five-membered ring in each compound (3 and 4) (see Figure S8). The above NMr data evidence proved the structures of compounds 3 and 4, namely, davotremulane C (3) and D (4) (see Figure 1). In the NoeSY spectrums of compounds 3 and 4, only the cross peaks of H-7/H-14 were observed, suggesting that the H-7 and H-14 groups lied at the same side of the five-membered ring in compounds 3 and 4 (see Figure S9).
The chromatograph behaviour of davotremulane C (3) and D (4) showed complete separation status on the common C-18 column in HPLC analysis. After the 0.5 mL of CDCl 3 was added to the two dry samples for NMr testing, we found that the 1 H NMr spectrum of davotremulane C (3) was identical to that of davotremulane D (4). Firstly, the acidity influence of the CDCl 3 solvent was taken into account, and the solvent was changed to CD 3 CoCD 3 -d 6 , but we obtained the same results, the 1 H NMr spectrum of davotremulane C (3) was still identical to the 1 H NMr spectrum of davotremulane D (4). Thus, we supposed that the davotremulane C (3) and D (4) can be enantiotropic each other in room temperature, which meant that the enantiotropic energy barrier between davotremulane C (3) and D (4) was much lower in room temperature.
Compounds 1-8 were evaluated to inhibiting the A549, Cask, HepG2 and MCF-7 cells with MTT method, and the results indicated that compounds 1, 2 and 5 exhibited selectively moderate inhibiting activity at IC 50 15.3, 25.2, 35.3 μg/mL to A549 cell line, respectively, and showed no activities to Caski, Hep G2 and MCF-7 cell lines. The antimicrobial activities against to two plant-pathogenic microbes, Erwinia carotovora sub sp. Carotovora (Jones) Bersey et aland Sclerotium rolfsii Sacc. were evaluated, but the compounds 1-8 displayed no significant inhibiting activity to two plant-pathogenic microbes.

General experimental procedures
UV spectra were obtained on a SCINCo Spectrometer, Ir spectrums were recorded on Nicoler Auatar Spectrometer series FT360 spectrophotometer. 1D and 2D NMr spectra were recorded on a Bruker Ultrashield-400 MHz NMr spectrometer (Switzerland). Mass spectra were obtained on a API 4000 mass spectrometer (USA). Silica gel GF254 (10-40 μm) prepared for TLC and silica gel (200-300 mesh) for column chromatography (CC) were obtained from Qingdao Marine Chemical Factory (Qingdao, People's republic of China). Fractions were monitored by TLC, and the spots were visualised under ultraviolet lamp at 254 nm and in iodine cylinder. Semipreparative high-performance liquid chromatography (HPLC) was performed on a Dionex Ultra-3000 (USA) using a Cosmosil C-18 column (10 μm × 20 × 250 mm and 5 μm × 4.6 × 250).

Isolation and identification of the plant-associated fungus
The healthy leaves and twigs of D. involucrate Baill were collected in the Dalaoling National Forest Park in Yichang and washed by running water, then immersed in the 5-10% NaClo solution for 5 min and 75% ethanol solution for 3 min, finally, washed by aseptic water three times and removed the aseptic water by sterile filter papers. The sterilised leaves and twigs were cut into small pieces with 0.5 × 0.5 cm 2 , and those pieces were placed on the PDA plates in the incubator under 28 °C. After the microbes on the PDA plates were enough for purification, the growing fungi were repeatedly purified by Streak method, until to obtain the single colonies. The pure single colonies were deposited on the slants under 4 °C. Totally, 25 plant-associated fungi were obtained from the leaves and twigs of D. involucrate Baill., then those fungi were fermented in 200 mL PDB liquid medium with 500-mL erlenmeyer flasks. The fungus X1-2 was selected to further chemical constituents investigation according to the results of bioactivity assays by MTT method and chemical structural diversity analysis by HPLC-DAD. The X1-2 was cultured in 40 L PDB liquid medium with 200 500-mL erlenmeyer flasks on the electronic oscillator under 28 °C with the speed at 120 rpm/min for 20 days.
The plant samples were identified as D. involucrate Baill. by Prof Yubing Wang, and the sample was disposed in Hubei Key Laboratory of Natural Product research and Development with the voucher specimen No. 20120828A During the identification of the fungus X1-2, we found that the 18S rDNA results shared less similarity to the other known fungi, only 25% similarity, and thus, the genus of X1-2 was unidentified.

