Neo-actinomycins A and B, natural actinomycins bearing the 5H-oxazolo[4,5-b]phenoxazine chromophore, from the marine-derived Streptomyces sp. IMB094

Neo-actinomycins A and B (1 and 2), two new natural actinomycins featuring an unprecedented tetracyclic 5H-oxazolo[4,5-b]phenoxazine chromophore, were isolated from the marine-derived actinomycete Streptomyces sp. IMB094. Their structures were elucidated by spectroscopic analyses. The presence of this ring system was proposed to originate from a condensation between actinomycin D (3) with α-ketoglutarate and pyruvate, respectively. Compound 1 showed potent cytotoxic activities against human cancer HCT116 and A549 cell lines in the nanomolar range (IC50: 38.7 and 65.8 nM, respectively) and moderate antibacterial activities against methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococci (VRE) strains.

The absolute configurations of amino acids were determined using advanced Marfey's method after acid hydrolysis 17,18 . Comparison of retention times between l-and d/l-FDLA (1-fluoro-2,4-dinitrophenyl-5-leucine amide) adducts (Supplementary Figure S2) assigned l configurations for Thr, MeVal, and Pro residues and d configuration for Val. With its structure fully resolved, 1 was established as a new member of the actinomycin family and named neo-actinomycin A.
The molecular formula of neo-actinomycin B (2) was elucidated as C 64 H 88 N 12 O 16 by HRESIMS, with one C 2 H 2 O 2 unit less than 1. The UV and NMR data (Supplementary Tables S3 and S4) of 2 were similar to those of 1. The major difference between the NMR data of 1 and 2 was the absence of signals corresponding to the carboxyethyl group in 2 that were present in 1. A new methyl signal was observed at δ H /δ C 2.63/14.4 in 2, suggesting that the carboxyethyl group in 1 was replaced by methyl in 2, which was confirmed by 2D NMR data analysis (Supplementary Figure S3). This is the first report on naturally occurring compound 2, which was previously synthesized as a derivative of actinomycin D 14 . It should be noted that prior to this study, no NMR data have been reported for this compound. Compound 2 was assigned the trivial name neo-actinomycin B since it had not been described from nature previously.
A plausible biosynthetic pathway (Fig. 3) for neo-actinomycins A and B (1 and 2) is proposed, starting from actinomycin D (3) and the tricarboxylic acid (TCA) cycle intermediates 19 , α-ketoglutarate (α-KG) and pyruvate (PA), respectively. Nucleophilic addition of the free 2-amino group of 3 to the α-keto function of α-ketoglutarate (or pyruvate), followed by a cyclization, would give rise to 1 (or 2). The cyclization process might proceed by a concerted intramolecular hydrogen shift mechanism as proposed previously 20,21 . To test the biosynthetic hypothesis we investigated the possibility of precursor-directed in situ synthesis of 1 and 2 by adding the proposed precursors α-ketoglutaric acid and pyruvic acid (1 mg/mL) after cultivation of Streptomyces sp. IMB094. LC-MS analysis indicated 12-fold increase in the production of 1 in α-ketoglutaric acid-supplemented cultures compared to unsupplemented control ( Figure S4). It is interesting to note that the yield of 1 and 2 both increased about 6-fold 24 h after pyruvic acid was added into the cultures. A possible explanation is that the exogenous pyruvic acid is converted into α-ketoglutaric acid through the in vivo tricarboxylic acid (TCA) cycle biosynthesis pathway during cultivation, but this remains to be demonstrated.
We further explored the possibility of in vitro transformation of the precursors in a variety of solvents, including the fermentation M8 media, H 2 O, and MeOH ( Figures S5 and S6). After incubation of 3 and α-ketoglutaric acid at 28 °C for 36 h, we observed approximately 10% conversion of 3 to 1 in H 2 O and in M8 media, but no production of 1 in MeOH ( Figure S5). Incubation with pyruvic acid under the same conditions lead to about 50% conversion of 3 to 2 in H 2 O and M8 media, and only 5% conversion in MeOH ( Figure S6). Further investigation revealed that the conversion rates in MeOH and H 2 O varied slightly under pH 1.0 and 2.0 conditions, but dramatically decreased under pH 4.0 (Table S6). The low conversion rate in MeOH could be explained by the strong hydrogen bond between the 2-amino group and the pentapetidolactone 14,22 . These results suggest that 1 and 2 were formed by a condensation of actinomycin D with α-ketoglutarate and pyruvate, respectively. H and 13 C NMR data were recorded at 600 and 150 MHz, respectively.
In a preliminary investigation of the biosynthetic origin of α-ketoglutarate and pyruvate, strain IMB094 was cultured in parallel in the M8 medium and in the chemically defined medium, the galactose-glutamate-mineral salts (GGM) medium, which was proved to be good for actinomycin production for S. antibioticus 3720 23 . After cultivation at 28 °C for six days, the marine-derived strain IMB094 grew considerably better in the M8 medium containing 3.0% artificial sea salts than in the GGM media. LC-MS analysis of the metabolites showed that 1 and 2 were not present in the GGM fermentation broth. Compounds 3 and 4 were produced in the cultures grown in GGM medium; but their yield rates were only 7% of those in the M8 medium ( Figure S7 and Table S7). Although α-ketoglutarate and pyruvate could not be detected in both media prior to cultivation by LC-MS analysis, they were found to be present in the extracts of both cultures grown in the M8 and GGM media ( Figure S8 and Table S7). This suggested that α-ketoglutarate and pyruvate were formed during cultivation. Since galactose and l-glutamic acid were the only two organics in the GGM medium, the presence of pyruvate in the cultures (21.4 μg/mL, even higher than in the M8 cultures) indicated that α-ketoglutarate and pyruvate in the cultures did not directly derive from the exogenous alanine and glutamate in the medium by deamination. A plausible origin for the two precursors was from the TCA cycle biosynthesis pathway of the strain. Further biosynthesis studies of neo-actinomycins A and B were being undertaken.
Compounds 1 and 2 were evaluated for their in vitro antibacterial activity against a series of Gram-positive and Gram-negative drug-resistant pathogens ( Table 2) and for cytotoxicity against two human cancer cell lines (Table 3)

