Introduction

Natural products and their derivatives are major sources in the discovery of novel drug candidates.1 Natural products displayed a unique and vast chemical diversity, and also diversity in biological activities, making natural products libraries favorable and attractive in drug discovery.2 In recent years, natural products-based drug discovery is gaining attention, which implies that natural products serve an evolving role in future drug discovery.3 Microorganisms were known to possess immense capacity to produce various bioactive secondary metabolites4 with diverse potential activities. Hence, microbial metabolites have been a significant source of pharmaceutical leads and therapeutic agents.5, 6 They are of great interest for human to isolate them and develop as novel drug candidates. These metabolites are also utilized as potential bioprobes7 in chemical biology studies, mainly as a tool for investigation of biological functions.8 To date, many secondary metabolites have been reported. Nevertheless, not all of them have been isolated nor have their latent useful activities examined due to their wide range of physicochemical properties and relatively low abundance. Thus, we have attempted to construct a microbial metabolites fraction library by a systematic separation method and develop a spectral database, namely NPPlot (Natural Products Plot), based on photodiode array (PDA) detector-attached LC/MS analysis.9 On the basis of NPPlot comparison method, structurally unique metabolites could be identified efficiently and rapidly. Also, the efforts for identifying known compounds would be reduced significantly. As previously reported, our methodology of constructing a fraction library led us to identify and isolate several novel compounds, including verticilactam10 from Streptomyces spiroverticillatus JC-8444, a tautomycin producer,11, 12, 13 spirotoamides A and B14 from Streptomyces griseochromogenes JC82-1223, a tautomycetin producer,15, 16 novel furaquinocins I and J17 and 6-dimethylalylindole-3-carbaldehyde18 from Streptomyces reveromyceticus SN-593, a reveromycin producer.19 These findings have revealed the advantages of utilizing fraction libraries for the isolation and probing of the activities of natural products.

The microbial metabolites fraction library was prepared following the scheme as described in the previous paper.9 To take advantage of a fraction library and to facilitate the distinction of novel metabolites from known compounds rapidly, we have developed a unique spectral database, namely NPPlot, based on PDA-LC/MS analysis to obtain the UV absorption and mass spectra of each metabolite within each fraction in the fraction library. These information were important for us to discriminate known compounds and to discover structurally novel metabolites. The open spectral database was unable to distinguish metabolite groups between microbial strains as compared with NPPlot, thus making it a useful tool for the discovery of novel metabolites.

NPPlot is a distribution map of metabolites plotted on a two-dimensional (2D) area by retention time of HPLC on x axis and m/z values on y axis. On the basis of NPPlot screening, we could discover novel compounds and also search for specific metabolites group for each microbial strain through the comparison of distribution patterns generated from several microbial strains. On the basis of this novel method, we have generated NPPlot from five different microbial strains and compared their distribution patterns. A group of compounds showing a characteristic distribution in the region of molecular mass around 1150 and retention time around 25 min was detected in strain RK88-1355 (Figure 1). On screening for structurally unique secondary metabolites using the NPPlot comparison method, we identified two unknown peaks and one known peak with identical UV absorption pattern, but slightly different mass spectra in fractions of Streptomyces sp. RK88-1355. The related fractions were purified by a C18-HPLC to afford compounds 1 (8.8 mg) and 2 (1.5 mg), together with one known compound 3 (5.7 mg), respectively (Table 1). We report herein the structures and biological activities of these compounds, designated RK-1355A (1) and B (2), as shown in Figure 2.

Figure 1
figure 1

NPPlot for Streptomyces sp. strain RK88-1355 showing the characteristic distribution of metabolites (highlighted region) plotted by retention time of HPLC on x axis and m/z values on y axis. Light color spot on NPPlot indicates one single metabolite and dark color spot implies an overlapped spot containing more than one or few metabolites, which are closely related. A full color version of this figure is available at The Journal of Antibiotics journal online.

Full size image
Table 1 Physicochemical properties of compounds 1 and 2
Full size table
Figure 2
figure 2

Structures of compounds 1, 2 and 3, UK-63,598.

