Direct catalytic asymmetric aldol reaction of hydroxyketones: Asymmetric Zn catalysis with a Et2Zn/linked-BINOL complex

Article abstract of DOI:10.1021/ja028926p

Full details of our newly developed catalyses with asymmetric zinc complexes as mimics of class II zinc-containing aldolase are described. A Et₂Zn/(S,S)-linked-BINOL complex was developed and successfully applied to direct catalytic asymmetric

Full text of DOI:10.1021/ja028926p

Published on Web 02/01/2003
Direct Catalytic Asymmetric Aldol Reaction of
Hydroxyketones: Asymmetric Zn Catalysis with a Et2Zn/
Linked-BINOL Complex
Naoya Kumagai,† Shigeki Matsunaga,† Tomofumi Kinoshita,† Shinji Harada,†
Shigemitsu Okada,† Shigeru Sakamoto,‡ Kentaro Yamaguchi,‡ and
Masakatsu Shibasaki*,†
Contribution from the Graduate School of Pharmaceutical Sciences, The UniVersity of Tokyo,
Hongo, Bunkyo-ku, Tokyo 113-0033, Japan, and Chemical Analysis Center, Chiba UniVersity,
Yayoicho, Inage-ku, Chiba-shi 263-0022, Japan
Received October 14, 2002 ; E-mail: mshibasa@mol.f.u-tokyo.ac.jp
Abstract: Full details of our newly developed catalyses with asymmetric zinc complexes as mimics of
class II zinc-containing aldolase are described. A Et2Zn/(S,S)-linked-BINOL complex was developed and
successfully applied to direct catalytic asymmetric aldol reactions of hydroxyketones. A Et2Zn/(S,S)-linked-
BINOL 1 ) 2/1 system was initially developed, which efficiently promoted the direct aldol reaction of
2-hydroxy-2′-methoxyacetophenone (7d). Using 1 mol % of (S,S)-linked-BINOL 1 and 2 mol % of Et2Zn,
we obtained 1,2-dihydroxyketones syn-selectively in high yield (up to 95%), good diastereomeric ratio (up
to 97/3), and excellent enantiomeric excess (up to 99%). Mechanistic investigation of Et2Zn/(S,S)-linked-
BINOL 1, including X-ray analysis, NMR analysis, cold spray ionization mass spectrometry (CSI-MS)
analysis, and kinetic studies, provided new insight into the active oligomeric Zn/(S,S)-linked-BINOL 1/ketone
7d active species. On the basis of mechanistic investigations, a modified second generation Et2Zn/(S,S)-
linked-BINOL 1 ) 4/1 with molecular sieves 3A (MS 3A) system was developed as a much more effective
catalyst system for the direct aldol reaction. As little as 0.1 mol % of (S,S)-linked-BINOL 1 and 0.4 mol %
of Et2Zn promoted the direct aldol reaction smoothly, using only 1.1 equiv of 7d as a donor (substrate/
ligand ) 1000). This is the most efficient, in terms of catalyst loading, asymmetric catalyst for the direct
catalytic asymmetric aldol reaction. Moreover, the Et2Zn/(S,S)-linked-BINOL 1 ) 4/1 system was effective
in the direct catalytic asymmetric aldol reaction of 2-hydroxy-2′-methoxypropiophenone (12), which afforded
a chiral tetrasubstituted carbon center (tert-alcohol) in good yield (up to 97%) and ee (up to 97%), albeit
in modest syn-selectivity. Newly developed (S,S)-sulfur-linked-BINOL 2 was also effective in the direct
aldol reaction of 12. The Et2Zn/(S,S)-sulfur-linked-BINOL 2 ) 4/1 system gave aldol adducts anti-selectively
in good ee (up to 93%). Transformations of the aldol adducts into synthetically versatile intermediates
were also described.
Introduction
classified as class I (peptide base) and class II (zinc metallo-
enzyme), has attracted chemists very much as a potentially
The aldol reaction is generally regarded as one of the most
powerful and efficient carbon-carbon bond-forming reactions.
Many efforts have been devoted to the development of catalytic
and asymmetric aldol reactions,1 but almost all of these reactions
require a preconversion of the ketone or ester moiety into a
more reactive species, such as an enol silyl ether or a ketene
silyl acetal, using no less than stoichiometric amounts of
reagents. On the other hand, living organisms utilize aldolases,
which directly catalyze aldol reactions of unmodified ketones
under mild conditions. The feature of bifunctional aldolases,
advantageous strategy to develop efficient and atom-economic2
artificial asymmetric catalysis.3,4 Since our success in performing
direct catalytic asymmetric aldol reactions with unmodified
ketones using the heterobimetallic multifunctional LaLi3tris-
(binaphthoxide) (LLB) complex,5a several groups, including
ours, have reported various methodologies for direct catalytic
asymmetric aldol reactions over the past 5 years. Our group,
(2) (a) Trost, B. M. Science 1991, 254, 1471. (b) Trost, B. M. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 259.
