racemic cycloaddition product (()-12a. In contrast, mi-
˚
crowave-promoted reaction of 6a (200 C, chlorobenzene,
30 min), in the presence of a radical inhibitor, gave smooth
conversion to (()-12a in very good yield (86%) but with
poor diastereoselectivity (1:2 trans to cis). When the
product was redissolved in ethanol and subjected to micro-
˚
wave conditions (120 C), product 12a (dr ∼1:2) equili-
brated completely to cis-12a (dr > 30:1). Formation of the
predominantly cis-product suggested R-epimerization of
12a, similar to that observed during IMDA of allylic
sorbates to bicyclic γ-butyrolactones.13 The stage was set
to explore optimzed conditions for asymmetric IMDA.
The precursor asymmetric IMDA was prepared in two
steps by methanolysis (K2CO3, MeOH) of the acetate ester
6a to the corresponding allylic alcohol which was oxidized
(MnO2, CH2Cl2) to aldehyde 5a14 in good overall yield
(53% over two steps). Exposure of 5a to either catalyst 11a
or 11b (Scheme 2) gave slow [4 þ 2] cycloaddition, a
disappointing yield of cycloadduct 4a (∼4%) and poor
recovery of starting material, probably due to the tendency
of the diene to polymerize.15 Reasoning that a terminally
substituted diene may fare better in the IMDA, aldehyde
5b was prepared using the same sequence of reactions and
replacement of pentadienoic acid with sorbic acid 11b. In
the presence of catalyst 11b, aldehyde 5b underwent clean
asymmetric IMDA in good yield (Table 1, Entry 2, 84%),
exclusively in endo mode, to give mostly (þ)-(4S,5R,8R)-
4b16 with a lesser amount of (4R,5R,8R)-4b (dr > 20:1),
albeit in modest enantiomeric excess (42% ee). The relative
configurations of the separated pure isomers (HPLC) were
determinedfrom extensive1D-NOE experiments (Figure 2
and Supporting Information).
Figure 2. 1D-NOE of (4S)- and (4R)- isomers of 4b (mixing time,
m = 400 mS). Numbering follows that of 1 (see ref ).
t
workers17 as an intermediate in the preparation of
copolymer-supported catalysts, gave the best outcomes.18
Under conditions similar to those used with 11b (Entry 2),
˚
IMDA of 5b in the presence of 11c (20 mol %, 23 C, Entry
5) gave (þ)-4b with almost double the enantioselectivity
(72% ee), albeit with lower diastereoselectivity (73%, dr =
3.8:1). Optimal conditions for IMDA of 5b (Entry 9, 20
˚
mol
% 11c, 0 C, 73 h) gave 4b (84% yield,
dr = 6:1, 88% ee).19 Base treatment of 4b (DBU, C6D6, 23
C, 13 h, Scheme 3) epimerized C4 and inverted the 4R:4S
ratio to >20:1 in favor of the configuration required for
1.20 Thus, pure (þ)-(4R,5R,8S,11S)-4b was obtained in
70% yield over two steps after preparative HPLC.
˚
Scheme 3. Base-Promoted Isomerization of (4S,8R)-4b
Optimization of the asymmetric IMDA (Table 1) was
undertaken and, similar to observations by MacMillan,8 it
was found that the catalyst structure, counterion, and
temperature all played important roles in affecting the
yield, diastereoselectivity and enantioselectivity. Catalyst
11a (HCl salt) gave poorer yields of 4b (Entry 1, 5%), even
after 72 h. A slight gain in enantioselectivity in formation
of the major epimer 4S-4b (Entry 3, 50% ee) was seen with
catalyst 11b (HClO4 salt) when the temperature was
The steric bulk of the N-protecting group influences the
outcome of the IMDA reaction through torsional strain
that also populates the required s-cis conformation of the
tertiary amide. Replacement of the N-PMB protecting
group with a 2,4-dimethoxybenzyl group (N-DMB) was
investigated to determined the effect on yield, enantio- and
stereoselectivity. Compound 5c, prepared from 9 using a
similar sequence for 5b, was treated with 11c (20 mol %,
3 °C, 54 h) to afford 4c in 67% yield, with slightly lower
enantioselectivity (84% ee) but with high 4S:4R diaster-
eoselectivity (dr > 20:1).21
˚
˚
lowered from 23 C to 10 C, however, at the expense of
lower yield, diastereoselectivity (60%, dr 6:1) and reaction
time (57 h instead of 4.5 h).
Gratifyingly, N-(2-hydroxy-1-ethyl)-imidazolidone 11c,
a MacMillan-type catalyst reported by Kristensen and co-
The kinetics of base-equilibration of purified (4S,8R)-4b
(DBU, C6D6, 23 °C, Scheme 3 and Supporting Informa-
tion) were briefly investigated. H NMR revealed rapid
conversion of the (4S,8R)-4b to the more stable isomer
(4R,8R)-4b,22 and finally, slower conversion of the latter to
a third isomer, (4R,8S)-4b.
(13) (a) Wu, J.; Yu, H.; Wang, Y.; Xing, X.; Dai, W.-M. Tetrahedron
Lett. 2007, 48, 6543. (b) Guy, A.; Lemaire, M.; Guiette, M. Tetrahedron
Lett. 1985, 26, 3575. (c) Boeckman, R. K., Jr.; Demko, D. M. J. Org.
Chem. 1982, 47, 1789. (d) Martin, S. F.; Williamson, S. A.; Gist, R. P.;
Smith, K. M. J. Org. Chem. 1983, 48, 5170.
1
(14) Contained ∼5ꢀ10% of the Z-isomer.
(15) Inclusion of a radical inhibitor C (Scheme 1) did not suppress
formation of polymer; therefore, the mechanism of polymerization is
likely ionic in nature.
(19) The hydroxyethyl side chain in 11c may play an important role in
stabilizing either the transition state by hydrogen bonding or through
intramolecular nucleophilic capture of the incipient iminium ion.
(20) The relative configurations of (4S,8R)-4b and (4R,8R)-4b were
assigned by 1D NOESY studies.
(16) The absolute configuration follows from the expected sense of
asymmetric induction demonstrated by McMillan and co-workers.
(17) Kristensen, T. E.; Vestli, K.; Jakobsen, M. G.; Hansen, F. K.;
Hansen, T. J. Org. Chem. 2010, 75, 1620–1629.
(21) See Table S1, Supporting Information, for optimization of
conditions for IMDA of 5c to give 4c.
(22) The cyclohexene ring conformation changed to a pseudoboat.
(18) Catalyst 11c is more conveniently prepared from L-phenylala-
nine methyl ester and inexpensive ethanolamine than 11a,b, which
requires the more expensive “controlled substance” MeNH2.
3934
Org. Lett., Vol. 13, No. 15, 2011