Synthesis of alkaloids as glycosidase inhibitors from natural chiral sources

The main purpose of this work was the synthesis of polyhydroxylated pyrrolizidines and aza-C-disaccharide derivatives. These compounds, as azasugars show a great synthetic interest due to their biological activity, associated to their glycosidase and glycosyltransferase inhibition properties. The first goal of this work was to synthesize polyhydroxylated pyrrolizidines via 1,3-dipolar cycloadditions of chiral azomethine ylides. Azomethine ylides are generated in situ from the condensation reaction of chiral pyrrolidines with ethyl glyoxalate, through unstable iminium species. Two chiral pyrrolidines were used for this purpose, which have been synthesized from L-diethyl tartrate. In the cycloaddition reaction dimethyl fumarate, dimethyl maleate and N-phenyl maleimide were used as dipolarophiles. Interestingly the cycloadditions with dimethyl fumarate and dimethyl maleate gave the same cycloadduct. This result could be explained by the observed isomerisation of dimethyl maleate to dimethyl fumarate during the reaction. In all cases stucture elucidation of all cycloadducts was based on spectroscopic study, in particular 1H-NMR, and in one case on X-ray crystal structure analysis. Generally, the generated ylides react as anti-dipoles and give the desired pyrrolizidine derivatives through exo transition states with high diastereoselectivity. The second part of this work deals with the synthesis of aza-C-disaccharide derivatives, carrying a pyrrolidine unit. Our approach is based on the cycloaddition reaction of a chiral open chain nitrone and alkenes obtained from sugar precursors. Starting from D-ribose, we achieved nitrone synthesis with determined stereochemistry in all chiral centers. As dipolarophiles we chose alkenes that were easily prepared from D-glucose and D-galactose in good yields. Cycloaddition reactions of cis-alkenes proceeded in a diastereoselective manner and gave isoxazolidines suitable for building pyrrolidine rings. By this approach, two aza-C-disaccharide derivatives have been synthesized in high yields. Structure elucidation was based on spectroscopic data 1H-NMR, 13C-NMR and also on NOE, double resonance, COSY and ROESY experiments. Comparing to cis-alkenes obtained D-glucose and D-galactose, the cycloaddition reaction of a trans-alkene prepared from D-glucose, does not give the desired diastereoselectivity, and a mixture of four diastereomeric cycloadducts was obtained. It was possible to isolate and study only one of those cycloadducts, which was transformed to a new aza-C-disaccharide derivative. The third part of this work presents efforts to synthesize Hyacinthacine A3 and its epimers using as key reaction the Cope cyclisation of alkenyl-hydroxylamines. Unfortunately double nucleophilic substitution of dimesylate-derivatives by hydroxylamine did not give the desired alkenyl-hydroxylamines. The dimesylates were prepared from D-ribose and D-arabinose lactones using Grignard reagent addition reaction. In a variation of this procedure alkenyl-hydroxylamines were obtained from oxime’s reduction, without any significant diastereoselectivity. Aza-Cope cyclization of hydroxylamines gave the corresponding N-oxides, in a very complex reaction. At this point N-oxide isomers could not be isolated and our method was so aborted. The forth section presents efforts to synthesize aza-C-disaccharide derivatives, with a six- or a seven-member azasugar unit, using intermolecular 1,3-dipolar cycloaddition reaction of an appropriate-nitrone having an alkenyl substitute. The key step for nitrone synthesis was a vinyl-epoxide ring opening reaction by a sugar derived oxime. Unfortunately the intermolecular 1,3-dipolar cycloaddition reaction did not proceed and the alkenyl-nitrone did not give the desired isoxazolidines, in different reaction conditions, possibly for steric reasons