Anti-selective catalytic asymmetric nitroaldol reaction via a heterobimetallic heterogeneous catalyst

Article abstract of DOI:10.1021/ja905885z

Full details of an anti-selective catalytic asymmetric nitroaldol reaction promoted by a heterobimetallic catalyst comprised of Nd₅O(O ⁱPr)₁₃, an amide-based ligand, and NaHMDS (sodium hexamethyldisilazide) are described.

Full text of DOI:10.1021/ja905885z

Published on Web 09/09/2009
anti-Selective Catalytic Asymmetric Nitroaldol Reaction via a
Heterobimetallic Heterogeneous Catalyst
Tatsuya Nitabaru,† Akihiro Nojiri,† Makoto Kobayashi,‡ Naoya Kumagai,*,† and
Masakatsu Shibasaki*,†
Graduate School of Pharmaceutical Sciences, The UniVersity of Tokyo, 7-3-1 Hongo, Bunkyo-ku,
Tokyo 113-0033, Japan, and Process Chemistry, R&D, Kissei Pharmaceutical Company, Ltd.,
197-5 Kamiyoshi, Kubiki-ku, Joetsu, Niigata 942-0145, Japan
Received July 15, 2009; E-mail: mshibasa@mol.f.u-tokyo.ac.jp; nkumagai@mol.f.u-tokyo.ac.jp
Abstract: Full details of an anti-selective catalytic asymmetric nitroaldol reaction promoted by a hetero-
bimetallic catalyst comprised of Nd5O(OiPr)13, an amide-based ligand, and NaHMDS (sodium hexameth-
yldisilazide) are described. A systematic synthesis and evaluation of amide-based ligands led to the
identification of optimum ligand 1m, which provided a suitable platform for the Nd/Na heterobimetallic
complex. During the catalyst preparation in THF, a heterogeneous mixture developed and centrifugation
of the suspension allowed for separation of the precipitate, which contained the active catalyst and which
could be stored for at least 1 month without any loss of catalytic performance. The precipitate promoted a
nitroaldol (Henry) reaction for a broad range of nitroalkanes and aldehydes under heterogeneous conditions,
affording the corresponding 1,2-nitroalkanol in a highly anti-selective (up to anti/syn ) >40/1) and
enantioselective manner (up to 98% ee). Inductively coupled plasma (ICP) and X-ray fluorescence (XRF)
analyses revealed that the precipitate indeed included both neodymium and sodium, which was further
supported by high-resolution ESI TOF MS spectrometry.
Introduction
perfect atom efficiency of the reaction as well as the obvious
synthetic utility of the products have substantially advanced this
The nitroaldol (Henry) reaction (eq 1) is a valuable methodol-
ogy for carbon-carbon bond construction, and, despite initially
being discovered at the end of the 19th century,1,2 nearly 100
years passed before the reaction was rendered enantioselective
in a catalytic manner. In 1992, LaLi3(binaphthoxide)3 (LLB)
was disclosed as an efficient asymmetric catalyst for the
nitroaldol reaction of nitromethane.3 Since then, the catalytic
asymmetric nitroaldol reaction has emerged as a powerful
protocol for the rapid assembly of enantiomerically enriched
1,2-nitro alkanols. Facile reduction of 1,2-nitro alkanols allows
for direct access to 1,2-amino alcohols, a structural motif that
occurs ubiquitously in a wide range of natural products,
therapeutics, chiral ligands, and their key intermediates.4 The
area over the last two decades, uncovering a number of metal
catalysts and organocatalysts that promote the nitroaldol reaction
in a highly enantioselective manner.5,6 Despite the intense
demand for a diastereo- and enantioselective nitroaldol reaction
using nitroalkanes other than nitromethane, progress toward this
aim has been limited to a few sporadic reports. Since our report
of a syn-selective catalytic asymmetric nitroaldol reaction in
1995,7 several catalytic systems such as chiral guanidine and
Cu-diamine catalysts have achieved the transformation.8 anti-
Diastereoselectivity in the catalytic asymmetric nitroaldol reac-
(5) For selected examples for catalytic asymmetric nitroaldol reaction using
aldehydes and nitromethane, see: (a) Christensen, C.; Juhl, K.;
Jørgensen, K. A. Chem. Commun. 2001, 2222. (b) Trost, B. M.; Yeh,
V. S. C. Angew. Chem., Int. Ed. 2002, 41, 861. (c) Evans, D. A.;
Seidel, D.; Rueping, M.; Lam, H. W.; Shaw, J. T.; Downey, C. W.
J. Am. Chem. Soc. 2003, 125, 12692. (d) Kogami, Y.; Nakajima, T.;
Ashizawa, T.; Kezuka, S.; Ikeno, T.; Yamada, T. Chem. Lett. 2004,
33, 614. (e) Palomo, C.; Oiarbide, M.; Laso, A. Angew. Chem., Int.
Ed. 2005, 44, 3881. (f) Choudary, B. M.; Ranganath, K. V. S.; Pal,
U.; Kantam, M. L.; Sreedhar, B. J. Am. Chem. Soc. 2005, 127, 13167.
(g) Saa´, J. M.; Tur, F.; Gonza´lez, J.; Vega, M. Tetrahedron: Asymmetry
2006, 17, 99. (h) Marcelli, T.; van der Haas, R. N. S.; van Maarseveen,
J. H.; Hiemstra, H. Angew. Chem., Int. Ed. 2006, 45, 929. (i) Arai,
T.; Watanabe, M.; Fujiwara, A.; Yokoyama, N.; Yanagisawa, A.
Angew. Chem., Int. Ed. 2006, 45, 5978. (j) Mandal, T.; Samanta, S.;
Zhao, C.-G. Org. Lett. 2007, 9, 943. (k) Xiong, Y.; Wang, F.; Huang,
X.; Wen, Y.; Feng, X. Chem.-Eur. J. 2007, 13, 829. (l) Ma, K.; You,
J. Chem.-Eur. J. 2007, 13, 1863. (m) Bandini, M.; Piccinelli, F.;
Tommasi, S.; Umani-Ronchi, A.; Ventrici, C. Chem. Commun. 2007,
616. (n) Bandini, M.; Benaglia, M.; Sinisi, R.; Tommasi, S.; Umani-
Ronchi, A. Org. Lett. 2007, 9, 2151. (o) Blay, G.; Domingo, L. R.;
Herna´ndez-Olmos, V.; Pedro, J. R. Chem.-Eur. J. 2008, 14, 4725. (p)
Ube, H.; Terada, M. Bioorg. Med. Chem. Lett. 2009, 19, 3895.
† The University of Tokyo.
‡ Kissei Pharmaceutical Co., Ltd.
(1) Henry, L. C. R. Hebd. Seances Acad. Sci. 1895, 120, 1265.
(2) For recent reviews, see: (a) Shibasaki, M.; Gro¨ger, H. In ComprehensiVe
Asymmetric Catalysis; Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.;
Springer: Berlin, Germany, 1999; Vol. III, pp 1075-1090. (b) Luzzio,
F. A. Tetrahedron 2001, 57, 915. (c) Palomo, C.; Oiarbide, M.; Mielgo,
A. Angew. Chem., Int. Ed. 2004, 43, 5442. (d) Shibasaki, M.; Gro¨ger,
H.; Kanai, M. In ComprehensiVe Asymmetric Catalysis, Supplement
1; Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer:
Hidelberg, Germany, 2004; pp 131-133. (e) Boruwa, J.; Gogoi, N.;
Saikia, P. P.; Barua, N. C. Tetrahedron: Asymmetry 2006, 17, 3315.
