Catalysis of 3-pyrrolidinecarboxylic acid and related pyrrolidine derivatives in enantioselective anti-Mannich-type reactions: Importance of the 3-acid group on pyrrolidine for stereocontrol

Article abstract of DOI:10.1021/ja074907+

The development of enantioselective anti-selective Mannich-type reactions of aldehydes and ketones with imines catalyzed by 3-pyrrolidinecarboxylic acid and related pyrrolidine derivatives is reported in detail. Both (3R,5R)-5-methyl-3-pyrrolidinecarboxylic acid and (R)-3-pyrrolidinecarboxylic acid efficiently catalyzed the reactions of aldehydes with α-imino esters under mild conditions and afforded anti-Mannich products with high diastereo- and enantioselectivities (anti/syn up to 99:1, up to >99% ee). For the reactions of ketones with α-imino esters, (R)-3-pyrrolidinecarboxylic acid was an efficient catalyst (anti/syn up to >99:1, up to 99% ee). Evaluation of a series of pyrrolidine-based catalysts indicated that the acid group at the β-position of the pyrrolidine ring of the catalyst played an important role in forwarding the carbon-carbon bond formation and in directing anti-selectivity and enantioselectivity.

Full text of DOI:10.1021/ja074907+

Published on Web 12/29/2007
Catalysis of 3-Pyrrolidinecarboxylic Acid and
Related Pyrrolidine Derivatives in Enantioselective
anti-Mannich-Type Reactions: Importance of the 3-Acid
Group on Pyrrolidine for Stereocontrol
Haile Zhang,† Susumu Mitsumori,† Naoto Utsumi,† Masanori Imai,†
Noemi Garcia-Delgado,† Maria Mifsud,† Klaus Albertshofer,†
Paul Ha-Yeon Cheong,§ K. N. Houk,§ Fujie Tanaka,*,† and Carlos F. Barbas, III*,†
The Skaggs Institute for Chemical Biology and the Departments of Chemistry and Molecular Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, and
Department of Chemistry and Biochemistry, UniVersity of California, Los Angeles, California 90095-1569
Received July 3, 2007; E-mail: ftanaka@scripps.edu; carlos@scripps.edu
Abstract: The development of enantioselective anti-selective Mannich-type reactions of aldehydes and
ketones with imines catalyzed by 3-pyrrolidinecarboxylic acid and related pyrrolidine derivatives is reported
in detail. Both (3R,5R)-5-methyl-3-pyrrolidinecarboxylic acid and (R)-3-pyrrolidinecarboxylic acid efficiently
catalyzed the reactions of aldehydes with R-imino esters under mild conditions and afforded anti-Mannich
products with high diastereo- and enantioselectivities (anti/syn up to 99:1, up to >99% ee). For the reactions
of ketones with R-imino esters, (R)-3-pyrrolidinecarboxylic acid was an efficient catalyst (anti/syn up to
>99:1, up to 99% ee). Evaluation of a series of pyrrolidine-based catalysts indicated that the acid group at
the â-position of the pyrrolidine ring of the catalyst played an important role in forwarding the carbon-
carbon bond formation and in directing anti-selectivity and enantioselectivity.
Introduction
compounds that afford products with high enantioselectivities
have been performed using Zn-catalysts,2a-e Y-catalysts,2f Cu-
Mannich and Mannich-type reactions are important carbon-
carbon bond-forming reactions for the synthesis of amino acids,
amino alcohols, amino carbonyls, and their derivatives that
contain two adjacent stereocenters; accordingly there is a
demand for the direct catalytic reactions that afford syn- or anti-
Mannich products with high diastereo- and enantioselectivities.1-5
Because reactions that use unmodified carbonyl compounds as
nucleophile sources are more atom economical than those that
use preactivated carbonyl compounds, such as silyl enol ethers
or preformed enamines, development of the reactions that use
unmodified carbonyl compounds has been of interest. syn- or
anti-Selective Mannich-type reactions of unmodified carbonyl
catalysts,2g,h Pd-catalysts,2i cinchona alkaloids,2j,k and proline
and related amine-based organocatalysts.3 In Mannich-type
reactions of R-hydroxyketones, Zn-catalysts selectively afforded
anti- or syn-products depending on the protecting group on the
reactant imine nitrogen.2a-e Cu-catalysts,2g,h Pd-catalysts,2i and
cinchona alkaloids2j,k have been used for Mannich-type reactions
of â-ketoesters. Proline and related amine-based enamine
catalysis have been used for enantioselective Mannich-type
reactions of aldehydes and ketones including simple alkylalde-
hydes and alkanones, in which in situ-formed enamine inter-
mediates act as nucleophiles. When proline and related pyrro-
lidine derivatives that possess acidic groups at the R-position
of pyrrolidine are used as catalysts for these Mannich-type
reactions, typically syn-isomers are obtained as the major
products.3 Development of enamine-based Mannich-type reac-
tions of aldehydes and of ketones that afford anti-products with
high diastereo- and enantioselectivities is a current challenge.4,5
Since Mannich-type reactions with R-imino esters are especially
useful for syntheses of a variety of amino acid derivatives
(Scheme 1),1a-c,2g,3a-g,j-l,q,r-t development of anti-Mannich-type
reactions with R-imino esters is considerably important.4,5a-c
We have recently reported in communications that pyrrolidine
derivatives (3R,5R)-5-methyl-3-pyrrolidinecarboxylic acid and
(R)-3-pyrrolidinecarboxylic acid catalyze anti-Mannich-type
reactions of aldehydes4a and of ketones,4b respectively. To
provide information for the further development of efficient,
† The Scripps Research Institute.
§ University of California, Los Angeles.
(1) Enantioselective syn- or anti-selective Mannich-type reactions that use
silyl enolates or glycine imines as nucleophiles: (a) Ferraris, D.; Young,
B.; Dudding, T.; Lectka, T. J. Org. Chem. 1998, 63, 4584. (b) Ferraris, D.;
Young, B.; Cox, C.; Drury, W. J., III; Dudding, T.; Lectka, T. J. Org.
Chem. 1998, 63, 6090. (c) Ferraris, D.; Young, B.; Cox, C.; Dudding, T.;
Drury, W. J., III; Ryzhkov, L.; Taggi, A. E.; Lectka, T. J. Am. Chem. Soc.
2002, 124, 67. (d) Kobayashi, S.; Ishitani, H.; Ueno, M. J. Am. Chem.
Soc. 1998, 120, 431. (e) Kobayashi, S.; Hamada, T.; Manabe, K. J. Am.
Chem. Soc. 2002, 124, 5640. (f) Kobayashi, S.; Matsubara, R.; Nakamura,
Y.; Kitagawa, H.; Sugiura, M. J. Am. Chem. Soc. 2003, 125, 2507. (g)
Nakamura, Y.; Matsubara, R.; Kiyohara, H.; Kobayashi, S. Org. Lett. 2003,
5, 2481. (h) Hamada, T.; Manabe, K.; Kobayashi, S. J. Am. Chem. Soc.
2004, 7768. (i) Akiyama, T.; Itoh, J.; Yokota, K.; Fuchibe, K. Angew.
Chem., Int. Ed. 2004, 43, 1566. (j) Ooi, T.; Kameda, M.; Fujii, J.; Maruoka,
K. Org. Lett. 2004, 6, 2397. (k) Okada, A.; Shibuguchi, T.; Ohshima, T.;
Masu, H.; Yamaguchi, K.; Shibasaki, M. Angew. Chem., Int. Ed. 2005,
44, 4564. (l) Salter, M. M.; Kobayashi, J.; Shimizu, Y.; Kobayashi, S. Org.
Lett. 2006, 8, 3533. (m) Carswell, E. L.; Snapper, M. L.; Hoveyda, A. H.
Angew. Chem., Int. Ed. 2006, 45, 7230.
9
10.1021/ja074907+ CCC: $40.75 © 2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 875-886
875

