Recently, we have constructed a library of isolated natural products (CB library) primarily of microbial origin, in order to perform efficient screenings.1 During the exploration of rare microbial products with promising biological and pharmacological properties, we have developed an advanced system for compound identification based on accumulated HPLC-MS profiling data and strain information designated as ‘MBJ’s special selection’. Using this system, we have already succeeded in discovering novel eremophilane derivatives MBJ-0009 and MBJ-0010 from Nectria sp. f26111,2 MBJ-0011, MBJ-0012 and MBJ-0013 from Apiognomonia sp. f24023,3 cytotoxic chaetoglobosin derivatives MBJ-0038, MBJ-0039 and MBJ-0040 from Chaetomium sp. f24230,4 and a cytotoxic hydroxamate MBJ-0003 from Micromonospora sp. 29867.5 During the course of further screening, we discovered two aziridine-containing linear peptides MBJ-0034 (1) and MBJ-0035 (2) from the culture of Streptosporangium sp. 32552 (Figure 1a). Herein, we report the fermentation, isolation, structure elucidation and in brief, the biological activities of 1 and 2.

Figure 1
figure 1

(a) Structures of 1 and 2. (b) Structure determination of 1. COSY and HMBC (1H to 13C) correlations are shown as bold lines and arrows, respectively. (c) Structure determination of 2. COSY and HMBC (1H to 13C) correlations are shown as bold lines and arrows, respectively.

Full size image

The producing microorganism Streptosporangium sp. 32552 was isolated from a plant sample collected in Iwata, Shizuoka 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 Assoc., Hokkaido, Japan), 2% glucose (Junsei Chemical, Tokyo, Japan), 2% soy bean powder (SoyPro, J-Oil Mills, Tokyo, Japan), 0.5% yeast extract powder (Oriental Yeast, Tokyo, Japan), 0.25% NaCl (Junsei Chemical), 0.32% CaCO3 (Wako Pure Chemical Industries, Osaka, Japan), 0.0005% CuSO4·5H2O (Wako Pure Chemical), 0.0005% ZnSO4·7H2O (Wako Pure Chemical Industries) and 0.0005% MnCl2·4H2O (Junsei Chemical) (pH 7.4). The flasks were shaken on a rotary shaker (220 r.p.m.) at 28 °C for 3 days. Then, the aliquots (0.5 ml) of the broth were transferred to 500-ml Erlenmeyer flasks containing 50 ml of the production medium of the same composition and were cultured on a rotary shaker (220 r.p.m.) at 28 °C for 4 days.

The whole culture broth (2 l) was extracted with an equal volume of n-BuOH. After the n-BuOH layer was evaporated in vacuo, the resulting residue was suspended in brine (350 ml) and then extracted with EtOAc (350 ml × 3) and n-BuOH (300 ml × 2), successively. The n-BuOH extract (3.1 g) was subjected to reversed-phase medium-pressure liquid chromatography (Purif-Pack ODS-30, size: 60 (39 g), Shoko Scientific Co., Ltd., Yokohama, Japan) with the UV detection wavelength set at 254 nm. A H2O–CH3CN stepwise gradient system (100 ml each of 10, 20, 30, 40 and 50% CH3CN) was used to eluate 1 (63.2 mg, 40% CH3CN). The EtOAc extract (2.4 g) was fractionated by medium-pressure silica gel column chromatography (Purif-Pack SI-30, size: 60 (27 g), Shoko Scientific) using an n-hexane–EtOAc linear gradient system (0–25% EtOAc over 12 min and was kept at 25% for 3 min, flow rate: 20 ml min–1) followed by a CHCl3–MeOH stepwise gradient system (100 ml each of 0, 2, 5, 10, 20, 30 and 100% MeOH). The 20% MeOH fraction (288.1 mg) was subjected to gel filtration chromatography (Sephadex LH-20, GE Healthcare BioSciences AB, Uppsala, Sweden; 2.5 × 45 cm) and eluted using CHCl3–MeOH (1:1) to afford crude 2 (63.5 mg). The final purification of 2 (11.9 mg) was carried out by reversed-phase HPLC using a CAPCELL PAK C18 MGII column (5.0 μm, 20 × 150 mm; Shiseido, Tokyo, Japan) with 35% aqueous CH3CN containing 0.1% formic acid (flow rate: 10 ml min–1, retention time: 12.5 min).

