Furan and Lactam Jadomycin Biosynthetic ... - ACS Publications

Article pubs.acs.org/jnp

Furan and Lactam Jadomycin Biosynthetic Congeners Isolated from Streptomyces venezuelae ISP5230 Cultured with Nε‑Trifluoroacetyl‑L‑lysine Stephanie M. Forget,† Andrew W. Robertson,† David P. Overy,§ Russell G. Kerr,§ and David L. Jakeman*,†,‡ †

Department of Chemistry and ‡College of Pharmacy, Dalhousie University, Halifax, Nova Scotia B3H 4R2, Canada § Department of Chemistry, University of Prince Edward Island, Charlottetown, Prince Edward Island C1A 4P3, Canada S Supporting Information *

ABSTRACT: Angucycline antibiotics are composed of a classical four-ring angularly linked polyaromatic backbone. Differential cyclization chemistry of the A- and B-rings in jadomycin biosynthesis led to the discovery of two new furan analogues, while oxidation led to a ring-opened form of the jadomycin Nε-trifluoroacetyl-L-lysine (TFAL) congener. The compounds were isolated from Streptomyces venezuelae ISP5230 cultures grown with TFAL. Biosynthetic incorporation using D-[1-13C]-glucose in cultures enabled the unambiguous assignment of the aldehyde, alcohol, and amide functionalities present in these new congeners through NMR spectroscopy. Tandem mass spectrometry analysis of cultures grown with 15Nα- or 15Nε-lysine demonstrated the incorporation of Nα exclusively into the angucycline backbone, contrasting results with ornithine [J. Am. Chem. Soc. 2015, 137, 3271]. Compounds were evaluated against antimicrobial and cancer cell panels and found to possess good activity against Gram-positive bacteria.

A

containing jadomycin (JdOct) as a result of preferential cyclization of δ-amine over the α-amine.11 The jadomycin derived from L-lysine (JdK) was hypothesized to possess either a five-membered oxazolone E-ring (JdKα) or a nine-membered ring (JdKε), arising from incorporation of the α-amine or the εamine, respectively (Figure 1B). Herein, we resolve the conundrum regarding the structure of JdK using tandem mass spectrometry and utilize Nε-trifluoroacetyl-L-lysine (TFAL) as a protected lysine to access novel additional unanticipated lysine-derived congeners.

ctinobacteria, particularly the genus Streptomyces, are prolific producers of secondary metabolites, and natural products isolated from these bacteria or their derivatives have contributed significantly to the current arsenal of drugs.1−3 While nature generates a rich diversity of complex structures, our ability to tweak the chemical functionality of bacterial metabolites for structure−activity studies or for lead optimization is typically very narrow due to the challenges associated with either the synthesis of these complex molecules or the limited scope of genetic engineering.4 Type II polyketide synthase (PKS-II)-derived natural products, such as the chemotherapeutic doxorubicin, are an important subset of these bioactive bacterial secondary metabolites.5 The jadomycins are a family of PKS-II secondary metabolites produced by Streptomyces venezuelae ISP5230 that are encoded within an inducible cryptic biosynthetic pathway.6 An exploitable feature of the jadomycins is a nonenzyme-mediated biosynthetic step in which an amino acid is incorporated into the structure, forming the E-ring (Figure 1).7 Taking a precursor-directed approach, culturing S. venezuelae in the presence of single amino acids has generated a series of jadomycin analogues with structural diversity about the E-ring (Figure 1A).8−11 However, not all amino acids are incorporated.12 Previously, S. venezuelae cultures with L-ornithine produced an eight-membered ring © 2017 American Chemical Society and American Society of Pharmacognosy

■

RESULTS AND DISCUSSION The isolation of JdK has remained elusive in our hands due to significant breakdown to L-digitoxylphenanthroviridin during purification.13 To determine the E-ring size, we analyzed culture extracts by tandem liquid chromatography mass spectrometry (LC-MS2) using 15Nα-L-lysine and 15Nε-L-lysine as selectively labeled precursors, where the fragmentation patterns in MS2 experiments were anticipated to be different depending on which amino group resided within the phenanthroviridin core. From 15Nα-L-lysine incorporation, an Received: February 21, 2017 Published: May 18, 2017 1860

DOI: 10.1021/acs.jnatprod.7b00152 J. Nat. Prod. 2017, 80, 1860−1866

Journal of Natural Products

Article

signals, arising from interconversion at the hemiaminal ether (3a) position to produce diastereomers. 1D-NOE experiments (Figure S17) identified the major diastereomer as 3aS in a 1.7:1 ratio to 3aR. The same ratio was observed in the 19F NMR spectrum, with δ −77.28 and −77.29 for the 3aS and the 3aR diastereomers, respectively (Figure 2A). Cyclization of the αamino group lysine to produce a five-membered oxazolone ring (E-ring) was confirmed by the observation of an HMBC correlation between H3a and C1.