A R T I C L E S
Scheme 1
Zhang et al.
Scheme 2
highly diastereo- and enantioselective organocatalytic methods
of chemical transformations, here we report the details of the
design of pyrrolidine-derived catalysts bearing acid groups at
the 3-position of pyrrolidine for anti-Mannich-type reactions
and the scope of the catalyzed anti-Mannich-type reactions,
including reactions with R-imino esters.
Results and Discussion
Design and Evaluation of Catalysts for anti-Mannich-Type
Reactions of Aldehydes. (S)-Proline catalyzes Mannich-type
reactions between unmodified aldehydes and N-p-methoxyphe-
nyl (PMP)-protected imine of ethyl glyoxylate and affords
(2S,3S)-syn-products with high enantioselectivities.3b,c The ster-
eochemical outcome of the proline-catalyzed reactions can been
explained by the mechanism shown in Scheme 2a.3c,6 With
proline, (E)-enamines predominate. The s-trans-enamine con-
formation (A) of the (E)-enamine is used for the C-C bond-
forming transition state (C); the re face of the enamine reacts
with the si face of the imine. The C-C bond-forming transition
(2) Enantioselective syn- or anti-selective Mannich-type reactions of hydrox-
yketones, â-ketoesters, a trichloromethyl ketone, or an imide that use other
than enamine catalysis: (a) Matsunaga, S.; Kumagai, N.; Harada, S.;
Shibasaki, M. J. Am. Chem. Soc. 2003, 125, 4712. (b) Matsunaga, S.;
Yoshida, T.; Morimoto, H.; Kumagai, N.; Shibasaki, M. J. Am. Chem. Soc.
2004, 126, 8777. (c) Yoshida, T.; Morimoto, H.; Kumagai, N.; Matsunaga,
S.; Shibasaki, M. Angew. Chem., Int. Ed. 2005, 44, 3470. (d) Trost, B.;
Terrell, L. R. J. Am. Chem. Soc. 2003, 125, 338. (e) Trost, B. M.;
Jaratjaroonphong, J.; Reutrakul, V. J. Am. Chem. Soc. 2006, 128, 2778. (f)
Sugita, M.; Yamaguchi, A.; Yamagawa, N.; Handa, S.; Matsunaga, S.;
Shibasaki, M. Org. Lett. 2005, 7, 5339. (g) Juhl, K.; Gathergood, N.;
Jorgensen, K. A. Angew. Chem., Int. Ed. 2001, 40, 2995. (h) Foltz, C.;
Stecker, B.; Marconi, G.; Bellemin-Laponnaz, S.; Wadepohl, H.; Gade, L.
H. Chem. Commun. 2005, 5115. (i) Hamashima, Y.; Sasamoto, N.; Hotta,
D.; Somei, H.; Umebayashi, N.; Sodeoka, M. Angew. Chem., Int. Ed. 2005,
44, 1525. (j) Lou, S.; Taoka, B. M.; Ting, A.; Schaus, S. E. J. Am. Chem.
Soc. 2005, 127, 11256. (k) Ting, A.; Lou, S.; Schaus, S. E. Org. Lett. 2006,
8, 2003. (l) Tillman, A. L.; Ye, J.; Dixon, D. J. Chem. Commun. 2006,
1191. (m) Morimoto, H.; Lu, G.; Aoyama, N.; Matsunaga, S.; Shibasaki,
M. J. Am. Chem. Soc. 2007, 129, 9588. (n) Cutting, G. A.; Stainforth, N.
E.; John, M. P.; Kociok-Kohn, G.; Willis, M. C. J. Am. Chem. Soc. 2007,
129, 10632.
(3) Enamine-based enantioselective syn-selective Mannich or Mannich-type
reactions that use aldehydes or ketones as nucleophiles: (a) Cordova, A.;
Notz, W.; Zhong, G.; Betancort, J.; Barbas, C. F., III. J. Am. Chem. Soc.
2002, 124, 1842. (b) Cordova, A.; Watanabe, S.; Tanaka, F.; Notz, W.;
Barbas, C. F., III. J. Am. Chem. Soc. 2002, 124, 1866. (c) Notz, W.; Tanaka,
F.; Watanabe, S.; Chowdari, N. S.; Turner, J. M.; Thayumanuvan, R.;
Barbas, C. F., III. J. Org. Chem. 2003, 68, 9624. (d) Chowdari, N. S.;
Ramachary, D. B.; Barbas, C. F., III. Synlett 2003, 1906. (e) Notz, W.;
Watanabe, S.; Chowdari, N. S.; Zhong, G.; Betancort, J. M.; Tanaka, F.;
Barbas, C. F., III. AdV. Synth. Catal. 2004, 346, 1131. (f) Chowdari, N.;
Suri, J.; Barbas, C. F., III. Org. Lett. 2004, 6, 2507. (g) Chowdari, N.;
Ahmad, M.; Albertshofer, K.; Tanaka, F.; Barbas, C. F., III. Org. Lett.
2006, 8, 2839. (h) List, B. J. Am. Chem. Soc. 2000, 122, 9336. (i) List, B.;
Pojarliev, P.; Biller, W. T.; Martin, H. J. J. Am. Chem. Soc. 2002, 124,
827. (j) Zhuang, W.; Saaby, S.; Jorgensen, K. A. Angew. Chem., Int. Ed.
2004, 43, 4476. (k) Westermann, B.; Neuhaus, C. Angew. Chem., Int. Ed.
2005, 44, 4077. (l) Enders, D.; Grondal, C.; Vrettou, M.; Raabe, G. Angew.
Chem., Int. Ed. 2005, 44, 4079. (m) Enders, D.; Vrettou, M. Synthesis 2006,
2155. (n) Enders, D.; Grondal, C.; Vrettou, M. Synthesis 2006, 3597. (o)
Yang, J. W.; Stadler, M.; List, B. Angew. Chem., Int. Ed. 2007, 46, 609.
(p) Fustero, S.; Jimenez, D.; Sanz-Cervera, J. F.; Sanchez-Rosello, M.;
Esteban, E.; Simon-Fuentes, A. Org. Lett. 2005, 7, 3433. (q) Janey, J. M.;
Hsiao, Y.; Armstrong, J. D., III. J. Org. Chem. 2006, 71, 390. (r) Cobb, A.
J. A.; Shaw, D. M.; Ley, S. V. Synlett 2004, 558. (s) Cobb, A. J. A.; Shaw,
D. M.; Longbottom, D. A.; Gold, J. B.; Ley, S. V. Org. Biomol. Chem.
2005, 3, 84. (t) Wang, W.; Wang, J.; Li, H. Tetrahedron Lett. 2004, 45,
7243.
state involving s-cis-enamine conformation B is disfavored
compared to transition state C. There are two main possiblities
for the predominant contribution of conformation A over B in
the C-C bond-forming transition state. One is position matching
between the nucleophilic carbon of the enamine and the
electrophilic carbon of the imine in A: When the imine is
protonated by the carboxylic acid, the enamine nucleophilic
carbon of conformation A is positioned within a suitable reaction
distance from the electrophilic carbon of the imine. The position
of the enamine nucleophilic carbon of conformation B, however,
(4) (a) Mitsumori, S.; Zhang, H.; Cheong, P. H.-Y.; Houk, K. N.; Tanaka, F.;
Barbas, C. F., III. J. Am. Chem. Soc. 2006, 128, 1040. (b) Zhang, H.;
Mifsud, M.; Tanaka, F.; Barbas, C. F., III. J. Am. Chem. Soc. 2006, 128,
9630. (c) Ramasastry, S. S. V.; Zhang, H.; Tanaka, F.; Barbas, C. F., III.
J. Am. Chem. Soc. 2007, 129, 288. (d) Cheong, P. H.-Y.; Zhang, H.;
Thayumanavan, R.; Tanaka, F.; Houk, K. N.; Barbas, C. F., III. Org. Lett.
2006, 8, 811. (e) Cordova, A.; Barbas, C. F., III. Tetrahedron Lett. 2002,
43, 7749.
(5) (a) Kano, T.; Yamaguchi, Y.; Tokuda, O.; Maruoka, K. J. Am. Chem. Soc.
2005, 127, 16408. (b) Franzen, J.; Marigo, M.; Fielenbach, D.; Wabnitz,
T. C.; Kjaersgaard, A.; Jorgensen, K. A. J. Am. Chem. Soc. 2005, 127,
18296. (c) Kano, T.; Hato, Y.; Maruoka, K. Tetrahedron Lett. 2006, 47,
8467. (d) Guo, Q.-X.; Liu, H.; Guo, C.; Luo, S.-W.; Gu, Y.; Gong, L.-Z.
J. Am. Chem. Soc. 2007, 129, 3790. (e) Cheng, L.; Wu, X.; Lu, Y. Org.
Biomol. Chem. 2007, 5, 1018.
(6) Bahmanyar, S.; Houk, K. N. Org. Lett. 2003, 5, 1249.
9
876 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

anti-Mannich-Type Reactions
A R T I C L E S
Chart 1
trans to avoid steric interactions between the 5-methyl group
of catalyst 1 and the imine in the C-C bond-forming transition
state. The trans-relationship between the 3-carboxylic acid and
5-methyl group should also block the approach of the imine to
the enamine from the undesired reaction face. Thus, catalyst 1
should catalyze the Mannich-type reactions via transition state
F that use s-trans-enamine conformation D of the (E)-enamine
intermediate (Scheme 2b). In this transition state, si-face of the
enamine reacts with the si-face of the imine. As we have
preliminary reported4a and further describe below, amino acid
1 is an excellent catalyst for the anti-selective Mannich-type
reactions of aldehydes and the product stereochemistry of the
1-catalyzed reaction accords with the C-C bond formation
through transition state F.
The acid group at the 3-position of 1 can be any functional
group that engages the proton transfer to the imine in the
transition state. The tetrazole derivative of (S)-proline, (S)-5-
pyrrolidine-2-yl-1H-tetrazole, has been used as an efficient
catalyst for Mannich reactions that can be catalyzed by proline,
and the reactions catalyzed by this catalyst affords Mannich
products with stereoselectivities similar to that obtained from
proline catalysis;3r therefore, we reasoned that ent-2 (Chart 1)
should also be an excellent anti-Mannich catalyst.
is not correctly positioned near the electrophilic carbon of the
imine for bond formation under the proton transfer from the
carboxylic acid to the imine. The other possibility is that
conformation B has unfavorable steric interactions between the
carboxylic acid and the enamine, and thus conformation A,
which does not have such unfavorable interactions, predominates
over conformation B. Although computational analysis of the
transition state of the C-C bond-forming step suggests transition
state C,6 it has not been clearly explained why enamine
conformation A is used more favorably than conformation B
in the C-C bond-forming step. Therefore, in the design of
catalysts of anti-selective Mannich-type reactions, we considered
both possibilities.
To alter the selectivity from syn to anti, the reaction face of
the enamine or imine must be reversed from that of the proline-
catalyzed reactions. Since the 2-carboxylic acid of proline
controls both enamine conformation and reaction faces of the
enamine and the imine, we reasoned that separation of the steric
and the acid roles of this group into two groups and placing of
the groups at appropriate positions on a pyrrolidine ring should
result anti-selectivity. We reasoned that a pyrrolidine derivative
bearing substituents at 2- and 4-positions (or at 3- and
5-positions) should afford anti-products in the Mannich-type
reactions: The acid group at the 3-position of pyrrolidine can
be any acid functional group that engages the proton transfer
to the imine in the transition state and the substituent at
5-position can be any functional group that cannot initiate proton
transfer. Based on these considerations, we designed (3R,5R)-
5-methyl-3-pyrrolidinecarboxylic acid (1) (Chart 1) as an anti-
selective catalyst for the Mannich-type reactions of aldehydes
(Scheme 2b). We hypothesized that the 5-methyl group of
catalyst 1 controlled the enamine conformation to be present
as s-trans conformation D rather than s-cis conformation E
because of unfavorable steric interactions between the 5-methyl
group and the enamine in conformation E. We reasoned that
the 3-carboxylic acid of catalyst 1 should control the reaction
faces of the enamine and the imine through the proton transfer
from the carboxylic acid to the imine nitrogen. The proton
transfer from the carboxylic acid to the imine nitrogen should
also accelerate the reaction by increasing the imine electrophi-
licity. The 3-carboxylic acid and the 5-methyl group should be
The 5-methyl group of catalysts 1 and ent-2 can be altered
upon the degrees of its contribution to the stereoselectivities of
the reactions. To determine the contribution of the carboxylic
acid at the 3-position and of the methyl group at the 5-position
of 1 to directing the stereochemical outcome of the reaction,
we reasoned to evaluate (2S,3R)-2-methyl-3-pyrrolidinecar-
boxylic acid (3), which has a methyl group at the 2-position on
the pyrrolidine, the unmethylated derivative (R)-3-pyrrolidin-
ecarboxylic acid (4), and the pyrrolidine derivative without
3-carboxylic acid, (R)-2-methylpyrrolidine (5). If the acid group
at the 3-position is the group that predominantly directs the
stereochemical outcome, the use of unmethylated derivative 4
should also afford anti-products with high diastereo- and
enantioselectivities. Because catalyst 4 is more readily accessed
than catalyst 1, this catalyst would be useful for anti-selective
Mannich-type reactions if this catalyst provide reactivities and
selectivities similar to catalyst 1. As described below, reactions
of many aldehyde nucleophiles catalyzed by 4 afforded anti-
Mannich products with high diastereo- and enantioselectivities.
With catalyst 4, both enamine conformations G and H may be
similarly present; that is, conformations G and H may have
similar free energies (Scheme 2c). However, only conformation
G should advance the C-C bond formation through transition
state I (Scheme 2c). The nucleophilic carbon of enamine G
should be properly positioned for the reaction with the elec-
trophilic carbon of the imine, whereas the nucleophilic carbon
of enamine H should be too far from the imine electrophilic
carbon to form a bond.
Catalysts 1-14 were evaluated in the Mannich-type reaction
of isovaleraldehyde and N-PMP-protected ethyl glyoxylate imine
(Table 1). The reaction catalyzed by (3R,5R)-5-methyl-3-
pyrrolidinecarboxylic acid (1) afforded (2S,3R)-anti-157 in a
good yield with excellent diastereo- and enantioselectivities
(7) Determination of the absolute stereochemistry of 15 generated by the
1-catalyzed reaction has been reported in our communication (ref 4a).
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 877

