Lewis acid-Lewis acid heterobimetallic cooperative catalysis: Mechanistic studies and application in enantioselective Aza-Michael reaction

Article abstract of DOI:10.1021/ja054066b

The full details of a catalytic asymmetric aza-Michael reaction of methoxylamine promoted by rare earth-alkali metal heterobimetallic complexes are described, demonstrating the effectiveness of Lewis acid-Lewis acid cooperative catalysis. First, enones we

Full text of DOI:10.1021/ja054066b

Published on Web 09/01/2005
Lewis Acid-Lewis Acid Heterobimetallic Cooperative
Catalysis: Mechanistic Studies and Application in
Enantioselective Aza-Michael Reaction
Noriyuki Yamagiwa, Hongbo Qin, Shigeki Matsunaga,* and Masakatsu Shibasaki*
Contribution from the Graduate School of Pharmaceutical Sciences, The UniVersity of Tokyo,
Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
Received June 20, 2005; E-mail: mshibasa@mol.f.u-tokyo.ac.jp; smatsuna@mol.f.u-tokyo.ac.jp
Abstract: The full details of a catalytic asymmetric aza-Michael reaction of methoxylamine promoted by
rare earth-alkali metal heterobimetallic complexes are described, demonstrating the effectiveness of Lewis
acid-Lewis acid cooperative catalysis. First, enones were used as substrates, and the 1,4-adducts were
obtained in good yield (57-98%) and high ee (81-96%). Catalyst loading was successfully reduced to
0.3-3 mol % with enones. To broaden the substrate scope of the reaction to carboxylic acid derivatives,
R,â-unsaturated N-acylpyrroles were used as monodentate, carboxylic acid derivatives. With â-alkyl-
substituted N-acylpyrroles, the reaction proceeded smoothly and the products were obtained in high yield
and good ee. Transformation of the 1,4-adducts from enones and R,â-unsaturated N-acylpyrroles afforded
corresponding chiral aziridines and â-amino acids. Detailed mechanistic studies, including kinetics, NMR
analysis, nonlinear effects, and rare earth metal effects, are also described. The Lewis acid-Lewis acid
cooperative mechanism, including the substrate coordination mode, is discussed in detail.
Introduction
The recent development and use of various types of coopera-
tive catalyses in asymmetric reactions have received much
attention.1 Since the early 1990s,2a we reported a series of rare
earth-alkali metal heterobimetallic complexes that enable
various catalytic asymmetric reactions.2 The structure of the rare
earth-alkali metal heterobimetallic complexes, determined by
X-ray, mass spectrometry, and NMR analysis, consisted of one
rare earth metal (RE), three 1,1′-bi-2-naphthol (BINOL), and
three alkali metals (M) (REMB, Figure 1).2b,c The asymmetric
environment of REMB is finely tunable by varying the
combination of rare earth metals and alkali metals. Mechanistic
studies of asymmetric reactions with the REMB complex
revealed that REMB functions as a cooperative Lewis acid-
Brønsted base catalyst. The REMB catalyst functions not only
as a Lewis acid to activate electrophiles but also as a Brønsted
base to deprotonate nucleophiles to form activated metal
nucleophiles. In previous reports of the REMB heterobimetallic
catalysts, only nucleophiles with protons with a relatively low
pKa value (ca. 10-19 in H2O), such as nitroalkane, malonate,
ketone, and thiol, were used due to the limitation of the Brønsted
basicity of the catalysts.1c Nucleophiles with a higher pKa value
Figure 1. Structures of (S,S,S)-REMB, (S,S,S)-YLB 1a, and (S,S,S)-DyLB
1b.
were not applicable to REMB catalysts. To broaden the substrate
scope of nucleophiles in REMB catalysis, we planned to use
the same REMB heterobimetallic catalysts in a different reaction
mode: Lewis acid-Lewis acid cooperative catalysis.3,4
We hypothesized that if the Lewis acidic part of the chiral
catalyst controls both the orientation of the electrophile and
nucleophile properly, with only a little, if any, deactivation of
nucleophiles, high enantio-induction would be achieved (Figure
2). To evaluate this concept, we chose catalytic asymmetric 1,4-
addition of O-alkylhydroxylamines. 1,4-Addition of amines to
(1) Reviews on bifunctional asymmetric catalysis: (a) Ma, J.-A.; Cahard, D.
Angew. Chem., Int. Ed. 2004, 43, 4566. (b) Rowlands, G. J. Tetrahedron
2001, 57, 1865. (c) Kanai, M.; Kato, N.; Ichikawa, E. Shibasaki, M. Synlett
2005, 1491.
(2) (a) Sasai, H.; Suzuki, T.; Arai, S.; Arai, T.; Shibasaki, M. J. Am. Chem.
Soc. 1992, 114, 4418. (b) Sasai, H.; Suzuki, T.; Itoh, N.; Tanaka, K.; Date,
T.; Okamura, K.; Shibasaki, M. J. Am. Chem. Soc. 1993, 115, 10372.
Reviews on rare earth-alkali metal bifunctional catalysts: (c) Shibasaki,
M.; Yoshikawa, N. Chem. ReV. 2002, 102, 2187. (d) Shibasaki, M.; Sasai,
H.; Arai, T. Angew. Chem., Int. Ed. Engl. 1997, 36, 1236.
(3) Review for combined acid catalysis including Lewis acid-assisted Lewis
acid catalysis, see: (a) Yamamoto, H.; Futatsugi, K. Angew. Chem., Int.
Ed. 2005, 44, 1924. For representative examples of Lewis acid-assisted
Lewis acid catalysis, see: (b) Ishihara, K.; Kobayashi, J.; Inanaga, K.;
Yamamoto, H. Synlett 2001, 394 and references therein. (c) Hanawa, H.;
Hashimoto, T.; Maruoka, K. J. Am. Chem. Soc. 2003, 125, 1708. (d) Ooi,
T.; Takahashi, M.; Yamada, M.; Tayama, E.; Omoto, K.; Maruoka, K. J.
