Introduction

Antimicrobial resistance (AMR) has become a major threat to the global public health especially in developing countries and it is anticipated to be responsible for 10 million death cases every year and a reduction of 2 to 3.5% in Gross Domestic Product (GDP) by 2050 [1]. Therefore, the development of new types of antibiotics against microbe-related diseases remains an urgent requirement. Among all antibiotic-producing microbes, the Streptomyces species are considered to be the most important resource and account for the production of 50ā€“55% known antibiotics [2, 3]. Angucyclines and actinomycins are two classes of important antibiotics produced by Streptomyces sp., with a broad spectrum of biological properties especially anticancer and antibacterial activities [4,5,6,7]. Angucyclines are characterized in the tetracyclic benz[a]anthracene core and most of them exist as O- or C- glycosides in nature. Actinomycins are a family of chromopeptide lactones, and as the most well-known member of them, actinomycin D has been widely used in clinic as an anti-tumor drug for the treatment of childhood rhabdomyosarcoma and Wilmsā€™ tumor [5, 6]. Meanwhile, actinomycin D also exhibited excellent activity against gram-positive bacteria [5, 6].

As a continuation of our effort to discover new antibacterial agents from actinomycetes based on a bioassay-guided strategy, the Streptomyces sp. XZHG99T from the soil sample collected from Color desert, Dengpa District, Tibet (29Ā°66ā€² N, 84Ā°55ā€² E) showed well inhibitory activity against Staphylococcus aureu. Our investigation on the crude extract led to the isolation and identification of three new angucycline-type derivatives grincamycins Lā€“N (1ā€“3), four structurally related co-metabolites including rabelomycin (4) [8, 9], moromycin B (5) [10], fridamycin D (6) [11] and saquayamycin B1 (7) [12], and two actinomycin analogs actinomycin X2 (8) [5] and actinomycin D (9) [5] (Fig.Ā 1). Grincamycin L was the first example of angucycline family having a rhodinosyl C-glycoside linkage at C-9, implying the likely participation of a specific glycosyltransferase in the biosynthetic process. Herein we reported the isolation, structure elucidation and bioactivity of compounds 1ā€“9.

Fig. 1
figure 1

Chemical structures for compounds 1ā€“9

Results and discussion

Grincamycin L (1) was isolated as a yellow powder, and the molecular formula was determined to be C31H30O10 based on the HR-ESIMS ion at m/z 561.1764 [Mā€‰āˆ’ā€‰H]ā€“(calcd for C31H29O10, 561.1766). An anthraquinone skeleton was suggested by the maximal UV absorption at 210, 265, and 471ā€‰nm, the conjugated carbonyl 13C signals at Ī“ 188.2 and 187.7, as well as the significantly downfield shifted hydroxyl 1H signals at 13.14 and 13.08 due to OH bonding [13]. Analysis of HMBC correlations (Fig.Ā 2) from H-12 to C-1, C-11, C-4a, and C-5a, H-13 to C-2, C-3, and C-4, H-4 to C-4a and C-5, and H-2 to C-1 and C-12a, along with three aromatic protons (Ī“H 7.88, d, 7.8ā€‰Hz; 7.93, d, 7.8ā€‰Hz; 8.49, s), revealed that 1 had a linear tetracyclic anthraquinone aglycone as grincamycins E [14] and G [15]. Meanwhile, two doublet methyl signals (Ī“H 1.38 & 1.40) and two anomeric methine signals (Ī“H 4.85 & 5.34; Ī“C 73.3 & 95.1) indicated the presence of two sugars in 1. The 1Hā€“1H COSY correlations across H-1ā€™ to H3-6ā€™ spin coupling fragment and HMBC correlation from H-5ā€™ to C-1ā€™ revealed a rhodinose moiety, while 1Hā€“1H COSY correlations of H-2ā€ with H-1ā€ and H-3ā€, and H-5ā€ with H3-6ā€, as well as HMBC correlations from H-2ā€ and H3-6ā€ to C-4ā€, and H-1ā€ to C-5ā€, enabled the assembly of an aculose moiety (Fig.Ā 2). The HMBC correlation from H-1ā€ to C-4ā€™ established the disaccharide fragment as aculose-(1ā€‰ā†’ā€‰4)-rhodinose, while the C-glycosyl bond between C-9 and C-1ā€™ was supported by the HMBC correlations from H-1ā€™ to C-8, C-9, and C-10. The coupling constants of J1ā€™,2ā€™ (11.1ā€‰Hz) and J1ā€,2ā€ (3.5ā€‰Hz) implied the relative configurations of the two anomeric carbons were Ī² for rhodinose and Ī± for aculose, respectively [12]. The NOESY correlations from H-1ā€² to H-5ā€² and H-1ā€³ to H-4ā€² further confirmed the relative configurations of the disaccharide part as shown (Fig.Ā 2). The configuration of C-3 was considered to be consistent with that of N05WA963C [13] on the basis of excellent NMR data comparison, which was further supported by the proposed biosynthetic relation with the known co-metabolites as described later. Although aculose and rhodinose are usual sugar moieties in anthraquinone family, rhodinosyl C-glycoside at C-9 was reported for the first time.