Figure 2. 19F{1H} NMR spectra (470 MHz) in methanol-d4 of (A) jadomycin TFAL 1, a mixture of 3aS/3aR diastereomers, and (B) jadomycin lactam 2.

A second compound appeared as an orange spot by TLC. It was initially isolated in quantities insufficient for full characterization (submilligram), but possessed a 1H NMR spectrum displaying features of a jadomycin family natural product, including the characteristic aromatic doublet−triplet−doublet pattern of the H9−H11 spin system, all of the L-digitoxose signals, as well as signals consistent with the presence of the amino acid side chain. A singlet 19F NMR signal at δ −77.3 was consistent with the expected chemical shift for the trifluoroacetamide group (Figure 2B). To obtain sufficient quantities of 2 for full characterization, a 1.5 L culture supplemented with 33% isotopically labeled glucose, D-[1-13C]-glucose, was performed. D-[1-13C]-Glucose is metabolized via glycolysis, generating acetyl-CoA, the polyketide synthase substrate for the jadomycin biosynthetic pathway (Scheme S1). A more expedient purification procedure was developed using preparatory HPLC, resulting in a 10-fold improvement in isolated yield (5.9 mg L−1). 13C-Enriched signals were found at the predicted positions in the 13C NMR spectrum (Figure 3). Percent 13C-incorporation, calculated from the ratio of the proton satellite signals to the corresponding uncoupled signal, was found to be 27% for the anomeric carbon (C1″) of Ldigitoxose and 14% for the PKS-II assembled angucycline carbon C4, in close agreement with the predicted values. Spectroscopic data of 13C-labeled 2 showed signals characteristic of a jadomycin with an intact angucycline backbone (A−Drings), with evident deshielding of H4 and H6, indicating the presence of an electron-withdrawing group (EWG) associated with the A-ring. There was no doubling of signals corresponding to a diastereomeric pair observed in the 1H NMR spectrum and, consistently, no signal corresponding to the hemiaminal ether proton (H3a). 2D-HMBC correlations from 13C-labeled 2 showed a correlation between H1 of the amino acid side chain and a 13C-shift δ 165, consistent with an oxidized carbon at position 3a (Figure 4). The chemical shift

Figure 1. (A) JdOct and JdNon from previous studies. This work: (B) two proposed structures for the jadomycin incorporating L-lysine, JdKα or JdKε (not observed). (C) Jadomycin TFAL 1 and lactam 2 and furans 3 and 4.

MS2 fragment of the parent ion 566 to 307 was observed for the phenanthroviridin, whereas from 15Nε-L-lysine incorporation an MS2 fragment of 306 was observed. This provided conclusive evidence that the five-membered E-ring containing jadomycin was the natural product (Figure S1). Given that the instability of JdK likely arose from the nucleophilicity of the Nε amine group, protection of the amine was anticipated to furnish a stable analogue. TFAL was selected for the inclusion of an NMR-sensitive 19F handle. Furthermore, a large proportion of synthetically derived drugs are designed to include a fluorine substituent;14 by contrast, there are very few fluorine-containing natural products.15 S. venezuelae VS1099 was cultured in the presence of TFAL as the sole nitrogen source following standard procedures for jadomycin production.16 Isolation of compounds was accomplished by a series of chromatographic steps as detailed in the Experimental Section. Isolation of a reddish-purple compound that possessed spectroscopic and HRMS data consistent with the anticipated jadomycin TFAL 1 was achieved with an isolated yield of 4.5 mg L−1. The 1H NMR spectrum showed two distinct sets of 1861



Article

(160−165 ppm).17 These data supported oxidation of position 3a to give lactam 2. Isolation of a bright yellow solid, initially collected as a precipitate from methanol-d4, resulted in the discovery of compounds 3 and 4, isolated as a mixture. Attempts to separate 3 and 4 were unsuccessful using our preparatory HPLC methods, and as such, these compounds were characterized and subsequently assayed as a mixture. The 1H NMR revealed that the sample contained two compounds in a 2:1 ratio both possessing angucycline aromatic protons characteristic of jadomycin family natural products. The absence of a sidechain methylene functionality in the 1H and 19F resonances in the NMR spectra indicated that the amino acid moiety was not incorporated. L-Digitoxylation on the D-ring was confirmed by 1 H NMR, HMBC analysis, and HRMS. Furthermore, analysis of the NMR spectra showed that the two compounds were closely related structurally but were not diastereomers. The major compound 3 possessed a proton (H1) at δ 11.37 with a corresponding 13C-signal (C1) at δ 195.2 (Figure 5), indicating an aldehyde group. 2D correlations positioned the aldehyde group on the A-ring; as was the case for compound 2, the Aring aromatic protons of 3 were deshielded in the 1H NMR spectrum, due to the EWG effects of the aldehyde. The minor compound 4, however, did not appear to have an EWG at position 1 as the A-ring aromatic protons were between δ 7.2 and 7.4. An edited HSQC experiment identified a methylene unit, with a 13C-chemical shift of δ 65.7 (Figure 6A), consistent with an analogue of 3 with reduction of the aldehyde to an alcohol. A COSY correlation between the CH2 and a hydroxy group confirmed the presence of a primary