(f) Palomo, C.; Oiarbide, M.; Laso, A. Eur. J. Org. Chem. 2007, 2561.
(3) Sasai, H.; Suzuki, T.; Arai, S.; Arai, T.; Shibasaki, M. J. Am. Chem.
Soc. 1992, 114, 4418.
(4) For reviews on asymmetric synthesis of 1,2-amino alcohols, see:(a)
Kolb, H. C.; Sharpless, K. B. In Transition Metals in Organic
Synthesis; Beller, M., Bolm, C., Eds.; Wiley-VCH: Weinheim, 1998;
p 243. (b) Reetz, M. T. Chem. ReV. 1999, 99, 1121. (c) Bergmeier,
S. C. Tetrahedron 2000, 56, 2561.
9
13860 J. AM. CHEM. SOC. 2009, 131, 13860–13869
10.1021/ja905885z CCC: $40.75  2009 American Chemical Society

anti-Selective Catalytic Asymmetric Nitroaldol Reaction
A R T I C L E S
tion, however, has remained a formidable task, likely because
a simple chelation model prefers syn-diastereoselectivity. A
highly anti-selective catalytic asymmetric nitroaldol reaction
utilizing a chiral P-spiro triaminoiminophosphorane as an
effective organocatalyst was recently reported by Ooi et al.,
although low temperature (-78 °C) was required.9 In our
continuing studies on the catalytic asymmetric nitroaldol reac-
tion, we investigated a bimetallic strategy to achieve high anti-
diastereoselectivity.10 We reported that a Nd/Na/amide-based
ligand heterobimetallic catalytic system promoted an anti-
selective catalytic asymmetric nitroaldol reaction of
nitroethane,10a where the presence of the two distinct metal
cations was key for anti-diastereoselectivity. Herein, we describe
a significant advance of the Nd/Na heterobimetallic catalytic
system for a scalable and efficient anti-selective nitroaldol
reaction. Systematic design of the amide-based ligand and
careful observation of the catalyst preparation procedure led us
to identify a highly diastereo- and enantioselective heteroge-
neous Nd/Na heterobimetallic catalyst, which was characterized
by inductively coupled plasma (ICP), X-ray fluorescence (XRF),
and ESI TOF MS analyses. The broad substrate generality and
operational simplicity of the present heterogeneous anti-selective
nitroaldol protocol are particularly valuable for the synthesis
of enantioenriched 1,2-amino alcohols, which constitute a
versatile substructure in medicinal chemistry.
Figure 1. Diastereoselectivity in nitroaldol reactions. (a) Preferential
formation of anti-diastereomer in the reaction using silyl nitronates. (b)
Preferential formation of syn-diastereomer in the presence of monometallic
catalyst. (c) Preferential formation of anti-diastereomer by heterobimetallic
catalyst.
nitronates and aldehydes by Seebach et al.11 An antiparallel
orientation of a silyl nitronate and an aldehyde was proposed
to explain the observed anti-diastereoselectivity (Figure 1a). This
strategy for anti-diastereoselectivity was valid in a catalytic
enantioselective nitroaldol reaction using silyl nitronates as
reported by Jørgensen et al.12 and Maruoka et al.,13 affording
the anti-1,2-nitro alkanols in a highly enantioselective manner.
The use of silyl nitronates, however, requires an extra process
for their preparation and produces unwanted silicon-derived
wastes, severely limiting the utility of the nitroaldol reaction
as an atom-economical, waste-free, and environmentally benign
protocol. We envisioned attaining the antiparallel transition state
with an in situ generated metal nitronate, thereby allowing the
anti-selective catalytic asymmetric nitroaldol reaction to occur
for a simple set of substrates (nitroalkanes and aldehydes) under
waste-free proton-transfer conditions.14 In contrast to the
reaction with silyl nitronates, a nitroaldol reaction using a metal-
based catalyst predominantly affords the syn product, likely
because a cyclic transition state in which both the nitronate and
the aldehyde coordinate to one metal cation M is involved
(Figure 1b). We hypothesized that a heterobimetallic complex
comprising a ligand with an appropriate spatial arrangement for
metal coordination sites would provide a suitable chiral platform
for the antiparallel transition state. In this catalyst design, each
distinct metal cation, M1 and M2, works independently as a
Lewis acid to activate an aldehyde and a Brønsted base (metal
phenoxide) to form metal nitronates; therefore, the reaction
Results and Discussion
Catalyst Design. High anti-diastereoselectivity was reported
in an elegant study of the racemic nitroaldol reaction using silyl
(6) For selected examples for catalytic asymmetric nitroaldol reaction using
ketones and nitromethane, see: (a) Christensen, C.; Juhl, K.; Jørgensen,
K. A. Chem. Commun. 2001, 2222. (b) Misumi, Y.; Bulman, R. A.;
Matsumoto, K. Heterocycles 2002, 56, 599. (c) Du, D. M.; Lu, S. F.;
Fang, T.; Xu, J. X. J. Org. Chem. 2005, 70, 3712. (d) Choudary, B. M.;
Ranganath, K. V. S.; Pal, U.; Kantam, M. L.; Sreedhar, B. J. Am.
Chem. Soc. 2005, 127, 13167. (e) Li, H.; Wang, B.; Deng, L. J. Am.
Chem. Soc. 2006, 128, 732. (f) Tosaki, S.-y.; Hara, K.; Gnanadesikan,
V.; Morimoto, H.; Harada, S.; Sugita, M.; Yamagiwa, N.; Matsunaga,
S.; Shibasaki, M. J. Am. Chem. Soc. 2006, 128, 11776. (g) Tur, F.;
Saa´, J. M. Org. Lett. 2007, 9, 5079. (h) Takada, K.; Takemura, N.;
Cho, K.; Sohtome, Y.; Nagasawa, K. Tetrahedron Lett. 2008, 49, 1623.
(i) Blay, G.; Herna´ndez-Olmos, V.; Pedro, J. R. Org. Biomol. Chem.
2008, 6, 468. (j) Chen, X.; Wang, J.; Zhu, Y.; Shang, D.; Gao, B.;
Liu, X.; Feng, X.; Su, Z.; Hu, C. Chem.-Eur. J. 2008, 14, 10896. (k)
Hara, K; Tosaki, S.-y.; Gnanadesikan, V.; Morimoto, H.; Harada, S.;
Sugita, M.; Yamagiwa, N.; Matsunaga, S.; Shibaski, M. Tetrahedron
2009, 65, 5030.
(7) Sasai, H.; Tokunaga, T.; Watanabe, S.; Suzuki, T.; Itoh, N.; Shibasaki,
M. J. Org. Chem. 1995, 60, 7388.
(8) (a) Sohtome, Y.; Hashimoto, Y.; Nagasawa, K. Eur. J. Org. Chem.
2006, 2894. (b) Arai, T.; Watanabe, M.; Yanagisawa, A. Org. Lett.
2007, 9, 3595. (c) Sohtome, Y.; Takemura, N.; Takada, K.; Takagi,
R.; Iguchi, T.; Nagasawa, K. Chem. Asian J. 2007, 2, 1150. (d) Arai,
T.; Takashita, R.; Endo, Y.; Watanabe, M.; Yanagisawa, A. J. Org.