Am. Chem. Soc. 2004, 126, 1150 and references therein. For other examples,
see review (ref 3a).
9
10.1021/ja054066b CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 13419-13427
13419

A R T I C L E S
Yamagiwa et al.
Table 1. Optimization of Catalytic Asymmetric Aza-Michael
Reaction of 2a with 3a
catalyst
time
(h)
yield
(%)
ee
entry
additive
(x mol %)
(%)
1
2
3
4
5
6
7
8
9
none
10
10
10
10
10
10
5
3
1
0.5
0.3
1
24
24
24
24
24
24
42
42
48
80
80
48
94
29
85
67
44
94
94
97
95
96
96
98
97
93
96
96
97
97
96
95
96
96
92
95
H2Ob
MS 3 Å
MS 4 Å
MS 5 Å
Drierite
Drierite
Drierite
Drierite
Drierite
Drierite
Drierite
Figure 2. Working hypothesis of Lewis acid-Lewis acid cooperative
catalysis in enantioselective 1,4-addition of alkoxylamine.
R,â-unsaturated carbonyl compounds provides a direct and
attractive strategy for the construction of optically active
â-amino carbonyl compounds,5-8 which are often found in
biologically interesting compounds. Among them, a few highly
enantioselective Lewis acid catalyses for 1,4-additions of
O-alkylhydroxylamines were recently reported,7 providing
versatile chiral building blocks. The products were readily
transformed into chiral aziridines. We hypothesized that the
oxygen atom of O-alkylhydroxylamine would coordinate to the
alkali metal of a heterobimetallic complex without a significant
decrease in the nucleophilicity of nitrogen (Figure 2). Here we
report the complete details of our trials to utilize the heterobi-
metallic complex in Lewis acid-Lewis acid cooperative ca-
talysis for enantioselective 1,4-addition of methoxylamine. The
(S,S,S)-YLi3tris(binaphthoxide) complex (YLB 1a, Figure 1)
efficiently promoted the addition of methoxylamine to enones.
The substrate scope was further broadened to R,â-unsaturated
carboxylic acid derivatives using R,â-unsaturated N-acylpyrroles
as substrates. The mechanistic studies of the reaction to verify
10
11
12a
a Reaction was performed in 10 g scale at 2.1 M. b30 mol % of H2O
was added.
the Lewis acid-Lewis acid bifunctional reaction mode are also
described in detail.
Results and Discussion
A. Addition to Enones.9 Initially, we screened various
heterobimetallic complexes containing rare earth metals and
alkaline metals for the addition of 3a to 2a, and YLB 1a gave
the best reactivity and enantioselectivity.10 The reaction pro-
ceeded smoothly with 10 mol % of 1a at -20 °C to give 4a in
94% yield and 97% ee (Table 1, entry 1). In contrast to the
catalytic asymmetric ethoxycarbonylation reaction of aldehydes
using the YLB-H2O complex,11,12 the addition of H2O some-
what retarded the reaction. Anhydrous conditions gave the best
reaction rate, although H2O had no adverse effects on enanti-
oselectivity (entry 1 vs entry 2).10 We speculated that even a
small amount of H2O derived from a possible side reaction
(oxime formation) would be problematic for reducing the
catalyst loading. Thus, various desiccants were screened (entries
3-6), and Drierite (CaSO4) gave the best results (entry 6).13
The low chemical yield produced using molecular sieves (entries
3-5) is due to the absorption of methoxylamine to the molecular
sieves.13 Under the optimal conditions (YLB, 3a: 1.2 equiv,
Drierite), catalyst loading was reduced. As summarized in entries
7-11, the reaction proceeded without any problem with as little
as 0.5-1 mol % of 1a, giving 4a in good yield and ee (entry 9,
(4) For selected recent examples of other bifunctional chiral metal catalysis,
see: (a) France, S.; Shah, M. H.; Weatherwax, A.; Wack, H.; Roth, J. P.;
Lectka, T. J. Am. Chem. Soc. 2005, 127, 1206. (b) Knudsen, K. R.;
Jørgensen, K. A. Org. Biomol. Chem. 2005, 3, 1362. (c) Kanemasa, S.;
Ito, K. Eur. J. Org. Chem. 2004, 4741. (d) Evans, D. A.; Seidel, D.;
Rueping, M.; Lam, H. W.; Shaw, J. T.; Downey, C. W. J. Am. Chem. Soc.
2003, 125, 12692. (e) Sammis, G. M.; Danjo, H.; Jacobsen, E. N. J. Am.
Chem. Soc. 2004, 126, 9928. (g) Josephsohn. N. S.; Kuntz, K. W.; Snapper,
M. L.; Hoveyda, A. M. J. Am. Chem. Soc. 2001, 123, 11594. (h) Trost, B.
M.; Terrell, L. R. J. Am. Chem. Soc. 2003, 125, 338 and references therein.
(i) Matsunaga, S.; Yoshida, T.; Morimoto, H.; Kumagai, N.; Shibasaki,
M. J. Am. Chem. Soc. 2004, 126, 8777 and references therein. For other
examples, see review in refs 1-3.
(5) Reviews for enantioselective conjugate addition: (a) Sibi, M. P.; Manyem,
S. Tetrahedron 2000, 56, 8033. (b) Krause, N.; Hoffmann-Ro¨der, A.
Synthesis 2001, 171.
(6) Review for aza-Michael reaction: Xu, L.-W.; Xia, C.-G. Eur. J. Org. Chem.
2005, 633.
(7) Alkoxyamines as nucleophiles: (a) Sibi, M. P.; Shay, J. J.; Liu, M.; Jasperse,
C. P. J. Am. Chem. Soc. 1998, 120, 6615. (b) Jørgensen, K. A.; Falborg,
L. J. Chem. Soc., Perkin Trans. 1 1996, 2823. (c) Sugihara, H.; Daikai,
K.; Jin, X. L.; Furuno, H.; Inanaga, J. Tetrahedron Lett. 2002, 43, 2735.