Figure 3. Red circles (●) indicate positions with 14% isotopic incorporation, derived from the PKS-II substrate malonyl-CoA, the blue circle (○) indicates 27% incorporation derived from glucose 1phosphate, and the pink diamond (⧫) indicates 11% incorporation, arising from shuffling the isotopic label from C1 to C6 during glycolysis and gluconeogenesis. *Solvent signal (methanol-d4).

Figure 4. Key HMBC correlations showing amide connectivity in 13Clabeled 2 (700 MHz, methanol-d4).

Figure 5. Key HMBC correlation demonstrating the presence of an aldehyde functional group in 3 (700 MHz, chloroform-d3).

Figure 6. (A) Expansion of an edited-HSQC spectrum of the yellow precipitate containing compounds 3 and 4; CH and CH3 signals are colored in blue and CH2 signals are colored in fushia. (B) Expansion of a COSY NMR spectrum of the yellow precipitate showing the hydroxyl group present at C1 in the minor compound 4 (700 MHz, chloroform-d3).

data were similar to synthetic cyclic amides of similar architecture, possessing 13C-chemical shifts in this range 1862



Article

Scheme 1. Proposed Mechanistic Origin of Jadomycins 1−4 Based on Common Biosynthetic Intermediate 6a

a

It is plausible that 7 is the common biosynthetic intermediate for 1−4.

observed (data not shown), and the production of compounds 3 and 4 was not observed. Compounds 1 and 2 were evaluated against a five-strain antimicrobial panel with three Gram-positive bacteria, methicillin-resistant Staphylococcus aureus (MRSA), Staphylococcus warneri, and vancomycin-resistant Entrococcus faecium (VRE), two Gram-negative bacteria, Proteus vulgaris and Pseudomonas aeruginosa, and a strain of yeast, Candida albicans. Oxazolone-ring-containing 1 was active against MRSA (minimum inhibitory concentration, MIC, 3 μg mL−1), S. warneri (3 μg mL−1), and VRE (13 μg mL−1), comparable to the activity of previously tested jadomycins against MRSA.25 In contrast, 2 was much less active (MIC ≥ 100 μgmL−1) against all three Gram-positive strains. The evaluated compounds were not active against Gram-negative bacteria or the yeast strain. Both compounds were not toxic against human fibroblast and kidney cells (MIC > 128 μgmL−1). The enhanced antibiotic activity of 1 in comparison to 2 implies that the hemiaminal ether functionality plays an important role in the antimicrobial properties of the jadomycins. Compounds 2−4 were evaluated against the National Cancer Institute’s 60 human cancer cell line panel, but did not show sufficient activity from the initial one-dose screen to warrant further evaluation. In this study, the isolation and characterization of three novel jadomycin congeners differing in the functionality of the E-ring are described. This demonstrates that the nonenzymatic process to form the jadomycin oxazolone E-ring is amenable to different functional group manipulations. To the best of our knowledge, these are the first reports of amide- and furancontaining jadomycins and demonstrate the surprising plasticity of the jadomycin biosynthetic machinery to deliver unanticipated derivatives. Furthermore, we demonstrate using MS the presence of a five-membered E-ring size in the jadomycin derived from L-lysine, which contrasts with the congeners isolated from L-ornithine, and utilize TFAL as a means to deliver and observe the L-lysine-derived congener.