Fig. 2
figure 2

Key 2D NMR correlations for 1

Grincamycin M (2) was obtained as a yellow powder, and the molecular formula of C32H32O11 was deduced from HR-ESIMS analysis (m/z 593.2026 [Mā€‰+ā€‰H]+, calcd for 593.2017) and NMR data. The 1H and 13C NMR data (TableĀ 1) of 2 showed high similarity to those of 5 with the only difference attributable to the appearance of a methoxy group (Ī“H 4.04, Ī“C 56.6) in 2 instead of the aromatic proton (H-5) in 5, which was further confirmed by the remarkably downfield shifted C-5 signal (Ī“C 160.8) and the HMBC correlations from the methoxy protons and H-4 to C-5 (Fig.Ā 3). It is also worth to note that H2-4 in 5 was resolved as a singlet signal while geminal coupling was observed for H2-4 in 2, which could be caused by the change of chemical environment around H2-4 (Supplementary material, TableĀ S1). Moreover, the structure and C-3 configuration was determined by the consistent NMR data of 2 with 5 and with N05WA963A [13] which had the same aglycone as 2, as well as biogenetic consideration (Fig.Ā 4) [4].

Table 1 1H (600ā€‰MHz) and 13C (150ā€‰MHz) NMR data for compounds 1 and 2 in CDCl3
Fig. 3
figure 3

Key 2D NMR correlations for 2

Fig. 4
figure 4

Putative biosynthetic pathway for 1ā€“7

Grincamycin N (3) was obtained as a dark-red powder with the UV maximal absorption at 265 and 497ā€‰nm. The molecular formula of 3 was suggested as C31H28O10 according to the HR-ESIMS ion at m/z 559.1600 ([Mā€‰āˆ’ā€‰H]ā€“, calcd for C31H27O10, 559.1610). Detailed analysis of 1H and 13C NNR data (TableĀ 2) of 3 implied a linear C-glycosylated tetracenequinone skeleton as galtamycin B [16], while the only difference was attributable to the disaccharide side chain at C-9. As with 2, the sugar moiety of 3 was also assigned to be an Ī±-cinerulose B-(1ā€‰ā†’ā€‰4, 2ā€‰ā†’ā€‰3)-Ī²-olivose unit based on decent comparison between the NMR data of the two co-metabolites, which was further corroborated by careful examination of 2D NMR correlations (Supplementary material, Figs. S19ā€“S22). Particularly, the location of 10-OH was established by its HMBC correlations to C-9, C-10 and C-10a. Although no HMBC correlations of 5-OH were observed, its dramatic downfield chemical shift (Ī“H 15.06, brs), due to intramolecular hydrogen bonding with C-6 ketone, enabled its assignment as shown. The structure of 3 was thus characterized.