(3) Recent reviews for bifunctional asymmetric catalysis: (a) Shibasaki, M.;
Kanai, M.; Funabashi, K. Chem. Commun. 2002, 1989. (b) Shibasaki, M.;
Yoshikawa, N. Chem. ReV. 2002, 102, 2187. (c) Gro¨ger, H. Chem.-Eur. J.
2001, 7, 5246. (d) Rowlands, G. J. Tetrahedron 2001, 57, 1865. (e)
Shibasaki, M.; Kanai, M. Chem. Pharm. Bull. 2001, 49, 511. (f) Shibasaki,
M. Chemtracts-Org. Chem. 1999, 12, 979. (g) Shibasaki, M.; Sasai, H.;
Arai, T. Angew. Chem., Int. Ed. Engl. 1997, 36, 1236. (h) Sawamura, M.;
Ito, Y. Chem. ReV. 1992, 92, 857.
† The University of Tokyo.
‡ Chiba University.
(1) For a general review on the catalytic enantioselective aldol reaction, see:
(a) Palomo, C.; Oiarbide, M.; Garc´ıa, J. M. Chem.-Eur. J. 2002, 8, 37. (b)
Johnson, J. S.; Evans, D. A. Acc. Chem. Res. 2000, 33, 325. (c) Nelson, S.
G. Tetrahedron: Asymmetry 1998, 9, 357. (d) Gro¨ger, H.; Vogl, E. M.;
Shibasaki, M. Chem.-Eur. J. 1998, 4, 1137. (e) Carreira, E. M. In
ComprehensiVe Asymmetric Catalysis; Jacobsen, E. N., Pfaltz, A., Yama-
moto, H., Eds.; Springer: Heidelberg, 1999; Vol. 3, Chapter 29.1.
(4) Reviews for direct catalytic asymmetric aldol reactions: (a) Alcaide, B.;
Almendros, P. Eur. J. Org. Chem. 2002, 1595. (b) List, B. Tetrahedron
2002, 58, 5573. See also refs 3a, 3b, and 15b.
9
10.1021/ja028926p CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 2169-2178
2169

A R T I C L E S
Kumagai et al.
Trost et al., and others reported chiral La,5b,d Ba,5c Zn,6 Ca,7
and Ti8 catalysts, respectively, as class II aldolase mimics. List
et al.,9 Barbas et al.,9 and others10 reported direct aldol reactions
with L-proline and its derivatives as class I aldolase mimics.
Recently, several groups reported enantio- and diastereoselective
direct aldol reactions using biological-type catalysts and/or small
molecular catalysts, which considerably widened the scope of
the direct aldol reaction.11,12 In particular, the direct aldol
reaction of hydroxyketones afforded a new method to effectively
synthesize either syn- or anti-1,2-diols.13 In many cases,
however, more than 5 mol % catalyst loading was required to
ensure the completion of the aldol reactions, leaving room for
improvement in terms of catalyst loading and reactivity.
Moreover, mechanistic studies to clarify the structure of the
active species are still lacking in many direct catalytic asym-
metric aldol reactions, despite the fact that knowledge of the
active species structure would aid in logically designing studies
to further improve the catalytic activity, selectivity, and substrate
scope. In this article, we report full details of our new
approach.14 Asymmetric zinc catalysis as a mimic of class II
zinc metalloenzymes was developed and successfully applied
to the direct catalytic enantio- and diastereoselective aldol
reaction of hydroxyketones, giving a new efficient and atom-
economic method for the synthesis of optically active 1,2-diols.
Development of the first generation Et2Zn/linked-BINOL 1 )
2/1 system for the direct aldol reaction of 2-hydroxy-2′-
methoxyacetophenone (7d), mechanistic investigations, includ-
ing revision of the structure of Zn/linked-BINOL complex, and
modification into the second generation Et2Zn/linked-BINOL
1 ) 4/1 system based on those mechanistic studies are discussed.
In the best system, as little as 0.1 mol % chiral ligand (substrate/
chiral ligand ) 1000) efficiently promoted the aldol reaction
Figure 1. (S,S)-Linked-BINOL 1 and its derivatives 2-5.
in excellent enantioselectivity. The direct aldol reaction of
2-hydroxy-2′-methoxypropiophenone (12) using either linked-
BINOL (Figure 1, 1)15,16 or a newly developed sulfur-linked-
BINOL (Figure 1, 2), which afforded aldol adducts with a tert-
OH group, is also described.
Results and Discussion
(A) First Generation Et2Zn/Linked-BINOL ) 2/1 System.