(d) Jin, X. L.; Sugihara, H.; Daikai, K.; Takeishi, H.; Jin, Y. Z.; Furuno,
H.; Inanaga, J. Tetrahedron 2002, 58, 8321. (e) Cardillo, G.; Gentilucci,
L.; Gianotti, M.; Kim, H.; Perciaccante, R.; Tolomelli, A. Tetrahedron:
Asymmetry 2001, 12, 2395.
(9) A part of this article (addition to enones, section A) was reported previously
as a preliminary communication. Yamagiwa, N.; Matsunaga, S.; Shibasaki,
M. J. Am. Chem. Soc. 2003, 125, 16178.
(10) Effects of rare earth metals and alkali metals on the reaction rate and
enantioselectivity are described in detail in section D. Adverse effects of
H2O are also discussed in detail in section D.
(8) For recent highly enantioselective catalytic asymmetric 1,4-additions of
other nitrogen nucleophiles to afford â-amino carbonyl compounds, see:
N-Benzylhydroxylamine: (a) Sibi, M. P.; Prabagaran, N.; Ghorpade, S.
G.; Jasperse, C. P. J. Am. Chem. Soc. 2003, 125, 11796 and references
therein. Aromatic amine: (b) Zhuang, W.; Hazell, R. G.; Jørgensen, K. A.
Chem. Commun. 2001, 1240. (c) Fadini, L.; Togni, A. Chem Commun.
2003, 30. (d) Li, K.; Hii, K. K. Chem Commun. 2003, 1132. (e) Li, K.;
Cheng, X.; Hii, K. K. Eur. J. Org. Chem. 2004, 959. (f) Hamashima, Y.;
Somei, H.; Shimura, Y.; Tamura, T.; Sodeoka, M. Org. Lett. 2004, 6, 1861.
Azide ion: (g) Myers, J. K.; Jacobsen, E. N. J. Am. Chem. Soc. 1999, 121,
8959. (h) Guerin, D. J.; Miller, S. J. J. Am. Chem. Soc. 2002, 124, 2134.
Carbamate: (i) Palomo, C.; Oiarbide, M.; Halder, R.; Kelso, M.; Go´mez-
Bengoa, E.; Garc´ıa, J. M. J. Am. Chem. Soc. 2004, 126, 9188. For ligand-
controlled asymmetric addition of lithium amide, see: (j) Doi, H.; Sakai,
T.; Iguchi, M.; Yamada, K.-i.; Tomioka, K. J. Am. Chem. Soc. 2003, 125,
2886 and references therein.
(11) In the asymmetric cyanation reaction, YLB-H2O complex was essential
to achieve good reactivity and enantioselectivity. Anhydrous YLB 1a
complex gave cyanation adducts in poor ee; see: (a) Yamagiwa, N.; Tian,
J.; Matsunaga, S.; Shibasaki, M. J. Am. Chem. Soc. 2005, 127, 3413. (b)
Tian, J.; Yamagiwa, N.; Matsunaga, S.; Shibasaki, M. Angew. Chem., Int.
Ed. 2002, 41, 3636. (c) Tian, J.; Yamagiwa, N.; Matsunaga, S.; Shibasaki,
M. Org. Lett. 2003, 5, 3021. (d) Abiko, Y.; Yamagiwa, N.; Sugita, M.;
Tian, J.; Matsunaga, S.; Shibasaki, M. Synlett 2004, 2434.
(12) For preparation and characterization of the structure of anhydrous YLB:
(a) Aspinall, H. C.; Dwyer, J. L. M.; Greeves, N.; Steiner, A. Organome-
tallics 1999, 18, 1366. (b) Aspinall, H. C.; Bickley, J. F.; Dwyer, J. L. M.;
Greeves, N.; Kelly, R. V.; Steiner, A. Organometallics 2000, 19, 5416.
(13) Ability of Drierite as desiccant toward THF was checked by Karl Fisher
experiment. See Supporting Information. Absorption of methoxylamine 3a
to molecular sieves was confirmed by NMR analysis of methoxylamine
solution in THF using 2,4,6-trimethylbenzene as an internal standard.
9
13420 J. AM. CHEM. SOC. VOL. 127, NO. 38, 2005

Lewis Acid−Lewis Acid Heterobimetallic Catalysis
A R T I C L E S
Table 2. Catalytic Asymmetric Aza-Michael Reaction of Various Enones
enone
YLB
time
(h)
yield
(%)
ee
entry
R1
R2
product
(mol %)
(%)
1
2
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
2a
2a
2b
2b
2c
2c
2d
2e
2f
2g
2h
2i
2i
2j
4a
4a
4b
4b
4c
4c
4d
4e
4f
4g
4h
4i
4i
4j
3
1
3
1
3
1
3
3
3
3
3
3
1
3
3
1
3
3
3
3
3
3
3
3
3
3
3
3
42
48
42
46
54
65
48
74
48
48
78
48
78
48
82
74
42
48
122
84
48
60
84
48
78
48
96
84
97
95
96
92
97
91
96
91
96
95
96
92
97
96
85
85
98
95
92
80
96
91
96
95
97
98
57
91
95
96
96
96
96
96
94
96
92
94
93
92
93
96
95
95
81
92
82
92
95
85
84
93
86
82
82
95
3
4-Cl-C6H4
4
4-Cl-C6H4
5a
4-F-C6H4
6a
4-F-C6H4
7
4-Me-C6H4
8a
4-MeO-C6H4
9
3-Me-C6H4
10
11
12a
13
14a
15
16a
17a
18a
19
20a
21b
22
23
24
25
26
27
28
2-furyl
2-thienyl
Ph
4-Cl-C6H4
4-Cl-C6H4
4-Me-C6H4
4-MeO-C6H4
4-MeO-C6H4
3-NO2-C6H4
3-ClC6H4
2-Cl-C6H4
2-furyl
Ph
Ph
Ph
2k
2k
2l
4k
4k
4l
Ph
Ph
Ph
2m
2n
2o
2p
2q
2r
2s
4m
4n
4o
4p
4q
4r
4s
Ph
Ph
Ph
2-thienyl
Ph
4-pyridyl
Ph
n-C5H11
Ph
i-PrCH2
Ph
i-Pr
2t
4t
Ph
cyclo-hexyl
t-Bu
2u
2v
2w
4u
4v
4w
Ph
Ph
trans-PhCHdCH
a 2 equiv of 3a were used. b 3 equiv of 3a were used.