Chem. 2008, 73, 4903. For a partially successful example of syn-
selective catalytic asymmetric nitroaldol reaction, see: (e) Yoshimoto,
J.; Sandoval, C. A.; Saito, S. Chem. Lett. 2008, 37, 1294. See also ref
10c.
(10) (a) Nitabaru, T.; Kumagai, N.; Shibasaki, M. Tetrahedron Lett. 2008,
49, 272. (b) Handa, S.; Nagawa, K.; Sohtome, Y.; Matsunaga, S.;
Shibasaki, M. Angew. Chem., Int. Ed. 2008, 47, 3230. (c) Sohtome,
Y.; Kato, Y.; Handa, S.; Aoyama, N.; Nagawa, K.; Matsunaga, S.;
Shibasaki, M. Org. Lett. 2008, 10, 2231.
(11) (a) Colvin, E. W.; Seebach, D. J. Chem. Soc., Chem. Commun. 1978,
689. (b) Seebach, D.; Beck, A. K.; Lehr, F.; Weller, T.; Colvin, E.
Angew, Chem., Int. Ed. Engl. 1981, 20, 397. (c) Seebach, D.; Beck,
A. K.; Mukhopadhyay, T.; Thomas, E. HelV. Chim. Acta 1982, 65,
1101.
(9) Uraguchi, D.; Sakaki, S.; Ooi, T. J. Am. Chem. Soc. 2007, 129, 12392.
For partially successful anti-selective catalytic asymmetric nitroaldol
reactions, see refs 5o,p. For anti-selective nitroaldol reaction of
benzaldehyde catalyzed by hydroxynitrile lyase, see: (a) Purkarthofer,
T.; Gruber, K.; Gruber-Khadjawi, M.; Waich, K.; Skranc, W.; Mink,
D.; Griengl, H. Angew. Chem., Int. Ed. 2006, 45, 3454. (b) Gruber-
Khadjawi, M.; Purkarthofer, T.; Skranc, W.; Griengl, H. AdV. Synth.
Catal. 2007, 349, 1445.
(12) Risgaard, T.; Gothelf, K. V.; Jørgensen, K. A. Org. Biomol. Chem.
2003, 1, 153.
(13) Ooi, T.; Doda, K.; Maruoka, K. J. Am. Chem. Soc. 2003, 125, 2054.
(14) For a theoretical study on the diastereoselectivity in nitroaldol reaction,
see: Lecea, B.; Arrieta, A.; Morao, I.; Coss´ıo, F. P. Chem.-Eur. J.
1997, 3, 20.
9
J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009 13861

A R T I C L E S
Nitabaru et al.
Table 1. Evaluation of Ligands 1a-c for the Construction of
Heterobimetallic Catalyst for Nitroaldol Reactiona
Figure 2. Amide backbone as a platform for bimetallic complex.
proceeds predominantly through an anti-periplanar transition
state to afford the anti-nitroaldol product (Figure 1c). The
arrangements of the two metals would be key to overriding the
reaction through the undesired cyclic transition state.
Identification of an Appropriate Amide-Based Ligand for
a Heterobimetallic System. To attain the anti-periplanar transi-
tion state, our primary focus was directed to exploit an amide
backbone as a platform for two distinct metal cations.15 On the
basis of transition state analysis, we anticipated that a ligand
bearing a binary metal coordination site linked by a stereogenic
R-amino acid through two amide bonds would provide a suitable
three-dimensional arrangement for the anti-periplanar transition
state (Figure 2). Taking the acidity of nitroalkanes into account,
a phenol is a feasible functional group to incorporate metal
cations on the amide-based ligand where facile deprotonation
by the action of metal phenoxide is expected. We assumed that
the combination of a rare earth (RE) metal (M1) cation as a
Lewis acid and an alkali metal (M2) phenoxide as a Brønsted
base would provide the desired transition state.16 We then
planned to investigate the heterobimetallic catalytic system
comprising an amide-based ligand 1, RE(OiPr)3, and sodium
hexamethyldisilazide (NaHMDS). Initial attempts were made
to evaluate three ligands, 1a-c, bearing an L-valine core and
differently installed phenol groups in a nitroaldol reaction of
benzaldehyde (2a) and nitroethane (3a). The catalyst was
prepared by mixing Pr(OiPr)3 and ligand 1 in a 1:2 ratio at room
temperature in THF, followed by the addition of NaHMDS
(equimolar amounts to Pr) at 0 °C. As we expected, the reaction
proceeded to afford the product 4aa with a slight preference
for the anti-diastereomer in the presence of 9 mol % of the
catalyst (based on Pr), although both the yield and the
diastereoselectivity were unsatisfactory (Table 1, entries 1-3).
a 2a, 0.1 mmol; 3a, 1.0 mmol. b Determined by 1H NMR with
1,4-dioxane as an internal standard. c Determined by 1H NMR analysis.
d Approximate value due to low chemical yield.
Figure 3. A possible site-selective coordination of Pr3+ and Na+ in ligand
1.
(15) Recent selected examples of asymmetric catalysis by peptide-based
catalysts: (a) Miller, S. J.; Copeland, G. T.; Papaioannou, N.; Horst-
mann, T. E.; Ruel, E. M. J. Am. Chem. Soc. 1998, 120, 1629. (b)
Fierman, M. B.; O’Leary, D. J.; Steinmetz, W. E.; Miller, S. J. J. Am.
Chem. Soc. 2004, 126, 6967. (c) Zhao, Y.; Rodrigo, J.; Hoveyda, A. H.;
Snapper, M. C. Nature 2006, 443, 67. For recent reviews on small
peptide-based asymmetric catalysis, see: (d) Hoveyda, A. H.; Hird,
A. W.; Kacprzynski, M. A. Chem. Commun. 2004, 1779. (e) Blank,
J. T.; Miller, S. J. Biopolymers (Pept. Sci.) 2006, 84, 38. (f) Colby
Davie, E. A.; Mennen, S. M.; Xu, Y.; Miller, S. J. Chem. ReV. 2007,
107, 5759.
The catalyst prepared with two molar equivalents of NaHMDS
to Pr(OiPr)3 exhibited higher catalytic activity (entries 4-6),
and the highest diastereo- and enantioselectivity were observed
with the reaction using the ligand 1b (entry 5), where m-
hydroxybenzoyl and o-aminophenol groups were furnished as
metal coordination sites. It is likely that the Pr3+ cation was
more prone to be localized at the aminophenol part in a seven-
membered chelate with a neighboring amide carbonyl while Na+
was located at the other side, leading to the appropriate spatial
arrangements for the desired anti-periplanar transition state
(Figure 3b). In contrast, poor stereoselectivity was observed with
ligand 1a, presumably because Pr3+ and Na+ showed no
preference for the coordination sites of ligand 1a; for both
metals, chelate formation was possible at either the o-hydroxy-
benzoyl or the o-aminophenol moieties (Figure 3a). Whereas
the preferential localization of Pr3+ and Na+ was also plausible
(16) For the properties on RE complexes, see: (a) Lanthanide and Actinide
Chemistry; Cotton, S., Ed.; Wiley: New York, 2006. For reviews, see:
(b) Aspinall, H. C. Chem. ReV. 2002, 102, 1807. (c) Tsukube, H.;
Shinoda, S. Chem. ReV. 2002, 102, 2389. (d) Parker, D. Chem. Soc.