Considering (1) the effectiveness of class II aldolase, in which
the Zn moiety has a crucial role in promoting the direct aldol
reaction of hydroxyketone, and (2) the effectiveness of the
previously reported La/Zn/linked-BINOL complex in asym-
metric catalysis,16b we initiated screening of the catalyst system
using various metals including Zn and lanthanides, linked-
BINOL 1, and its derivatives (Figure 1, 3-5)16e,f as chiral
ligands, 2-hydroxyacetophenone (7a), and aldehyde 6a. Initial
screening revealed that a complex prepared from 20 mol % of
Et2Zn and 10 mol % of linked-BINOL 1 in THF was most
promising.17 As shown in Table 1, entry 1, the Et2Zn/(S,S)-
linked-BINOL 1 complex promoted an aldol reaction of 6a and
7a in THF at -20 °C to give 8 in 66% yield and 70% ee. At
-40 °C with prolonged reaction time, 8 was obtained in good
yield and in moderate dr and ee (entry 2, yield 81%, syn/anti
) 67/33, syn: 78% ee). Neither BINOL itself (entry 3) nor
other bridged BINOL ligands with carbon linkers (Figure 1,
3-5; Table 1, entries 4-6) were effective, suggesting the
importance of a heteroatom in the linker to construct proper
asymmetric space.
Our previous results5b suggested that substituents on the
aromatic ring of acetophenones would affect both the diaste-
reoselectivity and the enantioselectivity. To improve the present
direct catalytic asymmetric aldol reaction, we chose methoxy-
substituted acetophenones, considering the following back-
ground: from a synthetic point of view, the use of aryl ketones
is potentially advantageous over the use of dialkyl ketones such
as acetone and hydroxyacetone,18 because the aromatic ring
functions as a placeholder for further conversions via regio-
(5) (a) Yamada, Y. M. A.; Yoshikawa, N.; Sasai, H.; Shibasaki, M. Angew.
Chem., Int. Ed. Engl. 1997, 36, 1871. (b) Yoshikawa, N.; Yamada, Y. M.
A.; Das, J.; Sasai, H.; Shibasaki, M. J. Am. Chem. Soc. 1999, 121, 4168.
(c) Yamada, Y. M. A.; Shibasaki, M. Tetrahedron Lett. 1998, 39, 5561.
(d) Yoshikawa, N.; Shibasaki, M. Tetrahedron 2001, 57, 2569.
(6) (a) Trost, B. M.; Ito, H. J. Am. Chem. Soc. 2000, 122, 12003. (b) Trost, B.
M.; Silcoff, E. R.; Ito, H. Org. Lett. 2001, 3, 2497. For a partially successful
attempt, see: (c) Nakagawa, M.; Nakao, H.; Watanabe, K.-I. Chem. Lett.
1985, 391.
(7) Suzuki, T.; Yamagiwa, N.; Matsuo, Y.; Sakamoto, S.; Yamaguchi, K.;
Shibasaki, M.; Noyori, R. Tetrahedron Lett. 2001, 42, 4669.
(8) Mahrwald, R.; Ziemer, B. Tetrahedron Lett. 2002, 43, 4459.
(9) (a) List, B.; Lerner, R. A.; Barbas, C. F., III. J. Am. Chem. Soc. 2000, 122,
2395. (b) List, B.; Pojarliev, P.; Castello, C. Org. Lett. 2001, 3, 573. (c)
Sakthivel, K.; Notz, W.; Bui, T.; Barbas, C. F., III. J. Am. Chem. Soc.
2001, 123, 5260. Unmodified aldehydes as donors: (d) Co´rdova, A.; Notz,
W.; Barbas, C. F., III. J. Org. Chem. 2002, 67, 301.
(10) (a) Saito, S.; Nakadai, M.; Yamamoto, H. Synlett 2001, 1245. Unmodified
aldehydes as donors: (b) Northrup, A. B.; MacMillan, D. W. C. J. Am.
Chem. Soc. 2002, 124, 6798. (c) Bøgevig, A.; Kumaragurubaran, N.;
Jørgensen, K. A. Chem. Commun. 2002, 620.
(11) For impressive direct catalytic asymmetric cross-aldol reaction of unmodi-
fied aldehydes by MacMillan et al., see ref 10b. Excellent direct catalytic
diastereoselective aldol reaction using chiral auxiliary was reported by Evans
et al., see: (a) Evans, D. A.; Tedrow, J. S.; Shaw, J. T.; Downey, C. W.
J. Am. Chem. Soc. 2002, 124, 392. (b) Evans, D. A.; Downey, C. W.; Shaw,
J. T.; Tedrow, J. S. Org. Lett. 2002, 4, 1127.
(12) Review for biological and chemical methods: (a) Machajewski, T. D.;
Wong, C.-H. Angew. Chem., Int. Ed. 2000, 39, 1352. For the use of catalytic
antibodies, see: (b) Turner, J. M.; Bui, T.; Lerner, R. A.; Barbas, C. F.,
III; List, B. Chem.-Eur. J. 2000, 6, 2772 and references therein.
