Natural products are specialized small molecules, that typically provide distinct biological functions for the producing organism. These compounds often exhibit remarkable pharmacological activities and are thus of great importance for human health purposes. Natural products are metabolically derived from the primary metabolite pool and assembled by distinct enzyme families, generating an outstanding structural complexity. Typical natural product classes, such as terpenoids, polyketides or non-ribosomal peptides are derived from oligoprenyl diphosphates, activated C2-building blocks like malonyl-CoA, or amino acids. These primary metabolite building blocks are converted to complex carbon frameworks by designated core biosynthetic enzymes, such as terpene synthases, polyketide synthases, or non-ribosomal peptide synthetases, acting as gatekeeping enzymes that link primary with secondary metabolism¹.

In our study, we aimed to elucidate the genetic and enzymatic basis for the biosynthesis of the structurally highly unusual compounds altemicidin (1), SB-203207 (2), and SB-203208 (3) isolated from actinomycete bacteria²; the latter two exhibiting potent isoleucyl-tRNA synthetase inhibitory activities (Fig.1a). Their carbon backbones are comprised of a unique 6-azatetrahydroindane scaffold, that sparked our interested as to how nature assembles these unusual compounds. In preceding work, we identified the biosynthetic gene cluster of 1-3 by resistance gene-guided genome mining and initial studies revealed the function of encoded enzymes involved in the side chain tailoring steps, namely the generation and installation of the sulfamoylacetic acid and ß-methylphenylalanine side chains³. The seven remaining gene products (Fig.1b), presumably responsible for the generation of the 6-azatetrahydroindane scaffold, however, do not exhibit significant sequence homology to common natural product biosynthetic enzymes, raising the question of how 1-3 are biosynthesized. 

In order to identify possible primary metabolite precursors, a series of isotopic labelling experiments were conducted, which revealed significant label incorporation for L-aspartic acid (fragment A and C) and glycerol (fragment B), initially prompting us to assume an amino acid and phospho-sugar metabolic origin (Fig.1c). However, all attempts to reconstitute potential enzymatic steps in vitro with various substrates were unsuccessful, urging us to find a different experimental approach to solve the biosynthetic riddle.

The key experiment in our study was based on a simple consideration: that one of the remaining seven gene products would act as the gatekeeping enzyme of the pathway, hereby converting specific building blocks from the primary metabolite pool into the first pathway specific intermediate. Thus, we constructed single gene expression strains in the heterologous host Streptomyces lividans TK21 and subjected the culture extracts to an untargeted metabolomics analysis. Metabolic profiling under various analytical conditions, finally enabled us to identify a specific metabolite of high polarity, accumulating in the SbzP (PLP-dependent aminotransferase) expression strain. Structure elucidation of the accumulating metabolite led to the surprising identification of nucleoside 4, already harbouring the complete 6-azatetrahydroindane scaffold and indicating an unexpected nucleotide metabolic origin for 1-3 (Fig.1d).

However, the actual substrates for SbzP remained elusive. Re-evaluation of the feeding experimental data suggested that ß-nicotinamide mononucleotide (β-NMN), itself metabolically derived from L-aspartic acid and glyceraldehyde 3-phosphate⁴, might be utilized and that annulation of the cyclopentane ring is facilitated by a nucleophilic driven, stepwise (3+2)-cycloaddition⁵. We hypothesized that the required nucleophile, on the other hand, might be generated by a PLP-mediated formation of an intermediary β,γ-unsaturated quinonoid species from an aspartate-derived α-amino acid carrying a suitable Cγ-leaving group (Fig.1d)⁶. Thus, various candidate α-amino acids with Cγ-substitution were incubated with β-NMN and recombinant SbzP. However, no detectable substrate consumption was observed.

Although the structure of 4 indicated a mononucleotide metabolic origin, the experimental findings prompted us to hypothesize, that instead of ß-NMN, the dinucleotide analogs ß-nicotinamide adenine dinucleotide (ß-NAD, 5) or ß-nicotinamide adenine dinucleotide phosphate (ß-NADP), exhibiting the same reactive pyridinium moiety, might be utilized instead. Based on these considerations, we indeed detected selective conversion of ß-NAD (5) in combination with S-adenosylmethionine (6) into the 6-azatetrahydroindane dinucleotide 7. Subsequent successful reconstitution of the complete downstream biosynthetic pathway revealed the function of several novel dinucleotide specific enzymes, further decorating the dinucleotide framework and subsequently removing the adenosine diphosphoribosyl moiety to generate the 6-azatetrahydroindane product 1 (Fig.2).

These intriguing findings demonstrated for the first time, that the pivotal metabolite ß-NAD functions as a building block for natural product biosynthesis. Furthermore, the identified gatekeeping enzyme SbzP, catalyses an unprecedented PLP-mediated tandem Cα/Cγ-alkylation reaction, leading to a cyclopentane annulation at the pyridinium moiety of ß-NAD by a (3+2)-cycloaddition reaction and thus represents the first enzyme, able to specifically tailor ß-NAD. Intriguingly, functional SbzP homologs were found to be widely distributed in the bacterial kingdom and incorporated in diverse biosynthetic gene clusters.

We are excited what the future will hold for the newly discovered natural product class and believe that our findings expand our understanding of the chemical biology of ß-NAD and will provide the basis to investigate ß-NAD-derived secondary metabolism.

For the complete story please visit: https://www.nature.com/articles/s41586-021-04214-7

For more behind the scenes, you can follow Lena here: https://twitter.com/barra_lena

Reference

1. Walsh, C. T. & Tang, Y. Natural Product Biosynthesis: Chemical Logic and Enzymatic Machinery (RSC, 2017).
2. Awakawa, T., Barra, L. & Abe, I. Biosynthesis of sulfonamide and sulfamate antibiotics in actinomycete. J. Ind. Microbiol. Biotechnol. 48, 1–8 (2021).
3. Hu, Z., Awakawa, T., Ma, Z. & Abe, I. Aminoacyl sulfonamide assembly in SB-203208 biosynthesis. Nat. Commun. 10, 184 (2019).
4. Walsh, C. T. & Tang, Y. The Chemical Biology of Human Vitamins (RSC, 2019).
5. Bull, J. A., Mousseau, J. J., Pelletier, G. & Charette, A. B. Synthesis of pyridine and dihydropyridine derivatives by regio- and stereoselective addition to n-activated pyridines. Chem. Rev. 112, 2642–2713 (2012).
6. Du, Y. L. & Ryan, K. S. Pyridoxal phosphate-dependent reactions in the biosynthesis of natural products. Nat. Prod. Rep. 36, 430–457 (2019).