95% yield and 96% ee; entry 10, 96% yield and 96% ee). The
reaction also proceeded with 0.3 mol % of 1a, although the
enantiomeric excess was slightly decreased (92% ee, entry 11).
In entry 12, the reaction was performed on a 10 g scale at 2.1
M. 4a was obtained without any problem in large scale under
concentrated conditions.
We then examined the substrate scope of the reaction. For
convenience, most reactions were performed with 3 mol % of
1a. For selected substrates, including 2k with an electron
donating substituent on the aromatic ring, reactions were also
performed with 1 mol % of 1a, and similar results were obtained
(Table 2, entries 1, 3, 5, 12, 15 vs 2, 4, 6, 13, 16). â-Aryl
substituted enones with various substituents (entries 1-19)
afforded 1,4-adducts in 81-96% ee and 85-97% yield. For
the less reactive substrate, 2 to 3 equiv of 3a were used.
â-Heteroaromatic- (entries 20-22) and â-aliphatic- (entries 23-
27) enones and dienone (entry 28) also gave products in good
ee.14
B. Addition to R,â-Unsaturated N-Acylpyrroles. To broaden
the substrate scope to R,â-unsaturated carboxylic acid deriva-
tives, we investigated R,â-unsaturated N-acylpyrroles as elec-
trophiles. The concept for the use of R,â-unsaturated N-acylpyr-
role as a monodentate ester surrogate is summarized in Figure
3.15,16 Because the lone electron pair of the nitrogen in the
pyrrole ring is delocalized in an aromatic system, the property
of a carbonyl group is similar to that of a phenyl ketone.16a We
supposed that the relatively low LUMO energy of R,â-
unsaturated N-acylpyrrole due to the aromaticity of pyrrole
Figure 3. Property of N-acylpyrrole.
would facilitate the 1,4-addition of 3a. In addition, the mono-
dentate coordination mode of N-acylpyrrole is the same as that
of enones. Thus, we hypothesized that the chiral environment
optimized for enones would also be effective for N-acylpyr-
roles.
Although we previously reported the efficient synthesis of
various R,â-unsaturated N-acylpyrroles using the Wittig
reaction,15a the method was not suitable for the preparation of
â-alkyl-substituted R,â-unsaturated N-acylpyrroles due to modest
E/Z selectivity. Thus, we first developed an improved synthetic
procedure for â-alkyl-substituted R,â-unsaturated N-acylpyr-
roles, utilizing the HWE reaction. The HWE reagent was readily
synthesized by the reaction of carbonyl dipyrrole 516a with
lithiated 6 (Table 3). HWE reaction under Masamune-Roush
(15) Use of R,â-unsaturated N-acylpyrrole in catalytic asymmetric 1,4-addi-
tions: (a) Matsunaga, S.; Kinoshita, T.; Okada, S.; Harada, S.; Shibasaki,
M. J. Am. Chem. Soc. 2004, 126, 7559. (b) Mita, T.; Sasaki, K.; Kanai,
M.; Shibasaki, M. J. Am. Chem. Soc. 2005, 127, 514. Application to
diastereoselective 1,4-addition reactions: (c) Arai, Y.; Kasai, M.; Ueda,
K.; Masaki, Y. Synthesis 2003, 1511.
(16) Properties of N-acylpyrroles and carbonyl dipyrrole 5 were reported by
Evans and co-workers in detail. (a) Evans, D. A.; Borg, G.; Scheidt, K. A.
Angew. Chem., Int. Ed. 2002, 41, 3188. (b) Evans, D. A.; Johnson, D. S.
Org. Lett. 1999, 1, 595. (c) Evans, D. A.; Scheidt, K. A.; Johnston, J. N.;
Willis, M. C. J. Am. Chem. Soc. 2001, 123, 4480.
(14) Aliphatic ketones (R1 ) alkyl) are not suitable substrates for the present
system. For example, no reaction proceeded with benzalacetone.
9
J. AM. CHEM. SOC. VOL. 127, NO. 38, 2005 13421

A R T I C L E S
Yamagiwa et al.
Scheme 1. Transformations of aza-Michael Adductsa
Table 3. Synthesis of â-Alkyl-Substituted R,â-Unsaturated
N-Acylpyrroles
entry
aldehyde R
product
time(h)
yield(%)
E/Z
1
2
3
4
5
(CH3)2CHCH2s
n-Pr
PhCH2CH2-
CH2dCH(CH2)8s
i-Pr
8a
9a
9b
9c
9d
9e
13
12
22
15
10
96
90
89
94
90
>20/1
>20/1
>20/1
>20/1
>20/1
8b
8c
8d
8e
Table 4. Catalytic Asymmetric Aza-Michael Reaction of Various
R,â-Unsaturated N-Acylpyrroles
a
Reagents and conditions: (a) TiCl4, CH2Cl2; then Et3N, -30 °C to
catalyst
time
(h)
yield
(%)
ee
-10 °C. (b) NaOMe, MeOH, 4 °C, 10 min. (c) Pd/C, H2 (1 atm), MeOH,
rt; then Boc2O. (d) NaOt-Bu, THF, rt. (e) Zn(BH4)2, Et2O, -78 °C. (f)
K-Selectride, THF, -78 °C.