A R T I C L E S
Zhang et al.
Table 1. Evaluation of Catalysts for the anti-Selective Mannich-Type Reaction of Isovaleraldehyde to Afford 15a
catalyst load
(equiv)
time
(h)
yieldb
(%)
drc
anti/syn
major
anti-enantiomer
eed
(%)
entry
catalyst
1
2
3
4
1
0.05
0.05
0.05
0.05
0.05
0.2
0.2
0.2
0.2
0.2
3
3
6
4
3
85
87
75
81
87
16
20
21
19
9
20
83
93
35
<20
51
82
62
98:2
98:2
97:3
99:1
(2S,3R)
(2R,3S)
(2S,3R)
(2S,3R)
(2R,3S)
(2R,3S)
(2R,3S)
(2R,3S)
(2R,3S)
99
99
91
94
ent-2
3
4
ent-4
ent-5
5
99:1
93
6e
7e
8e
9e
10e
11e
12
13
14
15
16e
17f
18
72
72
72
72
72
72
2.5
4
24
14
30
14
4
75:25
80:20
67:33
64:36
63:37
89:11
94:6
87:13
93:7
75:25
78:22
1:1.4
10:90
29 (19)
40 (18)
54 (40)
36 (46)
ent-5 + CF3CO2H
ent-5 + 6
ent-5 + 7
7
pyrrolidine
ent-8
9
10
11
12
13
14
0.2
0.05
0.05
0.3
0.3
0.3
0.3
0.3
(2R,3S)
(2S,3R)
(2S,3R)
85
94 (95)
20
<5 (20)
36 (12)
98 (>99)
76 (80)
(2S,3R)
(2S,3R)
(2S,3R)
a Typical reaction conditions: To a solution of N-PMP-protected R-imino ester (0.25 mmol, 1.0 equiv) and aldehyde (0.5 mmol, 2.0 equiv) in anhydrous
DMSO (2.5 mL), catalyst (0.0125 mmol, 0.05 equiv, 5 mol % to the imine) was added, and the mixture was stirred at room temperature (25 °C). b Isolated
yield containing anti- and syn-15. c The diastereomeric ratio (dr) was determined by 1H NMR or HPLC. d The ee of anti-15 was determined by chiral-phase
HPLC analysis. The ee of syn-15 is indicated in parenthesis. e When catalyst 0.05 equiv was used, no product formation was detected on TLC after 3 h.
f Taken from ref 4d.
(anti/syn 98:2, 99% ee, entry 1).8 The reaction using tetrazole
derivative ent-2 afforded the corresponding product enantiomer,
(2R,3S)-anti-15, with the same degree of diastereo- and enan-
tioselectivities as the reaction using catalyst 1. Catalysts 1 and
ent-2 showed essentially the same level of catalytic efficiency.
Reactions catalyzed by 1- and by ent-2 were approximately 2-
to 3-fold faster than the corresponding proline-catalyzed reaction
that afforded the syn-product. These results indicate that the
tetrazole group at the 3-position of catalyst ent-2 functions as
efficiently as the carboxylic acid at the 3-position of catalyst 1.
Although the reaction catalyzed by 3, which has a methyl
group at the 2-position on the pyrrolidine, was slightly slower
than the reaction catalyzed by 1, there was not a significant
difference in rates of the reactions (entry 3 versus entry 1). The
reaction catalyzed by 3 afforded the same anti-product (2S,3R)-
15 as that by catalyst 1 with the same degree of diastereose-
lectivity and slightly lower enantioselectivity (91% ee, entry 3)
compared to the reaction catalyzed by 1 (99% ee, entry 1). These
results suggest that both s-trans and s-cis-enamine conformations
are present in the reaction catalyzed by 3 and that the C-C
bond formation is controlled by proton transfer from the
3-carboxylic acid group to the imine (Scheme 2d). Regardless
of the position of the methyl group of the catalysts, either at
the 5-position or at the 2-position, the reaction faces of the
enamine and the imine were the same in the formation of the
major product. These results indicate that a methyl group at
either the 5-position or the 2-position of the pyrrolidine ring of
the catalyst does not significantly effect the enamine conforma-
tion. Although the 5-methyl group of catalyst 1 contributes to
afford almost perfect anti-selectivity and enantioselectivity in
the 1-catalyzed reaction, no significant steric interaction is
present between the enamine moiety and the methyl group at
either the 5-position or the 2-position. On the other hand, the
results indicate that the 3-acid group of the catalysts plays an
important role in directing the stereochemical outcome. Sig-
nificant contribution of the 3-acid group on the pyrrolidine ring
of the catalysts in the stereocontrol of the reaction was also
supported from results of the reaction catalyzed by unmethylated
catalyst (R)-3-pyrrolidinecarboxylic acid (4). The reaction with
catalyst 4 afforded the same anti-product (2S,3R)-15 as did the
reaction with 1, and the enantiomeric excess of the anti-product
(2S,3R)-15 obtained using catalyst 4 was only slightly lower
(94% ee, entry 4) than that with catalyst 1. The major product
obtained from the 4-catalyzed reaction was in accord with a
mechanism of C-C bond formation through transition state I
(Scheme 2c).
The 3-acid group was essential not only for directing the
product stereochemistry, but also for reaction progress: The
reaction using (S)-2-methylpyrrolidine (ent-5), which lacked an
acid group, did not afford the Mannich product after 3 h when
0.05 equiv (i.e., 5 mol %) of catalyst to the imine was used.
Note that the reaction with 5 mol % of either 1 or ent-2 afforded
the anti-Mannich product in a good yield within a few hours.
Reaction with a higher loading of ent-5 (0.2 equiv) afforded
the Mannich product, but the diastereo- and enantioselectivities
were moderate (entry 6). The low yield of the reaction with
ent-5 was not improved by longer reaction time: A longer
reaction time resulted in decomposition of the imine and
Mannich product and formation of byproducts. Addition of acid
(equivalent to the catalyst), such as CF3CO2H, methyltetrazole
(6), or cyclopentanecarboxylic acid (7), to the ent-5-catalyzed
(8) To obtain the Mannich products generated from the reactions of aldehyde
nucleophiles in an excellent dr, quick and careful workup and purification
procedures were required. For example, the dr values of the products slightly
(up to 4% variation) changed during silica gel flash column chromatography.
Diastereomers of 15 were inseparable by typical silica gel flash column
chromatography (see Supporting Information).
9
878 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

