emetine was synthesized by the same method. This paper
describes these results.
With a practical amount of compound 2 in hand, we
decided to functionalize the obtained allyl group. After the
protection of an amino group of 2, reaction of 3 with various
monosubstituted alkenes using the second-generation Grubbs’
catalyst was carried out, and it was found that the cross-
metathesis products were obtained in high yields and good
stereoselectivity.13 Without the Boc protecting group, the
metathesis reaction did not proceed.
Although there are a few papers on the catalytic asym-
metric allylation of imines,8 there has been no report
concerning the catalytic allylation of cyclic imines.9 Although
Yamamoto et al. recently reported a general method for the
allylation of various kinds of imines, they showed the
incompatibility of cyclic imines to their reaction system.8d
Recently, a new allylation reaction of ketones and aldehydes
has been published by Shibasaki et al.10 using allyltrimethox-
ysilane and a catalytic amount of Cu(I) salt. We applied their
reaction system to the allylation of 6,7-dimethoxy-3,4-
dihydroisoquinoline and found that the reaction proceeded
in a stereoselective manner as shown in Table 1.
By using the ethyl acrylate, sufficient stereoselectivity was
obtained to give an adequate amount of a functionalized (E)-
alkene derivative 4 (Scheme 1).
Scheme 1
Table 1
yield of ee of 212
entry solvent
chiral ligand
time
2 (%)
(%)
1
2
3
4
5
6
7
8
THF
THF
THF
THF
THF
DMF
ether
20 h
1 d
1 d
72
91
35
78
31
21
37
35
(R)-tol-BINAP
(R)-BINAP
(R,R)-DIPAMP
71 (S)
71 (S)
5
21 (S)
50 (S)
47 (S)
67 (S)
1 d
The deprotection of 4 followed by Michael addition of 5
with methyl vinyl ketone afforded an N-(3-oxobutyl) deriva-
tive, which was then cyclized to 6 in a completely diaste-
reoselective manner (Scheme 2). In our first plan, the acetyl
group would be transformed to the corresponding ethyl group
according to the reported method,14 but our attempt to reduce
the keto group to give 7 resulted in a very low yield under
various conditions.15
(R,R)-CHIRAPHOS 1 d
(R)-tol-BINAP
(R)-tol-BINAP
1 d
1 d
1 d
dioxane (R)-tol-BINAP
Various phosphine derivatives were investigated as chiral
ligands, and it was found that tol-BINAP in THF at room
temperature afforded the best result for the present reaction.
The yield of 2 was lowered to 21% by the reaction at 10 °C
without an increase of the ee, and the reaction did not proceed
at 0 °C. Other allylation reagents such as allyltributyltin
afforded a racemic product. Although the stereoselectivity
is moderate, this is the first example that a cyclic imine is
adopted as a catalytic allylation reaction.
The product 2 thus obtained was treated with (-)-
dibenzoyl-L-tartaric acid to form a mixture of the diastere-
omeric salts, which was recrystallized from acetonitrile/H2O
(20:1) to give optically pure 2 (97% ee) in 67% yield based
on the starting material.11
Thus, we changed the synthetic procedure as follows
(Scheme 3). Although Michael addition of acrolein to
compound 5 resulted in a complex mixture of the products,
slowing the addition of acrolein considerably improved the
reaction yield to a practical level. That is, the addition of
acrolein to 5 over 5 h followed by treatment with pyrrolidine
afforded a ring-closing product 8 in good yield and complete
stereoselectivity. Although the compound 8 was obtained at
first as its epimer at the C-3 position (according to emetine
numbering), the epimeric compound rapidly isomerized to
8 under the reaction conditions. The formyl derivative 8 thus
(11) With three times of careful recrystallization, the tartrate salt of
racemic 2 afforded the enantiomeric 2 in 97% ee.
(12) The absolute configuration of the compound 2 was determined by
the transformation to the known compound 10.
(13) Nagata, K.; Itoh, T.; Fukuoka, H.; Nakamura, S.; Ohsawa, A.
Heterocycles 2005, 65, 1283.
(14) Hirai, Y.; Terada, T.; Hasegawa, A.; Yamazaki, T. Chem. Pharm.
Bull. 1998, 36, 1343.
(15) Other than the reported method that used ethanedithiol-TFA
followed by Raney Ni, several reduction systems were tested which involve
various variants of Wolff-Kishner or Clemmensen reduction, but the
product 7 was not obtained in more than 7% yield.
(8) (a) Nakamura, H.; Nakamura, K.; Yamamoto, Y.J. Am. Chem. Soc.
1998, 120, 4242. (b) Fang, X.; Johannsen, M.; Yao, S.; Gathergood, N.;
Hazell, R. G.; Jorgensen, K. A. J. Org. Chem. 1999, 64, 4844. (c) Gastner,
T.; Ishitani, H.; Akiyama, R.; Kobayashi, S. Angew. Chem., Int. Ed. 2001,
40, 1896. (d) Fernandes, R. A.; Stimac, A.; Yamamoto, Y. J. Am. Chem.
Soc. 2003, 125, 14133.
(9) There are a few reports which claimed the asymmetric allylation using
a stoichiometrical amount of chiral compound; see: Nakamura, M.; Hirai,
A.; Nakamura, E. J. Am. Chem. Soc. 1996, 118, 8489 and references therein.
(10) Yamasaki, S.; Fujii, K.; Wada, R.; Kanai, M.; Shibasaki, M. J. Am.
Chem. Soc. 2002, 124, 6536.
1296
Org. Lett., Vol. 8, No. 7, 2006