Endophytes inhabiting the normal plant tissues without causing apparent pathogenic symptoms are attracting considerable attention as sources of biologically active secondary metabolites for pharmaceutical and agricultural industries.1, 2, 3, 4, 5, 6, 7 Phomopsis spp. are prolific endophytic fungi that produce many different types of compounds, such as six new 10-phenyl-[11] cytochalasans, named cytochalasins N–S, together with the four known compounds, epoxycytochalasins H and J, and cytochalasins H and J.8 Three new antibiotic and cytotoxic dicerandrols were isolated from an endophytic fungus P. longicolla of the endangered mint Dicerandra frutescens.9, 10 During our ongoing chemical investigations of endophytic fungi as sources of new bioactive natural products, a subculture of an isolate of P. chimonanthi was grown on solid-substrate fermentation culture. Purification of its crude extract led to the isolation of three compounds, including a new euphane triterpenoid, 3S, 22R, 26-trihydroxy-8, 24E-euphadien-11-one (1), together with two known compounds, 3-hydroxy-3-phenylpropanoic acid (2) and 3-hydroxy-3-(4-hydroxyphenyl) propanoic acid (3) (Figure 1). In this paper, the planar structure, relative and absolute configurations of the new compound are presented in detail. Also, compound 1 was evaluated for cytotoxic activities against three human cancer cell lines.

Figure 1
figure 1

Structures for compounds 1–3. A full color version of this figure is available at The Journal of Antibiotics journal online.

The culture of P. chimonanthi was isolated from branches of the salt-tolerant wild medicinal plant Tamarix chinensis in the yellow river delta, Dongying, in May 2011. The isolate was identified as P. chimonanthi by one of the authors (AJ) based on the sequence (GenBank Accession number JQ234963) analysis of the internal transcribed spacer (ITS) region of the recombinant DNA, and deposited in the Biology Institute, Shandong Academy of Sciences (Registration No. YG-1), Jinan. The fungal strain was cultured on slants of potato dextrose agar at 28 °C for 10 days. The agar plugs were inoculated in 250 ml Erlenmeyer flasks, each containing 50 ml of media (0.4% glucose, 1% malt extract and 0.4% yeast extract), and the final pH of the media was adjusted to 6.5 before sterilization. Flask cultures were incubated at 28 °C on a rotary shaker at 170 r.p.m. for 5 days. Fermentation was carried out in four 500 ml Fernbach flasks each containing 80 g of rice. Spore inoculum was prepared by suspension in sterile, distilled H2O to give a final spore/cell suspension of 1 × 106 ml−1. Distilled H2O (100 ml) was added to each flask and the contents were soaked overnight before autoclaving at 15 lb in−1 for 30 min.2 After cooling to room temperature, each flask was inoculated with 5.0 ml of the spore inoculum and incubated at 28 °C for 40 days.

The fermented rice substrate was extracted repeatedly with ethyl acetate (EtOAc) (4 × 500 ml), and the organic solvent was evaporated to dryness under vacuum to afford the crude extract (5.2 g). The extract was fractionated by Silica gel vacuum liquid chromatography using petroleum ether–EtOAc gradient elution. The fraction (125 mg) eluted with 80% EtOAc was separated by Sephadex LH-20 column chromatography and repeatedly eluted with 1 : 1 CHCl3–CH3OH to give compound 1 (5 mg). The fraction (236 mg) eluted with 8% EtOAc was separated by Sephadex LH-20 column chromatography eluted with 1 : 1 CHCl3–CH3OH to obtain compound 2 (10 mg). The fraction (180 mg) eluted with 20% EtOAc was separated by Sephadex LH-20 column chromatography eluted with 1 : 1 CHCl3–CH3OH to produce compound 3 (7 mg).