anti-Mannich-Type Reactions
A R T I C L E S
Scheme 3
Table 2. Isomerization of syn-15 to anti-15 Catalyzed by
Pyrrolidinesa
catalyst load
(equiv)
catalyst
solvent
time
syn-15/anti-15b
pyrrolidine
0.2
0.2
0.1
0.1
CDCl3
0 min
4 h
24 h
0 min
4 h
24 h
0 min
3 h
24 h
0 min
2 h
96:4c
1:1
1:2.5d
96:4c
2:1
ent-5
CDCl3
1:2d
4
4
CDCl3
96:4c
95:5
83:17e
96:4c
96:4
95:5
DMSO-d6
reaction did not meaningfully enhance the reaction rate; the
diastereo- and enantioselectivities of the reactions with these
acids were also moderate (entries 7-9). Reaction catalyzed by
cyclopentanecarboxylic acid (7) or by pyrrolidine also afforded
the anti-product as the major isomer (entries 10 and 11,
respectively). Since enantioselectivity was observed (although
low) in the ent-5-catalyzed reactions, the 2-methyl group plays
some role in the stereocontrol. The stereodirecting effect of the
2-methyl group of ent-5 was in accord with that of the 3-acid
group of catalyst 4. A pyrrolidine derivative with a bulky
2-substituent has been demonstrated as an anti-selective catalyst
in Mannich-type reactions of aldehydes,5b but catalyst ent-5,
pyrrolidine with a 2-methyl group alone, did not provide high
levels of stereocontrolling effects. These results indicate that
the 3-acid group of 1 plays a significant role in directing the
stereochemistry and in efficient reaction progress through proton
transfer to the imine intramolecularly. The 5-methyl group of
catalyst 1 cooperates with the stereodirecting effect of the 3-acid
group to provide perfect anti-selectivity and enantioselectivity.
The anti-product can be generated by isomerization of the
syn-product at the R-position of the aldehyde group (the
3-position of 15). To test whether the isomerization of the
product occurs under the Mannich-type reaction conditions, syn-
rich 15 generated by the (S)-proline-catalyzed reaction was
treated with catalysts. First, syn-rich 15 was treated with
pyrrolidine under conditions similar to the catalyzed reactions
in Table 1, entry 11, and the ratios of the diastereomers and
enantiomers were analyzed. After 1 day, the anti-isomer became
the major isomer (anti/syn ) 1.6:1) as shown in Scheme 3;
(2S,3S)-syn-15 was converted to (2S,3R)-anti-15 in the presence
of pyrrolidine. Pyrrolidine isomerized the Mannich products
either through enamine formation with the product and/or by
acting as a base for enolization. This result indicates that the
anti-isomer is more thermodynamically stable than the syn-
isomer. The anti-isomer was also more thermodynamically
favored in imidazole isomerization:4a,9 In the presence of
imidazole, the isomerization rate of 15 from syn to anti was
faster than that from anti to syn. The ratio of the diastereomers
obtained from the pyrrolidine-catalyzed reaction that afforded
the anti-isomer as the major product may reflect the conse-
quences of isomerization; the pyrrolidine-catalyzed reaction may
not form the anti-product by kinetic control in the C-C bond-
forming step.
7 h
a Conditions: Mannich product 15 (0.1 M) and catalyst (0.2 or 0.1 equiv
to 15 as indicated). b The ratio of syn-15/anti-15 was determined by 1H
NMR. c Before addition of catalyst. d Includes decomposed products ∼20%.
e Includes decomposed products ∼5%.
Diastereomeric ratio of 15 obtained from the ent-5-catalyzed
reaction may also reflect the result of isomerization of the
kinetically formed product. On the other hand, isomerization
with 4, which possesses 3-carboxylic acid, was significantly
slower than that with pyrrolidine or with ent-5. Isomerization
with 4 was negligible over the time range of the 4-catalyzed
Mannich-type reaction. The negligible isomerization of the
Mannich product in the presence of 4 also supports that the
anti-selectivity of the 1- or 4-catalyzed reaction is the result of
the kinetic control at the C-C bond-forming step. In the
isomerization with pyrrolidine or with ent-5, basic environs may
contribute to acceleration of the isomerization rates.
To further understand the contribution of the 3-acid group
of the catalysts, 3-substituted pyrrolidines, including ent-8 and
9, 3-hydroxypyrrolidine (10), and 3-pyrrolidinecarboxylic acid
methyl ester (11), were also evaluated. The sulfonamide group
that is present in catalysts ent-8 and 9 has been used as an acid
functionality in the enamine-based catalysts for Mannich
reactions.3t,5a,c Catalysts ent-8 and 9 (entries 12 and 13) had
catalytic activity similar to 1, 2, 3, and 4, indicating that the
sulfonamide group is also a good acid for catalysis in the anti-
selective Mannich-type reaction. Since the anti-selectivity and
enantioselectivity varied depending on the catalyst, the results
also indicate that the distance and orientation between the 3-acid
group of a catalyst, enamine nucleophilic carbon, and imine
electrophilic carbon in the transition state affect to the anti-
selectivity and enantioselectivity. 3-Hydroxypyrrolidine (10) was
not a good catalyst (entry 14), indicating that hydroxyl group
at the 3-position was not an efficient acid for this reaction.
Reaction with methyl ester 11 was slow (entry 15), also
supporting the importance of the acid group on pyrrolidine for
catalysis.
Reaction catalyzed by 3-piperidinecarboxylic acid (12), a six-
membered ring catalyst, was slow and afforded the product with
moderate anti-selectivity and enantioselectivity (entry 16),
indicating the importance of the five-membered pyrrolidine ring
structure for efficient catalysis and high selectivity. We previ-
ously reported the (S)-pipecolic acid (13)-catalyzed reaction that
afforded both anti- and syn-isomers (2S,3R)-15 and (2S,3S)-15
(anti/syn 1:1.4) with high enantioselectivities (entry 17).4d
Computational analysis of the transition states of the C-C bond
formation step of the 13-catalyzed reaction suggested that the
both s-cis- and s-trans-enamines were similarly used for the
Next, isomerization of syn-15 to anti-15 in the presence of
pyrrolidine, ent-5, or 4 in CDCl3 or in DMSO-d6 was analyzed
by 1H NMR (Table 2). The isomerization rate with ent-5, which
possesses 2-methyl group, was slower than that with pyrrolidine
but the anti-isomer became the major isomer after 1 day.
(9) Ward, D. E.; Sales, M.; Sasmal, P. J. Org. Chem. 2004, 69, 4808.
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 879

A R T I C L E S
Zhang et al.
Table 3. Evaluation of Catalysts for the anti-Selective
Scheme 4
Mannich-Type Reaction of 3-Pentanonea
time
(h)
yieldb
(%)
drc
anti/syn
major
anti−enantiomer
eed
(%)
entry
catalyst
1
2
3
4
1
4
72
29
72
17
<10
75
83
94:6
94:6
75:25
(2S,3R)-16
(2R,3S)-16
(2S,3R)-16
97
85
ent-8
9
98
97 (syn 97)
a To a solution of N-PMP-protected R-imino ester (0.5 mmol, 1.0 equiv)
and 3-pentanone (2.0 mL, 19 mmol, 38 equiv) in anhydrous DMSO (3.0
mL), catalyst (0.1 mmol, 0.2 equiv, 20 mol % to the imine) was added,
and the mixture was stirred at room temperature (25 °C). b Isolated yield
containing anti- and syn-16. c Determined by HPLC before purification.
d Determined by chiral-phase HPLC for anti-16.
ketone reaction originated from relatively slow formation of
the enamine intermediates due to steric interactions with the
methyl group of the catalyst (Scheme 4a). We reasoned that
unmethylated catalyst 4 should afford anti-Mannich products
in the ketone reactions. Although enamine conformations J and
K may have similar free energies, only J should advance to
the C-C bond formation via transition state L (Scheme 4b).
When proton transfer occurs from the acid at the 3-position of
the catalyst to the imine nitrogen, the nucleophilic carbon of
enamine J should be properly positioned to react with the imine,
whereas the nucleophilic carbon of enamine K should be too
far from the imine electrophilic carbon to form a bond. Since 4
does not have an R-substituent on the pyrrolidine, neither
enamine J nor K has a disfavored steric interaction like the
one shown in Scheme 4a and enamine formation of ketones
with 4 should be faster than that with 1.
In fact, the reaction catalyzed by 4 was significantly faster
than the 1-catalyzed reaction and afforded anti-Mannich product
(2S,3R)-1610 in good yield with high diastereo- and enantiose-
lectivities (entry 2, anti/syn 94:6, anti 97% ee). Catalyst ent-8,
which possesses a sulfonamide group, also catalyzed the
formation of the anti-product (entry 3), but the reaction catalyzed
by 4 was faster and showed higher enantioselectivity than the
ent-8-catalyzed reaction. Catalyst 9, which also has a sulfona-
mide group, catalyzed the reaction more efficiently than ent-8
did, and the ee of the product anti-16 obtained from the
9-catalyzed reaction was excellent (entry 4). These results
indicate that the acid functionality at the 3-position on the
pyrrolidine ring plays an important role in directing stereo-
chemistry in the reactions of ketones as it did in the reactions
of aldehydes. The acid group contributes to proper positioning
of the imine for efficient catalytic activity and high anti-
selectivity and enantioselectivity. Because catalyst 4 afforded
the best results with respect to anti-selectivity and enantiose-
lectivity, this catalyst was further investigated for the anti-
Mannich-type reactions of ketones.
C-C bond formation in the 13-catalyzed reaction and that thus
both syn and anti products formed.4d Reaction catalyzed by
2-azetidinecarboxylic acid (14), a four-membered ring catalyst,
afforded syn-products as the major product. Stereoselectivity
of the product depended on the ring size of the amine catalyst
and on the position of the acid.
The results of the catalyst evaluation indicate that the acid
group at the 3-position of catalyst 1 plays a dominant role in
forwarding reaction (efficient catalysis) and in directing stere-
ochemistry through proton transfer to the imine. The 5-methyl
group of catalyst 1 cooperatively contributes to the stereocon-
trolling effect of the 3-acid group to provide anti-products with
excellent enantioselectivity as almost single enantiomer. The
5-methyl group of catalyst 1 has little effect on controlling the
enamine conformation: There is no significant steric role of
R-methyl group on pyrrolidine to control enamine conformation.
The (E)-enamine of 1 (or 4) is present as either s-trans D or
s-cis E (or s-trans G or s-cis H), but only D (or G) advances
to the C-C bond formation through F (or I). The preference
of conformation D over E in the transition state of the C-C
bond-forming step originates mainly from the proton transfer
from the 3-acid group to the imine. Proper positioning of the
enamine and the imine for efficient catalytic activity, high anti-
selectivity, and high enantioselectivity is achieved by the proton
transfer from the 3-carboxylic acid to the imine nitrogen.
Of the catalysts evaluated, catalysts 1 and ent-2 were the most
efficient catalysts for the anti-Mannich-type reactions of alde-
hydes with respect to the catalytic efficiency, anti-selectivity,
and enantiostereoselectivity. When accessibility of catalysts was
also considered, 4 had advantages. Therefore, catalysts 1 and 4
were further studied for the anti-Mannich-type reactions of
aldehydes.
Design and Evaluation of Catalysts for anti-Mannich-Type
Reactions of Ketones. Compared to the reactions of aldehydes,
the reactions of ketones require additional consideration of steric
interactions in formation of enamine intermediates. Although
amino acid 1 was an excellent anti-selective catalyst for the
Mannich-type reactions of aldehydes, the reaction of 3-pen-
tanone in the presence of catalyst 1 was very slow (Table 3,
entry 1). Whereas the R-methyl group on pyrrolidine did
not hinder the formation of enamine intermediates with alde-
hydes, we reasoned that the low efficiency of catalyst 1 in the
Mannich-Type Reactions of Aldehyde Donors Catalyzed
by 1. Solvents were evaluated for the 1-catalyzed Mannich
reaction to afford (2S,3R)-anti-15 in a good yield with high
diastereo- and enantioselectivities in a short reaction time (Table
(10) Determination of the absolute stereochemistry of 16 generated by the
4-catalyzed reaction has been reported in our communication (ref 4b).
9
880 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