Full size image

Results and discussion

We have isolated novel quinomycin derivatives, RK-1355A (1) and B (2) from Streptomyces sp. RK88-1355, which was isolated from a soil sample collected in Tohoku area, Japan in 1988. This strain was deposited in RIKEN NPDepo (Natural Products Depository). The molecular formula of compound 1 was established to be C53H62N10O15S2 by HRESIMS (found: m/z 1143.3923 [M+H]+, calcd for C53H63N10O15S2 1143.3916), which required the addition of an oxygen atom to 3 (UK-63,598).20 The UV absorption spectrum showed λmax (MeOH) values at 218, 230, 298 and 356 nm (Table 1), suggesting the presence of 3-hydroxyquinaldic acid.20, 21, 22, 23 On the basis of interpretation of IR spectrum, the presence of hydroxyl group (3370 cm−1), an ester group (1735 cm−1) and an amide group (1655 and 1520 cm−1) were confirmed by the appearance of corresponding bands in the spectrum. These observations were characteristic of a depsipeptide with an aromatic 3-hydroxyquinaldic acid chromophore. The C–H stretching absorptions were observed at 2920 cm−1 and 2850 cm−1, respectively. There was an additional IR band observed at 1014 cm−1, implying the presence of a sulfoxide in 1.24 The 1H NMR spectrum in chloroform-d suggested that 1 contained nine methyl groups. It also contained six exchangeable protons at δ 11.29 (OH) and δ 8.85 (NH) with 2H integral values, respectively, and at δ 6.49 (NH) and δ 6.40 (NH) based on integration values and HSQC interpretation. These exchangeable protons were confirmed by the addition of D2O in the 1H NMR spectrum. Two of the exchangeable protons were observed as a single sharp signal at δ 11.29 with 2H integral value, suggesting that 1 had two hydroxyl groups, which were overlapped in 1H NMR spectrum. At δ 8.85 (br d, J=9.2 Hz), two exchangeable protons were observed with 2H integral value, which confirmed the presence of two secondary amides that were overlapped. The 13C NMR spectrum of 1 showed 51 signals, which did not match with the molecular formula obtained from HR-MS/MS. On the basis of detailed interpretation of 13C DEPT, HSQC and HMBC spectra, one methyl signal at δ 11.8 and one quaternary signal at δ 153.7 were found, each of which was observed as a signal with strong intensity, respectively, were overlapped completely. These observations led to the identification of 53 carbon signals, which were attributed to 9 methyls, 5 methylenes, 19 methines that bore 10 olefinic carbons and 20 quaternary carbons, including 10 carbonyls. These assignments were verified by the 13C DEPT experiment and HSQC spectral data (Table 2).