1H and 13C NMR data were acquired with Bruker-600 spectrometers using solvent signals (DMSO; δH 2.50/δC 39.5) as references. The HMQC and HMBC experiments were optimized for 145.0 and 8.0 Hz, respectively. High resolution electrospray ionization mass spectrometry (HRESIMS) data were acquired using a LTQ Orbitrap XL mass spectrometer (Thermo Fisher, Bremen, Germany). Sephadex LH-20 (Pharmacia Biotech AB, Uppsala, Sweden), silica gel (300–400 mesh) and silica gel GF254 sheets (0.20–0.25 mm) (both from Qingdao Haiyang Chemical Group, Shandong Province, China) were used for column chromatography and TLC, respectively.

Preparation of (S)-MTPA (α-Methoxy-α-Trifluoromethylphenylacetic acid) Ester (1a) and (R)-MTPA Ester (1b)

The complete powder of 1 (1.5 mg) was transferred to two small clean bottles, respectively, and then pyridine-d5 (0.5 ml) and (R)-MTPACl/(S)-MTPACl (8.0 μl) were quickly added into the clean bottles, and all contents were mixed thoroughly by shaking the clean bottles carefully. The reaction was performed at room temperature, and the solution was allowed to stand for 24 h. The S-MTPA ester and R-MTPA ester (1a/1b) were obtained by HPLC (Lumiere Tech Ltd., Berlin, Germany); YMC-Pack ODS-A column; 10 μm; 250 × 10 mm; 2 ml min−1, 90% MeOH in H2O for 2 min, and followed by 90–100% for 10 min) to obtain 1a and 1b (tR 23.8 and 23.5 min), respectively.

3S, 22R, 26-trihydroxy-8, 24E-euphadien-11-one (1), was obtained as a colorless powder. It was assigned a molecular formula of C30H48O4 (seven degrees of unsaturation) on the basis of HRESIMS analysis (m/z 495.3462 (M+Na)+; Δ –1.2 mmu) and NMR data (Table 1). Analysis of the 1H, 13C and HMQC–NMR spectroscopic data (Supplementary Figures S1, S2 and S4) of 1 revealed the presence of seven methyl groups, nine methylene units (one of which is oxygenated), five methines (two of which are oxygenated), four quaternary carbons, four olefinic carbons (one of which is protonated) and one carbonyl carbon. These data accounted for all 1H and 13C resonances, and required 1 to be tetracyclic, which is consistent with the molecular formula C30H48O4. The planar structure was completely determined by 2D NMR including 1H–1H COSY and HMBC correlation spectra (Figure 2, Supplementary Figures S3 and S5).

Table 1 NMR data of 1 (in DMSO-d6)
Figure 2
figure 2

1H–1H COSY and Key HMBC correlations for compound 1.

Analysis of the 1H–1H COSY led to obtain four isolated fragments corresponding to C-1–C-3 (3-OH), C-6–C-7, C-15–C-17–C-20 (21-CH3)–C-22 (22-OH)–C-24 and C-26–26-OH (Figure 2). The connection of the four fragments was solved by HMBC correlations. The correlations from Me-19 to C-1, C-5 and C-10, from Me-28 and Me-29 to C-3, C-4 and C-5 established one cyclohexane ring (A ring). The HMBC correlation of Me-19 with C-5, C-10 and C-11, and of H2-6 with C-5 and C-10, H2-7 with C-8 and C-9 led to form one cyclohexene ring (B ring) fused with A ring at C-5 and C-10. The distinct cross peaks in HMBC spectra from H2-12 to C-11, C-9, from Me-18 to C-12, C-13 and C-14, and from Me-30 to C-8, C-14 and C-15 completed the 2, 3, 3a, 4-tetrahydro-1H-inden-5(7aH)-one ring (C and D rings), which was fused with the B ring at C-8 and C-9. The HMBC correlations from Me-27 to C-24, C-25 and C-26 confirmed that C-25 was connected with C-24, C-26 and C-27. Thus the planar structure for compound 1 was determined as shown in Figure 1.