MBJ-0034 (1) was isolated as a colorless amorphous powder: [α]25D +0.6 (c 0.1, MeOH); UV λmax nm (log ɛ): 241 (4.2) and 305 (3.8) in MeOH; IR (ATR) νmax 3300 and 1650 cm−1 (hydroxy and carbonyl). The molecular formula of 1 was established as C26H37N7O9 by HR-ESIMS (m/z 590.2560 [M–H]–, calcd for C26H36N7O9: 590.2575). The peptide-like nature of 1 was evident from the NMR resonances corresponding to amide NH protons (δH 8.49–7.91) and the resonances corresponding to the carbonyl carbons (δC 165.1–173.7) in the 1H and 13C NMR spectra, respectively. The planar structure of 1 was determined by a series of 2D NMR analyses including double quantum filtered COSY (DQF-COSY), heteronuclear single quantum coherence and constant-time heteronuclear multiple-bond correlation6 (CT-HMBC). The 13C and 1H NMR data of 1 is listed in Table 1.

Table 1 13C and 1H NMR spectroscopic data for MBJ-0034 (1) in DMSO-d6 and MBJ-0035 (2) in CD3OD
Full size table

The CT-HMBC spectrum showed a 1H–13C long-range coupling from a nitrogen-bearing methyl proton H3-26 (δH 2.20, δC 34.4) to a methine carbon C-7 (δC 63.9), in which proton H-7 (δH 2.86) was long-range coupled to an amide carbonyl carbon C-6 (δC 173.7). Together with these HMBC correlations, the 1H sequence from the H-7 through aliphatic methylene protons H2-8 (δH 1.47, 1.40) and H2-9 (δH 1.58) to nitrogen-bearing methylene protons H2-10 (δH 3.46, δC 47.2) were observed in the DQF-COSY spectrum, indicating the presence of an N-methylornithine moiety (Figure 1b). On the other hand, observation of COSY correlations from a doublet amide NH proton 2-NH (δH 8.09) through an α-methine proton H-2 (δH 4.31), aliphatic methylene protons H2-3 (δH 1.93, 1.68) and H2-4 (δH 1.88) to nitrogen-bearing methylene protons H2-5 (δH 3.45, δC 51.4) and HMBC correlations from H-2, H2-3 and H2-5 to an amide carbonyl carbon C-1 (δC 165.1) revealed the presence of a δ-lactam of an ornithine moiety.

The presence of a β-alanine moiety was ascertained by 1H–1H spin systems from α-methylene protons H2-12 (δH 2.51) through β-methylene protons H2-13 (δH 3.25, 3.24) to a triplet amide NH proton 13-NH (δH 7.91) and 1H–13C long-range couplings from H2-12 and H2-13 to an amide carbonyl carbon C-11 (δC 171.1). The structure of a glycine moiety was determined by 1H–1H couplings between a triplet amide NH proton 15-NH (δH 8.49) and α-methylene protons H2-15 (δH 3.74, 3.65) together with 1H–13C long-range couplings from all the protons to an amide carbonyl carbon C-14 (δC 168.5).