anti-Mannich-Type Reactions
A R T I C L E S
Table 4. Solvent Effects on the anti-Mannich-Type Reaction of an
Aldehyde Catalyzed by 1a
Table 5. anti-Mannich-Type Reactions of Aldehydes Catalyzed
by 1a
catalyst load time
yieldb
(%)
drc
anti/syn
eed
(%)
entry
R
(equiv)
(h)
product
time
(h)
yieldb
(%)
drc
anti/syn
eed
(%)
1
2
3
4
Me
0.05
0.05
0.3
0.05
0.02
0.01
0.05
0.02
0.02
1
3
1
(2S,3R)-17 70 94:6 >99e
(2S,3R)-15 85 98:2 99
entry
solvent
i-Pr
i-Pr
n-Bu
(2S,3R)-15 86 93:7 >99
1
2
3
4
5
6
7
8
9
DMSO
DMF
CH3CN
EtOAc
dioxane
CH2Cl2
THF
3
3
3
6
5
85
78
76
80
77
85
45
40
35
98:2
97:3
96:4
94:6
97:3
91:9
91:9
90:10
94:6
99
98
96
96
95
93
83
84
77
0.5 (2S,3R)-18 54 97:3
99
99
5f n-Bu
6f n-Bu
1
2
3
3
3
(2S,3R)-18 71 97:3
(2S,3R)-18 57 97:3 >99
(2S,3R)-19 80 97:3 >99
(2S,3R)-20 72 96:4 >97
(2S,3R)-21 84 99:1 >99
7
n-Pent
8f CH2CHdCH2
3
9
CH2CH
CH(CH2)4CH3
20
20
5
Et2O
[bmim]BF4
10 t-Bu
0.1
48
(2S,3R)-22 57 86:14
77
a Typical reaction conditions: To a solution of N-PMP-protected R-imino
ester (0.25 mmol, 1.0 equiv) and aldehyde (0.5-1.0 mmol, 2-4 equiv) in
anhydrous DMSO (2.5 mL), catalyst 1 (0.0125 mmol, 0.05 equiv, 5 mol %
to the imine) was added, and the mixture was stirred at room temperature.
b Isolated yield containing anti- and syn-diastereomers. c The dr was
determined by 1H NMR or HPLC. d The ee of anti-product was determined
by chiral-phase HPLC analysis. e The ee was determined by HPLC analysis
of the corresponding oxime prepared with O-benzylhydroxylamine. f The
reaction was performed in a doubled scale.
a Typical reaction conditions: To a solution of N-PMP-protected R-imino
ester (0.125 mmol, 1.0 equiv) and aldehyde (0.5 mmol, 4.0 equiv) in
anhydrous solvent (1.25 mL), catalyst (0.0063 mmol, 0.05 equiv, 5 mol %
to the imine) was added, and the mixture was stirred at room temperature.
b Isolated yield containing anti- and syn-diastereomers. c The dr was
determined by 1H NMR. d The ee of (2S,3R)-anti-15 was determined by
chiral-phase HPLC analysis.
4). The reaction in DMSO provided the best anti-selectivity and
enantioselectivity among those solvents tested (entry 1). Reac-
tions in DMF, CH3CN, EtOAc, and dioxane were as efficient
with respect to reaction rate as that in DMSO and afforded high
anti-selectivity and enantioselectivity (anti/syn 94:6-97:3, 95-
97% ee, entries 2-5). Although ionic liquids were good solvents
for the proline-catalyzed Mannich reaction affording syn-15,3d
the 1-catalyzed reaction in ionic liquid [bmim]BF4 (bmim )
1-butyl-3-methylimidazolium) was slow (entry 9).
Table 6. anti-Mannich-Type Reactions of Aldehydes Catalyzed by
1: Variation of R-Imino Glyoxylatesa
The scope of the 1-catalyzed Mannich-type reaction with ethyl
glyoxylate imine was examined using a series of aldehydes
(Table 5). The reactions catalyzed by 1 afforded a series of anti-
Mannich products (2S,3R)-15 and (2S,3R)-17-21 with excellent
diastereo- and enantioselectivities (anti/syn 94:6-99:1, anti
>97->99% ee).8 With 5 mol % catalyst loading, the reaction
rates with catalyst 1 were approximately 2- to 3-fold faster than
the corresponding proline-catalyzed reactions that afford the syn-
products. Because of the high catalytic efficiency of 1, the
reactions catalyzed by only 1 or 2 mol % of 1 afforded the
desired products in reasonable yields within a few hours (entries
5, 6, 8, and 9). For the reaction of a bulky aldehyde 3,3-
dimethylbutanal, a higher loading of catalyst and a longer
reaction time were required to afford the Mannich product anti-
22 (entry 10) than for the reactions of R-unbranched aldehydes.
These anti-Mannich products were generally more resistant
to the isomerization at the R-position of the aldehyde carbonyl
group than corresponding syn-Mannich products, which were
often isomerized to anti-isomer3b,c or to R,â-unsaturated carbonyl
compounds11 (if applicable) during purification.8 Thus, forma-
tion of highly enantiomerically enriched Mannich products in
an anti-selective fashion was beneficial to access to amino acid
derivatives in more pure forms than formation of corresponding
syn-Mannich products.
time
(h)
yieldb
(%)
drc
anti/syn
eed
(%)
entry
R1
R2
product
1
2
3
4
i-Pr
n-Pent i-Pr
i-Pr
i-Pr
i-Pr
1
1
(2S,3R)-23
(2S,3R)-24
92
85
85
87
97:3
96:4 >99
>99:1
98:2
98
t-Bu
CH2CHdCH2
2.5 (2S,3R)-25
(2S,3R)-26
98
97
3
a Reaction conditions: To a solution of N-PMP-protected R-imino ester
(0.25 mmol, 1.0 equiv) and aldehyde (0.5-1.0 mmol, 2-4 equiv) in
anhydrous DMSO (2.5 mL), catalyst 1 (0.0125 mmol, 0.05 equiv, 5 mol %
to the imine) was added, and the mixture was stirred at room temperature.
b Isolated yield containing anti- and syn-diastereomers. c The dr was
determined by 1H NMR or HPLC. d The ee of anti-product was determined
by chiral-phase HPLC analysis.
The use of catalyst 1 was also examined in anti-selective
Mannich-type reactions of a series of N-PMP-protected gly-
oxylate imines to afford anti-products 23-26 (Table 6). The
anti-products were obtained in good yields and excellent
diastereo- and enantioselectivities. The reaction with a glyoxy-
late imine possessing a bulky group, tert-butyl, also proceeded
smoothly and afforded the Mannich product with excellent anti-
selectivity and enantioselectivity (entry 3).
Mannich-Type Reactions of Aldehyde Donors Using
Catalysts 4 and ent-4. Evaluation of solvents for the 4-catalyzed
Mannich-type reaction that affords (2S,3R)-anti-15 showed that
DMSO was the best among solvents tested with respect to yield,
anti-selectivity, and enantiomeric excess of the anti-Mannich
product (Table 7). Whereas the reaction in aprotic solvents
(11) Utsumi, N.; Zhang, H.; Tanaka, F.; Barbas, C. F., III. Angew. Chem., Int.
Ed. 2007, 46, 1878.
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 881