The relative configuration was established by detailed analysis of NOESY correlation (Figure 3; Supplementary Figure S6). The NOESY correlation from the olefinic-proton H-24 to Me-27 confirmed the Z-configuration of the double bond. The correlation of H-3 with H-5 and Me-28, of Me-29 with Me-19 and 3-OH implied that H-3, H-5 and Me-28 had the same orientation, whereas Me-19 and Me-29 were on the other side of the cyclohexane ring. Thus the relative configuration for the left part was solved. The cross peaks between Me-18 and Me-21, and between H-17 and H-12a, and Me-30 and H-22 in the NOESY spectra suggested that these protons were in closeness in space. Thus, the relative configuration for the right-part of 1 was determined. The correlations from H-5 to H-7b and from Me-30 to H-7a led to the connection between the left and right parts, revealing that 1 possessed the same relative configuration as its analogs except for the additional 22-hydroxyl group (Figure 3).11 From the biosynthetic pathway, compound 1 might possess the same absolute configuration as its analogs. To support this hypothesis, modified Mosher’s reaction (Figure 4) was used to determine the stereochemistry for C-3 and C-22.12 Treatment of 1 with (R)-MTPACl and (S)-MTPACl affords the S-MTPA ester (1a) and R-MTPA ester (1b), respectively. The 1H NMR data (Supplementary Figures S7 and S9) revealed three methoxyl groups present in the target products, and the chemical shift values of H-3, H-22 and H2-26 were significantly down-field, which confirmed that 3-OH, 22-OH and 26-OH all reacted with MTPACl. Analyzing the 1H–1H COSY data (Supplementary Figures S8 and S10) of the S-MTPA ester (1a) and R-MTPA ester (1b) gave the chemical shift values for H2-1, H2-2, H-3, H-5, H2-6, H2-7, H-17, H-20, Me-21, H-21, H2-22, H-23 and H-24, whereas 1H–1H COSY spectra could not afford the chemical shift values of the Me-18, Me-19, Me-28, Me-29 and Me-30, directly. After careful analysis and comparison of the 1H NMR from 1, 1a and 1b, the 1H NMR signals of these methyl groups were obtained according to the following procedures. The basic mechanism of Mosher’s reaction to determine the absolute configuration is due to the diamagnetic effect of the benzene ring, which leads to the down-/up-field change of nearby protons. Compared with other protons such as H2-1, H2-2, Me-28, Me-29, Me-21, H-21, H2-22, H-23 and H-24 (Supplementary Figure S11), the protons of Me-18, Me-19 and Me-30 were in long distance from MTPA esters at C-3 and C-22. Thus, the chemical shift values for these methyls (Me-18, Me-19 and Me-30) changed little compared with other protons (above-mentioned) in the 1H NMR spectra, whereas the distance between Me-28, Me-29 and the benzene ring in MTPA esters at C-3 was short, and the chemical shift values for these two methyls changed significantly. In addition, though the chemical shift values for these methyls changed in the 1H NMR spectra of MTPA esters, the shapes (sharp or broad) of these did not change in 1 and 1a/1b (Supplementary Figure S11). Thus according to these procedures, the chemical shift values of Me-18, Me-19, Me-28, Me-29 and Me-30 were finally obtained (Supplementary Figure S11). The difference in chemical shift values (Δδ δS−δR) for the diastereomeric esters 1a and 1b was calculated in order to assign the absolute configuration at C-3 and C-22. Analysis of the relevant signals suggested S and R absolute configuration at C-3 and C-22 in 1, respectively (Figure 4). Considering that the relative configuration was determined by NOESY correlation, the stereochemistry of compound 1 was determined. In addition, the cytotoxic activities of 1 was assessed against A549, MDA-MB-231, PANC-1, human cancer cell lines by the MTT method with 5-fluorouracil (5-Fu) as the positive control. 3S, 22R, 26-trihydroxy-8, 24E-euphadien-11-one (1) displayed weak cytotoxic activities against the three cancer cell lines with IC50 values of 20.32, 19.87 and 30.45 μM, whereas IC50 values for 5-Fu were 0.47, 0.12 and 0.67 μM, respectively.

Figure 3
figure 3

Key NOESY correlations for compound 1.

Figure 4
figure 4

Δδ values (in p.p.m.)=δS−δR obtained for (S)- and (R)-MTPA esters 1a and 1b. A full color version of this figure is available at The Journal of Antibiotics journal online.