Conclusion
In summary, we have isolated two new natural actinomycins, neo-actinomycins A and B (1 and 2) from a marine-derived Streptomyces sp. strain. Neo-actinomycin A represents the first reported instance of natural actinomycins possessing an unusual 5H-oxazolo[4,5-b]phenoxazine ring system. Investigation of biosynthesis indicated that neo-actinomycins A and B are formed by a condensation of α-ketoglutarate and pyruvate, with the actinocin chromophore of actinomycin D.

Methods
General Experimental Procedures. Optical rotations were determined using a Perkin-Elmer model 343 polarimeter. UV and CD spectra were recorded on an Applied Photophysics Chirascan spectropolarimeter. IR spectra were recorded on a Nicolet 5700 FT-IR microscope spectrometer (FT-IR microscope transmission). 1Dand 2D-NMR spectra were obtained at 600 MHz for 1 H and 150 MHz for 13  Isolation. The culture broth (30L) was separated into the mycelia and the supernatant by centrifugation.
The mycelial cake was extracted four times with acetone. The supernatant was subjected to an XAD-7HP macroporous adsorbent resin column (6L). The column was washed with H 2 O, and eluted successively with 50% and 90% aqueous acetone. The two latter fractions were combined with the mycelial acetone extracts and then concentrated under reduced pressure to afford a crude extract (25 g). The extract was applied to silica gel column chromatography (CC) eluting with a step-gradient of CH 2 Cl 2 −MeOH (50:1-0:100, v/v) to give 10 fractions (F 1 -F 10 ) on the basis of TLC results. The fractions F 5 (20:1, 0.36 g), F 6 (9:1, 0.54 g), and F 7 (9:1, 0.65 g) were separately applied to Sephadex LH-20 CC eluting with 90% aqueous MeOH and then separated on a preparative reversed-phase (RP) C 18        Comparison analysis of the metabolites in M8 and GGM media. Spores of strain IMB094 were inoculated into 500 mL Erlenmeyer flasks containing 100 mL of TCG medium (3 g of tryptone, 5 g of casitone, 4 g of glucose, and 30 g of artificial sea salt in 1 L of H 2 O) 26 . The cultures were grown at 28 °C for 3 days. 5 mL of seed medium was inoculated to 500 mL Erlenmeyer flasks containing 100 mL of the M8 and GGM (10 g of galactose, 2 g of l-glutamic acid, 1 g of K 2 HPO 4 , 0. Antibacterial Assay. The antibacterial assay was performed by using the agar dilution method as described previously 13 . Organisms used in this study included strains from the ATCC collection and clinical isolates. The test medium was Mueller-Hinton broth, and the inoculum was 10,000 colony forming units (CFU)/spot. The final concentrations of compounds ranged from 0.03 to 128 μg/mL. Culture plates were incubated at 35 °C for 18 h and MICs were then recorded. The MIC was defined as the lowest concentration that prevented visible growth of the bacteria.
Data Availability. All data generated or analysed during this study are included in this published article (and its Supplementary Information files).