Attempts to find the correct structure of uniflorine A

Publication Year:
2008
Usage 535
Downloads 302
Abstract Views 233
Repository URL:
http://ro.uow.edu.au/theses/143
Author(s):
Davis, Andrew Stewart
Publisher(s):
School of Chemistry - Faculty of Science
thesis / dissertation description
The alkaloid uniflorine A was isolated in 2000 from the leaves of the tree Eugenia uniflora L, together with two other water soluble alkaloids, uniflorine B and the known alkaloid (+)-(3α,4α,5β)-1-methylpiperidine-3,4,5-triol piperidine. Uniflorine A was found to be an inhibitor of the α-glucosidases, rat intestinal maltase and sucrase, with IC50 values of 12 and 3.1 μM, respectively, and its structure was deduced from NMR analysis to be structure 1. Uniflorine B was also found to be an inhibitor of the above α-glucosidases and its structure was determined from NMR analysis to be structure 2. The initial goal of this study was to complete the total synthesis of 1 and determine the validity of its proposed structure. In the event, an efficient 9-step diastereoselective synthesis of 1 was achieved by using the Petasis borono-Mannich reaction, ring-closing metathesis and stereoselective cis-dihydroxylation as key steps. The structure of our synthetic 1 was unequivocally established by a single-crystal X-ray crystallographic study of its pentaacetate derivative. However, the 1H and 13C NMR data for synthetic 1 did not match with those reported for uniflorine A; the latter showed many more downfield peaks in the 1H NMR, perhaps consistent with the amine salt. The 1H NMR of the hydrochloride salt of synthetic 1, however, did not match the literature spectroscopic data either. We therefore concluded that the structure assigned to uniflorine A was not correct. We also found that the coupling constant J1,8a of 4.5 Hz for uniflorine A, was more consistent with the relative syn-H-8a, H-1 configuration, suggesting that uniflorine A, if it was an indolizidine alkaloid, had the same H-1 configuration as castanospermine. Our attempts to prepare 2-epi-1 and 1,2-di-epi-1 were unsuccessful due to unexpected competing side-reactions. In addition, the diasteresoselective synthesis of the C-1, C-2 di-epimer of 1 was achieved. This synthesis employed a novel pyrrolo[1,2-c]oxazin-1-one precursor to allow for the reversal of π-facial diastereoselectivity in an osmium(VIII)-catalysed syn-dihydroxylation (DH) reaction. The NMR spectroscopic data of this epimeric compound and that of related isomers did not match that of the natural product. From a comparison of the NMR data of uniflorine A and uniflorine B with that of casuarine and the known synthetic 1,2,6,7-tetrahydroxy-3-hydroxymethylpyrrolizidine isomers we concluded unequivocally that uniflorine B is the known alkaloid casuarine. Although we cannot unequivocally prove the structure of uniflorine A, without access to the original material and data, the published data suggest that the natural product is also a 1,2,6,7-tetrahydroxy-3-hydroxymethylpyrrolizidine with the same relative C-7-C-7a-C-1-C-2-C-3 configuration as casuarine. We thus suggest that uniflorine A is 6-epi-casuarine.