Table 2 13C NMR chemical shifts for compounds 1 and 2 in CDCl3
Full size table

The planar structure of 1 was determined by 2D NMR analysis, as shown in Figures 3, 4 and 5. The connections between protons and carbons were established by HSQC spectrum. The structure of quinomycin with quinoline moiety for 1 was constructed based on interpretation of DQF-COSY, HSQC and HMBC spectra and confirmed by MS/MS spectrum. The COSY and HMBC spectra showed the presence of four partial structures as shown in Figure 3. The partial structure A, 3-hydroxyquinaldic acid moiety was implied by the presence of nine sp2 carbons (C-2′ to C-8a′) and five related aromatic protons (δ 7.47, δ 7.48, δ 7.67, δ 7.68, δ 7.69) with a proton of hydroxyl group (δ 11.29) and the COSY correlations of H-5′ to H-6′, H-6′ to H-7′ and H-7′ to H-8′. This aromatic moiety was confirmed by the following HMBC correlations: from H-7′ to C-8a′, from H-6′ to C-4a′, from H-5′ to C-8a′, from H-8′ to C-4a′, from H-4′ to C-4a′, C-5′, C-3′, C-8a′ and C-2′, and from OH-3′ to C-2′, C-3′ and C-4′. The COSY spectrum of partial structure B showed a methyl–methine–methylene spin system and it was determined as a 2-methyl-1-methylaminocyclopropanecarboxylic acid moiety through the long-range couplings from H2-27, H-29 and H-30 to C-2 and from H2-27 to C-1. The partial structure C was indicated to be an alanine residue by the COSY correlations of H-8 with NH-9 and H-32 to H-8, and confirmed by the HMBC correlations from H-8 to C-7 and C-32, from H-32 to C-7 and C-8. The COSY correlations from H-11 to NH-33 and H-12, and the HMBC correlations from H-11 to C-10 and H-12 to C-11 confirmed the partial structure D as a serine residue. The COSY and HMBC correlations revealed the presence of two alanine residues (Ala and Ala′), two serine residues (Ser and Ser′), two 2-methyl-1-methylaminocyclopropanecarboxylic acid moieties and two 3-hydroxyquinaldic acid chromophores in compound 1. The HMBC correlations from H-12 to C-14 and from H-25 to C-1 confirmed the ester linkages between the serine residues and the 2-methyl-1-methylaminocyclopropanecarboxylic acid moieties. The connections of the alanyl–serinyl residues were revealed by the long-range correlations from NH-9 to C-10 and from NH-22 to C-23, respectively. The attachments of the aromatic 3-hydroxyquinaldic acid chromophore moieties to the serine residues were clearly indicated by the long-range correlations from NH-33 and NH-41 to C-34 and C-42 of the carbonyl carbons for the chromophore units. Furthermore, the connectivities of these partial structures were established through the long-range correlations from H-30 to C-4, from NH-9 to C-10, from H2-12 to C-14, from H-38 to C-17, and from NH-22 to C-23. For the connection of sulfur-containing intramolecular cross-linkage, which resembled cysteine sulfoxide moiety, it was determined by the COSY correlations of H-5 to H-5a and H-18 to H-18a, and the HMBC correlations from H-18 to C-39 and C-17, from H-5a and H-5 to C-4, from H-31 to C-5 and C-7, and from H-5c and H-18a to C-5a. This connection was confirmed by the 13C NMR chemical shifts of C-5a (71.3 p.p.m.), C-18a (50.0 p.p.m.) and C-5c (18.7 p.p.m.) (Figure 4). Owing to the acquisition of an additional oxygen atom in 1, HR-MS/MS analysis was conducted to confirm the site of oxidation. The fragmentation showing a loss of S-methyl (m/z 1077 [M+H]+) from the parent compound was observed, indicating that the oxidation in 1 was resided at the Cys′ sulfur atom. Thus, the existence of a sulfoxide moiety in 1 was confirmed, which was supported by the IR band at 1014 cm−1 described previously. The cyclic depsipeptide skeleton of 1 was confirmed by HR-MS/MS analysis, in which a small fragment ion peak was observed on m/z value of 1077 [M+H−H2O−CH3SH]+, indicating the dehydration and loss of fragment S-methyl from the protonated compound, which confirmed the presence of S-methyl group in 1 (Figure 5). A fragment ion peak also appeared on m/z value of 685 [M+H−H2O−C22H24N4O6]+, which indicates the dehydration and elimination of fragment C22H24N4O6 from the protonated compound 1. Another fragment ion peak was observed on m/z value of 814 [M+H−C16H15N3O5]+, implying the elimination of fragment C16H15N3O5 from the protonated compound 1. These fragmentation patterns give rise to the skeleton of quinomycin with 3-hydroxyquinaldic acid chromophore. On the basis of the above NMR and MS spectroscopic analyses, the structure of 1 was determined as shown in Figure 2.

Figure 3
figure 3

Partial structures of compound 1.

Full size image
Figure 4
figure 4

Selected DQF-COSY (bold lines) and HMBC correlations (solid arrows) observed for 1.

Full size image
Figure 5
figure 5

MS/MS fragments pattern (dotted lines) and MS/MS spectrum observed for 1.