A R T I C L E S
Zhang et al.
Table 7. Solvent Effects on the anti-Mannich-Type Reaction of an
Aldehyde Using Catalyst 4a
catalyst 4, although the ee of the anti-product was moderate
(entry 9). By using 4 or its enantiomer ent-4, both enantiomers
of the anti-Mannich products were obtained in good yields with
high enantioselectivities.
Mannich-type reactions of R,R-disubstituted aldehydes cata-
lyzed by 4 afforded products 28 and 29 in good yields (Table
9), although the diastereo- and enantioselectivities were low to
moderate and the syn-product was the major diastereomer in
both cases. For the 4-catalyzed Mannich reactions of R-branched
aldehydes, use of 2-PrOH as the solvent provided good results
with respect to reaction rate and yield. These 4-catalyzed
Mannich-type reactions of R,R-disubstituted aldehydes were
significantly faster than the corresponding proline-catalyzed
reactions.3f
Enantioselective and asymmetric Mannich-type reactions
constitute key steps for simple access to enantiomerically
enriched â-lactams.3b,c,12 Using Mannich product (2R,3S)-anti-
15 obtained by the ent-4-catalyzed reaction, trans-â-lactam
(3S,4R)-30 was synthesized via 31 (Scheme 5). We previously
demonstrated that the corresponding syn-Mannich products
generated from (S)-proline-catalyzed reactions were easily
transformed to cis-â-lactams.3b,c With use of anti- or syn-
Mannich-type reactions, highly enantiomerically enriched â-lac-
tams with desired stereochemistries, either trans or cis, respec-
tively, can be readily obtained.
time
(h)
yieldb
(%)
drc
anti/syn
eed
(%)
entry
solvent
1
2
3
4
5
6
7
8
9
10
DMSO
DMF
CH3CN
2.5
3.5
4.5
8
8
7
6
8
2
10
87
84
81
65
72
78
68
68
52
70
95:5
95:5
92:8
93:7
91:9
92:8
88:12
93:7
90:10
92:8
93
91
88
87
82
79
78
83
53
86
N-methylpyrrolidone (NMP)
EtOAc
dioxane
CHCl3
toluene
2-PrOH
[bmim]BF4
a Reaction conditions: To a solution of N-PMP-protected R-imino ester
(0.125 mmol, 1.0 equiv) and aldehyde (0.25 mmol, 2.0 equiv) in anhydrous
solvent (1.25 mL), catalyst 4 (0.0125 mmol, 0.1 equiv, 10 mol % to the
b
imine) was added, and the mixture was stirred at room temperature. Isolated
yield containing anti- and syn-diastereomers. c The dr was determined by
HPLC. d The ee of (2S,3R)-anti-15 was determined by chiral-phase HPLC
analysis.
Mannich-Type Reaction of Ketone Donors Using Catalysts
4 and ent-4. Evaluation of solvents for the 4-catalyzed Mannich-
type of reactions between 3-pentanone and N-PMP-protected
ethyl glyoxylate imine showed that 2-PrOH was the best solvent
tested to afford (2S,3R)-anti-16 in good yield with high anti-
selectivity and enantioselectivity within a short reaction time
(Table 10, entry 9). The reactions in DMSO, DMF, and
N-methylpyrrolidone (NMP) also proceeded well and afforded
the anti-product with high selectivities. Whereas catalyst 4 was
soluble in DMSO, DMF, and NMP, catalyst 4 was not
completely soluble in CHCl3, CH3CN, dioxane, THF, and
EtOAc in the presence of excess 3-pentanone under conditions
of Table 10. Catalyst 4 was even less soluble in THF and EtOAc
(or THF-3-pentanone or EtOAc-3-pentanone). The slow reaction
rate and low yield in these solvents may originate from the low
solubility of the catalyst. Catalyst 4 was also not completely
soluble in alcohols, including 2-PrOH, but the reaction in
alcohols proceeded well. The reaction rate in 2-PrOH was
approximately 2-fold faster than that in DMSO and the reaction
in 2-PrOH provided less byproducts than that in DMSO. The
reaction in EtOH was also efficient, although the ee of the anti-
Mannich product was slightly lower than that in 2-PrOH. When
the reaction in MeOH (entry 11) was performed in the absence
of catalyst 4, no formation of the Mannich product was detected
after 48 h, indicating that the catalyst is necessary for the
reaction to proceed even in MeOH. Note that for the 4-catalyzed
reactions of aldehydes, 2-PrOH was not a good solvent; the ee
of the anti-Mannich product of the aldehyde reaction was
moderate in 2-PrOH (see Table 7).
afforded the anti-Mannich product with good enantioselectivity,
the reaction in 2-PrOH provided the anti-product with moderate
ee. When the reactions in entry 1 (in DMSO) and in entry 9 (in
2-PrOH) were performed in the absence the catalyst, no
formation of Mannich product 15 was detected after 2.5 h in
either reaction. After 6.5 h, in 2-PrOH, formation of trace
amounts of 15 (<5%, anti/syn ) 4:1) was detected by 1H NMR
of the crude mixture. Formation of 15 in the absence of a catalyst
was negligible for the time range of the catalyzed reaction shown
in Table 7. When purified syn-15 (syn/anti 96:4) in 2-PrOH
(0.1 M) was treated with catalyst 4 (0.1 equiv to 15), the syn/
anti ratio of 15 was 95:5 at 3 h and 87:13 at 6 h. Although the
isomerization rate in 2-PrOH was faster than that in CDCl3 or
in DMSO-d6 (see Table 2), changes in the dr in 2-PrOH were
negligible for the time range of the 4-catalyzed Mannich
reaction. These results suggested that background reaction
(product formation without involving the catalyst) and isomer-
ization of syn-15 with low ee (if any) to anti-15 were not main
reasons for the moderate ee of anti-15 obtained by the
4-catalyzed reaction in 2-PrOH and that anti-15 with moderate
ee in 2-PrOH was formed at the C-C bond-forming step. The
4-catalyzed reaction of isovaleraldehyde in 2-PrOH may partly
proceed without desired stereocontrolling participation of the
catalyst.
The scope of the Mannich-type reactions between aldehydes
and glyoxylate imines catalyzed by 4 and ent-4 was examined
in reactions that afforded products 15 and 17-27 (Table 8).
The anti-products were obtained with high diastereo- and
enantioselectivities in most cases (anti/syn 93:7-99:1, anti 90-
99% ee). In some cases, the enantioselectivity of the 4-catalyzed
reaction was slightly lower than that of the 1-catalyzed reaction.
However, reactions of aldehydes with a longer alkyl chain using
catalyst 4 afforded anti-Mannich products with excellent enan-
tioselectivities (entries 5-8). Reaction with a bulky aldehyde,
3,3-dimethylbutanal, proceeded efficiently in the presence of
The reactions of aldehydes in 2-PrOH may proceed without
desired stereocontrolling participation of catalysts in the C-C
bond-forming transition state. Enamine formation with ketones
(12) (a) Kobayashi, S.; Kobayashi, J.; Ishiani, H.; Ueno, M. Chem. Eur. J. 2002,
8, 4185. (b) Hata, S.; Iwasawa, T.; Iguchi, M.; Yamada, K.; Tomioka, K.
Synthesis 2004, 1471. (c) Iza, A.; Vicario, J. L.; Carrillo, L.; Badia, D.
Synthesis 2006, 4065.
9
882 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

anti-Mannich-Type Reactions
A R T I C L E S
Table 8. anti-Mannich-Type Reactions of Aldehydes Catalyzed by 4 and ent-4a
R1
R2
catalyst
time
(h)
product
yieldb
(%)
drc
anti/syn
eed
(%)
entry
1
2
3
4
5
6
7
8
Me
i-Pr
i-Pr
n-Bu
Et
Et
Et
Et
Et
Et
Et
Et
Et
i-Pr
i-Pr
t-Bu
4
4
4
4
3
2
3
3
3
1
16
3
(2S,3R)-17
(2S,3R)-15
(2R,3S)-15
(2S,3R)-18
(2S,3R)-19
(2S,3R)-20
(2S,3R)-21
(2S,3R)-27
(2S,3R)-22
(2S,3R)-23
(2R,3S)-23
(2S,3R)-25
(2S,3R)-26
75
81
79
60
80
78
83
87
88
82
78
82
85
93:7
99:1
99:1
99:1
99:1
99:1
98:2
96:4
95:5
98:2
98:2
99:1
98:2
96
94
93
ent-4
4
4
4
4
4
4
4
95
n-Pent
CH2CHdCH2
CH2CHdCH(CH2)4CH3
CH2Ph
t-Bu
i-Pr
i-Pr
i-Pr
>97
>97
99
99
63
91
90
94
95
9e
10
11
12
13
ent-4
4
4
3
2.5
3
i-Pr
CH2CHdCH2
a Typical reaction conditions: To a solution of N-PMP-protected R-imino ester (0.25 mmol, 1.0 equiv) and aldehyde (0.5 mmol, 2.0 equiv) in anhydrous
DMSO (2.5 mL), catalyst 4 or ent-4 (0.0125 mmol, 0.05 equiv, 5 mol % to the imine) was added, and the mixture was stirred at room temperature. b Isolated
yield containing anti- and syn-diastereomers. c The dr was determined by 1H NMR or HPLC. d The ee of anti-product was determined by chiral-phase HPLC
analysis. e Catalyst 4 (0.1 equiv) and DMSO (0.5 mL) were used.
Table 9. Mannich-Type Reactions of R,R-Disubstituted Aldehydes
Catalyzed by 4a
Table 10. Solvent Effects on the anti-Mannich-Type Reaction of
3-Pentanone Using Catalyst 4a
time
(h)
yieldb
(%)
drc
anti/syn
anti eed
(%)
syn eed
(%)
entry
R
solvent
product
yieldb
(%)
drc
anti/syn
eed
(%)
entry
solvent
time
1
2
3
Ph
2-PrOH
1
28
99
95
90
36:64
37:63
45:55
24
34
29
37
9
14
n-Pr 2-PrOH 28 29
n-Pr DMSO 72 29
1
2
3
4
5
6
7
8
DMSO
DMF
NMP
CHCl3
CH3CN
dioxane
THF
EtOAc
2-PrOH
EtOH
29 h
38 h
40 h
3 d
3 d
3 d
3 d
3 d
18 h
18 h
32 h
75
74
65
46
51
94:6
87:13
93:7
95:5
97:3
74:26
ND
97
97
96
97
95
71
ND
ND
98
92
87
a To a solution of N-PMP-protected R-imino ester (0.25 mmol, 1.0 equiv)
and aldehyde (0.5 mmol, 2.0 equiv) in solvent (2.5 mL), catalyst 4 (0.025
mmol, 0.1 equiv, 10 mol % to the imine) was added, and the mixture was
stirred at room temperature. b Isolated yield containing anti- and syn-
diastereomers. c The dr was determined by 1H NMR. d The ee was
determined by chiral-phase HPLC analysis.
30
<10
<10
90
91
79
ND
9
10
11
97:3
95:5
96:4
Scheme 5 a
MeOH
a Typical reaction conditions: To a solution of N-PMP-protected R-imino
ester (0.1 mmol, 1.0 equiv) and 3-pentanone (0.4 mL, 3.8 mmol, 38 equiv)
in anhydrous solvent (0.6 mL), catalyst 4 (0.02 mmol, 0.2 equiv, 20 mol %
to the imine) was added, and the mixture was stirred at room temperature.
b Isolated yield containing anti- and syn-diastereomers. c The dr of the
isolated product was determined by HPLC analysis. d The ee of (2S,3R)-
anti-16 was determined by chiral-phase HPLC analysis.
carboxylic acid of the catalysts to the imine in the C-C bond-
forming transition state of the reactions of ketones. 2-PrOH may
stabilize the charged transition states of enamine formation of
ketones and the transition state of the C-C bond formation of
the Mannich-type reactions of ketones, resulting in the faster
reaction rate in this solvent than in other non-alcohol solvents
tested.
The scope of 4-catalyzed Mannich-type reactions between a
variety of ketones and R-imino esters was examined in the
formation of anti-products 16 and 32-55 (Tables 11-13). In
most cases, anti-Mannich products were obtained in good yields
a Conditions: (a) (i) NaClO2, KH2PO4, 2-methyl-2-butene, t-BuOH/H2O;
(ii) TMSCHN2; (b) LHMDS; (c) previously reported, see ref 3b,c.
is difficult compared to that with aldehydes and the reactions
of ketones do not proceed without proper catalyst participation
even in alcohol solvents. As a result, the reactions of ketones
occur in 2-PrOH through the transition state that affords the
anti-Mannich product in a highly stereocontrolled manner. That
is, 2-PrOH does not interrupt the proton transfer from the
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 883