Extraction and isolation of compounds
The broth and mycelium were collected by eight layers of gauze, and the 40 L broth was extracted with 20 L ethyl acetate for five times, then the extract was condensed to 20 g crude under vacuum. The 20 g crude was subjected to silica gel CC and eluted from light petroleum ether to acetone, then to methanol. The subfractions were merged to six fractions according to TLC results. The fraction B (1.5 g) was further subjected to C-18 reverse-phase silica gel CC with the gradient elution from 20% methanol-H 2 o to 80% methanol-H 2 o, and the fraction B was fractionated into five subfractions. The subfraction B-3 was further purified by semipreparative reverse-phase HPLC with gradient elution from 25% to 42% acetronitrile-H 2 o to afford compound 1 (t r 14.5 min,4.5 mg), 2 (t r 17.8 min,4.2 mg), 3 and 4 (t r 25.3 min, 27.5 min, 8 mg), 5 (t r 32.6 min,10 mg). The subfraction B-4 and B-5 were purified by repeated silica gel CC, eluted with petroleum ether-acetone (4:1) and petroleum ether-acetone (3:1), to yield the known compound 5 (15 mg), 6 (30 mg), 7 (25 mg) and 8 (50 mg).

Cytotoxic activity to four cancer cell in vitro
A549, Caski, HepG2 and MCF-7 cells were cultured in rPMI 1640 medium (HyClone) supplemented with 10% FBS (HyClone). The cells were maintained in 5% Co 2 at 37 °C. The 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT, Sigma) colorimetric assay was used to evaluate cell proliferation in the presence of different chemicals. The cells were seeded in 96-well culture plates and treated with vehicle or desired concentrations of chemicals for a further 24 h. After treatment, cells were incubated at 37 °C with MTT (10 μL/ well, 5 mg/mL) for 4 h, and the cell growth response to the chemicals was determined by measuring the absorbance at 570 nm on a plate reader. Three replicates were used for each treatment. In the cytotoxic activity in vitro experiment, mitomycin was employed as positive control for cytotoxic activity assay (Nakata et al. 2007).

Antimicrobial assay
Antimicrobial assay against plant-pathogenic E. carotovora sub sp. Carotovora (Jones) Bersey et al. and S. rolfsii Sacc. were carried out using the well diffusion method. Amphotericin B was used as positive control for antifungal assay (Yakushiji et al. 2013).

Conclusion
In summary, the isolation and structural elucidation of eight secondary metabolites from endophytic fungus X1-2 and their cytotoxic and antimicrobial activity were reported. In bioactivity assay, compounds 1, 2, 5 displayed selectively moderate inhibiting activity to A549 cell lines, no inhibiting activity to Caski, Hep G2 and MCF-7 cell lines. Compounds 1-8 showed no significant inhibiting activity to two plant-pathogenic microbes. In those compounds, compounds 1-5 were tremulane derivatives. Tremulane derivatives belong to sesquiterpenes in structure, but they may not be derived from farnesyl pyrophosphate in biosynthesis (Ayer & Cruz 1993). Apparently, the new compounds 1-4 were derived from the known compound 5 through hydroxymethylation, oxidation, cyclisation and lactonising in biosynthesis. The enantiotropic behaviour of 3 and 4 may be caused by the lower energy barrier between davotremulane C (3) and D (4) in room temperature.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was financially supported by the Natural