mation). The influence of other protecting groups of imines
were also investigated (Table 1, entries 6–9). On replace-
ment of the tosyl group with the phenyl group, no ee value
was obtained, although the yield was satisfactory, which im-
plied that the interaction between the tosyl group and the
catalyst was crucial for stereo-induction (Table 1, entry 6).
On employing the 2-hydroxyphenyl and diphenylmethyl
protecting groups, no reaction occurred (Table 1, entries 7
and 8). The Mannich-type reaction of 4-methoxyphenyl-pro-
tected imine also proceeded with low stereoselectivity
(Table 1, entry 9).
surrounding to optimum. Moreover, when catalyst loading
was decreased from 10 mol% to 5 mol%, the enantioselec-
tivity of the reaction decreased slightly (Table 2, entry 6 vs.
2). Fortunately, when the dosage of additive 5b was in-
creased to 7.5 mol%, the results were generally maintained,
while further increasing the amount of acid 5b to 10 mol%
resulted in loss of yield (Table 2, entry 8 vs. 6, 7). The influ-
ence of the sulfonyl group of the imine was tested. Use of
the benzenesulfonyl group protected imine gave 92% ee,
whereas the sterically smaller mesyl group protected imine
afforded the product with greatly reduced enantioselectivity
(65% ee) and comparative yield (entry 9 vs. entries 8 and
10), which might result from the more flexible conformation
of the imine with a methanesulfonyl group. The poor results
of other imines without sulfonyl substituents (Table 1, en-
tries 6–9) clearly illustrated the important role of the sulfo-
nyl group on the stereoselectivity. Screening of other reac-
tion conditions identified that the optimal conditions were
5 mol% bisguanidine 1d, 7.5 mol% additive 5b, 0.1 mmol
a-isothiocyanato imide 2, and 0.2 mmol N-Ts-protected
imine 3a in 1.0 mL THF/CHCl3 (v/v, 1/1) at À208C (Table 2,
entry 8). In addition, the configuration of the product 4a
was assigned to be (4S, 5R) by comparison of the optical ro-
tation to the reported value of the corresponding enantio-
mer.[7c]
To further improve the enantioselectivity of the reaction,
some achiral additives were employed.[11] The Brønsted ba-
sicity of guanidine was crucial for the activation of the iso-
thiocyanato imide, since no product was detected if guani-
À
dine salt 1e with two BF4 counteranions was used. Interest-
ingly, weak acids were found to have positive effect on the
reaction. When p-CNC6H4CO2H (5b) was used as additive,
the enantioselectivity was improved, but the yield was re-
tained (94% yield, 97% ee, Table 2, entry 2). Compared
Table 2. Acidic additives screened in this reaction.
A series of representative N-Ts-protected imines were in-
vestigated under the optimized conditions (Table 3). A vari-
ety of aldimine substrates provided the corresponding prod-
ucts in good to excellent yields, with high diastereoselectiv-
ity (up to >95:5 d.r.) and excellent ee values (up to 99%).
The aromatic imines underwent the Mannich-type reactions
to yield the optically active adducts in excellent yields and
90–>99% ee. It is noteworthy that not only the electronic
properties of the substitutions at the aromatic ring, but also
the steric hindrance, had no obvious effect on the diastereo-
selectivity and enantioselectivity (Table 3, entries 1–17).
While the condensed-ring imines reacted smoothly with a-
isothiocyanato imide, giving the products with 96% ee and
93% ee (Table 3, entries 18 and 19), the a,b-unsaturated
imines showed a slightly reduced reactivity and diastereose-
lectivity (Table 3, entry 20). Heteroaromatic substituted var-
iants delivered the Mannich-type adduct with a high selec-
tivity of 95–98% ee (Table 3, entries 21 and 22). Both acyclic
and cyclic aliphatic imines could also be converted to the
corresponding adducts with excellent results (Table 3, en-
tries 23 and 24). Other substrates such as N-Ts-protected ke-
toimine also were explored, but the adducts were not detect-
ed.[13]
[b]
Entry[a] Acid
pKa
Yield
[%][c]
trans: cis[d] ee
[%][e]
1
C6H5CO2H (5a)
4.20
3.55
3.42
4.38
95 (4a)
94 (4a)
88 (4a)
96 (4a)
90 (4a)
92 (4a)
86 (4a)
92 (4a)
>95:5
>95:5
>95:5
>95:5
>95:5
>95:5
>95:5
>95:5
91
2
3
4
p-CNC6H4CO2H (5b)
p-NO2C6H4CO2H (5c)
p-MeC6H4CO2H (5d)
97
93
90
92
93
95
96
65
92
5
p-MeOC6H4CO2H (5e) 4.47
6[f]
7[g]
8[h]
9[h,i]
10[h,j]
p-CNC6H4CO2H (5b)
p-CNC6H4CO2H (5b)
p-CNC6H4CO2H (5b)
p-CNC6H4CO2H (5b)
p-CNC6H4CO2H (5b)
3.55
3.55
3.55
3.55
3.55
85 (4a’) >95:5
80 (4a’’) >95:5
[a] Unless otherwise noted, all reactions were carried out with
2
(0.1 mmol), 3a (0.2 mmol), 10 mol% additive, 10 mol% 1d in THF/
CHCl3 (v/v, 1/1, 1.0 mL) at À208C for 10 h. [b] Relative pKa in water.[12]
1
[c] Isolated yield of combined diastereomers. [d] Determined by H NMR
analysis. [e] ee of the trans isomer, measured by chiral HPLC using Chira-
cel ADH column. [f] 5 mol% catalyst loading, 5 mol% additive.
[g] 5 mol% catalyst loading, 10 mol% additive. [h] 5 mol% catalyst load-
ing, 7.5 mol% additive. [i] Mesyl group protected imine 3a’ was used.
[j] Benzenesulfonyl-protected imine 3a’’ was used.
with acid 5b, the slightly less acidic C6H5CO2H (5a), p-
MeC6H4CO2H (5d), and p-MeOC6H4CO2H (5e) led to less
improvement of the ee values (Table 2, entries 1, 4, 5, vs. 2),
whereas the more acidic p-NO2C6H4CO2H (5c) exhibited
lower reactivity and enantioselectivity (Table 2, entry 3 vs.
2). The influence of the protonic acid might by partly associ-
ated with the formation of a guanidinium salt, which could
activate the imines through hydrogen bonding.[10] The car-
boxylate anion adhering nearby might fine tune the stereo-
The low catalyst loading, mild reaction conditions, and
the inexpensive starting materials and available catalyst for
this Mannich-type reaction offered a practical way to scale-
up production. In the presence of 5 mol% of bisguanidine
1d, 7.5 mol% of p-CNC6H4CO2H (5b), 10 mmol of a-iso-
thiocyanato imide 2 reacted with 2.0 equivalents of N-Ts-
protected imine 3a to provide the desired product 4a in
82% yield without any loss in the enantioselectivity and dia-
stereoselectivity (Scheme 2).
2584
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2583 – 2586