alcohol (Figure 6B). With the absence of an amino acid side chain in compounds 3 and 4, a furan B-ring was proposed as a structure consistent with both the NMR and HRMS data. Jadomycin biosynthesis has been outlined through the use of blocked mutants and by sequence homology.18−21 We propose that 2, 3, and 4 are derived from characterized intermediate 5 (Scheme 1).22 In the generally accepted mechanism, the B-ring of 5 is enzymatically cleaved to yield reactive intermediate 6;23 subsequently, the spontaneous reaction of aldehyde 6 with an amino acid generates the proposed intermediate 7, which forms either 1 or 2 through competing pathways. Remarkably, the isolated yields of 1 and 2 were comparable, although a precise explanation for the production of lactam 2 as a major product when cultured with TFAL remains speculative. Alternately, 2 may be a product of the oxidation of 1, although our attempts to replicate this oxidation using H2O2 did not produce 2 (data not shown); enzyme-mediated oxidation is also plausible. Cyclization initiated by conjugate addition of the A-ring OH to the quinone produced the furan (B-ring) present in 3 and 4. The presence of compounds 2, 3, and 4 in these cultures, but not in detectable quantities in previous studies, suggests that the protected lysine plays a role directing product formation through the imine intermediate 7, plausibly due to a conformation of 7 that limits facile E-ring cyclization to give 1, enabling the formation of 2. A proposed long-lived imine intermediate, 7, may promote the production of 3 and 4 through a conjugate addition analogous to the chemistry that produces furan intermediate 8, followed by hydrolysis of the imine to unmask the aldehyde yielding 3. Studies observing the effect of trifluoroacetate on metabolic pathways have not been reported to the best of our knowledge, while fluoroacetate, a potent pesticide, is known to interfere with biological pathways in vivo by inhibiting the citric acid cycle.24 The effect of trifluoroacetate supplementation (30 mM final concentration) in a culture with a different amino acid, D-serine, was performed to investigate the effect on cell growth and natural product production in a well-characterized system; however, no change in cell growth and no change in the natural product profile were 1863



Article

Table 1. NMR Spectroscopic Data for Jadomycin TFAL Derivatives 1 3aR/S and 2 (700 MHz, Methanol-d4) 1 3aS (major) position

δH (J in Hz)

δC, type

1 3aR (minor) HMBC

1 2 3a

67.7, CH 176.3, C 89.4, CH

5.26, bs

3b 4

130.7, C 120.5, CH

6.71, s

5 5-CH3 6

141.4, C 20.8, CH3 120.3, CH

2.34, s 6.8, s

7 7a 7b 8 8a 9

155.9, 114.2, 130.8, 183.5, 136.1, 120.4,

C C Ca C C CH

7.52, d (8.3)

10

136.1, CH

7.69, t (8.3)

11

121.0, CH

7.80, d (8.3)

12 12a 13 13a 1′ 2′ 3′ 4′ 5′ 6′ 1″ 2″ 3″ 4″

155.5, CH 136.7 Ca 185.6, C 155.7, Ca 29.8, CH2 25.05, CH2 29.75, CH2 40.6, CH2 158.8, C CF3b 96.5, CH 36.2, CH2 68.2, CH 73.9, CH

2.36, 1.41, 1.61, 3.34,

2.14, m obs obs t (7.5)

2, 2′ 1′, 3′, 4′ 1′, 2′, 4′ 2′, 3′, 5′

5.93, 2.43, 4.07, 3.27,

bm dd (15.2,2.2); 2.21, dt (15.2, 3.5) m dd (9.9, 3.2)

5″

66.1, CH

3.95, m

5″-CH3 2-OH

18.2, CH3

1.21, d (6.2)

3″, 12 3″, 4″, 1″ 1″, 5″ 5″, 5″CH3 4″, 5″CH3 1″, 4″, 5″

5.67, s

1, 1′, 3b, 4, 7, 7a 3a, 3b, 5CH3, 6, 7

4, 5-CH3, 7, 7a

8, 8a, 10, 12 8a, 9, 11, 12 8a, 10, 12, 13

δC, type 64.4 177.4 85.8

2

δH (J in Hz) 5.37, bs 5.72, s

δC, type 64.0, CH 176.1, C 165.0, C

δH (J in Hz)

HMBC

5.57, bs

2, 3a, 7

135.6, Ca 121.5, CH

7.83, obs

3a, 6, 7, 8

130.6 120.1

6.76, s

141.4 20.8 120.5

2.36, s 6.83, s

142.6, C 21.1, CH3 124.9, CHc

2.44, s 7.15, d (1.5)

4, 5, 6, 7a 4, 5-CH3, 7, 7a

154.5 114.3 130.6a 183.9 136.2 121.2

7.52, obs

153.0, 118.7, 130.0, 187.1, 135.5, 121.5,

7.83, obs

135.7

7.66, t (8.0)

135.6, CH

7.71, t (8.3)

121.2

7.8, obs

121.2, CHc

7.6, d (8.3)

8, 8a, 10, 11 8a, 9, 11, 12 8a, 9, 10, 12, 13

155.5 136.5a 183.9 155.5 29.5 25.25 29.5 40.5 158.8 CF3 96.2 33.5 68.5 73.7

1.96, 2.24, 1.55, 1.58, 3.31,

2.24 obs obs obs obs

5.93, 2.55, 4.11, 3.31,

bm obs; 2.35, obs m obs

c

C Cc Ca Cc C CHc

153.0, CH 122.3, Ca,c 182.3, C 118.7, Ca,c 35.4, CH2 26.7, CH2 29.7, CH2 40.5, CH2 156, C 159, CF3 97.3, CHc 36.2, CH2 68.2, CH 74.4, CH

