all cases only one configuration at the newly formed
stereocenter was observed. Significantly, pyrrolidines 1c and
1l, which were formed from enantiomerically pure amines,
(S)-R-methylbenzylamine and (S)-phenylalaninol, respec-
tively, were obtained as a 1:1 mixture of diastereoisomers,
which clearly shows that the planar chirality of the diene
complex overrides the chirality of the amine. The standard
reaction conditions proved suitable for all substrates apart
from particularly basic amines such as methylamine and
ethanolamine, where the high reaction pH (>10) led to
exclusive pyrrole formation. In these cases, reducing the
reaction pH to 6 solved this problem and allowed the desired
pyrrolidine (1d, 1f) to be obtained in good yield.10 It proved
impossible to access the free pyrrolidine directly: the use
of NH3 led to a complex mixture of products, and when NH4-
HCO2 was employed, only the pyrrole was obtained.5e
However the free pyrrolidine 8 was readily accessed by
deallylation of 1h (Scheme 3).13
same, and in line with our hypothesis that the second
reductive amination proceeds on the iminium intermediate
in an s-trans conformation, with “hydride” attack from the
face remote from the bulky tricarbonyliron group.
Having established an efficient and highly stereoselective
route to 2-dienyl-substituted pyrrolidines, we were keen to
explore the possibility of employing the vinyl bromide
functionality present in 1i and 1k as an entry into the
biologically important pyrrolizidine skeleton.14
In the case of 1k, Cu(II)-mediated decomplexation af-
forded pyrrolidine 10 in 70% yield.15 While we expected a
vinyl radical formed from bromide 10 would cyclize
preferentially on to the pendant diene in a 5-exo-trig fashion,
the stereoselectivity of this cyclization and regioselectivity
in the subsequent H-abstraction step, could still afford a host
of isomeric products. It was therefore pleasing to observe
that reaction of 10 with Bu3SnH in the presence of AIBN,
provided pyrrolizidine 13 as a >10:1 mixture of vinylsilane
stereoisomers. The fact that the stereochemistry of the
vinylsilane in the major product is opposite to that in the
starting material is in accord with the known configurational
instability of vinyl radicals.16 The diastereoselectivity of the
reaction is consistent with cyclization proceeding on vinyl
radical 11 through the transition state shown in Scheme 4.
Scheme 3. Deallylation of 1h Followed by Acylation of the
Free Pyrrolidine Provided a Crystalline Product
Scheme 4. Formation of the Pyrrolizidine Skeleton through a
Radical Cyclization
Confirmation of the stereochemical outcome of the cy-
clization reactions came from crystal structures (see the
Supporting Information) of the pyrrolidine (Z)-1i and amide
9, formed from 1h after deallylation and acylation (Scheme
3). In both cases, the observed diastereoselectivity was the
(5) For some recent successful applications of this approach to pyrro-
lidines, see: (a) Laventine, D. M.; Davies, M.; Evinson, E. L.; Jenkins, P.
R.; Cullis, P. M.; Fawcett, J. Tetrahedron Lett. 2005, 46, 307-310. (b)
Gravier-Pelletier, C.; Maton, W.; Bertho, G.; Le Merrer, Y. Tetrahedron
2003, 59, 8721-8730. (c) Trost, B. M.; Schroeder, G. M.; Kristensen, J.
Angew. Chem., Int. Ed. 2002, 41, 3492-3495. (d) Baxendale, I. R.; Brusotti,
G.; Matsuoka, M.; Ley, S. V. J. Chem. Soc., Perkin Trans. 1 2002, 143-
154. (e) Momotake, A.; Mito, J.; Yamaguchi, K.; Togo, H.; Yokoyama,
M. J. Org. Chem. 1998, 63, 7207-7212. (f) Boga, C.; Manescalchi, F.;
Savoia, D. Tetrahedron 1994, 50, 4709-4722.
(6) For related reactions that also proceed on iminium ions appended to
diene complexes, see: (a) Takemoto, Y.; Takeuchi, J.; Matsui, E.; Iwata,
C. Chem. Pharm. Bull. 1996, 44, 948-955. (b) Ripoche, I.; Gelas, J.; Gre´e,
D.; Gre´e, R.; Troin, Y. Tetrahedron Lett. 1995, 36, 6675-6678.
(7) Vaidyanathan, R.; Kalthod, V. G.; Ngo, D. P.; Manley, J. M.; Lapekas,
S. P. J. Org. Chem. 2004, 69, 2565-2568.
(8) Turakhia, R. H.; Fischer, G. C.; Morrow, C. J.; Maschhoff, B. L.;
Toubbeh, M. I.; Zbur-Wilson, J. L. J. Org. Chem. 1986, 51, 1955-1960.
(9) Gre´e, D. M.; Martelli, J. T.; Gre´e, R. L.; Toupet, L. J. J. Org. Chem.
1995, 60, 2316-2317.
The position of the alkene in the side-chain can then be
rationalized by assuming H-abstraction proceeds on the less
hindered end of the resulting allyl radical intermediate 12.
In an alternative elaboration strategy, we envisaged that
treating vinyl bromide 1i with base should afford a carbenoid
(10) (a) Raiman, M. V.; Pukin, A. V.; Tyvorskii, V. I.; De Kimpe, N.;
Kulinkovich, O. G. Tetrahedron 2003, 59, 5265-5272. (b) Harmata, M.;
Ghosh, S. K.; Hong, X.; Wacharasindhu, S.; Kirchhoefer, P. J. Am. Chem.
Soc. 2003, 125, 2058-2059.
(11) Borch, R. F.; Bernstein, M. D.; Durst, H. D. J. Am. Chem. Soc.
1971, 93, 2897-2904.
(14) Liddell, J. R. Nat. Prod. Rep. 2002, 19, 773-781 and references
therein.
(15) Thompson, D. J. J. Organomet. Chem. 1976, 108, 381-383.
(16) Vinyl radicals are known to be configurationally unstable. For a
pertinent example, see: Miura, K.; Itoh, D.; Hondo, T.; Hosomi, A.
Tetrahedron Lett. 1994, 35, 9605-9608.
(12) Abdel-Magid, A. F.; Maryanoff, C. A.; Carson, K. G. Tetrahedron
Lett. 1990, 31, 5595-5598.
(13) Lemaire-Audoire, S.; Savignac, M.; Geneˆt, J.-P.; Bernard, J. M.
Tetrahedron Lett. 1995, 36, 1267-1270.
Org. Lett., Vol. 8, No. 20, 2006
4391