The presence of a salicylic acid moiety was established based on a 1H sequence from an aromatic proton H-22 (δH 6.99) to an aromatic proton H-25 (δH 7.63) through aromatic protons H-23 (δH 7.46) and H-24 (δH 6.94), in addition to HMBC correlations from H-23 and H-25 to an oxygen-bearing aromatic quaternary carbon C-21 (δC 159.2), from H-22 and H-24 to an aromatic quaternary carbon C-20 (δC 110.1), and from H-25 to an amide carbonyl carbon C-19 (δC 166.0). Furthermore, HMBC correlations from an α-methine proton H-17 (δH 5.00) and methylene protons H2-18 (δH 4.64, 4.51), which were mutually 1H spin coupled, to amide carbonyl carbons C-16 (δC 170.3) and C-19, revealed the presence of an aziridine moiety, which was acylated with the salicylic acid at the N-terminus.

The connectivity among the amino-acid units was determined by 1H–13C long-range couplings from H-2, H2-10, H2-13 and H2-15 to C-6, C-11, C-14 and C-16, respectively. The molecular formula indicated the presence of two hydroxamic acid groups in 1. The chemical shift value of C-26 (δC 34.4) suggested that the nitrogen atom connected to C-7 is protonated.7 Accordingly, the positions of two hydroxamic acid groups were determined to be C-1 (C-5) and C-10 (C-11). Thus, the gross structure of 1 was elucidated, as shown in Figure 1a.

MBJ-0035 (2) was obtained as a colorless amorphous powder: [α]25D –52 (c 0.1, MeOH); UV λmax nm (log ɛ): 240 (4.1) and 305 (3.7) in MeOH; IR (ATR) νmax 3300 and 1650 cm−1 (hydroxy and carbonyl). The molecular formula of 2 was established as C35 H46N8O10 from the HR-ESIMS data (m/z 739.3450 [M+H]+, calcd for C35H47N8O10: 739.3415).

Based on the analyses of 1D and 2D NMR spectra, the partial structure of 2 was found to be the same as that of 1, as shown in Figure 1c. The structural difference between 1 and 2 is an additional phenethylcarbamic acid moiety, whose presence was confirmed by 1H sequences from aromatic protons H-31/35 (δH 7.21) through aromatic protons H-32/34 (δH 7.27) to an aromatic proton H-33 (δH 7.17) and from nitrogen-bearing methylene protons H2-28 (δH 3.38, δC 43.6) to allylic methylene protons H2-29 (δH 2.80), together with HMBC correlations from H2-28 to carbonyl ketone carbon C-27 (δC 161.1) and from H2-29 to an aromatic quaternary carbon C-30 (δC 140.9) and an aromatic methine carbon C-31/35 (δC 129.9). Finally, HMBC correlations from an N-methyl proton H3-26 (δH 2.77) and a low-field shifted α-methine proton H-7 (δH 4.85) to C-27 indicated that the phenethylcarbamic acid moiety is connected to C-7 by an amide bond. Therefore, the structure of 2 was determined as shown in Figure 1a.

The cytotoxic activities of 1 and 2 against human ovarian adenocarcinoma SKOV-3 cells were examined by using the WST-8 [5-(2,4-disulfophenyl)-3-(2-methoxy-4-nitrophenyl)-2-(4-nitrophenyl)-2H-tetrazolium, monosodium salt] colorimetric assay (Cell Counting Kit; Dojindo, Kumamoto, Japan). After 72 h of treatment, 2 exhibited weak cytotoxic activity against SKOV-3 cells with the IC50 of 37 μM, whereas 1 exhibited no cytotoxicity (IC50>50 μM). Since the derivatives of 1 and 2, madurastatin A1 and its congeners were reported to show antibacterial activities against Micrococcus luteus,7 we evaluated the antimicrobial activities of 1 and 2. As the result, contrary to this report, 1 and 2 did not show antimicrobial activities against M. luteus nor Bacillus subtilis even at the concentrations of 100 μM. Compounds 1 and 2 did not show the antimicrobial activities against Escherichia coli. To the best of our knowledge, peptides with aziridine moiety is extremely rare in nature and have not been reported except for madurastatin A1. The planar structure of 1 appears to be same as madurastatin C1.8