We screened a library constructed by the advanced compound-identification system based on the accumulated HPLC-MS profiling data combined with strain information designated as ‘MBJ’s special selection’ for bioactive substances, and have isolated five novel eremophilane sesquiterpenoids MBJ-0009 and -0010 from the saprobic fungus Nectria sp. and MBJ-0011, -0012 and -0013 from the endophytic fungus Apiognomonia sp.1, 2 Further screening for cytotoxic compounds led to the identification of MBJ-0038 (1), -0039 (2) and -0040 (3), together with the known compounds chaetoglobosins C and F3, 4 from Chaetomium sp. f24230. Here, we describe the fermentation, isolation, structure elucidation, and, in brief, the cytotoxic activity of 13.

Chaetomium sp. f24230 was isolated from a soil sample collected in the Okinawa Prefecture, Japan. The strain was cultivated in 250-ml Erlenmeyer flasks, each containing 25 ml of a seed medium consisting of 2% potato starch (Tobu Tokachi Nosan Kako Agricultural Cooperative Association, Hokkaido, Japan), 1% glucose (Junsei Chemical, Tokyo, Japan), 2% soybean powder (Honen SoyPro, J-Oil Mills, Tokyo, Japan), 0.1% KH2PO4 and 0.05% MgSO4·7H2O. The flasks were shaken on a rotary shaker (220 r.p.m.) at 25 °C for 3 days. Aliquots (0.5 ml) of the broth were transferred to 500-ml Erlenmeyer flasks containing 50 ml of a production medium consisting of 2% potato starch (Tobu Tokachi Nosan Kako Agricultural Cooperative Association), 1% glucose (Junsei Chemical), 2% soybean powder (Honen SoyPro, J-Oil Mills), 0.1% KH2PO4 and 0.05% MgSO4·7H2O, and were cultured on a rotary shaker (220 r.p.m.) at 25 °C for 4 days.

The whole-culture broth (2 l) was extracted with an equal volume of n-BuOH and, following concentration, was successively partitioned between ethyl acetate (350 ml × 3) and water (350 ml). The ethyl acetate extract (4.56 g) was subjected to silica gel medium-pressure liquid chromatography (MPLC; Purif-Pack SI-30, Shoko Scientific Co., Yokohama, Japan) with a gradient system of n-hexane–ethyl acetate (0–25% ethyl acetate) followed by CHCl3–methanol (MeOH) with stepwise increases from 0 to 100% MeOH with monitoring by UV at 254 nm. The active 5% MeOH fraction (190.1 mg) was separated by gel filtration on a Sephadex LH-20 column (1:1, CHCl3/MeOH, GE Healthcare BioSciences AB, Uppsala, Sweden) followed by reversed phase-MPLC (Purif-Pack ODS-30; 70–100% aq. MeOH with 10% stepwise increments in the MeOH concentration) purification to yield 3 (90% MeOH, 6.1 mg). The active 10% MeOH fraction (563.9 mg) was subjected to Sephadex LH-20 column chromatography (1:1, CHCl3/MeOH) to yield a crude material containing 2 (98.9 mg). This material was purified using ODS MPLC (Purif-Pack ODS-30; 50–100% aq. MeOH, 10% stepwise) to yield 2 (80% MeOH, 12.3 mg). Another active fraction from the 10% MeOH eluate (1.70 g) was further purified using an ODS MPLC (Purif-Pack ODS-100; 20–100% aq. MeOH in 20% stepwise increments), and an 80% MeOH eluate (940 mg) was re-chromatographed using an ODS MPLC (Purif-Pack ODS-30; 50–100% aq. MeOH, 10% stepwise) to yield crude 1 (80% MeOH, 119.9 mg). Compound 1 (55.8 mg) was finally purified using Sephadex LH-20 column chromatography (1:1, CHCl3/MeOH).

MBJ-0038 (1) was isolated as a colorless amorphous powder: [α]25D –66 (MeOH; c 0.7); UV λmax (ɛ) in MeOH: 281 (7700) nm; IR (ATR) νmax: 3400 (hydroxy) and 1689 (carbonyl) cm−1. The molecular formula of 1 was established as C41H44N2O9 by HR-ESI-MS; (m/z 709.3122 [M+H]+, calcd. for C41H45N2O9 m/z 709.3125). The 13C and 1H NMR data for 1 are listed in Table 1. Structural information on 1 was obtained by a series of 2D NMR analyses such as double quantum-filtered COSY (DQF-COSY), HSQC and constant-time (CT)-HMBC.5

Table 1 13C and 1H NMR spectroscopic data for MBJ-0038 (1), MBJ-0039 (2) and MBJ-0040 (3).

