6-Methylsalicylic acid (6-MSA) is one of the oldest known polyketides. It is synthesised in vivo by the polyketide synthase 6-methylsalicylic acid synthase (6-MSAS), a multifunctional enzyme which uses its active sites iteratively. The stereochemistry of the hydroxyl produced from the single ketoreduction, as well as the order of dehydration, cyclisation and aromatisation steps, remain cryptic, despite extensive study. Holo 6-MSAS was heterologously expressed in E. coli and purified in two steps. A non-hydrolysable carba(dethia)malonyl-N-acetylcysteamine analogue was synthesised and used to off-load enzyme-bound intermediates from 6-MSAS. In assays with acetyl-CoA and acetoacetyl-CoA alone, diketide and triketide intermediates were off-loaded and detected by HPLC-HR-ESI-MS. In the presence of NADPH, the off-loaded triketide was reduced by the ketoreductase domain of 6-MSAS. A potential dehydrated intermediate was also observed. The dehydratase domain of 6-MSAS has recently been reassigned as a thioester hydrolase. To test this theory, the catalytic histidine residue in 6-MSAS was mutated to an alanine and the abolition of production of 6-MSA in vivo was observed. Mutated 6-MSAS was still able to produce the shunt product triacetic acid lactone. Incubation of mutated 6-MSAS with acetyl-CoA, malonyl-CoA, NADPH and carba(dethia)malonyl-N-acetylcysteamine saw only the off-loading of diketide and triketide analogues. To investigate the stereochemistry of ketoreduction in 6-MSA biosynthesis, steps were made to synthesise the resolved diastereomeric reduced-triketide CoAs which would be the substrates for the ketoreductase domain. Attempts to phosphopantetheinylate apo 6-MSAS in vitro with three different phosphopantetheinyltransferases were unsuccessful. Limited proteolysis of both holo and apo 6-MSAS found that the apo synthase rapidly lost a C-terminal fragment while holo 6-MSAS was much more stable under the same conditions. Attempts were made to express the acyl carrier protein domain from 6-MSAS to overcome these problems. These experiments represent the first use of the non-hydrolysable analogue methodology in a Type I iterative polyketide synthase and provide a framework for future experiments investigating intermediates in the biosynthesis of 6-MSA.

Heterocycles play an important role in many biological processes and are widespread among natural products. Oxazole-containing natural products possess a broad range of bioactivities and are of great interest in the pharmaceutical and agrochemical industries. Herein, the biosynthetic routes to the oxazole-containing phthoxazolins and the bis(benzoxaozle) AJI9561, were investigated. Phthoxazolins A-D are a group of oxazole trienes produced by a polyketide synthase-nonribosomal peptide synthetase (PKS-NRPS) pathway in Streptomyces sp. KO-7888 and Streptomyces sp. OM-5714. The phthoxazolin pathway was used as a model to study 5-oxazole and primary amide formation in PKS-NRPS pathways. An unusually large gene cluster for phthoxazolin biosynthesis was identified from the complete genome sequence of the producer strains and various gene deletions were performed to define the minimal gene cluster. PhoxP was proposed to encode an ATP-dependent cyclodehydratase for 5-oxazole formation on an enzyme-bound N-formylglycylacyl-intermediate, and its deletion abolished phthoxazolin production. In vitro reconstitution of the early steps of phthoxazolin biosynthesis was attempted to validate the role of PhoxP, but was unsuccessful. Furthermore, Orf3515, a putative flavin-dependent monooxygenase coded by a remote gene, was proposed to hydroxylate glycine-extended polyketide-peptide chain(s) at the α-position to yield phthoxazolins with the primary amide moiety. On the other hand, an in vitro approach was employed to establish the enzymatic logic of the biosynthesis of AJI9561, a bis(benzoxazole) antibiotic isolated from Streptomyces sp. AJ9561. The AJI9561 pathway was reconstituted using the precursors 3-hydroxyanthranilic acid and 6-methylsalicylic acid and five purified enzymes previously identified from the pathway as key enzymes for benzoxazole formation, including two adenylation enzymes for precursor activation, an acyl carrier protein (ACP), a 3-oxoacyl-ACP synthase and an amidohydrolase-like cyclase. Intermediates and shunt products isolated from enzymatic reactions containing different enzyme and precursor combinations were assessed for their competence for various steps of AJI9561 biosynthesis. Further bioinformatic analysis and in silico modelling of the amidohydrolase-like cyclase shed light on the oxazole cyclisation that represents a novel catalytic function of the amidohydrolase superfamily.