Formation of the Pyridazine Natural Product Azamerone Involves the Biosynthetic Rearrangement of an Aryl Diazoketone

All in the family

A variety of feeding experiments with ¹³C and ¹⁵N-labeled molecules established the phthalazinone core of azamerone is derived from the diazo chlorinated meroterpenoid SF2415A3 (see scheme). We propose that an oxidative rearrangement of the aryl diazoketone followed by rearomatization with the dinitrogen group occurs during the biotransformation. This unique biochemistry extends our limited knowledge of the biosynthesis of natural products containing N–N bonds.

Compounds containing N–N bonds have been made synthetically for over a hundred years. However, it wasn’t until 1951 that the first naturally occurring compound containing dinitrogen, the azoxy containing toxin macrozamin, was reported.^[1] Since that time, natural products containing hydrazine, nitrosamine, azoxy, diazo and other N–N bonded functional groups have been identified. The enzymes responsible for N–N bond formation could make valuable biocatalysts, but unfortunately, the biosynthesis of this rare structural unit is poorly understood.

Biosynthetic investigations of the azoxy containing antibiotic valanimycin (1)^[2] and the antifungal antibiotic pyridazomycin (2)^[3a] (Figure 1) have shown that both of these natural products are derived from the condensation of amino acids. Pyridazomycin^[3b] was the only natural product containing a pyridazine ring until the phthalazinone meroterpenoid azamerone (3) was isolated from the marine sediment-derived bacterium Streptomyces sp. CNQ-766.^[4] It was first speculated that the pyridazine ring moiety in 3 was also derived from the cyclization of amino acid residues, but identification of co-produced chlorinated meroterpenoids such as A80915C (4), 3″-hydroxy-7-methylnapyradiomycin B2 (5) and analogs possessing a diazo functional group (6–8)^[5] suggest that 3 may rather be derived from a naphthoquinone precursor (Scheme 1). Herein, we report the biosynthesis of 3 from a series of stable isotope tracer experiments, which validate its biosynthetic relationship to the napyradiomycin family of meroterpenoid antibiotics.

Figure 1.

Selected natural products containing a N–N bond.

Scheme 1.

Proposed biosynthetic pathway for the diazo chlorinated meroterpenoids 6–8^[8] and their relation to the chlorinated dihydroquinones such as 4 and 5. A summation of the two [1,2-¹³C₂]acetate labeling patterns are shown in black (bold lines and dots signify double and single enrichments, respectively). Enrichment from L-[methyl-¹³C]methionine is shown as a dashed line, and enrichment from [¹⁵N]nitrite or [¹⁵N]nitrate is shown in bold font.

The original isolation of 3 from S. sp. CNQ-766 generated only ~0.25 mg/L,^[4] which was insufficient for labeling experiments. Different fermentation conditions were thus evaluated to improve production yields that resulted in >30-fold increase in 3 at 8.5 mg/L with M1 medium.^[5a] In addition to exploring the biosynthesis of 3 with stable isotopes, we also examined 4 and the diazo meroterpenoids SF2415A3 (6) and A80915D (7) in order to probe the biosynthetic relationship of this extended structural family. Feeding experiments with [1,2-¹³C₂]acetate clearly revealed that the naphthoquinone core of 4 was derived from the symmetrical pentaketide 1,3,6,8-tetrahydroxynaphthalene (THN) and that the two isoprenoid units originated via the mevalonate pathway (Scheme 1).^[6] These observations were consistent with previous isotopic experiments in this compound series from Chainia rubra MG802-AF1^[7] and from the analysis of the 43-kbp napyradiomycin biosynthetic gene cluster (nap) from Streptomyces aculeolatus NRRL 18422 and Streptomyces sp. CNQ-525.^[8]