An indole moiety in 1 was identified based on 1H couplings from a doublet aromatic proton H-4′ (δH 7.53) through a doublet of doublets aromatic protons H-5′ (δH 7.03) and H-6′ (δH 7.10) to a doublet aromatic proton H-7′ (δH 7.34), and by following 1H–13C long-range couplings from a singlet aromatic proton H-2′ (δH 7.07) to aromatic quaternary carbon C-3′ (δC 111.2), C-3a′ (δC 128.7) and C-7a′ (δC 138.2); from H-4′ to C-3′ and C-7a′; from H-5′ to C-3a′; from H-6′ to C-7a′; and from H-7′ to C-3a′, together with their typical 13C NMR chemical shifts (Figure 1c).

Figure 1
figure 1

(a) Structures of 1 and 2. (b) Structure of 3. (c) NMR analysis of 1. COSY: bold line; HMBC (1H–13C): arrows (solid, strong; dashed, weak). (d) Partial structure of 2. COSY: bold line; HMBC (1H–13C): arrows (solid, strong; dashed, weak). (e) Partial structure of 3. COSY: bold line; HMBC (1H–13C): arrows.

The DQF-COSY spectrum revealed the following proton spin networks, from a methyl proton H3-11 (δH 0.62) through methine protons H-5 (δH 1.75), H-4 (δH 2.64) and H-3 (δH 3.85) to methylene protons H2-10 (δH 3.06 and 2.85); from a methine proton H-7 (δH 2.72) through a methine proton H-8 (δH 2.30), olefinic methine protons H-13 (δH 6.04) and H-14 (δH 5.15), methylene protons H2-15 (δH 2.14 and 1.72) and a methine proton H-16 (δH 2.42) to an olefinic proton H-17 (δH 4.89); and from a methine proton H-22 (δH 5.78) to a methine proton H-28 (δH 3.68) through a methine proton H-21 (δH 4.34). The presence of a six-membered ring (C-4 to C-9) was identified from HMBC correlations from a singlet methyl proton H3-12 (δH 1.18) to a methine carbon C-5 (δC 37.9), a quaternary carbon C-6 (δC 58.9) and a methine carbon C-7 (δC 63.3); and from the methine proton H-8 to a methine carbon C-4 (δC 52.8) and a quaternary carbon C-9 (δC 64.8). The high-field shift for the C-6 signal supported the presence of an epoxide ring between C-6 and C-7. HMBC correlations from H-4, H-8 and the nitrogen-substituted methine proton H-3 (δC 54.9) to a carbonyl carbon C-1 (δC 176.6) suggested the presence of a γ-lactam ring moiety, as shown in Figure 1c. Furthermore, the presence of a 13-membered carbon ring structure was revealed by following 1H–13C long-range couplings, from a methyl proton H3-30 (δH 0.89), which exhibited 1H spin coupling to H-16, to an olefinic methine carbon C-17 (δC 137.5); from a methyl proton H3-31 (δH 1.55) to C-17, an olefinic quaternary carbon C-18 (δC 131.8) and an oxymethine carbon C-19 (δC 82.2); from an oxymethine proton H-19 (δH 3.96) together with H-21 and H-22 to a ketone carbonyl carbon C-20 (δC 209.1); and from H-4, H-8, H-21 and H-28 to another ketone carbonyl carbon C-29 (δC 209.0). The direct connectivity between C-10 and C-3′ was evident from HMBC correlations from H2-10 to C-2′ (δC 124.6), C-3′ and C-3a′. Thus, a chaetoglobosin skeleton was established. The structure of the remaining part of the molecule was determined as follows. The structure of a tetrahydrofuran ring moiety was determined from long-range couplings from an oxymethine proton H-27 (δH 5.81, δC 82.7) to methine carbons C-21 (δC 51.2), C-28 (δC 58.2) and an oxymethine carbon C-22 (δC 82.1). HMBC correlations from an allylic methyl proton H3-32 (δH 2.28) to aromatic quaternary carbons C-25 (δC 145.0), C-26 (δC 111.3), C-26a (δC 135.1) and additionally to C-24 (δC 133.8) weakly; from H-27 to an aromatic quaternary carbon C-22a (δC 121.6) and C-26; from H-22 to aromatic quaternary carbons C-22a, C-23 (δC 138.8), C-24 and C-26a; and from H-28 to C-26a indicated the presence of an 8-methyl-1,2,3,4-tetrahydro-1,4-epoxynaphthalene-5,6,7-triol substructure. The proton chemical shift values at H-22 and H-27, together with low-field shifted 13C chemical shift values at C-22 and C-27 because of the presence of an ether bond agreed well with the corresponding proton chemical shifts of 1,4-epoxy-1,2,3,4-tetrahydronaphthalene-endo,cis-2,3-dicarboxylic acid (δH 5.48 and 5.44, in dimethyl sulfoxide-d6).6 A NOESY correlation between H-27 and H3-32 also supported these assignments. On the basis of the chemical shifts of C-23, C-24 and C-25, and the index of hydrogen deficiency deduced from the molecular formula, all three of these carbons were attached to phenolic hydroxy groups. An E geometry for the double bond at C-13–C-14 was deduced from a large coupling constant between H-13 and H-14 (15.0 Hz). The geometry of the position at C-17 was assigned for E based on the high-field shift for C-31 signal (δC 11.8) owing to a γ-effect. Therefore, the gross structure of 1 was identified to be that shown in Figure 1c.