2.74, 1.57, 1.67, 3.31,

m; 2.15, obs obs;1.72, obs m obs

2 3′, 4′ 2′, 4′ 2′, 3′, 5′

5.84, 2.67, 4.08, 3.25,

d (2.8) bd(14.8); 2.14, obs q (3.1) dd (9.6, 2.7)

4.00, dq (10.2, 6.4)

3″, 12 1″, 3″, 4″ 4″, 5″ 3″, 5″, 5″CH3 1″, 3″, 4″, 5″-CH3 1″, 4″, 5″

65.7

4.02, m

66.2, CH

18.2

1.19, d (6.1)

18.2, CH3c

1.19, d (6.2) 11.92, bsd

a

Assignment by 13C NMR only, resonances may be interchangeable. bSignal not identified. cIndicates a 13C-enriched position when cultured with D[1-13C]-glucose. dAcetone-d6 solvent.

■

5% HPLC grade CH3CN, pH 4.0) and CH3CN (D): a linear gradient from 90:10 C/D to 40:60 C/D over 8.0 min followed by a plateau at 40:60 C/D from 8.0 to 10.0 min at 1.0 mL min−1. NMR Experiments. Proton, carbon, and 2D NMR spectra were recorded at the Canadian National Research Council Institute for Marine Bioscience (NRC-IMB, Halifax, NS) using a Bruker AV-III 700 MHz spectrometer equipped with an ATMA 5 mm or 1.7 mm TCI cryoprobe. Fluorine NMR spectra and NOE experiments were acquired at the Nuclear Magnetic Resonance Research Resource (NMR-3, Dalhousie University, Halifax, NS) using a Bruker AV500 MHz spectrometer equipped with an ATMA-BBFO SmartProbe. Residual solvent signals were used as reference signals for calibration; methanol-d4 was calibrated to δ 3.31 and 49.0 and chloroform-d3 calibrated to δ 7.26 and 77.0 for 1H and 13C experiments, respectively. An external calibration to CCl3F δ 0.0 was used for 19F NMR spectra. The following abbreviations are used when describing splitting patterns in NMR spectra: singlet (s), doublet (d), triplet (t), quartet (q), multiplet (m), broad (b). Obscured (obs) may refer to overlap of

EXPERIMENTAL SECTION

General Experimental Procedures. Reagents and solvents were purchased from commercial suppliers and used without further purification. All solvents used in extraction and purification procedures were HPLC grade. Double-distilled water was used to prepare all buffers and culture media and in extractions. Visualization of pigmented compounds by TLC analysis was performed using glassbacked TLC plates (SiliCycle, 250 μM, F254 silica) and developed with an eluent system of 5:5:1 ethyl acetate/acetonitrile/water or 1:2 methanol/dichloromethane, as indicated. All normal-phase column chromatography was performed using an SPI 1 unit (Biotage), using prepacked columns (SiliCycle, 12 or 40 g). Preparatory-scale HPLC was performed using a C-18 reversed-phase column (Ultrasphere ODS, 5 μm particle size, 10 mm × 25 cm) with degassed methanol (A) and water (B) using the following method: a linear gradient from 5:95 A/B to 95:5 A/B over 30 min, followed by a hold at 95:5 A/B for 30 min at 5 mL min−1. Analytical HPLC was performed using a Hewlett-Packard Series 1050 instrument with an Agilent Zorbax 5 μM Rx-C18 column using buffer (C, 12 mM Bu4NBr, 10 mM KH2PO4, and 1864



Article

Table 2. NMR Spectroscopic Data for Jadomycin Furan Derivatives 3 and 4 (700 MHz, Chloroform-d3) 3

a

position

δC, type

1 1b 2 3 3-CH3 4 5 5a 5b 6 6a 7 8 9 10 10a 11 11a 1″ 2″ 3″ 4″ 5″ 5″-CH3 1-OH 3″-OH 4″-OH

195.2, C 132.8, C 127.8, CH 142.2, C 22.9, CH3 119.5, CH 157.8, C 123.1, C 116.4, C 180.6, C 119.9, C 123.1, CH 137.3, CH 121.7, CH 157.7, C 136.4, C 177, C Ca 95.2, CH 36.6, CH2 67, CH 73.7, CH 67.5, CH 19.0, CH3

4

δH (J in Hz) 11.37, s

δC, type

HMBC 1b, 2

8.02, s

1, 4, 3, 5a, 5-CH3

2.60, s 7.69, s

2, 3, 4 2, 3, 5, 5a, 5-CH3

7.91, d (7.5) 7.67, t (8.1) 7.55, d (8.4)

5b, 6, 6a, 8, 9 7, 9, 10, 10a 6a, 7, 8, 10, 11

5.94, 2.55, 4.19, 3.28, 3.87, 1.30,

3″, 10 1″, 3″,4″ 1″

d (2.9) obs; 2.24, obs obs obs m d (6.2)