¹³C-NMR analysis of [1,2-¹³C₂]acetate-enriched 3 also showed that all carbons except for the C8 methyl group originate from acetate (Figure 2).^[6] The bicyclic phthalazinone unit clearly harbored two distinct ¹³C-labeling patterns of equal proportions, thereby implying that it too derives from the symmetrical THN unit common to the napyradiomycins. Further analysis of L-[methyl-¹³C]methionine-enriched 3 revealed the 26% selective enrichment of C8, confirming that the acetyl methyl group of 3 is methionine-derived.^[6] Importantly, it was observed that if labeled materials were administered to S. sp. CNQ-766 during the initial production of 3, which is roughly five days after inoculation during stationary growth, no enrichment could be detected. Labeled materials had to be added at the time of inoculation in order for them to be efficiently incorporated, thereby suggesting that 3 originates from a methylated intermediate produced earlier in the fermentation process. Altogether these data suggest that 3 is biosynthetically linked to the napyradiomycin family of natural products.

Figure 2.

¹³C-labeling of azamerone (3) derived from [1,2-¹³C₂]acetate and L-[methyl-¹³C]methionine. Two equally independent labeling patterns 3a and 3b were evident from the incorporation of [1,2-¹³C₂]acetate (bold lines and dots signify double and single enrichments, respectively). The enrichment from L-[methyl-¹³C]methionine is shown as a dashed line. The numbering scheme for azamerone has been adopted from Ref. 5c and 8.

We turned our attention to the biosynthetic origin of the diazo functional group as seen in 6–8 and its relation to 3. Previous studies on the biosynthesis of the kinamycins with synthetic and natural intermediates revealed that its diazonium group is assembled from the step-wise addition of two nitrogen donors to a ketobenzo[b]fluorene precursor through an aminobenzo[b]fluorene intermediate into kinamycin D (9) (Figure 1).^[9] Considering a similar scenario, we administered different forms of ¹⁵N-enriched substrates, including ammonium sulfate, nitrite, and nitrate, and measured their incorporation into 3, 6 and 7 by MS and NMR (Table 1).^[6] MS analysis revealed that all ¹⁵N-labeled precursors were incorporated to different extents with the oxidized forms having the highest assimilation rates. While [¹⁵N]ammonium sulfate singly enriched 3 and 7 between 7–11%, ¹H-¹⁵N HMBC analysis of 3 showed that both nitrogen atoms were equally labeled. This result was in stark contrast to the enrichment by [¹⁵N]nitrite and [¹⁵N]nitrate, which not only were incorporated at a much greater extent at >70% but only labeled one of the two nitrogen atoms in the natural product. In the case of 7, ¹⁵N-NMR showed selective enrichment of the distal diazo nitrogen (-31 ppm).^{[9c,10] 15}N-labeling of 3 occurred at -2 ppm, which was assigned to the C6-bound nitrogen N2 in the pyridizine ring based upon the J_H-N coupling constants in the ¹H-¹⁵N HMBC experiment.^[6]

Table 1.

¹⁵N NMR data of 3 and 7 and incorporation percentages from nitrogen labeled precursors.

	Enrichment of 3[a]		Enrichment of 7[a]
Labeled Precursors	Singly/Doubly	δ15N ppm [b]	Singly/Doubly	δ15N ppm [c]
[15NO2]Na	77/6	−2	77/−	−31
[15NO3]Na	74/3	−2	79/−	−31
[15NH4]2SO4	11/−	−2 & 13	7/−	- [d]
[15N2H4]SO4	5/0.3	−2 & 13	0/−	- [d]
[2-15N, 9-13C]-6	11/6	−2	6/4	- [d]

^[a]Enrichment was calculated by m/z^[11];

^[b]1H-¹⁵N HMBC spectra were measured in CD₃OD at 600 MHz and referenced internally to nitromethane (δ 0 ppm);

^[c]15N NMR spectra were measured in CD₃CN at 30.4 MHz and referenced externally to nitromethane (δ 0 ppm);

^[d]chemical shift(s) were not measured due to inadequate sample size and low enrichment levels.