A R T I C L E S
Zhang et al.
Table 11. anti-Mannich-Type Reactions of Acyclic Ketones Catalyzed by 4a
entry
R1
R2
R3
time (h)
product
yieldb (%)
drc anti/syn
eed (%)
1
Et
Et
Et
Me
Me
Me
Et
Me
Me
Et
Et
20
48
20
96
5
16
16
32
33
34
ent-34
35
36
37
91
77
93
76
85f
81f
81f
85
68
80
78
97:3
97:3
>99:1
>99:1
97
98
95
2e
3
Et
t-Bu
Et
Et
Et
Et
Et
Et
Et
4
n-Pr
Me
Me
Me
Me
Me
Me
Me
82
5
∼10:1 (>99:1)g
∼10:1 (>99:1)g
∼10:1
90 (>99)g
6h
7
5
88 (99)g
92
10
14
14
24
24
8
9
10
11
CH2CHdCH2
(CH2)3Cl
(CH2)2CO2Et
(CH2)2CN
>95:5
91
84
93
90
>95:5
38
39
∼9:1
Et
∼9:1
a Typical conditions: To a solution of imine (0.5 mmol, 1.0 equiv) and ketone (5.0 mmol, 10 equiv) in 2-PrOH (1.0 mL), 4 (0.05 mmol, 0.1 equiv) was
added, and the mixture was stirred at 25 °C. b Isolated yield containing anti- and syn-diastereomers. c Determined by 1H NMR of isolated products. d Determined
by chiral-phase HPLC for the anti-product. e Ketone (4 equiv), 4 (0.05 equiv), at 4 °C. f Containing regioisomer (∼5-10%). g Data after crystallization are
shown in parentheses. The dr was determined by HPLC. h Catalyst ent-4 was used.
Table 12. anti-Mannich-Type Reactions of R-Functionalized
Acyclic Ketones Catalyzed by 4a
diastereo- and enantioselectivities of these products varied. Rates
of the 4-catalyzed reactions of these R-heteroatom-bearing
ketones were generally faster than those of the 4-catalyzed
reactions of alkylketones and than those of the (S)-proline-
catalyzed reactions of the same R-heteroatom-bearing ketones.
The 4-catalyzed reactions of methoxyacetone, benzyloxyacetone,
and hydroxyactone afforded anti-Mannich products as the major
products with high enantiomeric excess (entries 1, 2, and 5).
The diastereoselectivity of the reaction of hydroxyacetone was
poor compared to those of the reactions of alkylketones
catalyzed by 4. Use of DMSO as solvent for the reaction of
hydroxyacetone did not improve the result: When the reaction
was performed using the imine (1.0 mmol) and hydroxyacetone
(10 mmol) in the presence of catalyst 4 (0.2 mmol) in DMSO
(2.0 mL) at 25 °C, after 7 h product 42 was obtained in 67%
yield (anti/syn ) 1:2, anti-42 62% ee, syn-42 <5% ee).
To analyze the product formation in the absence of the
catalyst, the reactions of Table 12, entry 1 (the reaction of
methoxyacetone) and entry 3 (the reaction of hydroxyacetone)
were performed under the conditions of this table but in the
absence of the catalyst. After 2.5 h, no formation of the Mannich
product was detected for either reaction. When a mixture of
the imine (0.1 mmol) and hydroxyacetone (0.45 mmol) in
2-PrOH (0.2 mL), which was more concentrated than the
reaction in Table 12, entry 3, was stirred at room temperature
for 6.5 h, formation of a small amount of Mannich product 42
(∼10%, anti/syn ) 1:4) was observed by 1H NMR of the crude
mixture. Although the Mannich-type reaction of hydroxyacetone
proceeded in 2-PrOH without a catalyst, the reaction rate in
the absence of the catalyst was much slower than the rate of
the 4-catalyzed reaction. These results indicate that the stereo-
selectivities of 40 and 42 obtained in the 4-catalyzed reactions
shown in Table 12 were not affected by the background reaction
(product formation without involving the catalyst).
time
(h)
yieldb
(%)
drc
anti/syn
eed
(%)
entry
R1
R2
product
1
Me
Me
Me
Me
Me
Me
Me
Et
OMe
OBn
OH
OH
OH
OH
SMe
N3
OBn
OH
3
16
1
1
1
6
4
24
5
24
40
41
42
42
42
42
43
44
45
46
78
91
71
62
53
60
78
58
69
41
∼5:1
∼4:1
∼1:1
∼1:1
∼2:1
∼2:1
∼1:1
1.3:1
1.4:1
∼3:1
86
90
2e,f
3
80 (10)
84 (8)
90 (12)
78 (55)
48
69 (54)
45 (9)
8
4g
5h
6i
7f
8h,j,k
9f
CH2OBn
CH2OH
10f,l
a Typical conditions: To a solution of imine (0.25 mmol, 1.0 equiv)
and ketone (0.5 mmol, 2.0 equiv) in 2-PrOH (1.25 mL), 4 (0.025 mmol,
0.1 equiv) was added, and the mixture was stirred at 25 °C. b Isolated yield
containing anti- and syn-diastereomers. c Determined by 1H NMR of isolated
products. d Determined by chiral-phase HPLC for the anti-product. The ee
of syn-product is indicated in parenthesis. e Catalyst ent-4 was used. f 2-
PrOH (0.5 mL). g Ketone (10 equiv). h Reaction at 4 °C. i Catalyst 9 was
used. j Catalyst 4 (0.2 equiv). k imine (0.2 mmol), ketone (0.4 mmol), 4
(0.02 mmol), and 2-PrOH (2.0 mL). l Ketone (1.0 equiv).
with high diastereo- and enantioselectivities. For the reactions
of unsymmetrical methyl alkyl ketones, the reaction occurred
predominantly at the more substituted R-position of the ketones
(Table 11, entries 5-11). The regio-, diastereo-, and/or enan-
tiomeric purities of the anti-products were readily improved by
crystallization (Table 11, entries 5, 6; Table 13, entry 3). Catalyst
4 was also useful to catalyze the reactions of ketones bearing
functional groups, such as halide, ester, and cyano groups, in
their alkyl chains. The 4-catalyzed reactions of these function-
alized ketones afforded desired anti-Mannich products without
special reaction conditions (Table 11, entries 9-11).
Isomerization of Mannich product 42 in the reaction mixture
was not also the reason for the low dr of this product: When a
mixture of purified product 42 (0.2 mmol; anti/syn ) 1:2, anti-
42 62% ee, syn-42 <5% ee) and catalyst 4 (0.02 mmol) in
2-PrOH (0.4 mL) was stirred at 25 °C, the dr and ee values of
Reactions of R-heteroatom-bearing ketones were also cata-
lyzed by 4 (Table 12). In all cases, C-C bond formation
selectively occurred at the carbon bearing the heteroatom. The
9
884 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