entry
N-acylpyrrole R
(x mol %)
(%)
1
2
(CH3)2CHCH2s
9a
9a
9a
YLB 1a (5)
YLB 1a (10)
DyLB 1b (10)
38
38
92
96
92
91
97
86
94
84
89
94
91
53
63
49
92
94
94
83
86
83
85
89
89
86
84
81
80
82
3
62
4
CH3CH2CH2s
9b YLB 1a (10)
9b DyLB 1b (10)
YLB 1a (10)
DyLB 1b (10)
CH2dCH(CH2)5CH2s 9d YLB 1a (10)
9d DyLB 1b (10)
38
5
38
6
PhCH2CH2-
9c
9c
38
7
38
8
38
9
62
10
11
12
13
14
(CH3)2CHs
9e
9e
9f
YLB 1a (10)
DyLB 1b (10) 114
DyLB 1b (10) 165
106
Phs
m-ClsC6H4s
p-CF3sC6H4s
9 g DyLB 1b (10) 107
9h DyLB 1b (10) 107
conditions17 gave various â-alkyl substituted R,â-unsaturated
N-acylpyrroles in good yield and high E-selectivity.
The reaction conditions for the aza-Michael reaction were
optimized with substrate 9a (Table 4). YLB 1a (5 mol %)
promoted the reaction of 9a using 1.03 equiv of 3a at -30 °C,
and the product 10a was obtained in 92% yield and 92% ee
after 38 h (entry 1). Ee increased to 94% when using 10 mol %
of YLB (entry 2). DyLi3tris(binaphthoxide) (DyLB 1b, Figure
1) gave comparable results (92% yield, 94% ee; entry 3). Other
substrates with alkyl substituents were also applicable, and
products were obtained in high yield and good ee (entries 4-11).
For N-acylpyrroles 9b and 9c, DyLB 1b gave better enantiose-
lectivity than YLB 1a. The reactivity of â-aryl-substituted
substrates was only modest (entries 12-14). Products were
obtained in modest yield using 10 mol % of DyLB 1b.18
Figure 4. Substituent effects of chalcones 2a-e on initial rate.
performed (Scheme 1). 4 and 10 were readily converted into
chiral aziridines without loss of enantiomeric excess. Treatment
of 10c and 10f with TiCl4 and Et3N afforded 11c and 11f in
76% and 81% yield, respectively.19 The N-acylpyrrole unit was
converted into methyl ester by treatment with NaOMe for 10
min at 4 °C. The N-O bond in 13a and 13f was easily cleaved
with Pd/C under H2 (1 atm) in MeOH. After protection with a
Boc group, 14a and 14f were isolated in 94% and 96% yield,
respectively (two steps). For the aza-Michael adducts from
C. Transformation of 1,4-Adducts. To demonstrate the
utility of aza-Michael adducts, several transformations were
(17) Blanchette, M. A.; Choy, W.; Davis, J. T.; Essefeld, A. P.; Masamune, S.;
Roush, W. R.; Sakai, T. Tetrahedron Lett. 1984, 25, 2183.
(18) Similar tendency that â-aryl substituted carboxylic acid derivatives are much
less reactive than â-alkyl substrates was observed in asymmetric aza-
Michael reactions of alkoxylamines. See ref 7a and 7b. Trials to activate
substrates by introducing an electron-withdrawing group on a pyrrole ring
failed.
(19) For aziridine formation using TiCl4 and amine, see: Bongini, A.; Cardillo,
G.; Gentilucci, L.; Tomasini, C. J. Org. Chem. 1997, 62, 9148. For the use
of NaOt-Bu, see ref 7d. No loss of enantiomeric excess was observed during
transformations in Scheme 1 as confirmed by HPLC analysis.
9
13422 J. AM. CHEM. SOC. VOL. 127, NO. 38, 2005

Lewis Acid−Lewis Acid Heterobimetallic Catalysis
A R T I C L E S
Figure 5. Initial reaction rate kinetics of YLB 1a, chalcone (2a), and amine 3a.
Table 5. Effects of Structures of Amines 3 on Yield and
Enantiomeric Excess
determining step in the catalytic cycle. The initial reaction
profiles of chalcone derivatives are summarized in Figure 4.
Chalcones 2 with an electron-withdrawing group were more
reactive (F > 1), indicating that the carbon-nitrogen bond-
forming step is the rate-determining step. Initial rate kinetic
studies using 2a revealed that the reaction had (i) first-order
dependency on YLB 1a, (ii) first-order dependency on chalcone
2a, and (iii) zeroth-order dependency on amine 3a (Figure 5).
Results in Figures 4 and 5 indicated that the carbon-nitrogen
bond-forming step is the rate-determining step, while the free
amine 3a is not involved in the rate-determining step. Thus,
the active species is speculated to be a YLB-(MeONH2)n
complex. Formation of the postulated active YLB-(MeONH2)n
complex was supported by NMR analysis (Figure 6).21,22 A high
field shift of the amine proton indicated that the amine proton
interacted with the oxygen atom of YLB 1a. The low field shift
in the 13C NMR of 3a indicated coordination of the oxygen
atom of 3a with YLB 1a. Only one amine proton peak was
Figure 6. NH2 region {left (a-c)} 1H NMR spectra and CH3O region
{right (a-c)} 13C NMR spectra of methoxylamine 3a with and without
YLB 1a in THF at -20 °C; (a) without YLB 1a; (b) YLB 1a (60 mM),
MeONH2 3a (180 mM); (c) YLB 1a (60 mM), MeONH2 3a (600 mM).
enones, NaOt-Bu was effective to afford aziridines 15a and 15s
in 85% and 81% yield, respectively.19 Diastereoselective
synthesis of syn- or anti-1,3-amino alcohols 16 and 17 was also
achieved. Treatment of 4a with in situ generated Zn(BH4)2
afforded syn-1,3-methoxyamino alcohol 16 in 70% yield (syn/
anti ) 9/1), and treatment of 4a with K-Selectride afforded anti-
1,3-methoxyamino alcohol 17 in 93% yield (syn/anti ) 1/9).20
D. Mechanistic Studies. After developing synthetically useful
methods to provide optically active aziridines and â-amino
carbonyl compounds, we investigated the reaction mechanism
to evaluate whether the hypothetical Lewis acid-Lewis acid
cooperative mechanism functions.