Results from the ¹⁵N feeding studies suggested that the napyradiomycin diazo group is also formed in a step-wise manner. Oxidation of the naphthyl ring at C5 facilitates the introduction of the first nitrogen atom by a transamination reaction. The formation of the N–N bond arises from a nucleophilic attack of the aminodihydroquinone intermediate on nitrous acid (Scheme 1). The direct addition of a dinitrogen precursor was also probed with [^15N₂]hydrazine. MS analysis revealed that [¹⁵N₂]hydrazine was only incorporated singly and at a low level into 3 (Table 1). Furthermore, ¹H-¹⁵N HMBC analysis showed that both nitrogen atoms in 3 were labeled, yet unlike the [¹⁵N]ammonium-labeled product, N2 was enriched to a slightly higher extent than N1 suggesting that metabolism of hydrazine in S. sp. CNQ-766 results in non-equivalent forms of single nitrogen species.^[6]

The ¹⁵N labeling studies indirectly supported that the diazo group such as in 6 and 7 is a biosynthetic precursor to the pyridazine unit in 3 based on their similar enrichment characteristics. In order to unequivocally link the aryl diazoketones to 3, we biosynthetically prepared [2-¹⁵N, 9-¹³C]6 from [¹⁵N]nitrite and L-[methyl-¹³C]methionine and fed the labeled natural product back to S. sp. CNQ-766. HPLC-MS inspection of the other napyradiomycin derivatives indicated that 6 was converted into 7 and 3 with retention of both isotopes in the same molar ratio as in the substrate. Isolation and NMR characterization of [2-¹⁵N, 8-¹³C]6-enriched 3 showed ¹³C enrichment at C8 and ¹⁵N enrichment at N2,^[6] thereby confirming that not only is 6 a precursor of 3, but importantly that the diazo N2 atom is biosynthetically equivalent with N2 of the pyridazine group in 3.

These stable isotope experiments paint a picture uniting 3 with the napyradiomycin family of chlorinated meroterpenoids in which a Baeyer–Villiger-type oxidation of a diazonaphthoquinone such as 7 may initiate the biosynthetic interconversion (Scheme 2). Hydrolysis of the 7-membered heterocylic intermediate would open the ring to facilitate (1) the 1,2-alkyl shift of the monoterpene subunit and (2) the assembly of the pyridazine ring in which the diazo group forms a new linkage with the C8 diketide. Aromaticity of the ring would be restored by subsequent decarboxylation and dehydration reactions.

Scheme 2.

Proposed oxidative rearrangement of the aryl diazoketone SF2415A3 (6) via A80915D (7) to azamerone (3) in S. sp. CNQ-766. Isotopically labeled C and N atoms from [2-¹⁵N, 9-¹³C]6 are shown in bold font.

The molecular basis for the formation of the diazo group and its rearrangement, however, is presently unresolved even though the napyradiomycin biosynthetic gene cluster from the related Streptomyces sp. CNQ-525 has been cloned and sequenced.^[8] Further examination of this strain revealed that like CNQ-766, it too synthesizes 3. However inspection of the nap gene cluster did not reveal an obvious mechanism for diazo synthesis and rearrangement nor did its heterologous expression yield nitrogenated napyradiomycin analogs. The only gene that may encode a diazo biosynthetic enzyme is napB3, which codes for a putative aminotransferase and may facilitate the transamination to the aminodihydroquinone intermediate. The nap cluster furthermore harbors a number of oxygenases whose functions have not yet been clarified that may be involved in the nitrogen biochemistry of 3. It is quite likely though that some of the encoding genes are extraneous to the nap locus and may reside elsewhere in the S. sp. CNQ-525 genome given the expression of the cluster only yielded non-nitrogenated napyradiomycins.^[8]

In conclusion, we established with ¹³C- and ¹⁵N-labeled precursors that azamerone (3) is biosynthesized via the nap pathway through SF2415A3 (6) in which the aryl diazoketone undergoes a novel rearrangement wherein the aromatic ring is oxidatively cleaved to allow for its rearromatization with dinitrogen to give the unique phtalazinone core of 3. This unprecedented biochemistry extends our limited knowledge of the biosynthesis of natural products containing N–N bonds and opens the door to exploring and exploiting its molecular basis at the biochemical and genetic levels.