A "(+)-sparteine-like" chiral diamine, readily synthesized in three steps from (-)-cytisine, has been evaluated in four different asymmetric transformations; in each case, selectivity in an enantiocomplementary fashion to (-)-sparteine was observed.

Three chiral diamines were synthesised and evaluated as sparteine surrogates in the lithiation-substitution of N-(tert-butoxycarbonyl)pyrrolidine. The synthesis and attempted resolution of sparteine-like diamines [(1S*,2R*,8R*)-10-methyl-6,10-diazatricyclo[6.3.1.0(2,6)]dodecane and (1S*,2R*,9R*)-11-methyl-7,11-diazatricyclo[7.3.1.0(2,7)]tridecane] (via inclusion complex formation) are reported. Unfortunately, it was only possible to resolve the diazatricyclo[7.3.1.0(2,7)]tridecane compound. An alternative route to (1R,2S,9S)-11-methyl-7,11-diazatricyclo[7.3.1.0(2,7)]tridecane starting from the natural product, (-)-cytisine, is described. This simple three-step route furnished gram-quantities of a (+)-sparteine surrogate. X-Ray crystallography of an intermediate in the route, (1R,5S,12S)-3-methoxycarbonyldecahydro-1,5-methanopyrido[1,2-a][1,5]diazocin-8-one, enabled the stereochemistry of all of the tricyclic diamines described in this paper to be unequivocally established. Two other diamines, starting from (S)-proline and resolved 2-piperidine ethanol, were prepared using standard methods. These diamines lacked the bispidine framework of (-)-sparteine and were found to impart vastly inferior enantioselectivity. It was concluded that, for the asymmetric lithiation substitution of N-Boc pyrrolidine, a rigid bispidine framework and only three of the four rings of (-)-sparteine are needed for high enantioselectivity. Furthermore, it is shown that diamine (1R,2S,9S)-11-methyl-7,11-diazatricyclo[7.3.1.0(2,7)]tridecane is the first successful (+)-sparteine surrogate.

The enantioselective lithiation of N-Boc-pyrrolidine using sec-butyllithium and isopropyllithium in the presence of sparteine-like diamines has been studied experimentally and computationally at various theoretical levels through to B3P86/6-31G*. Of the (-)-cytisine-derived diamines (N-Me, N-Et, N-(n)Bu, N-CH(2)(t)Bu, N-(i)Pr) studied experimentally, the highest enantioselectivity (er 95:5) was observed with the least sterically hindered N-Me-substituted diamine, leading to preferential removal of the pro-R proton i.e., opposite enantioselectivity to (-)-sparteine. The experimental result with the N-Me-substituted diamine correlated well with the computational results: at the B3P86/6-31G* level, the sense of induction was correctly predicted; the lowest energy complex of isopropyllithium/diamine/N-Boc-pyrrolidine also had the lowest activation energy (DeltaH++ = 11.1 kcal/mol, DeltaG++= 11.5 kcal/mol) for proton transfer. The computational results with the N-(i)Pr-substituted diamine identified a transition state for proton transfer with activation energies of DeltaH++= 11.7 kcal/mol and DeltaG++= 11.8 kcal/mol (at the B3P86/6-31G* level). Although comparable to (-)-sparteine and the N-Me-substituted diamine, these DeltaH++ and DeltaG++ values are at odds with the experimental observation that use of the N-(i)Pr-substituted diamine gave no product. It is suggested that steric crowding inhibits formation of the prelithiation complex rather than increasing the activation enthalpy for proton transfer in the transition state. Three other ligands (N-H and O-substituted as well as a five-membered ring analogue) were studied solely using computational methods, and the results predict that the observed enantioselectivity would be modest at best.