|Abstract or Summary
- In previous work (Blakemore et al. Org. Lett. 2005, 7, 4721; Heterocycles 2006, 70, 609), successful elaborations of the lupine alkaloids (±)-α-isosparteine (dl-2) and (±)-β-isosparteine (dl-3) were realized from a common tetraoxobispidine precursor, 3,7-diallyl -2,4,6,8-tetraoxo-3,7-diazabicyclo[3.3.1]nonane (93). Herein, the tetraoxobispidine approach to lupine alkaloids was extended to a total synthesis of (±)-sparteine (dl-1), a second generation improved route to (±)-βisosparteine (dl-3) was established, and various novel synthetic manipulations of tetraoxobispidines were evaluated. The pivotal intermediate (93) was prepared in an optimized 16% overall yield (c.f. 9% previously reported) from dimethyl malonate and paraformaldehyde via acid-promoted cyclization of the Knoevenagel condensation adduct 1,1,3,3-propanetetracarboxamide (107), followed by N,N ́-diallylation of the resulting 2,4,6,8-tetraoxo-3,7diazabicyclo[3.3.1]nonane (108). Bisimide 93 was advanced to (±)-sparteine (dl1) in 12% overall yield via six operations: (a) monoreduction of the bisimide (93) with sodium borohydride, (b) Sakarai-type allylation of the hemiaminal (97), (c) nucleophilic addition of allylmagnesium bromide to the remaining imide (98), (d) double ring-closing olefin metathesis of the resulting tetraenyl bicyclic hydroxybislactam (104) to yield a tetracyclic diene intermediate (126), (e) hydrogenation of alkene moieties, and finally, (f) exhaustive reduction with lithium aluminum hydride. It was discovered that the sodium borohydride reduction [step (a)] also gave a C2-symmetric bishemiaminal (116) as a minor byproduct: double Sakarai-type allylation of this compound generated a tetraenylbislactam intermediate (100) identical to that employed in the original synthesis of (±)-β-isosparteine. As such, the tetraoxobispidine route to (±)-βisosparteine was effectively shortened to just five net steps from bisimide 93 (c.f. seven steps previously). A variety of known methods for the enantioselective addition of hydride and allyl nucleophiles to aldehydes were applied to C2Vsymmetric bisimide 93 in an attempt to realize an efficient asymmetric synthesis of sparteine alkaloids. None of the methods investigated, which included Keck allylation and Noyori transfer hydrogenation, gave any trace of addition adducts, reflecting the comparatively low intrinsic reactivity of imides. Finally, the reduction of N,N ́-dibenzyltetraoxobispidine (149) to N,N ́-dibenzylbispidine (150) was realized in 25% yield using sodium bis(methoxyethoxy)aluminum hydride (Red-Al). It was therefore established that tetraoxobispidines unsubstituted on the methylene bridge are viable precursors to potentially useful biologically active bispidines.