ReV. 2004, 33, 156. (e) Bari, L. D.; Salvadori, P. Coord. Chem. ReV.
2005, 249, 2854. (f) Bu¨nzli, J.-C. Z. Acc. Chem. Res. 2006, 39, 53,
and references cited therein. For examples of asymmetric catalysis
with the combination of amide-based ligand and RE, see: (g) Mashiko,
T.; Hara, K.; Tanaka, D.; Fujiwara, Y.; Kumagai, N.; Shibasaki, M.
J. Am. Chem. Soc. 2007, 129, 11342. (h) Nojiri, A.; Kumagai, N.;
Shibasaki, M. J. Am. Chem. Soc. 2008, 130, 5630. (i) Mashiko, T.;
Kumagai, N.; Shibasaki, M. Org. Lett. 2008, 10, 2725. (j) Nojiri, A.;
Kumagai, N.; Shibasaki, M. J. Am. Chem. Soc. 2009, 131, 3779.
9
13862 J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009

anti-Selective Catalytic Asymmetric Nitroaldol Reaction
A R T I C L E S
Table 2. Screening of Rare Earth Metals and Alkali Metals with
Ligand 1ba
Table 3. Evaluation of Ligands 1d-g Derived From Various
R-Amino Acids Based on the Structure of 1ba
entry RE alkali metal source H2O (mol %) yieldb (%) anti/sync ee (anti) (%)
1
2
3
4
5
6
7
8
9
La
Pr
NaHMDS
NaHMDS
NaHMDS
NaHMDS
NaHMDS
NaHMDS
NaHMDS
NaHMDS
LiHMDS
KHMDS
73
76
85
78
55
32
33
50
9
7.2/1
8.4/1
8.5/1
7.7/1
6.5/1
2.8/1
2.6/1
2.0/1
2.3/1
1.8/1
1.3/1
10.1/1
26
38
40
35
41
18
10
-7
4
Nd
Sm
Gd
Dy
Er
Yb
Nd
a 2a, 0.1 mmol; 3a, 1.0 mmol. b Determined by 1H NMR with
1,4-dioxane as an internal standard. c Determined by 1H NMR analysis.
d Average of three runs. 1.8 mol % of Nd5O(OiPr)13 (9 mol % based on
Nd) was used.
10 Nd
11 Nd
12 Nd
83
1
83
3
0
43
NaHMDS
9
a 2a, 0.1 mmol; 3a, 1.0 mmol. b Determined by 1H NMR with
1,4-dioxane as an internal standard. Determined by H NMR analysis.
c
1
m-hydroxybenzoyl part would contribute little to chelate forma-
tion as the C-C single bond between the m-hydroxy phenyl
group and amide carbonyl could freely rotate with two preferable
metastable conformers A and B, with dihedral angles of ca. 0°
and 180°, respectively (Figure 4). In the proposed transition
state model (Figure 1c), the relative location of the two distinct
metals is of prime importance, and possible bond rotation may
compromise the construction of a suitable heterobimetallic
architecture. Thus, we next focused on the restriction of the
bond rotation. o-Fluorobenzamide forms an intramolecular
hydrogen bond between the fluoro substituent and the amide
N-H hydrogen, thereby posing a restriction on the bond
rotation.20 The C-F· · ·H-N hydrogen bonding was exploited
in a ligand 1h, where the installed 2-fluoro substituent was
anticipated to prefer the conformation of the m-hydroxybenzoyl
part as conformer A (dihedral angle ca. 0°) (Figure 4).21 Ligand
1h outperformed the parent L-Leu ligand 1g in the standard
reaction conditions, affording 4aa in 97% yield with anti/syn
) 19/1 and 73% ee (anti) (Chart 1, entries 1, 2). The installation
of a 2-chloro substituent, which hardly formed a hydrogen bond
with amide NH,22 resulted in stereoselectivity inferior to that
obtained with 1g (entry 3). Ligand 1j, in which conformational
with ligand 1c (Figure 3c), the arrangement of two metals
(opposite to the case of ligand 1b) would not be suitable for
the desired anti-periplanar transition state. Screening of both
RE metals and alkali metals with the prototype ligand 1b
indicated that Nd(OiPr)3 was the best RE source in combination
with NaHMDS, exhibiting the highest diastereo- (anti/syn )
8.5/1) and enantioselectivity (40% ee (anti)) (Table 2, entry 3).
The use of LiHMDS or KHMDS as an alkali metal source
resulted in worse diastereoselectivity, and no enantioselection
was observed (entries 9, 10). Poor catalytic activity and
diastereoselectivity in the absence of an alkali metal verified
that the heterobimetallic system was crucial for constructing
the anti-periplanar transitions state (entry 11). Occasional
fluctuations in the stereoselectivity prompted us to investigate
the addition of water, which proved to be beneficial for both
diastereoselectivity and enantioselectivity (entry 12).17
Given the preliminary identification of reaction conditions
for the anti-selective nitroaldol reaction, we designed ligand
analogues bearing different amino acid residues 1d-g based
on the structure of 1b, and we evaluated these ligand analogues
in the nitroaldol reaction of 2a and 3a (Table 3). Among the
ligand analogues prepared from different R-amino acids, ligand
1g derived from L-Leu was superior in terms of both catalytic
activity and stereoselectivity (entries 1-5). Whereas some
fluctuation in chemical yield and stereoselectivity was occasion-
ally observed with the catalyst prepared from Nd(OiPr)3, the
use of Nd5O(OiPr)13 as the Nd source18 circumvented the
reproducibility problem, and a marginally better result was
obtained.19 Further structural manipulation of ligand substructure
was conducted by systematically varying conformational flex-
ibility and the electronic nature of the aromatic ring (Chart 1).
We anticipated that the conformation of the aminophenol part
would be relatively rigid upon complexation with the metal due
to the chelate formation with Nd3+ (Figure 3b). On the other
hand, the meta-orientation of the hydroxyl group of the
(19) Synthesis and characterization of Nd5O(OiPr)13, see: Kritikos, M.;
Moustiakimov, M.; Wijk, M.; Westin, G. J. Chem. Soc., Dalton Trans.
2001, 1931, and references cited therein. In contrast to the possible
incomplete complexation with ligand due to the non-ordered aggrega-
tion state of Nd(OiPr)3, reproducible formation of the desired complex
was anticipated with µ-oxo isopropoxide Nd5O(OiPr)13. Nd5O(OiPr)13
is available from Kojundo Chemical Co. Ltd. at http://www.kojun-
do.co.jp/english/index.html; fax, +81-49-284-1351; e-mail,
(20) (a) Zhao, X.; Wang, X.-Z.; Jiang, X.-K.; Chen, Y.-Q.; Li, Z.-T.; Chen,
G. J. J. Am. Chem. Soc. 2003, 125, 15128. (b) Li, C.; Ren, S.-F.;
Hou, J.-L.; Yi, H.-P.; Zhu, S.-Z.; Jiang, X. K.; Li, Z.-T. Angew. Chem.,
Int. Ed. 2005, 44, 5725. (c) Li, C.; Zhu, Y.-Y.; Yi, H.-P.; Li, C.-Z.;
Jiang, X.-K.; Li, Z.-T.; Yu, Y.-H. Chem.-Eur. J. 2007, 13, 9990. (d)
Zhu, Y.-Y.; Wu, J.; Li, C.; Zhu, J.; Hou, J.-L.; Li, C.-Z.; Jiang, X.-
K.; Li, Z.-T. Cryst. Growth Des. 2007, 7, 1490. (e) Tomita, K.; Oishi,
S.; Ohno, H.; Fujii, N. Biopolymers (Pept. Sci.) 2008, 90, 503.