Full size image

The compound 2 had the molecular formula of C54H64N10O15S2 based on HRESIMS interpretation (found: m/z 1157.4089 [M+H]+, calcd for C54H65N10O15S2 1157.4072), in which its molecular mass is 14 units higher than that of 1. The UV spectrum of 2 showed the characteristic absorptions at 218, 230, 298 and 356 nm (Table 1), suggesting the presence of a 3-hydroxyquinaldic acid chromophore,20, 21, 22, 23 which was identical to that of 1. The IR spectrum also indicated the presence of hydroxyl group (3380 cm−1), an ester group (1735 cm−1) and an amide group (1655 and 1510 cm−1) as observed in compound 1. The C–H stretching absorption was observed at 2920 cm−1 in the IR spectrum. A sulfoxide moiety was revealed by the presence of an IR band at 1014 cm−1, identical to that of 1.24 The similarities between compounds 1 and 2 suggested that 2 was an analog of 1. Compound 2 was analogous to 1 with S-ethyl derivative at C-5a. It was revealed by MS/MS spectra, which showed a fragment ion peak at m/z of 1093 [M−H−EtSH]−, indicating the elimination of fragment S-ethyl from the deprotonated compound, thus confirmed the existence of S-ethyl group. Similar to compound 1, HR-MS/MS experiment for 2 was carried out to determine the site of oxidation due to its additional oxygen atom. The observation of S-ethyl fragment lost (m/z 1093 [M−H]−) indicated that the oxidation for the parent compound was resided at the Cys′ sulfur atom, thus revealing a sulfoxide moiety in 2. The 1H NMR spectrum was very similar to that of 1, except for the disappearance of a singlet methyl signal instead of the observation of a new triplet methyl signal at 1.37 p.p.m. (J=7.4 Hz). The 13C NMR spectrum was nearly identical with that of 1, except for slight differences in chemical shifts of carbon signals around 18–28 p.p.m. The 13C DEPT experiment and HSQC spectral data revealed that 2 has an extra methylene signal implying the replacement of ethyl derivative at C-5a in 2 instead of methyl group in 1. Therefore, compound 2 was analogous to 1 with S-ethyl derivative at C-5a as shown in Figure 2. For compound 2, the presence of two alanine residues (Ala and Ala′), two serine residues (Ser and Ser′), two 2-methyl-1-methylaminocyclopropanecarboxylic acid moieties and two 3-hydroxyquinaldic acid chromophores were revealed by the COSY and HMBC spectra. The connection of intramolecular cross-linkage (cysteine sulfoxide moiety) was established by the COSY correlations of H-5 to H-5a and H-18 to H-18a, and the HMBC correlations from H-18a to C-5a, from H-5c to C-5a and from H-5d to C-5c, and was supported by the 13C NMR chemical shifts of C-5a (68.9 p.p.m.), C-18a (50.2 p.p.m.), C-5c (28.7 p.p.m.) and C-5d (17.0 p.p.m.). Thus, the structure of 2 was determined as shown in Figure 2 based on NMR and MS spectroscopic analyses.

Compounds 1–3 were subjected to several bioassays in vitro. Their cytotoxicities against various cancer cell lines, which included HL-60, HeLa, tsFT210 and srcts-NRK were evaluated. Also, their antibacterial activities against Staphylococcus aureus 209 (a gram-positive bacteria), and Escherichia coli HO141 (a gram-negative bacteria), and antifungal activities against Candida albicans JCM1542, Aspergillus fumigatus Af293 and Magnaporthe oryzae kita-1 were tested. All tested compounds showed potent antiproliferative activities against various cancer cell lines at the submicromolar range of their IC50 values (Table 3). They also displayed moderate bactericidal activities, but have relatively little effects against fungi (Table 4).

Table 3 In vitro cytotoxicities (IC50: μg ml−1) of compounds 1–3
Full size table
Table 4 Antimicrobial activities (IC50: μg ml−1) of compounds 1–3
Full size table

In this research, we have constructed a microbial metabolites fraction library and a novel type of spectral database named NPPlot based on PDA-LC/MS analysis, which is a 2D data plotted by physicochemical properties of metabolites for screening of novel compounds efficiently. On the basis of this screening method, we have discovered and isolated two novel quinomycin derivatives, 1 and 2, which contain a characteristic 3-hydroxyquinaldic acid chromophore and sulfoxide moiety. Their structures were determined by NMR and MS/MS analyses. Both compounds differ only on the sulfur-containing intramolecular cross-linkage, with 1 having S-methyl group and 2 having S-ethyl derivative. Although there have been several semisynthetic and natural products literatures reported the Cys and Cys′ sulfoxide analogs of echinomycin (quinomycin),25, 26 1 and 2 are the first natural products as a quinoline type quinomycin having a sulfoxide moiety. These compounds showed potent antiproliferative activities against various cancer cell lines and exhibited moderate antibacterial activity. Thus, these results demonstrated the advantages of fraction library and NPPlot for the discovery of novel metabolites efficiently and rapidly.