Table 2 1H (600ā€‰MHz) and 13C (150ā€‰MHz) NMR data for compound 3 in pyridine-d5

With seven angucycline analogs (1ā€“7) in hand, we were able to propose a common biogenesis for these co-metabolites (Fig.Ā 4). These angucyclines, belonging to the polyketide family, were apparently originated from the acetate biosynthetic pathway. The key intermediate involved could be the well-known UWM6 as clearly described previously [17]. A shunt product tetrangomycin from UWM6 [17] would yield moromycin (5) upon glycosylation and also produce 2 via methoxylation and glycosylation, while fridamycin D (6) could be obtained from tetrangomycin by Baeyerā€“Villiger oxidation, hydrolysis and glycosylation. Rabelomycin (4) could be easily produced from UWM6 through simple dehydration and isomerization, and a putative intermediate (I) from UWM6 [17] would afford saquayamycin B1 (7) via addition of the sugar moiety. Another important shunt intermediate (II) could be furnished from UWM6 as reported previously [18], and compounds 1 and 3 would derive from (II) via glycosylation and aromatization followed by glycosylation, respectively.

Compounds 1ā€“9 were screened against a panel of human cancer cell lines A549, H157, MCF7, MDA-MB-231, and HepG2, and all the isolates exhibited significant cytotoxicity with IC50 values ranging from 0.09ā€‰nM to 17.30ā€‰Ī¼M (TableĀ 3). Compounds 1ā€“5 and 8ā€“9 were also evaluated for their antimicrobial activity against two gram-positive strains Mycobacterium smegmatis ATCC 607 and Staphylococcus aureus ATCC 25923, two gram-negative strains Escherichia coli ATCC 8739 and Pseudomonas aeruginosa ATCC 9027 and a fungus Candida albicans ATCC10231 (6 and 7 were not tested owing to limited samples). Only compounds 4, 8, and 9 showed antibacterial activity against the two gram-positive strains M. smegmatis and S. aureus with IC50 values from 0.12 to 23.1ā€‰Ī¼M (TableĀ 3).

Table 3 Biological activity of compounds 1ā€“9 (IC50, Ī¼M)

Experimental section

General experimental procedures

NMR spectra were acquired on a Bruker Avance DRX600 NMR spectrometer (Bruker BioSpin AG, FƤllanden, Switzerland) with residual solvent peaks as references (CDCl3: Ī“H 7.26, Ī“C 77.16; pyridine-d5: Ī“H 8.71/7.55/7.18, Ī“C 149.7/135.3/123.3). Optical rotations were measured on a Rudolph VI polarimeter (Rudolph Research Analytical, Hackettstown, USA) with a 10ā€‰cm length cell. UV spectra were obtained on a Shimadzu UV-2600 spectrophotometer (Shimadzu, Kyoto, Japan) with 1ā€‰cm pathway cell. HR-ESIMS data were acquired on an Agilent 6545 Q-TOF mass spectrometer (Agilent Technologies Inc., Waldbronn, Germany). Semi-preparative HPLC separations were carried out on an Agilent 1260 series (Agilent Technologies Inc., Waldbronn, Germany) using an Agilent Zorbax SB-C18 column (250Ɨ9.4ā€‰mm, 5ā€‰Ī¼m). Column chromatography (CC) was performed on Silica gel (200ā€“300 mesh, Yantai Jiangyou Silica Gel Development Co., Yantai, China) and Sephadex LH-20 gel (GE Healthcare Bio-Sciences AB, Uppsala, Sweden). All solvents used for CC were of analytical grade (Tianjin Fuyu Fine Chemical Co. Ltd., Tianjin, China) and solvents used for HPLC were of HPLC grade (Oceanpak Alexative Chemical Ltd., Goteborg, Sweden).