(21) Indeed, a DFT calculation suggested that the rotational conformer A
(dihedral angle ca. 0°) was much more stable than that of B (dihedral
angle ca. 180°) in ligand 1h, whereas 1g displayed a small energy
difference between rotational conformers A and B. The energy profile
of the C-C bond rotation in a m-hydroxybenzoyl ester analogue was
similar to that of 1h, suggesting that the predominance of rotational
conformer A (dihedral angle ca. 0°) in 1h was due to the C-F· · ·H-
N hydrogen bond. See the Supporting Information for details.
(22) Zhu, Y.-Y.; Yi, H.-P.; Li, C.; Jiang, X.-K.; Li, Z.-T. Cryst. Growth
Des. 2008, 8, 1294.
(17) H2O may dissociate Nd(OiPr)3 aggregate or serve as an additional
ligand to Nd to acquire a more suitable chiral environment for the
reaction. See also ref 24.
(18) The amount was calculated on the basis of Nd content; cf., the molar
relationship of 3 mol % catalyst for 0.3 mmol of aldehyde is the
following: Nd5O(OiPr)13/1b/NaHMDS/aldehyde ) 1.8 µmol/18 µmol/
18 µmol/300 µmol, Nd/1b/Na/aldehyde ) 9/18/18/300 ) 1/2/2/33.3.
9
J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009 13863

A R T I C L E S
Nitabaru et al.
Chart 1
restriction was effected by lactam formation, gave the product
in low stereoselectivity, likely because the appended ethylene
linker would make the preferences of rotational conformers
along N-CR single bond unsuitable for the present reaction
(entry 4). Difluoro analogue 1k was expected to have no
preference for conformers A and B and exhibited poor stereo-
selectivity, although electronic perturbation of the m-hydroxy-
benzoyl group by two fluoro substituents may also be involved
(entry 5).23 Subsequent modification of the electronic nature of
the aminophenol part was achieved by introducing a fluoro
substituent, revealing that 5-fluoro analogue 1m further im-
proved the stereoselectivity to nearly perfect anti-selectivity with
84% ee (entries 6, 7). The moderate stereoselectivity obtained
from the reaction with related ligand 1n lacking the 2-fluoro
substituent on the m-hydroxybenzoyl part clearly illustrates the
importance of the conformational control of the m-hydroxy-
benzoyl part (entry 8).
Development of Heterogeneous Reaction Conditions. During
the catalyst preparation, the addition of Nd5O(OiPr)13 (0.2 M
in THF, based on Nd), NaHMDS (1.0 M in THF), and H2O
(0.2 M In THF) to the clear solution of ligand 1m in THF
produced a white precipitate, and the resulting mixture became
a suspension. Occasionally the precipitate partially dissolved
upon the addition of the substrates benzaldehyde (2a) and
nitroethane (3a), and the reaction mixture remained heteroge-
neous until completion of the nitroaldol reaction. The hetero-
geneous catalyst mixture led us to investigate whether the active
catalyst was in the solution phase or in the precipitate. Careful
observation of the transition of the mixture during the catalyst
preparation process revealed that (1) the addition of
Nd5O(OiPr)13 to the solution of 1m at room temperature gave a
clear mixture; (2) subsequent addition of NaHMDS at 0 °C
turned the mixture into a suspension; (3) upon the addition of
nitroethane (3a) (ca. 70 equiv to Nd) at room temperature, a
clear solution developed and the resulting mixture gradually
turned into a suspension again within 10 min; and (4) the use
of Nd5O(OiPr)13 as an Nd source instead of Nd(OiPr)3 allowed
us to prepare the catalyst without additional H2O (Figure 5).24
The mixture was left to stand at room temperature for 1 h, and
the resulting suspension was centrifuged (ca. 104 rpm) at room
temperature for 30 s. The separated supernatant and the
precipitate25 were individually evaluated in the nitroaldol
reaction of 2a and 3a (Scheme 1). Based on a comparison of
the stereoselectivities obtained in each case (Scheme 1, (a) vs
(b)), the active catalyst was in the precipitate, but the reason
for the difference in catalytic activity was not conclusive because
the mole fraction of distributed catalytically active components
to the precipitate and the supernatant was not determined. Of
particular note is the substantial enhancement of the enantio-
selectivity (94%) in the reaction using the precipitate beyond
what was obtained in the standard reaction conditions (Scheme
1, (a) vs (c)), where the whole catalyst mixture, including both
the precipitate and the supernatant, was used as catalyst. This
is likely due to the fact that the reaction partially proceeded
through the poor-performing catalyst in the supernatant to reduce
the overall stereoselectivity.
These findings prompted us to investigate the quantitative
distribution of catalyst components, Nd3+, ligand 1m, and Na+
in the precipitate and the supernatant. The catalyst was prepared
by following the procedure described in Figure 5 to give the
(23) Installation of multiple fluorine substitutes may weaken C-F· · ·H-
N hydrogen bonding, disfavoring the coplanarity of the aromatic ring
and planar amide. See also ref 20e.
(24) The beneficial effect of H2O was more prominent for the catalyst
prepared from Nd(OiPr)3 than that prepared from Nd5O(OiPr)13.
Nd(OiPr)3 may form Nd5O(OiPr)13 in the presence of H2O.
(25) The precipitate was washed by two cycles of the following procedure:
(1) re-suspended with dry THF; (2) mixed by vortex mixer; (3)
centrifuged; and (4) decantation of supernatant. See the Supporting
Information for details.
Figure 4. Metastable rotational conformers A and B of ligand 1g. (A)
Dihedral angle between the plane of m-hydroxybenzoyl group and the amide
plane is ca. 0°. (B) Dihedral angle between m-hydroxybenzoyl group and
the amide plane is ca. 180°.
9
13864 J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009

anti-Selective Catalytic Asymmetric Nitroaldol Reaction
A R T I C L E S
Figure 5. A schematic view of the catalyst preparation procedure.
Scheme 1. Nitroaldol Reaction with the Separated Precipitate and
Supernatant
Table 4. Elemental Analysis of the Precipitate and the
Supernatant by ICP and XRF Analyses
ICP precipitates
XRF precipitates
molar ratio
element
mass %
molar ratio
mass %
Nd
Na
F
15.8
5.81
ND
1
2.3
ND
16.2
4.23
3.91
1
1.8
1.92 (0.96 as 1m)
by ESI TOF MS analysis.26 Sample aliquots were withdrawn
at each specific stage during the catalyst preparation and
subjected to MS analysis in THF solution. Although no
prominent peaks appeared in the solution of 1m and
Nd5O(OiPr)13 (Figure 6a), the subsequent addition of NaHMDS
induced a significant change in the MS spectrum, affording a
series of peaks derived from [(1m)4Nd2Nan]2- (n ) 1,2),
[(1m)5Nd2Nan]2- (n ) 2-4), and [(1m)6Nd2Nan]2- (n ) 4,5)
having a matched isotopic pattern with less than 10 mmu
accuracy (Figure 6b).29 The appearance of these related peaks
suggested that the 1m/Nd fragment would, with the aid of Na+,
associate to form a white suspension before the addition of
nitroethane (3a) in Figure 5.30 The MS spectrum of the
precipitate obtained after the addition of nitroethane (3a) was
precipitate after washing with dry THF and drying under
vacuum, which was then subjected to ICP and XRF analyses
(Table 4).26 Both ICP and XRF analyses proved that ca. 16
mass % of the precipitate came from Nd3+, indicating that ca.
