Two new glutarimide antibiotics from Streptomyces sp. HS-NF-780

Abstract

Two new glutarimide antibiotics, 9-methylstreptimidone 2-α-d-glucopyranoside (1), and hydroxyiso-9-methylstreptimidone (2), along with a known compound, 9-methylstreptimidone (3), have been isolated from the broth of Streptomyces sp. HS-NF-780. Their structures were determined on the basis of spectroscopic analysis, including 1D and 2D NMR techniques as well as ESI-MS and comparison with data from the literature. By modified Mosher’s method and acid hydrolysis, the absolute configurations of compounds 1 and 2 were established. Compounds 1 and 2 exhibited moderate cytotoxic activity.

The antibiotics of the glutarimide group are structurally characterized by the presence of glutarimide ring bearing a side chain at the 1-position [1]. A number of distinct substances belonging to this class such as 9-methylstreptimidones [2], cycloheximide [3, 4], and migrastatin [5, 6] have been isolated from various species of Streptomyces [7]. Among them, 9-methylstreptimidone showed strong inhibitory activity against yeasts and filamentous fungi [1]. Besides, 9-methylstreptimidone inhibited NO production and iNOS expression in LPS-stimulated RAW264.7 cells, and induced apoptosis in Jurkat cells and adult T-cell leukemia cells, similar to other NF-κB inhibitors [8]. In the course of our screening program for novel microbe-derived bioactive secondary metabolites, two new glutarimide derivatives named 9-methylstreptimidone 2-α-d-glucopyranoside (1) and hydroxyiso-9-methylstreptimidone (2), along with a known compound, 9-methylstreptimidone (3) (Fig. 1), were isolated from the culture broth of Streptomyces sp. HS-NF-780. In this paper, we describe the fermentation, isolation, structure elucidation, and bioactivity of the two new compounds.

Fig. 1

Structures of compounds 1, 2, and 3

The producing strain HS-NF-780 was isolated from a soil sample collected from Linyi, Shandong province, China using the standard dilution plate method. The strain was identified as the genus Streptomyces because its 16S rRNA sequence (accession no: MH362834 in the GenBank) exhibited a high-sequence similarity of 100% with that of Streptomyces sp. NEAU-BGG209 (accession no: MG820043).

Strain HS-NF-780 was grown and maintained for 7 days at 28 °C on the YMS medium consisting of yeast extract 4.0 g, malt extract 10.0 g, glucose 4.0 g, CoCl₂·6H₂O 0.005 g and agar 20.0 g in 1.0 l tap water at pH 7.0. The stock culture was transferred into 1 l Erlenmeyer flasks containing 250 ml of the seed medium and incubated at 28 °C for 48 h on a rotary shaker at 250 r.p.m. The seed medium was composed of glucose 4.0 g, malt extract 10.0 g and yeast extract 4.0 g in 1.0 l tap water, pH 7.0. All of the media were sterilized at 121 °C for 20 min. The seed culture (5%) was transferred into 1 l Erlenmeyer flasks containing 25% volume of production medium. The production medium was composed of glucose 1%, soluble starch 4%, yeast extract 0.4%, malt extract 1%, CaCO₃ 0.2%, FeSO₄·7H₂O 0.1%, ZnSO₄·7H₂O 0.1%, and MnCl₂·4H₂O 0.1% at pH 7.2–7.4. The flasks were incubated at 28 °C for 7 days, shaken at 250 r.p.m.

