Scheme 2a
a (a) Bu4NOH, THF, 0 °C, 99%; (b) TMSCHN2, MeOH, rt, 99%; for 3 and 19: (c) AgNO3, (NH4)2S2O8. THF-H2O-MeCN, 60 °C,
53%; for 20: (d) CAN, H2O-MeCN, 0 °C, 91%, for 21.
Unfortunately, the lithium enolate of 2, generated by
treatment with n-BuLi in THF at -78 °C for 0.5 h, was
almost inert to aldimines 9 and 13 under the above optimized
conditions (for aldimine 9, 15% yield, 93:7 dr). We thus
focused on the use of a Lewis acid additive, anticipating the
pronounced activation of imine functionalities (Scheme 1).
The addition of 2 to aldimine 9 occurred effectively in the
presence of 0.5 equiv of R2Zn (R ) Me, Et) at -78 °C to
give â-aminoester 3 in 70% and 69% yields with diastere-
omeric ratios (dr) of 95.5:4.5. Using 1.0 equiv of Et2Zn was
generally ineffective in terms of chemical yield due to side
reactions, which in some cases generated a considerable
amount of ethylated products (entry 16).9 With the exception
of inertness with the 2-hexenal-derived aldimine (entry 11),
Table 2 shows that this method proved to be effective in
substrate scope, giving high dr up to 99:1. Comparable yields
and dr values were obtained whether a Lewis acid was added
after or before treatment of aldimine with the enolate. While
varying aldimines (entries 4-6) or Lewis acids for 9 (Me3-
Al, 90.5:9.5 dr; i-Bu3Al, 94.5:5.5 dr; Me3Ga, 97:3 dr) had
appreciable effect, absolute S configuration at the emerging
chiral center was induced throughout. o,o-Disubstituted
aniline-derived aldimines 16-18 underwent no addition to
give complete recovery of the acetate.
of (R,R)-2, adduct 3 was obtained (3S:3R ) 85:15 () dr)),
with the absolute configuration being consistent with the
other examples we tested (entries 1-3 and 12-16). It should
be pointed out that the reactions showed a general preference
for si face attack of aldimines, which is opposite to that for
the addition of aldehydes.6
The chiral auxiliary can readily be recovered (>99%) from
Mannich adduct 3 (95.5:4.5 dr) by treatment with NH4OH
in THF,13 followed by subsequent methylation of the
resulting acid with TMSCHN2. Oxidative cleavage of the
methoxyphenyl group was effected by catalytic AgNO3 (0.29
equiv) in the presence of excess (NH4)2S2O8 (7.0 equiv)14 to
afford aminoester 22 in 53% yield (90% ee) (Scheme 2).
When we used CAN as an alternative oxidant, aminoester
20 underwent considerable dimerization, resulting in negli-
gible formation of 22.15 This can be circumvented by use of
14 as an aldimine to give Mannich adduct 19 in 63% yield
(96:4 dr). Sequential hydrolysis and methylation were
followed by oxidation of 21 with CAN (4 equiv) to give
22 in 91% yield (92% ee), where no dimerization was
observed.
In conclusion, this work features a strategy that enables
effective addition of acetate enolates to aldimines and
expands this scope to the asymmetric process using a chiral
auxiliary. We propose that the successful asymmetric reaction
originated in part from control of the two possible enolate
conformations W-form and U-form,16 which have been
The equilibrium between (E)- and (Z)-aldimines is known
to be involved in the presence of Lewis acid and to affect
the stereochemistry of the asymmetric Mannich-type reac-
tion.10 Thus, there is a possibility that the isomerization to
(Z)-aldimines could have a potential effect on the rate
acceleration of the Mannich addition. In fact, when the
SnCl4-(Z)-9 complex12 was exposed to the lithium enolate
(12) Rasmussen, K. G.; Hazaell, R. G.; Jorgensen, K. A. Chem. Commun.
1997, 1103. (b) Rasmussen, K. G.; Juhl, K.; Hazaell, R. G.; Jorgensen, K.
A. J. Chem. Soc., Perkin Trans. 2 1998, 1347.
(13) Hasegawa, T.; Yamamoto, H. Synlett 1998, 882. See also ref 6.
(14) Bhattarai, K.; Cainelli, G.; Panunzio, M. Synlett 1990, 229.
(15) Although it has been reported that aminoester 20 was readily
oxidized by CAN to give 22 (ref 8), we were unable to detect the formation
of 22 using CAN by varying numerous reaction conditions. No attempt
was made here to characterize the exact structure of the dimer. Competitive
dimerization is a significant problem in certain cases using CAN, see: Jacob,
P., III; Callery, P. S.; Shulgin, A. T.; Castagnoli, N., Jr. J. Org. Chem.
1976, 41, 3627. We also tried oxidative removal of the o-fluorophenyl group
from methyl 3-(2-fluorophenyl)amino-3-phenylpropionate (entry 7, Table
1) under conditions similar to those employed for 20 and 21 using CAN.
However, product 22 was formed in 14% yield.
(16) It was proposed that the boron enolates of R-unsubstituted ketones
favor the more stable U-form by 1-2 kcal/mol than the W-form. Similarly,
it is conceivable that the bulky phenoxy group of the enolate of 2 renders
the U-form most likely and gives a twist-boat transition structure: (a)
Gennari, C.; Todeschini, R.; Beretta, M. G.; Favini, G.; Scolastico, C. J.
Org. Chem. 1986, 51, 612. (b) Hoffmann, R. W.; Ditrich, K.; Froech, S.
Tetrahedron 1985, 41, 5517. (c) Braun, M. Angew. Chem., Int. Ed. Engl.
1987, 26, 24.
(9) To explain the difference in the reaction profiles using 0.5 and 1.0
equiv of Et2Zn, distinctive zincate species (enolate)R2ZnLi (Li-enolate:R2-
Zn ) 1:1) and (enolate)2R2ZnLi2 (Li-enolate:R2Zn ) 2:1) might be invoked.
At present, we have no evidence for the formation of these species and
further research is needed. Regarding the search on the reactivity difference
between R3ZnLi and R4ZnLi2, see: (a) Uchiyama, M.; Kameda, M.;
Mishima, O.; Yokoyama, N.; Koike, M.; Kondo, Y.; Sakamoto, T. J. Am.
Chem. Soc. 1998, 120, 4934.
(10) For example, the reaction of the aldimine derived from 3-trimeth-
ylsilyl-2-propynal with a ketene silyl acetal exhibits reversal in the absolute
configuration, compared with that of the aldimine derived from benzalde-
hyde using a chiral boron reagent. This implies that the (Z)-structure is the
reactive form for the former aldimine, whereas it is (E)-isomer for the latter,
see ref 3j. In good contrast, the present reaction using the o-anisidine-derived
aldimines derived from these two types of aldehydes showed an identical
S configuration.
(11) See Supporting Information for experimental details.
Org. Lett., Vol. 2, No. 13, 2000
1893