85% of the added Nd3+ was incorporated in the precipitate. In
contrast, only ca. 39% of ligand 1m was incorporated based on
the fluorine atom content in XRF analysis.26 Of particular note
is that both Nd3+ and Na+ were incorporated in the precipitates
and worked as a highly anti-selective and enantioselective
heterogeneous heterobimetallic catalyst.27 Given the approximate
molar ratio of Nd/1m/Na ) 1/0.96/1.8 in XRF analysis, different
catalyst preparation procedures with varied ratios of Nd/1m/
Na were conducted to reduce the amount of the ligand 1m, ca.
60% of which remained in the supernatant under the standard
conditions. These attempts, however, failed to give the active
heterogeneous catalyst: either no precipitation or, if any
precipitate appeared, lower stereoselectivity in the nitroaldol
reaction.28
(27) For reviews for chiral catalysts incorporated into metal-organic
coordination frameworks, see: (a) Baiker, A. J. Mol. Catal. A 1997,
115, 473. (b) Yaghi, O. M.; Li, H.; Davis, C.; Richardson, D.; Groy,
T. L. Acc. Chem. Res. 1998, 31, 474. (c) Hagrman, P. J.; Hagrman,
D.; Zubieta, J. Angew. Chem., Int. Ed. 1999, 38, 2639. (d) Blake, A. J.;
Champness, N. R.; Hubberstey, P.; Li, W.-S.; Withersby, M. A.;
Schroder, M. Coord. Chem. ReV. 1999, 183, 117. (e) Moulton, B.;
Zaworotko, M. J. Chem. ReV. 2001, 101, 1629. (f) Kesanli, B.; Lin,
W. B. Coord. Chem. ReV. 2003, 246, 305. (g) Studer, M.; Blaser,
H. U.; Exner, C. AdV. Synth. Catal. 2003, 345, 45. (h) Yaghi, O. M.;
O’Keeffe, M.; Ockwig, N. W.; Chae, H. K.; Eddaoudy, M.; Kim, J.
Nature 2003, 423, 715. (i) Dai, L. X. Angew. Chem., Int. Ed. 2004,
43, 5726. (j) Mallat, T.; Orglmeister, E.; Baiker, A. Chem. ReV. 2007,
107, 4863. For selected leading examples of “self-supported” chiral
catalysts comprised of metal-organic coordination networks, see: (k)
Sawaki, T.; Aoyama, Y. J. Am. Chem. Soc. 1999, 121, 4793. (l) Seo,
J. S.; Whang, D.; Lee, H.; Jun, S. I.; Oh, J.; Jeon, Y. J.; Kim, K. A.
Nature 2000, 404, 982. (m) Hu, A.; Ngo, H. L.; Lin, W. J. Am. Chem.
Soc. 2003, 125, 11490. (n) Takizawa, S.; Somei, H.; Jayaprakash, D.;
Sasai, H. Angew. Chem., Int. Ed. 2003, 42, 5711. (o) Guo, H.; Wang,
X.; Ding, K. Tetrahedron Lett. 2004, 45, 2009. (p) Liang, Y.; Jing,
Q.; Li, X.; Shi, L.; Ding, K. J. Am. Chem. Soc. 2005, 127, 7694.
(28) This is presumably because an excess amount of ligand 1m was
essential for effective nucleation of the precipitate through bridging
Nd and Na cations. 1m on the surface of the heterogeneous catalyst
might be washed away. Excess 1m could be recovered from the
supernatant.
ESI TOF MS Analyses of the Heterobimetallic Complex in
the Precipitate. To gain further insight into the structure of the
active catalyst, the catalyst preparation procedure was traced
(29) Comparison of observed and theoretical isotopic pattern of the
prominent peaks was shown in the Supporting Information.
(26) See the Supporting Information for details.
9
J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009 13865

A R T I C L E S
Nitabaru et al.
Figure 6. ESI TOF MS spectra (negative ion mode) of the Nd/Na complex at each stage of catalyst preparation. (a) MS spectrum of 1m/Nd5O(OiPr)13
)
2/1 (based on Nd) solution in THF; (b) MS spectrum of 1m/Nd5O(OiPr)13/NaHMDS ) 2/1/2 (based on Nd) diluted solution in THF; and (c) MS spectrum
of 1m/Nd5O(OiPr)13/NaHMDS/nitroethane (3a) ) 2/1/2/excess (based on Nd) in THF.
Scheme 2. Nitroaldol Reaction with Nitroethane-d2 (3a-d2) Using
Stoichiometric Amount of the Precipitate
remarkably different. The peaks in Figure 6b mostly disap-
peared, and peaks derived from a series of 1m/Nd/Na fragments
including nitroethane (3a) appeared (Figure 6c).29,31 Based on
observations during the catalyst preparation, upon the addition
of NaHMDS, 1m/Nd/Na heterobimetallic complexes were
initially formed, which were then dissolved by the addition of
nitroethane (3a), and subsequently associated again incorporat-
ing nitroethane (3a) to afford the heterogeneous Nd/Na hetero-
bimetallic active catalyst. When the nitroaldol reaction using
10 equiv of nitroethane-d2 (3a-d2) was run with a stoichiometric
amount of thoroughly washed precipitate,32 11% of the product
was derived from nitroethane (3a), indicating that nitroethane
(3a) was tightly bound to the heterobimetallic complex in the
) 0.96/1/1.8, it can be assumed that (1m)2Nd2 fragments
associated with varying amounts of Na+ and nitroethane (3a)
to form an insoluble heterobimetallic entity.27,30 The predomi-
nance of (1m)4Nd2 fragments in the MS spectrum was presum-
ably due to the negative ion detection mode.
Nonlinear Effect. When the nitroaldol reaction with the
preformed precipitates was run in the presence of excess free
ent-1m, an enantiomer of 1m, the product 4aa was obtained
with stereoselectivity comparable to that obtained in Scheme
precipitate (Scheme 2). According to XRF analysis, which
showed that the composition of the precipitate was 1m/Nd/Na
(30) For an example of lanthanide coordination polymers from exo-
coordination of Na+, see:(a) Singh-Wilmot, M. A.; Richards-Johnson,
R. U.; Dawkins, T. N.; Lough, A. J. Inorg. Chim. Acta 2007, 360,
3727, and references cited therein. For an example of lanthanide
coordination polymer incorporating K+, see: (b) Bu¨rgstein, M. R.;
Roesky, P. W. Angew. Chem., Int. Ed. 2000, 39, 549. (c) Bu¨rgstein,
M. R.; Gamer, M. T.; Roesky, P. W. J. Am. Chem. Soc. 2004, 126,
5213, and references cited therein.
(31) The precipitates were dissolved with THF/DMSO ) 100/1 mixture
for ESI TOF MS analysis. The nitroaldol reaction of 2a and 3a using
the precipitates in THF/DMSO ) 600/1 afforded the comparable
reaction outcome (-40 °C, 20 h, 85% yield (determined by 1H NMR),
anti/syn ) 21/1, 86% ee (anti)).