(21) Active species of the aza-Michael reaction was proved to have three BINOL
units through preliminary kinetic studies. In the preliminary mechanistic
studies, however, the reaction mode (Lewis acid-assisted Lewis acid) and
coordination mode of substrates remained unclear; Yamagiwa, N.; Mat-
sunaga, S.; Shibasaki, M. Angew. Chem., Int. Ed. 2004, 43, 4493.
(22) For the full details of NMR analysis of YLB-3a complex, including YLB
1a region, see Supporting Information.
Substituent effects of chalcone derivatives on initial rate were
examined (2a, 2b, 2c, 2d, and 2e) to determine the rate-
(20) Pilli, R. A.; Russowsky, D.; Dias, L. C. J. Chem. Soc., Perkin Trans. 1
1990, 1213.
9
J. AM. CHEM. SOC. VOL. 127, NO. 38, 2005 13423

A R T I C L E S
Yamagiwa et al.
Figure 7. Supposed catalytic cycle of the aza-Michael reaction.
Figure 9. NMR spectra of (S,S,S)-YLB (60 mM in d8-tetrahydrofuran) at
-20 °C: (a) without H2O, (b) with H2O (60 mM), (c) with H2O (180 mM),
(d) with H2O (180 mM) and Drierite (CaSO4: 200 mg/mL d8-tetrahydro-
furan).
Exchange between 1,4-adduct 4a and amine 3a regenerates the
active YLB-(MeONH2)n complex (Cat-II).
The addition of H2O decreased the reaction rate, while still
maintaining similar enantioselectivity (Table 1, entry 2). Adverse
effects of H2O on the reaction rate were investigated in detail
(Figure 8). The tendency is in striking contrast to the asymmetric
cyano-ethoxycarbonylation using the same YLB complex,11 in
which the YLB-H2O complex was essential for high reactivity
and enantioselectivity. In the present aza-Michael reaction, the
initial rate was decreased to approximately half in the presence
of 10 mol % H2O and to a quarter with 20 mol % H2O. On the
other hand, reactivity recovered when Drierite (CaSO4) was
added. NMR spectroscopy revealed a reversible interaction
between YLB and H2O in solution (Figure 9). NMR spectra of
anhydrous YLB in THF at -20 °C showed six sharp signals
(Figure 9a), implying that the YLB complex has a C3-symmetric
structure in solution. The signals gradually broadened, however,
by adding H2O to YLB (Figure 9b and 9c). In Figure 9d, Drierite
(CaSO4) was added to the mixture of YLB and 3 equiv of H2O,
and the sharp signals of YLB 1a were observed again. These
observations suggested that there is equilibrium between
anhydrous-YLB and the YLB-H2O complex. Aspinall et al.
reported the X-ray crystallographic structure of both [Li(Et2O)]3-
[Eu(binol)3] (anhydrous-EuLB) and [Li(Et2O)]3[Eu(binol)3-
(H2O)] (EuLB-H2O).12 We assumed that H2O would coordinate
with the yttrium center in a similar manner as observed for the
europium complex. Considering the coordination number (six
or seven) of rare earth-alkali metal heterobimetallic complexes
obserVed in crystal structures,2d,12 only one substrate other than
tris(binaphthoxide) would coordinate with the yttrium center.
The H2O-induced deceleration is due to a reversible and
competitive occupation of the seventh coordination site of the
yttrium center. In other words, a vacant coordination site on
yttrium metal is essential to promote the conjugate addition
reaction.
Figure 8. Water effects on the reaction rate.
observed even in the presence of an excess amount of amine
3a at -20 °C (Figure 6c), indicating the rapid and reversible
coordination of 3a to YLB 1a at -20 °C. The adverse effects
of sterically crowded amines on reactivity and enantioselectivity
(Table 5, entries 1-3) also indicate that the complexation of
the YLB-amine nucleophile is important.23 Amines without the
oxygen group also gave poor results (entries 4-8), suggesting
that the interaction between the oxygen atom in 3a and YLB
1a is important.
The postulated catalytic cycle of the reaction to explain the
above observation is shown in Figure 7. Amine 3a reversibly
coordinates to YLB 1a (Cat-I) to form an active YLB-
(MeONH2)n complex (Cat-II). Zeroth-order dependency on
amine 3a (Figure 5) implies that amine 3a coordinated to YLB
1a reacts with chalcone 2a. YLB-(MeONH2)n complex (Cat-
II) is a dominant species among all the catalyst species (Cat-I,
Cat-II, Cat-III, and Cat-IV). 2a interacts with Cat-II, and
subsequent carbon-nitrogen bond formation affords an enolate
intermediate (Cat-III). The step from Cat-II to Cat-III is rate-
determining. Irreversible proton transfer affords Cat-IV.24
(23) The tendency observed in Table 5, entries 1-3 is different from results
obtained by Inanaga and co-workers using Sc chiral Lewis acid catalysis.
In their report, sterically crowded amine 3c gave the best enantioselectivity.
See ref 7d.
(24) The irreversibility of the present aza-Michael reaction was experimentally
confirmed. See Supporting Information for detail results.
9
13424 J. AM. CHEM. SOC. VOL. 127, NO. 38, 2005

Lewis Acid−Lewis Acid Heterobimetallic Catalysis
A R T I C L E S
Figure 11. Nonlinear effects on the catalytic asymmetric 1,4-addition
reaction of 3a to 2a.
Figure 10. Rare earth metal effects on reaction rate and enantioselectivity.
ionic radius (Tm, Yb, and Lu) would not accept a seventh ligand
for steric reasons. Thus, the desirable Lewis acid-Lewis acid
cooperative reaction would not proceed with Tm, Yb, and Lu
metals, resulting in a low reaction rate and low enantioselec-
tivity. In other words, coordination of the substrate to the center
metal of the REMB catalyst is essential to promote the reaction.