The final 20 l fermentation broth was filtered to separate mycelial cake and supernatant. The mycelial cake was washed with water (3 l) and subsequently extracted with MeOH (3 l) and the supernatant was subjected to a Diaion HP-20 resin (Mitsubushi Chemical, Tokyo, Japan) column eluting with 95% EtOH (5 l). The MeOH extract and the EtOH eluents were evaporated under reduced pressure at 55 °C to yield the crude extract. The crude extract was chromatographed on a silica gel (Qingdao Haiyang Chemical Group, Qingdao, China; 100-200 mesh) column and successively eluted with a stepwise gradient of CHCl₃/MeOH (100:0-50: 50, v/v) to give three fractions (Fr.1-Fr.3) based on the TLC profiles. TLC was performed on silica-gel plates (HSGF254, Yantai Chemical Industry Research Institute, Yantai, China) with solvent system of CHCl₃/MeOH (9:1, v/v) and the developed TLC plates were observed under a UV lamp at 254 nm or by heating after spraying with sulfuric acid-ethanol, 5:95 (v/v). The Fr.2 was subjected to another silica gel column eluted with CHCl₃/MeOH (95:5-60:40, v/v) to give six fractions (Fr.2-1-Fr.2-6). The Fr.2-4 was further purified by semi-preparative HPLC (Agilent 1100, Zorbax SB-C18, 5 μm, 250 × 9.4 mm inner diameter; 1.5 ml min⁻¹; 254 nm; Agilent, PaloAlto, CA, USA) eluting with CH₃CN/H₂O (20:80, v/v) to obtain compound 1 (t_R 24.68 min, 18.0 mg). The Fr.2-3 was separated by semi-preparative HPLC eluting with CH₃CN/H₂O (25:75, v/v) to afford compound 2 (t_R 20.13 min, 14.2 mg). The Fr.1 was subjected to a Sephadex LH-20 (GE Healthcare, Glies, UK) column eluted with CHCl₃/MeOH (1:1, v/v) and detected by TLC to give two subfractions (Fr.1-1-Fr.1-2). The Fr.1-2 was isolated by semi-preparative HPLC eluting with CH₃CN/H₂O (35:65, v/v) to give compound 3 (t_R 25.06 min, 23.5 mg). ¹H and ¹³C NMR spectra were measured with a Bruker DRX-400 (400 MHz for ¹H and 100 MHz for ¹³C) spectrometer (Bruker, Rheinstetten, Germany). The ESI-MS and HR-ESI-MS spectra were taken on a Q-TOF Micro LC-MS-MS mass spectrometer (Waters Co, Milford, MA, USA).

