conformers cannot be predicted easily. Preference among the
conformers is crucial for achieving highly selective reactions
because nucleophilic attack should be slow relative to
conformer interconversion.2b,5,6 The unsubstituted intermedi-
ate exists as two low-energy conformations, boat-chair-2 (1)
and boat-chair-3 (2), which are comparable in energy
(Scheme 1).7-10 When a remote substituent is appended to
Scheme 2
Scheme 1
the ring, conformational control becomes even more com-
plex. Selectivity upon nucleophilic attack depends not only
on the preferred substituent orientation but also on the relative
contribution of 1, 2, and other possible conformers to the
overall energy profile.7b Each of these conformers should
undergo attack by nucleophiles from the periphery as
observed for electrophilic approach to medium- and large-
ring olefins.11
To elucidate the influence of remote substitution on
selectivity, we compared nucleophilic substitution reactions
of eight-membered ring oxocarbenium ion precursors bearing
a C3-, C4-, or C5-alkyl or alkoxy substituent.12,13 Nucleo-
philic addition to a C4-methyl-substituted eight-membered
ring oxocarbenium ion, generated upon treatment of acetate
3a with a Lewis acid, is unselective (Scheme 2).14-16 The
lowest energy conformers of this intermediate, 6 and 7, bear
a strong resemblance to the unsubstituted cations shown in
Scheme 1. Although the steric preference for a methyl
substituent to adopt a pseudoequatorial orientation seems to
dominate, conformers 6 and 7 should be similar in energy.10
As a consequence, selectivity is low because both di-
astereomeric faces of the oxocarbenium ion are presented
to the nucleophile.
A C4-alkoxy substituent, however, does control the
conformation of the charged intermediate.10 The C4-benzyl-
oxy-substituted acetate 3b afforded high 1,4-trans di-
astereoselectivity upon nucleophilic substitution (Scheme
2).14-16 This selectivity can be explained by a through-space
electrostatic attraction17,18 between the remote electronegative
substituent and C-19a that stabilizes conformer 8 (Scheme
(14) Mixtures of diastereomeric acetates (the syntheses of acetates are
contained in Supporting Information) were used in these reactions. Control
experiments indicate that diastereomeric acetates give the same product with
largely the same degree of selectivity. These control experiments strongly
suggest that the reaction operates by a dissociative mechanism. If direct
displacement occurred, the selectivity should reflect the initial acetate ratio.
(15) In all cases, diastereoselectivities were determined by GC or single-
scan 1H NMR spectra of unpurified reaction mixtures. The relative
stereochemistry of the major product was not proven for unselective
reactions (C4- and C5-methyl) but was postulated to be cis on the basis of
computational results (ref 10). For nitrile 4b, the relative stereochemistry
was proven by X-ray crystallography.
(16) Control experiments indicate that cyanide addition is irreversible,
so these reactions are kinetically controlled. These reactions require the
presence of Lewis acid, and the selectivity of product formation is
independent of the solvent (CH2Cl2, toluene, or Et2O) and Lewis acid
(EtAlCl2 or TiCl4) employed. A table of comparative selectivities appears
in Supporting Information.
(5) Cremer, D.; Gauss, J.; Childs, R. F.; Blackburn, C. J. Am. Chem.
Soc. 1985, 107, 2435-2441.
(6) Under certain conditions, solvent cage effects can exert strong
influences on the selective reactions of oxocarbenium ions: Zhang, Y.;
Reynolds, N. T.; Manju, K.; Rovis, T. J. Am. Chem. Soc. 2002, 124, 9720-
9721.
(7) (a) Anet, F. A. L. In Conformational Analysis of Medium-Sized
Heterocycles; Glass, R. S., Ed.; VCH: New York, 1988, Chapter 2. (b)
Meyer, W. L.; Taylor, P. W.; Reed, S. A.; Leister, M. C.; Schneider, H.-J.;
Schmidt, G.; Evans, F. E.; Levine, R. A. J. Org. Chem. 1992, 57, 291-
298. (c) Yavari, I.; Tahmassebi, D.; Nori-Shargh, D.; Heydari, M. Monatsh.
Chem. 1996, 127, 1021-1025.
(8) Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry of Organic
Compounds; Wiley: New York, 1994; pp 765-771.
(9) (a) Dudley, T. J.; Smoliakova, I. P.; Hoffmann, M. R. J. Org. Chem.
1999, 64, 1247-1253. (b) Liang, G.; Sorensen, J. B.; Whitmire, D.; Bowen,
J. P. J. Comput. Chem. 2000, 21, 329-339.
(17) (a) Woods, R. J.; Andrews, C. W.; Bowen, J. P. J. Am. Chem. Soc.
1992, 114, 859-864. (b) Miljkovic´, M.; Yeagley, D.; Deslongchamps, P.;
Dory, Y. L. J. Org. Chem. 1997, 62, 7597-7604. (c) Jensen, H.; Bols, M.
Org. Lett. 2003, 5, 3419-3421.
(10) Details of calculations are provided as Supporting Information.
(11) Still, W. C.; Galynker, I. Tetrahedron 1981, 37, 3981-3996.
(12) Numbering in this paper considers the carbocationic carbon as C-1.
(13) For these reactions, a small nucleophile, trimethylsilyl cyanide
(Evans, D. A.; Carroll, G. L.; Truesdale, L. K. J. Org. Chem. 1974, 39,
914-917), was chosen to minimize steric effects in the transition state that
might perturb the inherent conformational preferences of the charged
intermediates. Lewis acid-mediated nucleophilic substitution of 10a and
15a with another small nucleophile, diethyl-2-phenylethynylalane, gave
selectivities comparable to those reactions using trimethylsilyl cyanide as
the nucleophile.
(18) The border between anchimeric assistance and electrostatics is not
a clear one. On the basis of B3LYP/6-31G* calculations, the oxygen of the
benzyloxy group and the carbocationic carbon are approximately 2.6 Å apart
in the optimized geometry of 8. Although this separation is less than the
van der Waals contact distance (3.2 Å, as found in: Bondi, A. J. Phys.
Chem. 1964, 68, 441-451), it is considerably greater than a standard
carbon-oxygen bond (1.4 Å) and longer than the bond distance in a trivalent
oxonium ion such as the Et3O cation (1.5 Å, as found in: Watkins, M. I.;
Ip, W. M.; Olah, G. A.; Bau, R. J. Am. Chem. Soc. 1982, 104, 2365-
2372). In addition, the C4-OBn bond and the C1dO+ bond are of standard
lengths (1.4 and 1.27 Å, respectively, ref 9b).
4740
Org. Lett., Vol. 6, No. 25, 2004