In contrast to Tm, Yb, and Lu metals, REMB complexes with
a large ionic radius are known to have the seventh coordination
site as confirmed by X-ray crystal analysis.26 When using Y,
Dy, Sm, Pr, and La, the seventh coordination site of the rare
earth metal center is available. Thus, the reaction proceeds
smoothly. Enantioselectivity depends on the size of the chiral
space ascribed to the ionic size of central metals.
The importance of the rare earth metal as a Lewis acid was
also supported by the results shown in Figure 10. aza-Michael
reaction of 2a was performed with various rare earth-Li-
BINOL ) 1/3/3 complexes (REMB, Figure 1, RE ) Y, La, Pr,
Sm, Dy, Er, Tm, Yb, Lu and M ) Li). Rare earth metals had
drastic effects on both reaction rate and enantioselectivity of
the reaction. The reaction rate and enantioselectivity correlated
with the ionic radius of the rare earth metals used. Initial reaction
rates were similar with metals larger than yttrium and drastically
dropped with smaller metals (Er, Tm, Yb, and Lu). Metals with
a small ionic radius (Tm, Yb, and Lu) had low enantioselectivity
and low reactivity. The best enantioselectivity was achieved with
Y and Dy metals. Metals with a large ionic radius (Sm, Pr, and
La) gave products with lower selectivity, although the reaction
proceeded smoothly. The observed rare earth metal effects are
explained nicely by considering the difference in coordination
number. Various investigations of rare earth-alkali metal
heterobimetallic complexes indicated that the coordination
number of central rare earth metals is different depending on
the ionic size of the rare earth metals.2,12,21 Salvadori and co-
workers reported that the YbLi3tris(binaphthoxide) (YbLB)
complex does not accept a seventh ligand for steric reasons,
which was suggested by NMR analysis of YbLB in the presence
of H2O.25 Based on their reports, rare earth metals with a small
Positive nonlinear effects27 for enantioselectivity and negative
nonlinear effects on reaction rate were observed for the present
aza-Michael reaction (Figure 11). A catalyst with 33% ee of
BINOL afforded products in 95% ee, which was comparable
to that obtained by an optically pure catalyst (96% ee). The
reaction rate, however, differed greatly depending on the optical
purity of the catalyst. For example, the reaction rate with a
catalyst with 33% ee was only 1/34 (v33% ) 2.98 × 10-4
ee
Mh-1) of that observed with an optically pure catalyst (v100% ee
(26) The crystal structures of La, Pr, Nd, Eu heterobimetallic complexes with
H2O as the seventh ligand are reported. See refs 2, 12 and references therein.
(27) Reviews: (a) Girard, C.; Kagan, H. B. Angew. Chem., Int. Ed. 1998, 37,
2922. (b) Kagan, H. B. AdV. Synth. Catal. 2001, 343, 227. For a review
discussing kinetic aspects of nonlinear effects, see: (c) Blackmond, D. G.
Acc. Chem. Res. 2000, 33, 402.
(25) Bari, L. D.; Lelli, M.; Pintacuda, G.; Pescitelli, G.; Marchetti, F.; Salvadori,
P. J. Am. Chem. Soc. 2003, 125, 5549.
9
J. AM. CHEM. SOC. VOL. 127, NO. 38, 2005 13425

A R T I C L E S
Yamagiwa et al.
Figure 13. Structures of (S,S,S)-YLB and (S,S,R)-YLB (THF is omitted
for clarity).
Table 6. Alkali Metal Effects on Aza-Michael Reaction
Figure 12. NMR spectra of homochiral (S,S,S)-YLB and heterochiral-YLB;
(a) (S,S,S)-YLB 1a, (b) (S,S,S)-YLB + (R,R,R)-YLB (1:1 mixture), (c)
(S,S,S)-YLB + (R,R,R)-YLB (5:1 mixture).
catalyst
time
(h)
yield
(%)
entry
(x mol %)
ee (%)
config
Scheme 2. Aza-Michael Reaction (a) with (S,S,S)-YLB (6.66 mol
%) + (R,R,R)-YLB (3.33 mol %) and (b) with Preformed
(S,S,R)-YLB (10 mol %)
1
none
42
42
42
42
42
trace
11
2
BuLi/BINOL (9/9)
Y(HMDS)3/BINOL (3/9)
YPB 1c (3)
12
16
12
95
R
R
R
S
3
29
4a
5
19
97
YLB 1a (3)
a (S,S,S)-YK3tris(binaphthoxide)(YPB 1c) was prepared from Y(HMDS)3/
KHMDS/(S)-BINOL in a ratio of 1/3/3.
) 1.01 × 10-2 Mh-1). The reaction rate with 0% ee catalyst
was 1/199 (v0% ee ) 5.07 × 10-5 Mh-1) of that with an optically
pure catalyst.
To explain the observed nonlinear effects, we continued
spectroscopic analysis of YLB. Aspinall et al. reported that the
(S,S,R)-heterochiral YLB complex is thermodynamically more
stable than the (S,S,S)-YLB complex due to the C-H to π
hydrogen bond.12 Because they synthesized the (S,S,R)-hetero-
chiral YLB [together with the (S,R,R)-complex] from racemic-
BINOL, it was uncertain whether ligand exchange occurred
under the 1,4-addition reaction conditions at -20 °C. To observe
the lability of BINOL in YLB under the reaction conditions,
we performed NMR analysis of a YLB solution prepared from
1 equiv of (S,S,S)-YLB and 1 equiv of (R,R,R)-YLB. NMR
analysis of the mixture solution revealed only the heterochiral
(S,S,R)-and (S,R,R)-YLB complex (Figure 12b) [no (S,S,S)-YLB
was detected by 1H NMR]. The spectrum was identical to that
reported by Aspinall et al.12 The mixture of (S,S,S)-YLB and
(R,R,R)-YLB in a 5:1 ratio resulted in an approximately 1:1
mixture of (S,S,S)-YLB and (S,S,R)-YLB (Figure 12c). The
results clearly indicated that the ligand exchange between
(S,S,S)-YLB and (R,R,R)-YLB occurred smoothly to exclusively
form a thermodynamically more stable (S,S,R)-YLB-type com-
plex. On the other hand, the reaction rate using 6.66 mol % of
(S,S,S)-YLB and 3.33 mol % of (R,R,R)-YLB mixed at -78
°C was as slow as that performed with 10 mol % of the
Figure 14. Proposed Lewis acid-Lewis acid cooperative transition-state
model for aza-Michael reaction.
