Doctor of Philosophy
Department of Chemistry
Tarrant, Gregory James, Alkaloid-based syntheses of potential new glycosidase inhibitors, Doctor of Philosophy thesis, Department of Chemistry, University of Wollongong, 1997. http://ro.uow.edu.au/theses/1109
The general aim of this work was to develop novel synthetic approaches to the 1- hydroxytropane and 1-hydroxynortropane class of compounds from the structurally advanced alkaloids, scopolamine 99 and castanospermine 7, which are isolated from Australian plants.
This project consists of two parts. The first part (Chapter 2) deals with the preparation of 3α-hydroxyphysoperuvine 103 from scopolamine. This was achieved in nine steps with an overall yield of 16%. The main steps in this synthesis included the deoxygenation of scopolamine and the Meisenheimer rearrangement of a tropane N-oxide. The deoxygenation of scopolamine was achieved using zinc-copper couple in a near quantitative conversion to give 6,7-dehydrohyoscyamine 111. The thermolysis of 3α-t-butyldimethylsiloxy-8-methyl-8-azabicyclo[3.2.1]oct-6-ene axial /V-oxide 161 proceeded quite readily, giving a 53 % isolated yield of the Meisenheimer rearrangement product (1R*, 3S*, 5S*) -3α - t- butyldimethylsiloxy - 9 - methyl - 8 - oxa - 9 - azabicyclo[3.2.2]non-6-ene 164. The thermolysis of 3α-t-butyldimethylsiloxy-8- methyl-8-azabicyclo[3.2.1]oct-6-ene equatorial N-oxide 162 did not proceed as readily, but it still gave the Meisenheimer rearrangement compound 164 in a 14 % isolated yield. The 3α-hydroxyphysoperuvine 103 did not show any glycosidase inhibition activity.
Chapter 3 deals with the mechanistic aspects of the Meisenheimer rearrangement of the tropane N-oxides 161 and 162. Rate studies showed that the N-oxide 161 rearranged approximately seven times faster than 162 at 100°C in DMSO-d6. Computer modelling studies failed to give a definitive answer as to why one N-oxide rearranged faster than the other.
The steps towards the total synthesis of the naturally occurring compound, 1- hydroxytropacocaine 15 were also investigated (Chapter 4). A Mitsunobu reaction carried out on the model compound 8-methyl-8-azabicyclo[3.2.1]oct-6-en-3α-ol 138 was successful in inverting the stereochemistry at C 3 to give 6,7-dehydrotropacocaine 183. However, using this procedure to invert the stereochemistry at C3 of [3S-(lα,5α,3β)]-heptafluorobutyric acid 3-hydroxy-8-methyl-8-azabicyclo[3.2.1]octan-l-yl ester 182 was not successful. Hence 1-hydroxytropacocaine was not synthesised.
The second part of this project involved the attempted conversion of castanospermine 7 via ring cleavage modification to give 2,3,4,6-epicalystegine C1 106. The ring cleavage of a mixture of TBDMS-protected castanospermine with excess methyl chloroformate gave a mixture of piperidine carbamates 202-205. The removal of the TBDMS groups with tetrabutylammonium fluoride gave in good yield (lS,6S,7R,8R,8aR)-l-(2'-chloroethyl)-l,5,6,7,8,8a-hexahydro-6,7,8-trihydroxy-3H- oxazolo[3,4-a]pyridin-3-one 216. The dechlorination of 216 gave in good yield the corresponding olefin 209. Hydrolysis of 209 yielded a new derivative 217 related to 1- deoxynojirimycin 5 which was tested for glycosidase inhibition. However, it showed no activity against a range of enzymes. This was the extent of the approach to 106. The next main step in this synthesis required the oxidation of a carbamate to insert what would eventually be the bridgehead hydroxyl. The oxidation of a model compound methyl 1- piperidinecarboxylate with ruthenium tetroxide gave the corresponding methyl 2-oxo-1-piperidinecarboxylate in good yield. However all attempts to oxidise a range of carbamates derived from castanospermine failed to yield any of the required oxidised products.
Although the olefin 217 did not show any glycosidase enzyme inhibition activity, it does provide the basis for the elaboration of a range of new polyhydroxylated amine derivatives in the future.