3″, 4″

5.26, d (10.0) 3.13, d (10.4)

65.7, CH2 119.4, C 130, CH 142.3, C 23.1, CH3 113.8, CH 158.6, C 118.4, C Ca 183.2, C 119.9, C 123.2, CH 137.1, CH 121.6, CH 157.6, C 136.2, C 176.6, C Ca 95.2, CH 36.3, CH2 67, CH 73.7, CH 67.0, CH 19.0, CH3

δH (J in Hz)

HMBC

4.97, bs

1b, 2

7.25, s

1, 3, 4, 5a, 5-CH3

2.52, s 7.37, s

2, 3, 4 2, 3, 5, 5a, 5-CH3

7.83, d (7.5) 7.59, t (8.0) 7.52, d (8.5)

6a, 9 10, 10a 6a, 7, 10, 10a, 11

5.93, 2.56, 4.19, 3.29, 3.90, 1.32, 4.44, 5.31, 3.13,

3″, 10

d (3.0) obs; 2.26, obs obs obs obs d (6.2) s d (10.2) obs

3″, 4″

Signal not identified.

NMR signals either from the same molecule, from a second species, or from a solvent signal. LC-MS2 and HRMS Conditions. For ESI− and ESI+ lowresolution LCMS experiments, high-performance liquid chromatography (HPLC, Agilent 1100) was coupled to a hybrid triple quadrupole mass spectrometer (Applied Biosystems, 2000Qtrap). The HPLC was equipped with a hydrophilic interaction liquid (HILIC) phase column (Phenomenex Kinetic 2.6 μM HILIC, 150 mm × 2.1 mm); elution of 5 μL of injected sample was accomplished with an isocratic gradient (7:3 CH3CN/NH4OAc(aq) (2 mM, pH 5.5)) with a flow rate of 120 μL min−1 over 10 min. For positive mode detection, the following settings were applied: capillary voltage +4500 kV, declustering potential +80 V, curtain gas 10 (arbitrary units). For negative mode detection, the following settings were applied: capillary voltage, 1000 kV, declustering potential, 100 V, curtain gas 20 (arbitrary units). LC-MS2 data were recorded using Analyst software (1.4.1, Applied Biosystems). High-resolution MS traces were recorded using a Bruker Daltonics MicroTOF Focus mass spectrometer using an ESI source. Bacterial Cultures. Recipes for all media can be found in the SI. Streptomyces venezuelae ISP5230 VS1099 (genotype ΔjadW2::aac(3)IV)26 spores were stored in a 25% glycerol stock at −70 °C; 100 μL of a thawed aliquot was plated to MYM agar with 50 μg mL−1 apramycin and incubated at 30 °C for 1−2 weeks. A 1 cm2 lawn of S. venezuelae VS1099 from these plates was used to inoculate MYM broth and grown overnight at 30 °C with shaking (250 rpm). After 16−18 h, the cloudy culture was centrifuged at 3750 rpm, the supernatant decanted, and the cell pellet resuspended in MSM broth. This process (pelleting, decanting, and resuspension) was repeated twice in order to remove traces of nutrient-rich media. Finally, the cells were resuspended in ∼20 mL of MSM, and this cell suspension was used to subculture MSM supplemented with 9 mM phosphate, 120 mM glucose, and 60 mM TFAL to an OD600 value 0.6 ± 0.1. Autoclaving the MSM media

supplemented with TFAL resulted in partial hydrolysis of the amino acid to L-lysine and trifluoroacetate, as observed by 19F NMR (Figure S6). Cultures were inoculated in triplicate, 3 × 25 mL for small-scale productions and 8 × 250 mL or 6 × 250 mL for large-scale productions, as indicated. Ethanol was added (7 to 250 mL of media) immediately to induce jadomycin production, and the cultures were shaken at 30 °C (250 rpm) for a period of 2 days. At 16 h, the pH of the solution was readjusted to 7.5 by titration with 5 M NaOH. Growth curves were generated by withdrawing a 200 μL aliquot from triplicate small-scale cultures (25 mL) at the indicated time intervals and measuring OD600 values and A526. Cells were removed by centrifugation prior to A526 readings. These curves can be found in the SI. 13 C-Supplementation Cultures. All jadomycin production procedures were as described with 33.3% of the glucose content of the MSM media substituted with D-[1-13C]-glucose. This fermentation was performed on a 1.5 L scale. Natural Product Purification. Large-scale (1.5 or 2 L) productions, initiated as described above, were left for 30 h prior to workup, the point at which A526 values plateaued. At this time, cells were removed by centrifugation (8500 rpm). The supernatant was filtered through 0.45 and then 0.22 μM Millipore filters. The clarified culture media was passed through a preconditioned (methanol followed by H2O) silica-phenyl column (12 g, SiliCycle). The column was washed with water (1 L total) to remove water-soluble impurities. Organic material was stripped from the column by washing with 50 mL of methanol. The highly pigmented methanol extract was concentrated to ∼2 mL, and the solution diluted with ethyl acetate (500 mL). The solution was extracted with water (3 × 300 mL). Both the aqueous and organic layers retained a similar pigmented maroon color. Concentration of the aqueous layer yielded 54 mg, and the organic layer yielded 67 mg. The organic-soluble material, containing the compounds of interest as determined by TLC, was fractionated 1865