preformed (S,S,R)-YLB-type complex (Scheme 2). Product 4a
was obtained in only 10% yield after 24 h (92% ee). The
reaction rate was much slower than that observed with (S,S,S)-
YLB. The results in Scheme 2 indicate that ligand exchange
occurs smoothly at -20 °C.28 On the basis of these results, the
nonlinear effects and the observed reaction rate tendency in
Figure 11 are explained as follows. The (S,S,R)-heterochiral
YLB complex is thermodynamically more stable than (S,S,S)-
YLB 1a. On the other hand, equilibrium between the (S,S,R)-
heterochiral YLB complex and (S,S,S)-YLB 1a occurs under
the reaction conditions. The (S,S,R)-heterochiral YLB complex
is not active in the 1,4-addition reaction, and only trace (S,S,S)-
YLB formed in equilibrium and promoted the reaction, when
the reaction was performed with ligands with 33% ee.The
(28) Recently, Salvadori and co-workers independently reported that binaph-
thoxide in YbK3tris(binaphthoxide) is labile in solution phase; see: Bari,
L. D.; Lelli, M.; Salvadori, P. Chem.sEur. J. 2004, 10, 4594.
9
13426 J. AM. CHEM. SOC. VOL. 127, NO. 38, 2005

Lewis Acid−Lewis Acid Heterobimetallic Catalysis
A R T I C L E S
enantiomeric excess of the product was the same with ligands
over 33% ee, while the reaction rate increased drastically by
increasing the ee of the ligands from 33% ee to >99% ee
because the amount of (S,S,S)-YLB increased accordingly.
A striking difference in the reaction rate between the
homochiral (S,S,S)-YLB and the heterochiral (S,S,R)-YLB would
also support the importance of the yttrium center. Structures of
(S,S,S)-YLB and (S,S,R)-YLB are shown in Figure 13. Aspinall
reported that the [Li(THF)2]3[Y{(S)-binol}2] fragment (Figure
13, dotted square) of homochiral and heterochiral YLB in crystal
structures does not highlight any gross structural change.12 Thus,
it is difficult to explain the large difference in the reaction rate,
if Li alone can promote the aza-Michael reaction. Based on the
observation, a transition state model where only Li metal
functions as a Lewis acid to activate chalcone 2a and controls
the orientation of amine 3a is unlikely. Both Y and Li metals
are required to promote the reaction. As shown in Table 6, the
BuLi/BINOL complex (entry 2) and Y(HMDS)3/BINOL com-
plex (entry 3) afforded product 4a only in poor yield and ee.
(S,S,S)-YK3tris(binaphthoxide) (YPB, 1c) also gave poor results
(entry 4). The reaction proceeded in high enantioselectivity only
when both Y and Li were used (entry 5).
ening the utility of heterobimetallic complexes. YLB 1a and
DyLB 1b complexes promoted catalytic asymmetric 1,4-addition
of methoxylamine 3a to enones and R,â-unsaturated N-acylpyr-
roles. Transformation of the 1,4-adducts from enones and R,â-
unsaturated N-acylpyrroles afforded corresponding aziridines.
Detailed mechanistic studies, including kinetics, NMR analysis,
nonlinear effects, and rare earth metal effects, revealed the
catalytic cycle. Examination for alkali metal effects and rare
earth metal effects revealed that both metals are essential for
catalyst activity. The results obtained in mechanistic studies
support the proposed Lewis acid-Lewis acid cooperative
mechanism.
Acknowledgment. We are thankful for the financial support
by Grant-in-Aid for Specially Promoted Research and Grant-
in-Aid for Encouragements for Young Scientists (B) (for S.M.)
from JSPS and MEXT. We thank Ms. M. Sugita for preparation
of R,â-unsaturated N-acylpyrroles.
Supporting Information Available: Experimental procedures,
characterization of the products, detail data for mechanistic
studies. This material is available free of charge via the Internet
at http://pubs.acs.org.
The proposed transition state model under the Lewis acid-
Lewis acid cooperative mechanism is shown in Figure 14.
Mechanistic studies in this section suggested that both Y and
Li function as a Lewis acid. On the basis of aza-Michael
reactions using conformationally rigid enones,29 we speculate
that enone 2a would coordinate to the yttrium center in a s-cis
form. NMR analysis of the YLB-3a complex (Figure 6)
suggested that 3a would coordinate with both the Li and O of
YLB 1a. Thus, the orientation of amine 3a is controlled nicely
to afford (S)-4a with high ee.30
JA054066B
(29) Conformationally rigid enones 4x, 4y, 4z and cyclohexenone were used.
4x and 4y with s-cis form afforded aza-Michael products in high ee,
although the reactivity was low probably due to steric hindrance. Reaction
did not proceed at all with cyclohexenone and enone 4z with s-trans form.
Thus, we speculate the reaction would proceed through s-cis conformation.
In summary, we demonstrated that the rare earth-alkali metal
heterobimetallic complex functions as Lewis acid-Lewis acid
cooperative catalysts. The reaction mechanism is different from
previous results obtained using related complexes,2 thus broad-
(30) The possibility that amine 3a coordinates to the yttrium center and enone
2a coordinates to Li cannot be excluded completely. Considering the
absolute configuration of aza-Michael products, however, we believe the
transition state model in Figure 14 is more appropriate.
9
J. AM. CHEM. SOC. VOL. 127, NO. 38, 2005 13427