Experimental procedure

General experimental procedures

All solvents and reagents were of analytical grade and were purchased from commercial sources. UV spectra and optical rotations were recorded on a BECKMAN DU 530 Life Science UV/Vis spectrophotometer (Brea, CA, USA) and a HORIBA SEPA-300 high sensitive polarimeter (HORIBA, Kyoto, Japan), respectively. IR spectra were recorded on a HORIBA FT-720 IR spectrometer with a DuraSampl IR II ATR instrument. NMR spectra were recorded on a JEOL ECA-500 FT-NMR spectrometer at 500 MHz for 1H NMR and 125 MHz for 13C NMR. Chemical shifts were reported in p.p.m. and referenced against the residual undeuterated solvent. Mass spectra were obtained on an AB Sciex Qtrap (ESIMS) and HRESIMS was accomplished on a Waters Synapt GII. PDA-LC/MS analysis was performed using a Waters Alliance 2965 HPLC system, attached to a Waters 2996 PDA detector, with a Waters Xterra C18-column (5 μm, 2.1 mm i.d. × 150 mm) that was connected to an AB Sciex Qtrap MS/MS system equipped with an ESI probe. Middle-pressure liquid chromatography was accomplished using a Teledyne ISCO CombiFlash Companion. Preparative HPLC was performed using a Waters 600E pump system with Senshu Pak Pegasil ODS column (5 μm, 20 mm i.d. × 250 mm).

Culture condition

Streptomyces sp. RK88-1355 was cultured in a 500 ml of cylindrical flask (K1 flask) containing 70 ml of culture medium (glucose 1%, soluble starch 2%, soybean meal 1.5%, malt extract 0.5%, vegetable extract 10%, potato dextrose 2.5%, KH2PO4 0.05% and MgSO4.7H2O 0.05%) for 96 h at 28 °C on a rotary shaker with agitation of 150 r.p.m. 140 ml of each preculture was used to inoculate two of 30-l jar fermentors that contained 15 l of the same culture medium, which were cultured with stirring speed at 100 r.p.m. and an aeration rate of 10 l min−1 for 4 days.

Construction of fraction library

The fraction library of microbial metabolites was constructed from 36.08 g of ethyl acetate extract, which was prepared from 30 l of culture broth. The methodology of fraction library construction was followed as in the previously described method.9

Construction of NPPlot

Each fraction was analyzed by PDA-LC/MS with an acetonitrile/0.05% formic acid aqueous gradient system (acetonitrile: 5–100% in 30 min, and hold for 15 min). The retention time of HPLC recorded on UV chromatogram and m/z values of each metabolite within the fraction were used to generate a spectral database, namely NPPlot. In NPPlot, each metabolite was plotted on a 2D area by retention time of HPLC on x axis and m/z values on y axis.9