Compound 1 was isolated as colorless oil with [α]$_{\mathrm{D}}^{25}$ + 125 (c 0.03, EtOH) and UV (EtOH) λ_max nm (log ε): 233 (4.02). Its molecular formula was determined to be C₂₃H₃₅NO₉ by HRESIMS at m/z 492.2207 [M + Na]⁺ (calcd as 492.2204 for C₂₃H₃₅NO₉Na) and NMR data (Table 1). In the IR spectrum of 1, absorption at 3447, 3197, and 1689 cm⁻¹ indicated the presence of hydroxyl, imide, and carbonyl groups, respectively. Analysis of ¹H NMR spectrum (Table 1) of 1 revealed the presence of three olefinic protons at δ_H 5.19 (1 H, br d, J = 9.8 Hz), 5.48 (1 H, m), 5.82 (1 H, br d, J = 11.6 Hz), one anomeric proton at δ_H 4.87 (1 H, d, J = 3.9 Hz), one oxygenated methylene at δ_H 3.64 (1 H, dd, J = 11.9, 5.3 Hz), 3.78 (1 H, dd, J = 11.9, 2.1 Hz), five oxygenated methine protons from δ_H 3.27 to δ_H 4.19, two olefinic methyls at δ_H 1.77 (3 H, dd, J = 7.2, 2.6 Hz), 1.86 (3 H, d, J = 1.2 Hz), one doublet aliphatic methyl at δ_H 1.14 (3 H, d, J = 6.8 Hz). The ¹³C NMR and DEPT135 spectra (Table 1) of 1 showed 23 resonances attributable to one carbonyl carbon at δ_C 212.1, two amide carbonyl carbons at δ_C 175.4 and 175.4, three sp² methines at δ_C 125.8, 129.7, 134.1, one sp² quaternary carbon at δ_C 136.7, one anomeric carbon at δ_C 99.1, five oxygen-bearing aliphatic methines between 71.6 and 75.0 ppm, one oxygenated methylene at δ_C 62.5, two aliphatic methines at δ_C 28.2, 48.2, four aliphatic methylenes at δ_C 37.9, 39.2, 41.4, 44.8, and three methyl carbons at δ_C 15.0, 16.7, 17.5. Comparison of the ¹H and ¹³C NMR data (Table 1) of 1 with those of 3 revealed significant similarities. The differences between 1 and 3 were that 1 showed six extra ¹³C resonances. The six extra ¹³C resonances were postulated to glucose moiety according to one doublet anomeric proton (δ_H 4.87), one oxygenated methylene (δ_H 3.64, 3.78) and four oxygenated methine protons. The relatively small ³J_HH value (3.9 Hz) of the anomeric proton suggested an α linkage [9]. The linkage of the glucose to the aglycone was established by the HMBC correlation (Fig. 2) from H-1″ to C-2. The NOESY correlations (Fig. 2) between H-5 and H₃-12, H-6 and H-8 demonstrated the geometry of Δ^6,7 was E. In the ¹H NMR spectrum of 1, the coupling constant between H-8 and H-9 was 11.6 Hz, indicating the geometry of Δ^8,9 was Z. Furthermore, the presence of glucose was evidenced by the acid hydrolysis. Compound 1 (2.5 mg) was dissolved in 2 ml of 2 M HCl and heated for 2 h at 80 °C, followed by neutralization with NaHCO₃. The reaction mixture was extracted with CHCl₃ to separate a sugar moiety-containing aqueous fraction from the aglycone-containing fraction. The aqueous fraction was identified by cochromatography with authentic glucose on TLC analysis using ethyl acetate/pyridine/glacial acetic acid/H₂O (8: 5: 1: 1.5, v/v) as a developing solvent. Spots were detected by heating after spraying with sulfuric acid-ethanol (5:95, v/v). The identical R_f values of the aqueous fraction with that of authentic glucose indicated that the sugar moiety of 1 was glucose. The aglycone-containing fraction was subjected to reverse phase HPLC for analysis, with a Zorbax B-C18 column (Agilent 1100, 250 × 9.4 mm inner diameter, 5 μm), mobile phase of CH₃CN/H₂O (35:65, v/v), flow rate at 1.5 ml min⁻¹, and detection wavelength at 220 nm. Under these conditions, the aglycone-containing fraction gives peak at t_R (min) = 25.06. The peak of the known compound 9-methylstreptimidone (3) was detected at t_R (min) = 25.11. The retention time of the aglycone moiety of 1 was in good agreement with 3, which suggested that the aglycone moiety of 1 was 9-methylstreptimidone. In order to determine the absolute configuration of the glucose moiety in 1, the sugar residue was dissolved in pyridine (3 ml) containing l-cysteine methyl ester hydrochloride (1 mg) and heated at 60 °C for 1 h. A total of solution of O-torylisothiocyanate (25 μl) was added to the mixture, which was heated at 60 °C for a further 1 h. The mixture was analyzed by reversed-phase HPLC (Amethyst C18-H, 5 μm, 250 × 4.6 mm inner diameter; 0.8 ml min⁻¹; 250 nm) at 35 °C eluting with CH₃CN/H₂O (25:75, v/v). Under these conditions, standard sugar gave peak at t_R (min) = 21.658 for d-glucose. The peak of the sugar residue was detected at t_R (min) = 21.789, which identified as d-glucose by comparison with the retention time of the authentic sample [10,11,12].

