Tailoring complex natural products by manipulating the biosynthetic machinery that builds them

Natural products are often referred to as 'privileged' structures in that they are predisposed for protein binding or other biological activity. This high propensity for bioactivity makes them an important source of drug scaffolds in modern drug design. Nonribosomal peptides represent an important class of secondary metabolite natural products and are biosynthesised by huge assembly line enzymes known as nonribosomal peptide synthetases. These synthetases are capable of producing extremely complex products, containing not just a huge range of natural and non-proteinogenic amino acids, but a large number of other modifications such as halogenations and glycosylations which would be practically unattainable from a purely synthetic standpoint. The genes responsible for these elaborate systems tend to be clustered together in the organisms that produce them. This gene organisation, coupled with the modular nature of the nonribosomal peptide synthetases, has led to the conclusion that it would be possible to gain access to this biochemical space through manipulation of the genes themselves. It is conceivable to envisage the redesign of such systems in order to carry out complex biotransformations and produce altered or 'unnatural' natural products which may have optimised or novel activity. In this work, natural product biosynthesis has been manipulated in three distinct ways: Firstly, it is shown how the backbone of a nonribosomal peptide antibiotic can be manipulated; in calcium dependent antibiotic biosynthesis, modification of the adenylation domain responsible for the incorporation of the polar 3-methyl-glutamate leads to the incorporation of the non-polar 3-methyl-glutamine, an unnatural non-proteinogenic amino acid not observed in nature, into the peptide backbone itself, and the new antibiotic variant is fully characterised by MS/MS (tandem mass spectrometry) and NMR. Secondly an attempt has been made to realise the biochemical potential of flavin-dependent halogenases as tools for aromatic halogenation of small molecules, and in this work methods of analysing the products of such enzymes have been investigated. A high pressure liquid chromatography coupled mass spectrometry (HPLC-MS) screen has been developed allowing the analysis of 120 assays a day. Additionally, a colourimetric assay is designed and tested. Finally the biosyntheses of the ramoplanins and enduracidins are investigated. In particular, this work concerns several tailoring enzymes; the action and portability of the aromatic halogenases dbv10 and clo-hal are tested by integrating them onto the chromosome of the enduracidin producing organism, Streptomyces fungicidicus ATCC 31731, in order to affect novel chlorination patterns in enduracidin, without success. Steps are made towards the chemoenzymatic synthesis of an amino acid bound to a peptidyl carrier protein (PCP), in order to generate potential substrates for halogenases that putatively act upon these systems, such as Ram20, the halogenase involved in ramoplanin biosynthesis. Novel enduracidin analogues were generated in vivo after the integration of a putative mannosyl transferase gene, ram29, taken from Actinoplanes ATCC 33076, onto the chromosome of the enduracidin producing organism, Streptomyces fungicidicus ATCC 31731. The new analogues were characterised by MS/MS.