Isolation of 1–3

The 33rd fraction of the third middle-pressure liquid chromatography fraction was purified by C18-HPLC at flow rate of 9 ml min−1 with acetonitrile/water (50:50) under isocratic elution to afford compounds 1 (8.8 mg) and 2 (1.5 mg) as pale yellow amorphous solids. The 37th fraction of the seventh middle-pressure liquid chromatography fraction was applied to the same C18-HPLC system and eluted with acetonitrile/water (70:30) in isocratic fashion to yield 5.7 mg of compound 3. The physicochemical properties of compounds 1 and 2 were summarized in Table 1. 1H and 13C NMR chemical shifts of 1 and 2 in chloroform-d were summarized in Tables 2 and 5. For compound 3, it was found to be identical with UK-63,598.20 The physicochemical properties of compound 3 were as followed: yellow amorphous solid; [α]58925 −120° (c 0.04, MeOH); UV (MeOH) λmax (log ɛ) 218 (4.79), 231 (4.79), 299 (4.00), 357 (3.96); IR vmax (ATR) cm−1 3370, 2930, 1735, 1655, 1510; ESIMS m/z 1127 [M+H]+; HRESIMS found m/z 1127.3949 [M+H]+ calcd for C53H63N10O14S2 1127.3967; 13C NMR, δC: 11.8 (C-37), 11.9 (C-29), 15.3 (C-5c), 16.7 (C-40), 18.0 (C-32), 24.6 (C-36), 25.1 (C-28), 26.0 (C-27), 26.2 (C-18a and C-35, 2C), 29.8 (C-39), 31.8 (C-31), 35.9 (C-30), 36.6 (C-38), 45.7 (C-8), 46.2 (C-21), 47.0 (C-2), 47.1 (C-15), 50.7 (C-11), 51.4 (C-5a), 51.7 (C-24), 54.3 (C-18), 60.5 (C-5), 63.9 (C-25), 64.3 (C-12), 121.2 (C-4′), 121.3 (C-4′′), 126.8 (C-5′ and C-5′′, 2C), 127.7 (C-7′), 127.7 (C-7′′), 128.6 (C-8′ and C-8′′, 2C), 128.9 (C-6′), 129.0 (C-6′′), 132.4 (4a′), 132.4 (4a′′), 133.7 (2′), 133.7 (2′′), 141.2 (8a′), 141.3 (8a′′), 153.7 (C-3′ and C-3′′, 2C), 167.3 (C-10), 167.7 (C-23), 168.5 (C-42), 168.6 (C-34), 169.3 (C-1), 169.6 (C-14), 170.4 (C-4), 171.6 (C-17), 172.9 (C-20) and 173.2 (C-7).

Table 5 1H NMR chemical shifts for compounds 1 and 2 in CDCl3
Full size table

In vitro cytotoxicity assay

The human promyelocytic leukemia cell line HL-6027 was cultured at 37 °C in RPMI-1640 medium (Invitrogen/Life Technologies, Carlsbad, CA, USA), supplemented with 10% fetal bovine serum (FBS; Sigma-Aldrich, St Louis, MO, USA). The human cervix epidermoid carcinoma cell line HeLa was cultured at 37 °C in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen/Life Technologies), supplemented with 10% FBS. tsFT210 cells,28 the mouse temperature-sensitive cdc2 mutant cell line of the mammary carcinoma FM3A, were cultured at 32 °C in RPMI-1640 medium supplemented with 5% calf serum (CS; PAA Laboratories GmbH, Buckinghamshire, UK). srcts-NRK cells,19 the rat kidney cells infected with ts25, a T-class mutant of Rous sarcoma virus Prague strain, were cultured at a permissive temperature (32 °C) in Minimum Essential Medium (MEM; Sigma-Aldrich) supplemented with 10% CS. Each cell line was seeded into a 96-well plate (1.5 × 104 cells per well for HL-60, 4 × 103 cells per well for HeLa, 1.6 × 104 cells per well for tsFT210 and 1.0 × 104 cells per well for srcts-NRK) and then exposed to test compounds for 48 h. Following 48-h exposures to test compounds, cell proliferation was determined using a Cell Count Reagent SF (Nacalai Tesque, Kyoto, Japan) according to the manufacturer’s instructions. Briefly, following a 48-h exposure, a 1/10 volume of WST-8 solution was added to each well, and the plates were incubated at 37 °C for 1 h. Then, cell growth was measured as the absorbance at 450 nm on a microplate reader (Perkin Elmer).

Antimicrobial activity assay

In broth microdilution assay, S. aureus 209, E. coli HO141, C. albicans JCM1542, A. fumigatus Af293, and M. oryzae kita-1 were used as test strains. For S. aureus and E. coli, 100 μl of cell suspension containing 1% of precultured broth was plated into a 96-well plate. Test compounds were added to the culture medium, and the plates were incubated at 37 °C for 24 h. For C. albicans and A. fumigatus, 200 μl of inoculum suspension containing 0.1% of a 0.5 McFarland standard suspension was plated into a 96-well plate. Test compounds were added to the culture medium, and the plates were incubated at 28 °C for 24 h (C. albicans) and 48 h (A. fumigatus). For M. oryzae, 200 μl of cell suspension containing 2% of precultured broth was plated into a 96-well plate. After added with test compounds, the plates were incubated at 28 °C for 48 h. The growths of these microorganisms were measured by absorbance at 600 nm.