Synthetic studies toward the four invariant stereogenic centres of the left side of the backbone of the fumonisins and AAL toxins

Abstract

The fumonisins are a class of polyketide mycotoxins produced by Fusarium verticilliodes (formerly Fusarium monoliforme) which commonly affects maize. Ingestion of these toxins has been associated with leukoencephalomalacia in equine species, pulmonary oedema in swine, hepatocarcinogenesis in rats and have been linked to oesophageal cancer in humans. The structurally related AAL toxins are host specific mycotoxins produced by Alternaria alternata f. sp. lycopersici, producing stem canker disease in susceptible tomato cultivars. Examination of the C-11-C-20 fragment of the fumonisin B1 backbone [(2S,3S,5R,10R,12S,14S, 15R,16R)-2-amino-3,5,10,14,15-hydroxy-12,16-dimethyleicosane] and the C-10-C-17 fragment of the AAL toxin TA backbone[(2S,4S,5R,11S,13S,14R,15R)-1-amino-2,4,5,13,14-hydroxy-11,15- dimethylheptadecane], reveals four common stereogenic centres, with the only difference between the two fragments being the length of the alkyl chain. It is thought that the position and configuration of these four stereogenic centres is conserved among all members of the fumonisin and AAL classes of toxins. Retrosynthetic analysis of the backbones reveals a common intermediate aldehyde, which can be synthesised from methyl (S)-3-hydroxy-2-methylpropionate. A simple synthetic route to access the C-11-C-20 fragment for the fumonisins and the C-10-C-17 fragment of the AAL toxins was devised utilising Sharpless asymmetric epoxidation and an Evans aldol reaction as key transformations. In practice, it was found that although the Sharpless asymmetric epoxidation produced the desired epoxide in low enantiomeric excess, the two diastereomers produced could be separated by two consecutive flash chromatography silica gel columns. In pursuit of a more efficient method for introduction of the stereogenic centre in the target, other synthetic routes and key transformations were considered. Jacobsen’s kinetic resolution of terminal racemic epoxides was explored, requiring a terminal alkene from which the racemic epoxide was synthesised. An attempt to synthesise the terminal alkene from the appropriate tosylate and vinyl-MgBr, mediated by copper (I) iodide, failed. The synthetic route was redesigned, and the terminal alkene was synthesised by two one-carbon additions: the first a nucleophilic substitution with cyanide, and the second a Wittig olefination. The resolution of the terminal epoxide was also unsuccessful with no significant kinetic resolution occurring. Sharpless asymmetric dihydroxylation was also investigated; however, this reaction too failed to produce products of high diastereomeric excess. As a consequence, it was decided to pursue the asymmetric epoxidation route as the diastereomeric products could at least be separated. The second key transformation, the Evans aldol reaction, also provided an interesting result. When the aldol reaction was attempted with benzaldehyde and enolates derived from (4R,5S)-3-butanoyl-4-methyl-5-phenyl-oxazolidin-2-one and (4R,5S)-3-hexanoyl-4-methyl-5-phenyl-oxazolidin-2-one, the butanoyl derivative was found to give the expected Evans syn product, while the hexanoyl derivative was found to give the non-Evans syn product, with proof provided by single crystal X-ray diffraction analysis. It is proposed that the aldol reaction with the hexanoyl derivative does not proceed through the expected Zimmerman-Traxler-type transition state, but rather through an open chain transition state similar to that seen for asymmetric alkylation reactions. Synthesis of the pentanoyl derivative, and subjecting it to the same aldol reaction gave the expected syn Evans product, as deduced from spectroscopic properties. When the aldol reaction was attempted with the appropriate aldehyde intermediate, it was found that the dibutylboron triflate in the reaction medium caused the cleavage of the O-TBS ether protection, resulting in the formation of (3S,5R)-3-(4-methoxybenzyloxy)-5-methyl-tetrahydropyran-2-ol, before the aldehyde could undergo the aldol reaction. In order to avoid this problem, it is suggested that an alternative protecting group strategy using a more robust protecting group, such as a benzyl group which is stable to Lewis acids, could be substituted for the O-TBS group. Copyright