(32) The precipitate was washed by four cycles of the following procedure:
(1) re-suspended with dry THF; (2) mixed by vortex mixer; (3)
centrifuged; and (4) decantation of supernatant. See the Supporting
Information for details.
9
13866 J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009

anti-Selective Catalytic Asymmetric Nitroaldol Reaction
A R T I C L E S
Scheme 3. Reaction Using the Precipitates in the Presence of
ent-1m
Scheme 4. Storage of the Heterobimetallic Catalyst
1a, suggesting that ligand 1m in the precipitate was not labile
and the architecture of the heterobimetallic complex in the
precipitate was rather stable (Scheme 3).33 Indeed, the precipitate
could be stored at -20 °C, and no degradation occurred at least
for 1 month in terms of the catalytic activity and stereoselectivity
observed in the reaction with the aged precipitate (Scheme 4).
Intriguingly, significant enantioenrichment was observed during
precipitate formation. When the catalyst was prepared with a
partially resolved 1m, the optical purity of 1m in the precipitate
was positively deviated, whereas that in the supernatant had a
negative deviation (Figure 7). The Nd/Na heterobimetallic
complex likely grew through the preferential nucleation of the
ligand with identical absolute configuration, and the enantio-
purity of 1m in the precipitate reached 86% ee and 64% ee
starting with 50% ee and 25% ee of 1m, respectively. Moreover,
prominent enantioamplification was observed in the nitroaldol
reaction promoted by the partially heterochiral precipitate,
affording 4aa in anti/syn ) >40/1 and 92% ee with 64% ee
precipitate, which was prepared from 25% ee of free 1m. It
can be deduced that the multiple homochiral 1m was involved
in the transition state as the heterobimetallic assembly in the
precipitate and the local heterochiral heterobimetallic site had
much lower catalytic activity. The double selection process of
homochirality (precipitate formation and catalysis) led to a
remarkable positive nonlinear effect.34
Figure 7. Precipitate formation and nitroaldol reaction using partially
resolved ligand 1m. Chemical yield of 4aa was determined by H NMR
1
analysis with 1,4-dioxane as an internal standard. Diastereomeric ratio was
1
determined by H NMR.
and washed twice with dry THF. The catalyst amount designates
the amount of Nd5O(OiPr)13 used (based on Nd) for catalyst
preparation. In the reaction with benzaldehyde (2a), catalyst
loading could be reduced to 1 mol % to complete the reaction
with marginal loss of stereoselectivity (entry 2).37 The reaction
proceeded smoothly with at least 1.5 equiv of nitroethane (3a)
(entries 3, 4). Nonpolar functionality on the aromatic ring hardly
affected the stereoselectivity (entries 5, 6). Aromatic aldehydes
bearing either electron-donating or -withdrawing substituents
were suitable substrates, and excellent stereoselectivities were
observed (entries 7-11) except for 4-nitrobenzaldehyde 2f, in
which the nitro functionality may competitively coordinate to
anti-Selective Asymmetric Nitroaldol Reaction Promoted
by the Heterobimetallic Catalyst. Having established a prepara-
tion procedure for the heterogeneous Nd/Na heterobimetallic
catalyst, we examined the substrate scope of the anti-selective
asymmetric nitroaldol reaction (Table 5).35,36 The heterogeneous
catalyst was prepared by mixing Nd5O(OiPr)13, 1m, and NaH-
MDS in a ratio of 1:2:2 in THF, and the following addition of
nitroethane (3a) led to the suspension, which was centrifuged
(35) The heterogeneous catalyst prepared from Sm5O(OiPr)13 was also
effective for nitroaldol reaction of 2a and 3a: 3 mol % of catalyst,
-40 °C, 20 h, 96% yield, anti/syn ) >40/1, 91% ee. ICP analysis of
the heterogeneous catalyst prepared from 1m/Sm5O(OiPr)13/NaHMDS
revealed that the catalyst was heterobimetallic containing both Sm
and Na like the catalyst prepared from Nd5O(OiPr)13. See the
Supporting Information for details.
(36) At higher temperatures (higher than -20 °C), retro-nitroaldol reaction
proceeded depending on the substrate used and may decrease the
stereoselectivity.
(33) Low chemical yield would be due to the competitive protonation by
phenol functionality of additional ent-1m. Low reaction rate was
observed when phenolic additives were employed in this catalytic
system.
(34) (a) Girard, C.; Kagan, H. B. Angew. Chem., Int. Ed. 1998, 37, 2922.
(b) Blackmond, D. G. Acc. Chem. Res. 2000, 33, 402. (c) Satya-
narayana, T.; Abraham, S.; Kagan, H. B. Angew. Chem., Int. Ed. 2009,
48, 456, and references cited therein.
(37) Additional 1 mol % of NaHMDS would scavenge the trace amount
of acid included in benzaldehyde (2a) without affecting the stereo-
selectivity.
9
J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009 13867

A R T I C L E S
Nitabaru et al.
Table 5. anti-Selective Catalytic Asymmetric Nitroaldol Reaction Promoted by the Nd/Na Heterobimetallic Complexa
aldehyde 2
R1
nitroalkane 3
entry
)
R2
)
equiv
x
solvent
product
time (h)
yieldb (%)
anti/sync
ee (anti) (%)
1
2d
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Ph
Ph
Ph
Ph
2a
2a
2a
2a
2b
2c
2d
2e
2f
2g
2h
2i
2j
2k
2l
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Et
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3b
3b
10
10
3
3
1
3
3
3
3
3
3
3
3
3
3
3
6
6
6
3
3
3
3
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
THF
DME
DME
DME
THF
THF
Et2O
THF
4aa
4aa
4aa
4aa
4ba
4ca
4da
4ea
4fa
4ga
4ha
4ia
4ja
4ka
4la
20
20
20
20
20
20
14
20
20
20
14
20
20
20
22
20
20
20
48
48
99
96
95
91
99
89
95
99
99
88
96
99
96
99
92
93
99
93
85
75
>40/1
>40/1
>40/1
33/1
>40/1
>40/1
37/1
>40/1
5.7/1
15/1
23/1
13/1
40/1
4.91
8.3/1
3.4/1
>40/1
19/1
13/1
19/1
92
91
93
89
98
97
94
97
86
94
89
82
97
77
95
87
90
90
90
95
1.5
10
10
10
10
10
10
10
10
10
10
10
10
10
10
5
2,4-Me2C6H3
4-BnOC6H4
4-BrC6H4
4-FC6H4
4-NO2
4-CN
4-MeO2CC6H4
2-furyl
trans-PhCHdCH
PhCH2CH2
cHex
nC8H11
Ph
4-FC6H4
Ph
Ph
2m
2a
2e
2a
2a
4ma
4ab
4eb
4ac
4ad
Et
TBSOCH2
BnOCH2
3c
3d
5
a 2, 0.3 mmol; 3, 3.0 mmol. b Isolated yield. c Determined by 1H NMR analysis. d 1 mol % of NaHMDS was added to the Nd/Na heterobimetallic
catalyst.
the coordination site of 3a to compromise the transition state.
delivering the desired products 4ac and 4ad with high stereo-
selectivity (entries 19, 20).40 It is intriguing that only nitroethane
(3a) formed the insoluble Nd/Na heterobimetallic catalyst,
allowing for the isolation of the active catalyst exhibiting high
stereoselectivity.