■

using normal-phase silica chromatography (40 g column) using a gradient of methanol/dichloromethane as follows: 5:95 1 CV, linear gradient increase to 50:50 over 5 CV, hold at 50:50 over 2 CV, and final stepwise increase to 100:0 for 2 CV. Two fractions of interest were obtained: A less polar fraction, fraction 1, containing 3 and 4 was eluted at 95:5 MeOH/CH2Cl2, and the more polar fraction 2 eluting at 20:80 contained compounds 1 and 2. The material from fraction 2 was resolved further by preparatory HPLC using the method outlined in the general methods. This gave 9.0 mg of 1 (from a 2 L production) and 8.8 mg of 2 (from a 1.5 L 13C-supplemented production). Complete 1H and 13C NMR chemical shift assignments are listed in Table 1. Jadomycin TFAL (1): deep plum colored, amorphous solid; 19F{1H} NMR (methanol-d4, 470 Hz) δ −77.28 (s, 1 3aS), −77.29 (s, 1 3aR) ppm; HRMS-ESI+ m/z 683.1808 (calcd for C32H31F3N2NaO10, 683.1823 found (error 2.13 ppm)); TLC Rf 0.7 (5:5:1 EtOAC/ CH3CN/H2O). Jadomycin TFAL lactam (2): deep orange colored, amorphous solid; 19F{1H} NMR (methanol-d4, 470 Hz) δ −77.3 ppm; HRMSESI+ m/z 699.1748 (calcd C32H31F3N2NaO11, 699.1772); TLC Rf 0.6 (5:5:1 EtOAC:/CH3CN/H2O). The material from fraction 1 was concentrated and taken up in 600 μL of methanol-d4. By NMR, there were multiple species present, and it was not possible to characterize this mixture. After a time, a yellow precipitate formed. The yellow precipitate was collected and washed with cold methanol to yield 1.3 mg of a mixture of 3 and 4. When the fermentation was repeated with D-[1-13C]-glucose, material from fraction 1 was instead purified using the preparatory HPLC method, to yield 0.5 mg of a less pure mixture of the same compounds. Complete 1 H and 13C NMR chemical shift assignments are listed in Table 2. Jadomcyin furan (3): bright yellow colored, amorphous solid; HRMS-ESI+ m/z 459.1050 (calcd C24H20NaO8, 459.1050); TLC Rf 0.42 (5:95 CH3OH/CH2Cl2). Jadomycin furan (4): bright yellow colored, amorphous solid; HRMS-ESI+ m/z 461.1189 (calcd C24H22NaO8, 461.1207); TLC Rf 0.36 (5:95 CH3OH/CH2Cl2). 15 N-Labeled Jadomycin K Productions and Sample Preparation for LC-MS2 Experiments. Jadomycin K production was performed as already described, with modification in the amino acid content of the MSM media, to contain 100% 15Nα-L-lysine (60 mM) or 15Nε-L-lysine (45 mM). Control fermentations with natural abundance L-lysine (60 and 45 mM) were simultaneously conducted. These productions were performed on a small scale (triplicate 25 mL cultures) following the usual procedures. After a fermentation period of 2 days, cells were filtered from the solution and samples of the clarified growth media were used without further purification for all qualitative LC-MS2 analyses. Samples prepared for MS analysis were dissolved and diluted in water.

■

ACKNOWLEDGMENTS X. Feng, Dalhousie University, is thanked for HRMS data collection. I. Burton, NRC-IMB, is thanked for assistance with NMR acquisition. S.M.F. is supported by the Killam Trusts and the CRTP trainee program funded in partnership with the Canadian Cancer Society, Nova Scotia Division. We thank NSERC and CIHR for funding.