MBJ-0039 (2) was obtained as a colorless amorphous powder: [α]24D –115 (MeOH; c 0.6). The UV absorption maxima in MeOH at 281 nm (ɛ 6300) and the IR spectrum (νmax: 3400 and 1689 cm−1) of 2 resembled those of 1. Furthermore, the molecular formula of 2 (C41H44N2O9) established based on HR-ESI-MS data (m/z 709.2923 [M+H]+) was the same as that of 1. Thus, 2 was considered to be a structural isomer of 1. The 1H and 13C NMR spectroscopic data of 2 were very similar to those of 1 (Table 1). Detailed analyses of the 2D NMR data revealed a chaetoglobosin skeleton identical to 1, including the exact assignments for the oxymethine protons H-22 (δH 5.63, δC 83.4) and H-27 (δH 5.97, δC 82.1). The structure of the 8-methyl-1,2,3,4-tetrahydro-1,4-epoxynaphthalene-5,6,7-triol moiety of 2 differed from that in 1. Observation of a NOESY correlation between an oxymethine proton H-22 and an allylic methyl proton H3-32 (δH 2.00) suggested that 2 had an inverted arrangement of 1 for this region. In addition, COSY and HMBC data supported these assignments as shown in Figure 1d. Thus, the structure determination of 2 was accomplished as shown in Figure 1a.

MBJ-0040 (3) features the following properties: [α]24D –110 (MeOH; c 0.3); UV λmax (ɛ) in MeOH: 281 (6000) nm; HR-ESI-MS: m/z 691.2986 [M–H], calcd. for C41H43N2O8 m/z 691.3019; and IR absorption (νmax) 3400 and 1689 cm−1. Analysis of NMR spectra revealed that the partial structure of 3 was the same as that of 2, including the geometries of 13E and 17E, with the exception of the cyclohexane ring moiety (C-4 to C-9) as described below. The DQF-COSY spectrum showed sequences from a methyl proton H3-11 (δH 0.94) through methine protons H-5 (δH 2.53) and H-4 (δH 2.62) to a nitrogen-substituted methine proton H-3 (δH 3.41) and from an olefinic methine proton H-7 (δH 5.36, δC 126.8) to a methine proton H-8 (δH 3.03). Furthermore, the CT-HMBC spectrum showed 1H–13C long-range correlations from H3-11 to methine carbons C-4 (δC 55.2), C-5 (δC 36.6) and an olefinic quaternary carbon C-6 (δC 141.7); from an allylic methyl proton H3-12 (δH 1.74) to C-5, C-6 and an olefinic methine carbon C-7; and from H-5 and H-7 to a quaternary carbon C-9 (δC 67.7), which revealed a 1,6-dimethylcyclohex-1-ene moiety (Figure 1e). Thus, the gross structure of 3 was determined as shown in Figure 1b.

The cytotoxic activities of novel compounds 13 and chaetoglobosins C and F against human ovarian adenocarcinoma SKOV-3 cells were tested using the WST-8 ((2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt) colorimetric assay (Cell Counting Kit; Dojindo, Kumamoto, Japan). After 72 h of treatment, all compounds exhibited moderate cytotoxic activity against SKOV-3 cells (IC50; 1: 14 μM, 2: 11 μM, 3: 14 μM, and chaetoglobosins C and F: 8 μM and 2 μM, respectively).

Although a number of chaetoglobosin derivatives have been isolated from fungi,3, 4, 7 13 are the first samples of chaetoglobosins containing an 8-methyl-1,2,3,4-tetrahydro-1,4-epoxynaphthalene-5,6,7-triol moiety.