A slight decrease in stereoselectivity was observed with
heteroaromatic aldehyde 2i (entry 12). The present protocol was
also effective for various aliphatic aldehydes, including linear
and branched aldehydes, albeit in lower anti-selectivity (entries
13-16). For aliphatic aldehydes, 6 mol % of catalyst loading
assured the high stereoselectivity, and the use of DME as the
solvent improved the enantioselectivity. To broaden the scope
of the present methodology, we next focused on the implemen-
tation of nitroalkanes other than 3a in this catalytic system
(entries 17-20). Because the catalyst preparation procedure
required 3a (Figure 5) and 3a was incorporated in the active
heterobimetallic catalyst, catalyst preparation was conducted
with nitropropane (3b) following the same procedure as
described in Figure 5 for the reaction of 2a and 3c. Unexpect-
edly, the precipitate did not appear when 3b was used for
catalyst preparation.38 Whereas the heterobimetallic catalyst
prepared with 3a included 3a as its components, the amount of
coordinated 3a in the catalyst was negligible under the condi-
tions of 3 mol % catalyst loading. Thus, the nitroaldol reaction
of 2a and 2e with 3b was performed with the catalyst prepared
using 3a, affording anti-4ab and anti-4eb in high enantiose-
lectivity without the formation of 4aa and 4ea (entries 17, 18).39
Although functionalized nitroalkanes 3c and 3d also failed to
form the heterogeneous heterobimetallic catalyst, they served
as viable nucleophiles with the catalyst prepared using 3a,
A large-scale demonstration of the nitroaldol reaction with
the heterogeneous catalyst highlights the utility and operational
simplicity of the present protocol. The nitroaldol reaction of
50 g of aldehyde 2c and nitroethane 3a (used as received from
a commercial source) was run with the heterogeneous hetero-
bimetallic catalyst prepared from 1 mol % Nd5O(OiPr)13 at -30
°C.41 24 h of stirring resulted in 85% conversion (anti/syn )
>40/1, 97% ee), and the following single recrystallization of
the crude mixture provided 51.2 g of the nitroaldol product 4ca
in pure form (76% yield) with anti/syn ) >40/1, 98% ee, which
is a key intermediate for the potent ꢀ3-adrenoreceptor agonist
5 (Scheme 5).42,43 The process is now under extensive inves-
tigation for an industrial application for the bulk production of
5.
Conclusion
A new anti-selective catalytic asymmetric nitroaldol reaction
was achieved with a heterogeneous Nd/Na heterobimetallic
(39) In the 1H NMR spectrum of the crude mixture, the signals derived
from 4aa or 4ea were not observed. Nitroethane (3a) incorporated in
the heterogeneous catalyst may be reluctant to be replaced with bulkier
nitroalkanes. If initially incorporated nitroethane (3a) reacted with an
aldehyde, the amount of the nitroaldol product derived from 3a would
be negligible with low catalyst loading.
(40) In the reaction using 3c in THF, the heterogeneous catalyst dissolved
and the reaction mixture became a clear solution, affording the product
with slightly lower stereoselectivity.
(38) The catalyst prepared from nitropropane (3b) did not form the
precipitate. When the resulting solution was used as catalyst in the
reaction of 2a and 3b, the reaction was sluggish (3 mol % of catalyst
loading, -40 °C, 21 h), affording <14% (determined by 1H NMR) of
the product 4ab in anti/syn ) 3.2/1 and 41% ee (anti). This is likely
due to the fact that the catalytically active Nd/Na heterobimetallic
catalyst would not be formed with 3b even in the solution phase.
(41) The reaction conditions were specifically optimized for this substrate,
and the heterobimetallic catalyst prepared from Nd5O(OiPr)13/1m/
NaHMDS ) 1/2/1.75 ratio gave the best performance.
9
13868 J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009

anti-Selective Catalytic Asymmetric Nitroaldol Reaction
A R T I C L E S
Scheme 5. Large-Scale Demonstration of anti-Selective Nitroaldol
Reaction Promoted by the Heterogeneous Nd/Na Heterobimetallic
Catalyst
arrangement of the two distinct metals. ICP, XRF, and ESI MS
analyses suggested that the 1m/Nd complex aggregated with
the aid of Na+ incorporating nitroethane to form a heterogeneous
active catalyst. The catalyst was storable and effective for a
broad range of aldehydes and nitroalkanes, showcasing the utility
of the present protocol.
Acknowledgment. This work was financially supported by a
Grant-in-Aid for Scientific Research (S) (M.S.) and a Grant-in-
Aid for Scientific Research on Innovative Areas (N.K.) from JSPS
and MEXT. T.N. thanks JSPS for a predoctoral fellowship. Mr. S.
Ikeda is gratefully acknowledged for XRF analysis of the hetero-
geneous catalyst. We thank Prof. S. Kobayashi, and Drs. R.
Matsubara, H. Miyamura, and R. Akiyama for technical assistance
in ICP analysis. Ms. K. Tanaka at JEOL Corp. is gratefully
acknowledged for ESI TOF MS analysis of the Nd/Na heterobi-
metallic catalyst.
Supporting Information Available: Detailed experimental
procedures and characterization of new compounds. This
material is available free of charge via the Internet at http://
pubs.acs.org.
complex. The present methodology allowed for easy access to
highly versatile enantioenriched anti-1,2-amino alcohols, which
are of particular importance in the enantioselective synthesis
of biologically active natural products and therapeutics. A
rational approach toward anti-diastereoselectivity through a
heterobimetallic strategy led to the identification of the hetero-
geneous Nd/Na heterobimetallic catalyst organized by amide-
based phenolic ligand 1m, exhibiting an exquisite spatial
JA905885Z
(43) In this specific case, the diastereomeric ratio of 4ca was monitored
during the progress of the nitroaldol reaction (conditions: precipitate
obtained from 2 mol % Nd5O(OiPr)13, 4 mol % of 1m, 3 mol % of
NaHMDS, 2 mol % H2O was used as catalyst, -30 °C, THF, 10 equiv
of 3a). The yield and stereoselectivity at the specified period of the
reaction time were as follows: 3 h, 55% yield, anti/syn ) >40/1, 97%
ee (anti); 6 h, 72% yield, anti/syn ) >40/1, 98% ee (anti); 12 h, 86%
yield, anti/syn ) 34/1, 97% ee (anti); 18 h, 89% yield, anti/syn )
31/1, 96% ee (anti); 24 h, 92% yield, anti/syn ) 29/1, 96% ee (anti).
At temperatures higher than -30 °C and with higher catalyst loading
(2 mol %), the retro-nitroaldol reaction competitively occurred to
decrease the stereoselectivity slightly.
(42) (a) Tanaka, N.; Tamai, T.; Mukaiyama, H.; Hirabayashi, H.; Muranaka,
M.; Sato, M.; Akahane, M. Patent WO00/02846, 1999. (b) Tanaka,
N.; Tamai, T.; Mukaiyama, H.; Hirabayashi, A.; Muranaka, H.;
Ishikawa, T.; Kobayashi, J.; Akahane, S.; Akahane, M. J. Med. Chem.
2003, 46, 105. (c) Tanaka, T.; Tamai, T. JP Patent JP 2002-64840; see
also ref 10b.
9
J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009 13869