■

REFERENCES

(1) Bérdy, J. J. Antibiot. 2012, 65, 385−395. (2) Moloney, M. G. Trends Pharmacol. Sci. 2016, 37, 689−701. (3) Butler, M. S.; Robertson, A. A. B.; Cooper, M. A. Nat. Prod. Rep. 2014, 31, 1612−1661. (4) Hillenmeyer, M. E.; Vandova, G. A.; Berlew, E. E.; Charkoudian, L. K. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 13952−13957. (5) Hertweck, C.; Luzhetskyy, A.; Rebets, Y.; Bechthold, A. Nat. Prod. Rep. 2007, 24, 162−190. (6) Ayer, S. W.; McInnes, A. G.; Thibault, P.; Wang, L.; Doull, J. L.; Parnell, T.; Vining, L. C. Tetrahedron Lett. 1991, 32, 6301−6304. (7) Rix, U.; Zheng, J.; Remsing Rix, L. L.; Greenwell, L.; Yang, K.; Rohr, J. J. Am. Chem. Soc. 2004, 126, 4496−4497. (8) Borissow, C. N.; Graham, C. L.; Syvitski, R. T.; Reid, T. R.; Blay, J.; Jakeman, D. L. ChemBioChem 2007, 8, 1198−1203. (9) Jakeman, D. L.; Farrell, S.; Young, W.; Doucet, R. J.; Timmons, S. C. Bioorg. Med. Chem. Lett. 2005, 15, 1447−1449. (10) Martinez-Farina, C. F.; Jakeman, D. L. Chem. Commun. 2015, 51, 14617−14619. (11) Robertson, A. W.; Martinez-Farina, C.; Smithen, D. A.; Yin, H.; Monro, S.; Thompson, A.; Mcfarland, S. A.; Syvitski, R. T.; Jakeman, D. L. J. Am. Chem. Soc. 2015, 137, 3271−3275. (12) Martinez-Farina, C. F.; Robertson, A. W.; Yin, H.; Monro, S. M. A.; McFarland, S. A.; Syvitski, R. T.; Jakeman, D. L. J. Nat. Prod. 2015, 78, 1208−1214. (13) Robertson, A. W.; Martinez-Farina, C. F.; Syvitski, R. T.; Jakeman, D. L. J. Nat. Prod. 2015, 78, 1942−1948. (14) Wang, J.; Sánchez-Roselló, M.; Aceña, J. L.; del Pozo, C.; Sorochinsky, A. E.; Fustero, S.; Soloshonok, V. A.; Liu, H. Chem. Rev. 2013, 114, 2432−2506. (15) Zhu, X. M.; Hackl, S.; Thaker, M. N.; Kalan, L.; Weber, C.; Urgast, D. S.; Krupp, E. M.; Brewer, A.; Vanner, S.; Szawiola, A.; Yim, G.; Feldmann, J.; Bechthold, A.; Wright, G. D.; Zechel, D. L. ChemBioChem 2015, 16, 2498−2506. (16) Jakeman, D. L.; Graham, C. L.; Young, W.; Vining, L. C. J. Ind. Microbiol. Biotechnol. 2006, 33, 767−772. (17) Khodade, V. S.; Sharath Chandra, M.; Banerjee, A.; Lahiri, S.; Pulipeta, M.; Rangarajan, R.; Chakrapani, H. ACS Med. Chem. Lett. 2014, 5, 777−781. (18) Wang, L.; White, R. L.; Vining, L. C. Microbiology 2002, 148, 1091−1103. (19) Kharel, M. K.; Pahari, P.; Shepherd, M. D.; Tibrewal, N.; Nybo, S. E.; Shaaban, K. A.; Rohr, J. Nat. Prod. Rep. 2012, 29, 264−325. (20) Kharel, M. K.; Rohr, J. Curr. Opin. Chem. Biol. 2012, 16, 150− 161. (21) Rix, U.; Wang, C. C.; Chen, Y. H.; Lipata, F. M.; Rix, L. L. R.; Greenwell, L. M.; Vining, L. C.; Yang, K. Q.; Rohr, J. ChemBioChem 2005, 6, 838−845. (22) Tibrewal, N.; Pahari, P.; Wang, G.; Kharel, M. K.; Morris, C.; Downey, T.; Hou, Y.; Bugni, T. S.; Rohr, J. J. Am. Chem. Soc. 2012, 134, 18181−18184. (23) Fan, K.; Pan, G.; Peng, X.; Zheng, J.; Gao, W.; Wang, J.; Wang, W.; Li, Y.; Yang, K. Chem. Biol. 2012, 19, 1381−1390. (24) Goncharov, N. V.; Jenkins, R. O.; Radilov, A. S. J. Appl. Toxicol. 2006, 26, 148−161. (25) Jakeman, D. L.; Bandi, S.; Graham, C. L.; Reid, T. R.; Wentzell, J. R.; Douglas, S. E. Antimicrob. Agents Chemother. 2009, 53, 1245− 1247. (26) Wang, L.; Vining, L. C. Microbiology 2003, 149, 1991−2004.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jnatprod.7b00152. Further experimental details, HPLC and TLC traces, MS2 data, and NMR spectra (PDF)

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: 1-902-494-7159. Fax: 1902-494-1396. ORCID

David L. Jakeman: 0000-0003-3002-3388 Notes

The authors declare no competing financial interest. 1866


Furan and Lactam Jadomycin Biosynthetic ... - ACS Publications

Recommend Documents