Communications
Table 1: Results of the Mannich reaction with 5 and 6 (see Scheme 2).
results, TFE was once again the solvent of
choice. After only 10 min 10 was obtained in
72% yield with high diastereo- (90:10 d.r.)
Entry Catalyst
Conc.
Solvent[a]
t [h] T [8C] Yield [%][b] d.r.[c]
ee [%][d]
[mol%]
and enantioselectivities (95% ee).
A
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
d,l-proline
30
30
30
30
30
30
30
30
30
30
20
10
5
DMSO
DMSO
H2O
acetonitrile
toluene
BMIMBF4
formamide
TFE
TFE/H2O (95:5)
TFE/H2O (90:10) 20
TFE
TFE
TFE
TFE
TFE
TFE
TFE
20
20
96
40
40
40
20
20
20
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
0
46
37
34
52
52
39
54
72
76
76
68
75
70
44
36
28
91:9
91:9
93:7
93:7
92:8
–
decrease in the irradiating power led to a
lower yield, although the selectivities
remained the same. The use of DMSO or
formamide as solvent led to a significant
decrease in the yield and diastereoselectiv-
ity.[15]
The Mannich-type reaction could also
be carried out as a three-component process
(Scheme 3). The reaction of glyoxylate 14,
p-anisidine (15), and 5 yielded 10 in 30%
yield with selectivities comparable with
those of the two-component reaction.
The relative configuration of all the
products was determined by NMR spec-
troscopy. The syn conformation of the main
diastereomer can be concluded based on the
3J coupling constant (CHO-CHN) of 2.3 Hz
l-proline
l-proline
l-proline
l-proline
l-proline
l-proline
l-proline
l-proline
l-proline
l-proline
l-proline
l-proline
l-proline
l-proline
DMTC (12)
82
34
80
42
89:11 n.d.
95:5
97:3
95:5
95:5
96:4
95:5
93:7
97:3
96:4
96
99
96
95
99
95
91
97
97
20
20
20
20
20
40
30
30
30
À20
RT
86:14 n.d.
diamine (13) 30
no reaction
[a] BMIMBF4 =1-butyl-3-methylimidazolium tetrafluoroborate; TFE=2,2,2-trifluoroethanol. [b] Yields of
isolated products. [c] d.r. determined by GC–MS. [d] ee determined by HPLC (Chiralcel OD-H); n.d.: not
determined.
allowed the isolation of the product in 72% yield with
reproducible diastereoselectivity (97:3 d.r.) and enantioselec-
tivity (99% ee). A slight decrease in the selectivities (95:5 d.r.
and 96/95% ee (Table 1, entries 9, 10) was observed in TFE/
H2O solvent mixtures (95:5; 90:10). To decrease the amount
of proline, we ran further experiments in TFE with decreasing
concentrations of catalyst (Table 1, entries 11–13). We proved
that the amount of catalyst could be decreased significantly.
Only at 5 mol% of l-proline did the stereoselectivity
decrease quite drastically. Temperature effects were not
observed (Table 1, entries 14, 15).
Other proline-derived catalysts proved to be unsuitable.
l-5,5-Dimethylthiazolidine-4-carboxylic acid (DMTC; 12)
resulted in poor selectivities. No reaction occurred in the
presence of (S)-1-(2-pyrrolidinylmethyl)pyrroline triflate (13)
(Table 1, entries 16, 17). Disappointingly, imines 7 and 8 and
hydrazone 9[14] did not react with 5.
Scheme 3. Three-component reaction to 10.
and on NOE interactions.[16] This is consistent with results
from other Mannich-type reactions. Therefore, it is possible to
assume the same transition state proposed by Cordova.[6]
In conclusion nitrogen-containing carbohydrate deriva-
tives were obtained with very good stereoselectivities from
DHA derivatives. Furthermore it is possible to improve
existing protocols (shorter reaction times, less catalyst) by
using 2,2,2-trifluoroethanol as solvent and microwave-
assisted procedure.[17]
One disadvantage of many proline-catalyzed reactions is
the long reaction time. We therefore tried to accelerate the
reaction under the action of microwaves. The results of the
microwave-assisted organocatalytic Mannich reactions are
listed in Table 2. Consistent with the previously observed
Experimental Section
Solvent (1 mL) and l-proline (35 mg, 0.3 mmol) were stirred at room
temperature for 3 min, before 5 (130 mg, 1.0 mmol) was added. After
15 min 6 (207 mg, 1.0 mmol) was added. The reaction mixture was
stirred for 20 h. A saturated solution of NH4Cl (1.0 mL) was added,
and the mixture was extracted with ethyl acetate (2 ꢀ 15 mL). The
combined organic extracts were dried with Na2SO4, filtered, and
concentrated. The residue was purified by column chromatography
through silica gel (petroleum ether/ethyl acetate 3:1) to give the
product 10 as an oil. Rf = 0.8 (petroleum ether/ethyl acetate 3:2);
HPLC (Daicel Chiralcel OD-H, hexane/2-propanol 90:10,
1.0 mLminÀ1, l = 254 nm); [a]2D0 = À107.0 (c = 0.2, ethyl acetate);
1H NMR (CDCl3, 300 MHz): d = 1.24 (t, 3J = 7.1 Hz, 3H; Me), 1.44 (s,
3H; Me), 1.49 (s, 3H; Me), 3.73 (s, 3H; Me), 4.02 (d, 2J = 16.5 Hz, 1H;
CHH, 5-H), 4.13 (dq, J = 7.1, 10.8 Hz, 1H; OCH2CH3), 4.24 (dq, J =
7.1, 10.8 Hz, 1H; OCH2CH3), 4.30 (dd, J1 = 1.6 Hz, J = 16.5 Hz, 1H;
CHH, 5-H), 4.59 (d, J = 2,3 Hz, 1H; 2-H), 4.74 (dd, J = 1.6, 2.3 Hz,
1H; 3-H), 6.7–6.8 ppm (m, 4H; Ar-H); 13C NMR (CDCl3, 75.5 MHz):
d = 14.2 (q), 23.3 (q), 24.2 (q), 55.5 (q), 58.8 (d), 61.5 (t), 67.0 (t), 76.4
Table 2: Results of the microwave-assisted Mannich reaction of 5 and 6
in the presence of l-proline (30 mol%).[a]
Solvent
t [min]
E [W][b]
Yield [%][c]
d.r.[d]
ee [%][e]
TFE
TFE
TFE
DMSO
formamide
formamide
5
10
10
10
5
300
300
100
300
300
300
63
72
64
18
40
38
89:11
90:10
90:10
92:8
83:17
80:20
94
94
95
n.d.
n.d.
n.d.
10
[a] All reactions were carried out in sealed 10-mL tubes. [b] Maximal
irradiated power. [c] Yields of isolated products. [d] Determined by GC–
MS. [e] Determined by HPLC (Chiralcel OD-H).
4078
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 4077 –4079