anti-Mannich-Type Reactions
A R T I C L E S
Table 13. anti-Mannich-Type Reactions of Cyclic Ketones Catalyzed by 4a
entry
X
R
catalyst (equiv)
product
yieldb (%)
drc anti/syn
eed (%)
1e
2
3
4
5
6
7
8
9
CH2
CH2
CH2
CH2
S
S
O
C(OCH2)2
C(OCH2)2
(CH2)2
(CH2)3
Et
i-Pr
t-Bu
CH2CHdCH2
Et
Et
Et
Et
Et
Et
Et
0.1
47
48
49
50
51
51
52
53
53
54
55
96
94
92
95
78
71
82
87
80
80
65
>99:1
>99:1
>99:1
>99:1
>99:1
>99:1
>95:5
>99:1
>99:1
>95:5
>95:5
96
94
0.05
0.05
0.05
0.1
0.05
0.1
0.1
0.05
0.1
95 (99)f
95
99
97
86
97
96
84
18
10
11
0.1
a Typical conditions: Imine (0.5 mmol, 1.0 equiv), ketone (1.0 mmol, 2.0 equiv), 4 (0.05 mmol, 0.1 equiv or 0.025 mmol, 0.05 equiv), 2-PrOH (1.0 mL),
25 °C. b Isolated yield. c Determined by 1H NMR of isolated products. d Determined by chiral-phase HPLC of the anti-product. e Ketone (5.0 mmol, 10
equiv). f Data after crystallization.
Scheme 6
42 were unchanged at 1.5 and 4 h. Decomposition of 42 in this
mixture was approximately <2% at 1 h and <5% at 4 h as
estimated by 1H NMR of the crude mixtures. The decomposition
rate of 42 in the presence of catalyst 4 was similar to that in
the absence of the catalyst, and these rates were faster than the
decomposition rate of the Mannich product of 3-pentanone, 16.
However, the dr of 42 was not affected by the decomposition
for the time range used for the catalyzed reaction. When a
mixture of ethyl glyoxylate imine (0.1 mmol), hydroxyacetone
(0.2 mmol), and Mannich product 42 (0.1 mmol; anti/syn )
1:2) in 2-PrOH (0.4 mL) was stirred at 25 °C for 4 h, the imine
remained and the dr of 42 was not altered although some
decomposed products formed (∼10%). These results indicate
that Mannich product 42 did not facilitate isomerization of 42
or equilibrium/exchange between 42 and the starting materials
and did not act as a catalyst to form 42 for the time range of
the 4-catalyzed reaction.
In the 4-catalyzed reaction of hydroxyacetone, formation of
(Z)-enamine intermediate4c may be possible, because catalyst 4
does not have R-substituent on the pyrrolidine ring and because
the hydroxy group is not a bulky group. Contributions of the
(Z)-enamine and the (E)-enamine and/or contributions of
enamine conformations K and J (Scheme 4) to the C-C bond-
forming transition state may cause the low diastereoseoectivity
of the 4-catalyzed reaction of hydroxyacetone.
The 4-catalyzed reactions of (methylthio)acetone, 1-azido-
2-butanone, and R,R′-dibenzyloxyacetone also afforded Mannich
products, although diastereo- and enantioselectivities were
moderate (entries 7-9). For the reaction of dihydroxyacetone,
although 4 catalyzed the reaction with much faster rate than
proline did, the anti-product of the 4-catalyzed reaction was
almost racemic (entry 10). For highly enantioselective anti-
Mannich-type reactions of R-azide ketones, R,R′-dibenzyloxy-
acetone, and dihydroxyacetone, we are developing other catalyst
systems that will be reported separately.
For the reactions of six-membered cyclic ketones, use of only
5 mol % of catalyst 4 and 2 equiv of ketone to the imine afforded
anti-Mannich products in good yields with high diastereo- and
enantioselectivities within approximately 12 h (Table 13). The
reaction of a seven-membered cyclic ketone was also efficiently
Scheme 7
catalyzed by 4 and afforded anti-Mannich product with high
diastereoselectivity and good enantioselectivity (entry 10). For
the reaction between 2,2-dimethyl-1,3-dioxan-5-one3k,l and
N-PMP-protected R-imino ethyl glyoxylate, however, use of
catalyst 4 afforded the Mannich product with moderate diaste-
reoselectivity as almost racemic form (Scheme 6).
Structural differences in enamine intermediates with catalyst
4 may cause deviation of positioning of reacting carbons of
enamine and imine from those of the most stable, desired
transition state that affords anti-Mannich products as a single
enantiomer, resulting in less differentiation of energies between
transition states that afford different diastereomers and enanti-
omers.
Reactions with Imines Other Than r-Imino Esters Cata-
lyzed by 4. As described above, pyrrolidine derivative 4 is useful
for catalyzing enantioselective, anti-selective Mannich-type
reactions between many of enolizable aldehydes or ketones and
N-PMP-protected R-imino esters. To further explore the ap-
plicability of this catalyst, Mannich reactions with imines other
than R-imino esters were performed.
As shown in Scheme 7, amino acid 4 efficiently catalyzed
the reaction between 3-pentanone and an R-imino amide that
was generated from glyoxyamide13 and p-anisidine in situ; anti-
(13) Evans, D. A.; Aye, Y.; Wu, J. Org. Lett. 2006, 8, 2071.
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 885

A R T I C L E S
Scheme 8
Zhang et al.
efficient acceleration of the C-C bond formation and for
stereocontrol as it engages proton transfer to the imine nitrogen
at the C-C bond-forming step. In the reactions of aldehydes,
the 5-methyl group of catalyst 1 cooperatively contributed to
the stereodirecting effect of the 3-acid group to provide perfect
anti-selectivity and enantioselectivity. Catalysis and stereocon-
trolling function provided by 1 and 4 originated in favored
positioning of the nucleophilic carbon of the enamine and the
electrophilic carbon of the imine under proton transfer from
the 3-acid group of the catalyst to the imine nitrogen.
For reactions between ketones and N-PMP-protected R-imino
esters or R-imino amide, catalyst 4 was useful to afford Mannich
products with high anti-selectivity and enantioselectivity. Ke-
tones applicable to the 4-catalyzed reactions included acyclic
and cyclic ketones and functionalized ketones, although anti-
selectivity and enantioselectivity varied. Catalyst 4 also ef-
ficiently catalyzed the reactions of hindered ketones; these
reactions are generally difficult with catalysis of pyrrolidine
derivatives possessing R-substituents, such as proline.
Mannich product 56 was obtained in 93% ee. Note that when
(S)-proline was used as catalyst for this reaction, the reaction
rate was very slow (<10% after 4 days).
In the 1- or 4-catalyzed reactions of reactants that did not
provide products with high stereoselectivities, bond lengths and
angles of the enamines and imines may vary from those of the
best reactants; causing deviations from the perfect positioning
of the reacting carbons. As a result, several transition states may
be used in the bond formation, leading reduced selectivities.
Importantly, our designed catalysts were useful for the reactions
with R-imino esters and amide to provide highly diastereomeri-
cally and enantiomerically enriched amino acid derivatives
containing two stereocenters.
Whereas 4-catalyzed reactions of R-imino esters generally
afforded anti-Mannich products with high diastereo- and enan-
tioselectivities, catalyst 4 was less optimal for reactions of imines
of arylaldehydes with respect to enantioselectivities as shown
in Scheme 8: The anti-Mannich products, 57, 58, and 59, were
obtained with high diastereoselectivities but ee values were
moderate (reaction conditions were not optimized for these
reactions). In the previously reported (S)-proline-catalyzed
reaction, 57 was obtained with anti/syn ) 1:2 (syn 84% ee).3h
The absolute stereochemistry of the major enantiomer of anti-
57 generated by the 4-catalyzed reaction was the same as that
of the isomerized (at the carbonyl R-position)9 product of major
enantiomer of syn-57 generated by (S)-proline-catalyzed reaction.3h
This suggested that the configuration of the carbon bearing the
amino group of the major enantiomer of anti-product 57
generated by the 4-catalyzed reaction was S as shown in Scheme
8 (absolute configuration of syn-57 generated by (S)-proline-
catalyzed reaction was assumed by analogy). Formation of this
major enantiomer is also in accord with transition state shown
in Scheme 4b when the ester group is altered to p-nitrophenyl
group. For imine that did not lead to high stereoselectivities,
bond lengths and angles of the imines may differ from those of
the glyoxylate imines, which lead to high anti-selectivity and
enantioselectivity. As a result, transition states other than the
most favored one may be involved in the C-C bond-forming
step, leading to a mixture of enantiomers.
Catalysts 1 and 4 are the smallest known catalysts for
enantioselective anti-selective Mannich reactions of enolizable
aldehydes and ketones. Compared to other larger catalysts
containing binaphthyl groups5a,d or metal-coordinating groups,2a-c,e
our catalyst designs are atom-economical. Our catalysts catalyze
enantioselective anti-selective Mannich reactions under mild
conditions using a low catalyst load. As we have shown here
and in previous reports, anti- or syn-Mannich products can be
enantioselectively generated by using catalysts that have an acid
group at the 3-position or 2-position on a pyrrolidine ring.
Further fine-tuning of the acid group at the 2- or 3-position of
pyrrolidine and attachment of substituents at remaining positions
of the acid-containing pyrrolidines should provide useful, highly
diastereoselective and enantioselective catalysts for other Man-
nich and Mannich-type reactions and for reactions other than
Mannich reactions, such as aldol and Michael reactions.
Computational studies concerning Mannich-type reactions cata-
lyzed by the series of pyrrolidine derivatives described here are
currently under investigation by the Houk laboratory4a and will
be reported in due course.
Conclusion
We have developed enantioselective anti-selective Mannich-
type reactions of enolizable aldehydes and ketones with imines
catalyzed by designed pyrrolidine derivatives bearing acid
functional groups at the 3-position of the pyrrolidine.
For reactions between aldehydes and N-PMP-protected
R-imino esters, catalysts 1 and 4 were efficient catalysts to afford
Mannich products with high anti-selectivity and enantioselec-
tivity. Catalyst 2 was also an efficient catalyst for the reaction
as catalyst 1. For the 1- and 4-catalyzed Mannich reactions of
aldehydes, the 3-acid group on pyrrolidine was essential for
Acknowledgment. This study was supported in part by The
Skaggs Institute for Chemical Biology.
Supporting Information Available: Detailed experimental
procedures and product characterization, including synthesis of
catalysts ent-2, 3, and 9, Mannich-type reactions, and synthesis
of â-lactams. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA074907+
9
886 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008