Biochemical and Structural Characterization of the Starter Module in the Saxitoxin Biosynthesis Pathway

Abstract

Natural products harbor great therapeutic potential and represent nearly 50% of FDA approved drugs. However, because of their structural diversity and complexity, many natural products cannot be chemically synthesized efficiently. Natural product biosynthesis pathways, especially the polyketide synthases (PKS) that condense acyl groups to form polyketides, are seen as powerful, chemoenzymatic tools to synthesize chemically challenging compounds. To harness the chemoenzymatic potential of PKS pathways, it is important to investigate their biochemical mechanisms and structural details. Saxitoxin interacts with the voltage-gated sodium channels and is one of the most potent neurotoxic alkaloids known. Saxitoxin derivatives have potential for development as selective neuropathic pain treatments. Through X-ray crystallography, biochemical assays, and mutagenesis, we further study the enzymes involved in the starter module of the saxitoxin biosynthesis pathway and expand their enzymatic capabilities. Three catalytic domains are found in the saxitoxin starter module: C-methyltransferase (C-MT), decarboxylase (DC) and 8-amino-7-oxonanoate synthase (AONS). The SAM- and metal-dependent SxtA MT initiates the biosynthesis. Structural and biochemical studies revealed essential catalytic residues and features that control the methylation extent in SxtA MT. Through a few simple amino acid substitutions, this monomethylating SxtA MT is converted into a dimethyltransferase. Interestingly, the two downstream enzymes, DC and AONS can process this MT product with an additional methyl group (dimethylmalonyl-ACP) to yield a methylated version of the saxitoxin precursor. This may be used to alter the chemical outcome of the saxitoxin pathway via domain modification, potentially resulting in the development of saxitoxin analogs with reduced toxicity and high selectivity as a pain medicine. SxtA DC was thought to be a gatekeeper because it does not act on MT methylation substrate (malonyl-ACP) and decarboxylates only the methylation product (methylmalonyl-ACP). Further investigation of SxtA DC demonstrates that it may not evolve a mechanism to filter another chemical group (such as dimethylmalonyl-ACP) that is uncommon in its biosynthesis environment. Substrate modeling with the structures of SxtA, CurA and GphF DCs from the saxitoxin, curacin, and gephyronic acid biosynthesis pathways, respectively, reveals that the size of the malonyl-binding pocket determines their substrate preference. A DC with a large pocket would prefer a bulkier malonyl-substrate and vice versa. Structural analysis of the final catalytic domain, AONS, was initiated with an AlphaFold prediction. The AONS performs condensation between an arginine and an acyl group to yield a propionylated arginine-like SxtA precursor. This structure and the AONS’s behavior in solution have suggested that AONS functions as a dimer. Characterization of MT, DC and AONS advances the understanding of the biosynthesis of chemically diverse natural products and provides valuable insights towards harnessing the biocatalytic potential of enzymes in PKS pathways.