Investigations into the isolation, structure elucidation and biosynthesis of bioactive natural products

摘要

This project is divided into two parts. The first part of the project involved the investigation of fish toxins produced by Streptomyces species. Samples for fishexposure experiments were prepared from an actinomycete belonging to the S. griseus clade isolated from a site of major fish kills, as well as from the type strain, Streptomyces griseus DSM 40236. Fish-exposure samples were prepared through a series of consecutive HPLC separations. We were able to narrow down the highest level of fish toxicity to four fractions, each containing only a handful of compounds. Comparison of the metabolic profile of the most toxic fractions showed that many of the compounds were common to all of them e.g. compounds yielding ions with m/z = 213, 241, 239, 301 and 309 were present in all the fractions. Some of these compounds were purified and analysed by high resolution mass spectrometry to determine their molecular formulae. A sample preparation and purification protocol for the fish toxins has been developed in this work. It was shown that the S. griseus type strain produces ichthyotoxic metabolites in addition to the environmental strain. This is a novel and unexpected observation. The second part of the project involved structural and biosynthetic investigations of the iron-chelator and natural Angiotensin Converting Enzyme (ACE) inhibitor foroxymithine. The high structural similarity of foroxymithine to the known siderophore coelichelin and the lack of unambiguous experimental evidence in the literature to support the proposed structure of foroxymithine provided the impetus for these investigations. Foroxymithine was purified from a novel source, the culture supernatant of Streptomyces narbonensis. The gallium complex of purified foroxymithine was prepared and analysed by one- and two-dimensional high-field NMR experiments (1H, COSY, HSQC, HMBC, NOESY, TOCSY and Difference NOE), which allowed the 1H and 13C NMR signals for the complex to be assigned. Careful inspection of the 1H NMR spectrum of Ga-foroxymithine revealed two signals (a major and a minor) for several of the protons. The origin of these signals was investigated using variable temperature and 71Ga NMR experiments, and by LC-MS analyses on a homochiral stationary phase. The duplicate signals were proposed to be associated with the existence of conformational isomers of Ga-foroxymithine. The absolute stereochemistry of the four amino acid constituents of Ga-foroxymithine was determined using Marfey’s method. Authentic standards of two of the anticipated acid-hydrolysis products of Ga-foroxymithine, D- and L-N5-hydroxyornithine were chemically synthesized to facilitate the Marfey’s analysis. The results confirmed that foroxymithine contains L-N5-hydroxyornithine. Similar analysis was performed using the authentic standards of D- and L-serine and the results confirmed that foroxymithine contains L-serine. Marfey’s method was also applied to elucidate the absolute stereochemistry (previously unknown) of coelichelin. Fe-coelichelin was purified from Streptomyces coelicolor M145. Marfey’s derivatised coelichelin hydrolysate and Marfey’s derivatives of L-Thr, L-allo-Thr, D-allo-Thr were analyed. The results showed that coelichelin contains D-allo-Thr. Similarly, analyses were carried out using the Marfey’s derivatives of chemically synthesised D- and L-N5- hydroxyornithine, however the results were inconclusive. The biosynthesis of foroxymithine in S. narbonensis was also investigated. A gene fragment proposed to be within the putative foroxymithine biosynthetic gene cluster was amplified by PCR from the genomic DNA of S. narbonensis. The gene fragment was cloned into a plasmid vector and sequenced. This confirmed that it encodes part of a putative formyl transferase that could be involved in foroxymithine biosynthesis. Fosmid libraries of S. narbonensis genomic DNA were prepared. Despite exhaustive efforts, a fosmid clone containing the putative formyl transferase encoding gene fragment could not be identified via PCR based screening of the library.