Table 1 The NMR spectroscopic data of compounds 1–3

Fig. 2

Key ¹H-¹H COSY, HMBC, and NOESY correlations of compounds 1 and 2

Compound 2 was isolated as pale yellowish oil with optical rotation of [α]$_{\mathrm{D}}^{25}$ + 105 (c 0.03, EtOH) and UV (EtOH) λ_max nm (log ε): 276 (4.10). HR-ESI-MS showed a molecular ion peak at m/z 346.1625 [M + Na]⁺ (calcd for C₁₇H₂₅NO₅Na, 346.1625), indicating a molecular formula of C₁₇H₂₅NO₅. In the IR spectrum of 2, absorption at 3419 and 1684 cm⁻¹ indicated the presence of hydroxyl and carbonyl groups, respectively. Analysis of ¹H NMR spectrum (Table 1) of 2 revealed the presence of two olefinic protons at δ_H 5.63 (1 H, d, J = 8.2 Hz), 6.99 (1 H, s), two oxygenated methine protons at δ_H 4.20 (1 H, m), 4.66 (1 H, m), two olefinic methyl carbons at δ_H 1.92 (3 H, s), 1.95 (3 H, s), one doublet aliphatic methyl at δ_H 1.28 (3 H, d, J = 6.3 Hz). The ¹³C NMR and DEPT135 spectra (Table 1) of 2 showed 17 resonances attributable to one carbonyl carbon at δ_C 203.2, two amide carbonyl carbons at δ_C 175.5 and 175.6, two sp² methines at δ_C 140.4, 144.7, two sp² quaternary carbons at δ_C 133.5, 137.2, one aliphatic methine at δ_C 28.7, two oxygenated methines at δ_C 65.2, 66.8, four methylenes at δ_C 38.0, 39.2, 42.9, 46.3, and three methyl resonances at δ_C 13.2, 16.7, 23.4. The complete assignment of all ¹H and ¹³C NMR spectral data of 2 was subsequently accomplished by the ¹H-¹H COSY, HSQC and HMBC spectra. The correlations between H-3′/H₂-1/H-2/H₂-3, H-8/H-9/H₃-10 protons in the ¹H-¹H COSY spectrum (Fig. 2) indicated the presence of two structural units of C-3′–C-3 and C-8–C-10. The observed HMBC correlations (Fig. 2) from H-2, H-6 to C-4, from H₃-11 to C-4, C-5, C-6, from H₃-12 to C-6, C-7, C-8 established the linkage of C-3′–C-10. A glutarimide ring was defined by two amide carbonyl carbons at δ_C 175.5 and 175.6, which were coupled long range in an HMBC experiment to the protons (Table 1) of a four protons pair of methylene signals at δ_H 2.36 (1 H, m), 2.38 (1 H, m), 2.68 (1 H, m), 2.75 (1 H, m), which were in turn coupled in a ¹H-¹H COSY spectrum to a single methine proton signal at δ_H 2.38 (1 H, m) [13]. On the basis of the above spectroscopic data, a gross structure of 2 was established (Fig. 1). The NOESY correlation (Fig. 2) between H₃-11 and H₃-12 demonstrated the geometry of Δ^5,6 was E. The geometry of Δ^7,8 was also assigned as E by the NOESY correlation between H-6 and H-8. The absolute configuration of 2 was determined by the modified Mosher’s method [14, 15]. To a solution of compound 2 (2.5 mg) in dry pyridine (200 μL) was added (−)-MTPA chloride (15 μl), and the solution was stirred at room temperature for 1 h. The reaction mixture was fractionated by semi-preparative HPLC eluting with CH₃CN/H₂O (80:20, v/v) to afford (S)-MTPA esters 1a (t_R 19.82 min, 1.2 mg) and 2a (t_R 20.39 min, 1.0 mg). In the same way, by using (+)-MTPA chloride, the compound 2 (2.5 mg) was converted into a mixture. The mixture was isolated by semi-preparative HPLC eluting with CH₃CN/H₂O (85:15, v/v) to obtain (R)-MTPA esters 1b (t_R 26.18 min, 1.0 mg) and 2b (t_R 27.43 min, 1.0 mg). Based on the MTPA determination rule, calculating △δ (ppm) = δ_{S 1a} − δ_{R 1b} (Fig. 3), the absolute configuration at C-2 was assigned as R and the absolute configuration at C-9 was assigned as R. However, calculating △δ (ppm) = δ_{S 2a} − δ_{R 2b} (Fig. 4), the absolute configuration at C-2 was assigned as R and the absolute configuration at C-9 was assigned as S. It showed that 2 was a C-9 epimeric mixture. Chiral HPLC analysis of 2 with a CElluose-C column and mobile phase of CO₂/MeOH further indicated that 2 was the epimeric mixture in a ratio of ~3:2.

Fig. 3

Δδ values for the MTPA esters (1a, 1b); Δδ (ppm) = δ_{S 1a} − δ_{R 1b}

Fig. 4

Δδ values for the MTPA esters (2a, 2b); Δδ (ppm) = δ_{S 2a} − δ_{R 2b}

Compound 3 was isolated as pale yellowish oil. Its structure was elucidated as 9-methylstreptimidone by analysis of its ¹H NMR and ESI-MS spectral data (Table 1) and comparison with literature values [16].

The cytotoxicity of 1 and 2 was assayed for growth-inhibition activity in vitro against three human tumor cell lines, human erythroleukemia cell line K562, human breast cancer cell line MCF-7, and human colon carcinoma cell line HCT-116 according to the CCK8 colorimetric method as reported in our previous papers [17, 18] using doxorubicin as positive control. The results (Table 2) demonstrated that the two new compounds possessed moderate cytotoxic activity towards the three tumor cell lines.

Table 2 Cytotoxic activity of 1 and 2 against selected human tumor cell lines