Biological material

The actinomycete XZHG99T strain was isolated from a soil sample collected from Color desert, Dengpa District, Tibet Autonomous Region, China (29Ā°66ā€² N, 84Ā°55ā€² E). Comparison of the 16S rRNA sequence of XZHG99T strain with the data from GenBank database (98.42% similarity to S. albiflavescens KCTC 29196T and 98.14% similarity to S. rungchingensis KCTC 29503T), in combination with the morphological traits, revealed that it might represent a new species of the genus Streptomyces. The BLAST sequenced data had been deposited at GenBank (accession no. MG272441). The strain was deposited in CGMCC center, Institute of Microbiology, Chinese Academy of Sciences.

Fermentation and extraction

Although the Streptomyces sp. XZHG99T showed well inhibitory activity against Staphylococcus aureu, HPLC analysis of the crude extract cultured in DS medium showed poor compound diversity. Therefore, different agents were tried to enrich the metabolites from Streptomyces sp. XZHG99T, and DS medium with 3% sea salt was found to yield more metabolites with better antimicrobial activity (Figs. S26ā€“S27).

The Streptomyces sp. XZHG99T was inoculated in 500ā€‰mL Erlenmeyer flasks containing 150ā€‰mL DS medium (0.5% soluble starch, 0.03% casein, 0.2% KNO3, 0.2% K2HPO4 3H2O, 0.005% MgSO4 7H2O, 0.002% CaCO3, and 0.001% FeSO4 7H2O) at 30ā€‰Ā°C on a rotary shaker at 140ā€‰rpm for 2 days as seed cultures. Then each of the seed cultures (10ā€‰mL) was inoculated into autoclaved 500ā€‰mL Erlenmeyer flasks containing 150ā€‰mL DS medium described above but containing 3% sea salt. The flasks were incubated at 30ā€‰Ā°C for 7 days on a rotary shaker (140ā€‰rpm).

The total 40ā€‰L fermentation broth was harvested and filtered to give filtrate and mycelia. The filtrate was extracted with an equal volume EtOAc three times, while the mycelia were extracted with 3.0ā€‰L 80% acetone three times. The acetone extract was evaporated under reduced pressure to afford an aqueous solution which was then extracted with EtOAc. The two organic layers were combined and dried to give a crude gum (9.8ā€‰g).

Isolation and purification

The whole EtOAc extract (9.8ā€‰g) was fractionated by a silica gel CC eluting with step gradient CH2Cl2-MeOH (v/v 100:0 to 0:100) to give seven fractions (Fr.1ā€“Fr.7) based on TLC and HPLC analysis. Fr.3 (1.9ā€‰g) was repeatedly separated by silica gel CC with step gradient CH2Cl2-(CH3)2CO (v/v 100:0 to 0:100) and divided into six subfractions (Fr.3-1ā€“Fr.3-6). Fr.3-3 (221.4ā€‰mg) was subjected to MPLC with an ODS column eluting with step gradient MeOH-H2O (v/v 20:80 to 0:100) to give three subfractions (Fr.3-3-1ā€“Fr.3-3-3) and Fr.3-3-2 (36.8ā€‰mg) was further purified by HPLC eluting with MeOH-H2O-AcOH (v/v/v 95:5:0.05, 3.0ā€‰mLā€‰mināˆ’1) to yield 3 (tRā€‰=ā€‰13.6ā€‰min, 5.2ā€‰mg); Fr.3-4 (150.2ā€‰mg) was divided by Sephadex LH-20 CC eluting with MeOH-CH2Cl2 (v/v 1:1) and then purified by HPLC eluting with MeCN-H2O (v/v 70:30, 3.0ā€‰mLā€‰mināˆ’1) to afford 1 (tRā€‰=ā€‰27.3ā€‰min, 2.3ā€‰mg); Fr.3-5 (314.8ā€‰mg) was also fractionated by Sephadex LH-20 CC eluting with MeOH-CH2Cl2 (v/v 1:1) to obtain four subfractions (Fr.3-5-1ā€“Fr.3-5-4) and Fr.3-5-2 (56.8ā€‰mg) was purified by HPLC eluting with MeOH-H2O (v/v 77:23, 3.0ā€‰mLā€‰mināˆ’1) to yield 7 (tRā€‰=ā€‰8.3ā€‰min, 0.6ā€‰mg), 5 (tRā€‰=ā€‰11.5ā€‰min, 1.3ā€‰mg), and 2 (tRā€‰=ā€‰14.6ā€‰min, 2.9ā€‰mg), while Fr.3-5-4 (9.3ā€‰mg) was purified by HPLC eluting with MeOH-H2O-AcOH (v/v/v 68:32:0.05, 3.0ā€‰mLā€‰mināˆ’1) to yield 4 (tRā€‰=ā€‰16.6ā€‰min, 2.2ā€‰mg). Fr.5 (567.4ā€‰mg) was first subjected to Sephadex LH-20 CC eluting with MeOH and then isolated by HPLC eluting with MeOH-H2O-AcOH (v/v/v 85:15:0.05, 3.0ā€‰mLā€‰min-1) to yield 8 (tRā€‰=ā€‰11.8ā€‰min, 7.2ā€‰mg) and 9 (tRā€‰=ā€‰14.9ā€‰min, 20.2ā€‰mg). Fr.6-2 (7.5ā€‰mg) obtained from Fr.6 (189.1ā€‰mg) via Sephadex LH-20 CC eluting with MeOH was further purified by HPLC eluting with MeOH-H2O-AcOH (v/v/v 85:15:0.05, 3.0ā€‰mLā€‰mināˆ’1) to give 6 (tRā€‰=ā€‰16.3ā€‰min, 0.8ā€‰mg).

