Discovery and characterization of biosynthetic enzymes for rare metallophore ligands

Abstract

Microorganisms are a rich source of chemical inspiration. Bacteria encode enzymes capable of performing challenging chemical reactions using mechanistic logic distinct from, or even inaccessible to, synthetic chemists. The discovery and characterization of biosynthetic enzymes from bacteria presents a ripe opportunity for unearthing chemical fundamentals employed in the natural world for the construction of architecturally complex molecules. Enzymatic heteroatom–heteroatom bond formation has been a particularly productive area of recent study, with the discovery of numerous novel biosynthetic enzymes in the last two decades. One class of natural products that has increasingly been found to contain functional groups with heteroatom–heteroatom bonds are metallophores, or small molecules produced for the acquisition and sequestration of biologically-relevant metal ions. However, there is still a lack of biochemical information as to how many of these metal-binding functional groups are forged, including N-hydroxy- and N-nitroso-containing ligands that are emerging as prevalent metal chelators. This dissertation describes the bioinformatic identification of an unexplored family of putative N-nitrosating enzymes as well as the biochemical characterization of a novel heme-dependent guanidine N-oxygenase and two unusual flavin-dependent acyl-CoA dehydrogenases (ACADs), each involved in the biosynthesis of a CuII-binding metallophore. Through these investigations, we have revealed new enzymatic logic for the construction of rare metallophore ligands, providing greater insight into the chemical strategies employed by nature for synthesis, and these findings can be further leveraged for new enzyme and metallophore discovery. Chapter 2 describes our bioinformatic exploration of the family of SznF-like N-nitrosating enzymes. We identified 426 unique proteins with homology to SznF, the first dedicated N-nitrosating enzyme to be characterized, encoded in bacterial genomes. Leveraging our mechanistic understanding of the metal-binding residues required by the multidomain enzyme SznF to perform both N-oxygenation and N-nitrosation, we utilized Sequence Similarity Networks and multiple sequence alignments to deduce which of the identified enzymes could perform either or both transformations. After examining the genomic context of sznF-like genes, we selected a biosynthetic gene cluster encoding the putative N-nitrosating enzyme from Streptomyces anulatus ATCC 11523 for further study. Our efforts characterizing the product of the biosynthetic gene cluster (chm) from S. anulatus ATCC 11523 as the CuII-binding metallophore chalkophomycin are described in Chapter 3. An analysis of gene annotations for the core and accessory biosynthetic genes led to the generation of a predicted metabolite scaffold containing a hydroxyaromatic, heterocyclic, and unknown N-nitroso-containing functional group, suggestive of a metallophore. Metabolomic analysis of native chm-encoding bacteria as well as an engineered heterologous expression strain allowed us to confirm that the chm gene cluster is responsible for chalkophomycin production. Furthermore, stable isotope feeding experiments and LC–MS/MS-based structural characterization confirmed the biosynthetic precursors for chalkophomycin’s rare metal-binding ligands, N-hydroxypyrrole and a diazeniumdiolate. The N-hydroxypyrrole functional group has been known in nature since the 1990s; however, the enzymes employed for its construction have never been biochemically characterized. In Chapter 4, we describe the first in vitro reconstitution of enzymatic N-hydroxypyrrole synthesis. In doing so, we discovered two flavin-dependent enzymes that convert carrier protein-bound L-proline to N-hydroxypyrrole, reminiscent of characterized pathways for functionalized pyrrole formation. However, these two enzymes, ChmG and ChmH, share low sequence homology to both known N-oxygenases and proline dehydrogenases, and they synthesize the N-hydroxypyrrole through an N-hydroxyprolyl intermediate. Characterization of these enzymes revealed a new biosynthetic login for functionalized pyrrole synthesis, in which derivatization of the heterocyclic core occurs prior to dehydrogenation, and a new class of N-oxygenases that act on proline and on protein-bound substrates. We also characterize the hydrolase ChmF, adenylase ChmK, and nonribosomal peptide synthetase ChmL, which are responsible for mobilization of N-hydroxypyrrole so it may be incorporated into the final natural product. Collectively, we reconstituted the entire biosynthetic pathway of N-hydroxypyrrole biosynthesis, in turn assigning function for 7 Chm enzymes. Our efforts to biochemically characterize diazeniumdiolate biosynthesis are described in Chapter 5. SznF-like enzymes have been implicated in diazeniumdiolate biosynthesis; however, the discrete transformations and enzymes that catalyze them have not been characterized. We discovered that ChmN, a heme-dependent enzyme, performs two hydroxylations of L-arginine to afford L-Nδ,Nω-dihydroxyarginine. As the third biochemically characterized member of an emerging metalloenzyme family, the discovery of its ability to perform two hydroxylation reactions is a new activity for this enzyme class. Efforts towards reconstituting N-nitrosation by the enzyme ChmM are discussed; however, all attempts have been unsuccessful thus far. We propose involvement of the NRPS ChmO for diazeniumdiolate biosynthesis, and work to purify this enzyme for in vitro studies is described.