Grincamycin L (1). Yellow powder; [Ī±]21D 65.1 (c 0.1, CHCl3); UV (MeOH) Ī»max (log Īµ) 210 (3.77), 265 (3.25), 471 (2.45) nm; 1H and 13C NMR data, TableĀ 1; IR (KBr) Ī½max 3457, 2931, 1698, 1636, 1420, 1237, 1045, 797ā€‰cmāˆ’1; (ā€“)-HR-ESIMS m/z [Mā€‰āˆ’ā€‰H]ā€“ 561.1764 (calcd for C31H29O10, 561.1766).

Grincamycin M (2). Pale yellow powder; [Ī±]21D 4.5 (c 0.18, CHCl3); UV (MeOH) Ī»max (log Īµ) 223 (3.64), 274 (3.00), 390 (2.02) nm; 1H and 13C NMR data, TableĀ 1; IR (KBr) Ī½max 3448, 1698, 1627, 1556, 1429, 1362, 1278, 1107ā€‰cmāˆ’1; (ā€‰+ā€‰)-HR-ESIMS m/z 593.2026 [Mā€‰+ā€‰H]+ (calcd for C32H33O11, 593.2017).

Grincamycin N (3). Dark-red powder; [Ī±]21D 12.5 (c 0.04, MeOH); UV (MeOH) Ī»max (log Īµ) 265 (5.04), 497.0 (4.40) nm; 1H and 13C NMR data, TableĀ 2; IR (KBr) Ī½max 3448, 2931, 1733, 1614, 1575, 1435, 1390, 1260, 1097, 1027, 905ā€‰cmāˆ’1; (ā€“)-HR-ESIMS m/z 559.1600 [Mā€‰āˆ’ā€‰H]ā€“ (calcd for C31H27O10, 559.1610).

Antimicrobial assay

The antimicrobial activity of compounds 1ā€“5 and 8ā€“9 was assayed against the gram-positive strains M. smegmatis ATCC 607 and S. aureus ATCC 25923, gram-negative strains E. coli ATCC 8739 and P. aeruginosa ATCC 9027 and fungus C. albicans ATCC10231, by the two-fold serial dilution method in 96-well microplates as described previously [19]. Penicillin was used as positive control in the current assay.

Cytotoxic assay

The cytotoxicity of compounds 1ā€“9 was evaluated toward A549, H157, MCF7, MDA-MB-231, and HepG2 cell lines using the SRB method as described previously [20], and vinorelbine was used as positive control.