(13) The use of hydroxyketones as a donor was pioneered by List et al. anti-
1,2-diol: (a) Notz, W.; List, B. J. Am. Chem. Soc. 2000, 122, 7386. (b)
Yoshikawa, N.; Suzuki, T.; Shibasaki, M. J. Org. Chem. 2002, 67, 2556.
See also refs 9c and 13d. syn-1,2-diol: (c) Trost, B. M.; Ito, H.; Silcoff, E.
R. J. Am. Chem. Soc. 2001, 123, 3367. (d) Yoshikawa, N.; Kumagai, N.;
Matsunaga, S.; Moll, G.; Ohshima, T.; Suzuki, T.; Shibasaki, M. J. Am.
Chem. Soc. 2001, 123, 2466. (e) Kumagai, N.; Matsunaga, S.; Yoshikawa,
N.; Ohshima, T.; Shibasaki, M. Org. Lett. 2001, 3, 1539.
(15) For the synthesis and application of linked-BINOL, see: (a) Matsunaga,
S.; Das, J.; Roels, J.; Vogl, E. M.; Yamamoto, N.; Iida, T.; Yamaguchi,
K.; Shibasaki, M. J. Am. Chem. Soc. 2000, 122, 2252. (b) Matsunaga, S.;
Ohshima, T.; Shibasaki, M. AdV. Synth. Catal. 2002, 344, 4. Linked-BINOL
is also commercially available from Wako Pure Chemical Industries, Ltd.
Catalog No. 152-02431 for (S,S)-ligand, No. 155-02421 for (R,R)-ligand.
Fax +1-804-271-7791 (USA), +81-6-6201-5964 (Japan), +81-3-5201-6590
(Japan).
(16) For other examples of catalytic asymmetric syntheses using linked-BINOL
as a chiral ligand, see: (a) Kim, Y. S.; Matsunaga, S.; Das, J.; Sekine, A.;
Ohshima, T.; Shibasaki, M. J. Am. Chem. Soc. 2000, 122, 6506. (b)
Matsunaga, S.; Ohshima, T.; Shibasaki, M. Tetrahedron Lett. 2000, 41,
8473. (c) Kumagai, N.; Matsunaga, S.; Shibasaki, M. Org. Lett. 2001, 3,
4251. (d) Takita, R.; Ohshima, T.; Shibasaki, M. Tetrahedron Lett. 2002,
43, 4661. For related compounds, see: (e) Vogl, E. M.; Matsunaga, S.;
Kanai, M.; Iida, T.; Shibasaki, M. Tetrahedron Lett. 1998, 39, 7917. (f)
Ishitani, H.; Kitazawa, T.; Kobayashi, S. Tetrahedron Lett. 1999, 40, 2161
and references therein.
(14) For the preliminary communication on this topic concerning the develop-
ment of the first generation Et2Zn/(S,S)-linked-BINOL 1 ) 2/1 complex,
see refs 13d and 13e.
(17) Other metal complexes, such as La-linked-BINOL16a and La-Zn-linked-
BINOL,16b gave less satisfactory results. Other solvents, such as Et2O, DME,
toluene, and CH2Cl2, gave less satisfactory results.
9
2170 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

Direct Catalytic Aldol Reaction of Hydroxyketones
A R T I C L E S
Table 1. Direct Aldol Reaction of 6a and 7 Using Various BINOL
Derivatives
Table 3. Direct Aldol Reaction of Various Aldehydes with
2-Hydroxy-2′-methoxyacetophenone (7d)a
temp
(°C)
time
(h)
yield
(%)
dra
(syn/anti) (syn/anti)
ee
entry
ligand
(x mol %)
1
(S,S)-linked-
BINOL 1
(S,S)-linked-
BINOL 1
10
-20
-40
12
48
66
81
67/33
67/33
56/44
60/40
70/60
78/76
32/47b
5/19
2
10
3
4
5
6
(S)-BINOL
20
10
10
10
-20
-20
-20
-20
24
24
24
24
84
trace
91
3
4
5
trace
a Determined by 1H NMR of crude mixture. b The syn (2S,3R) enantiomer
was obtained in major.
Table 2. Direct Aldol Reaction of 6a with Methoxy-substituted
2-Hydroxyacetophenones 7a
catalyst
(x mol %)
temp
(°C)
time
(h)
yieldb
(%)
drc
eed
entry
X
(syn/anti) (syn/anti)
1
2
3
4
5
6
7
H
7a
10
10
10
10
3
-40 48
-20 24
-30 12
81
73
85
93
94
94
94
67/33
60/40
70/30
89/11
90/10
89/11
87/13
78/76
86/86
77/77
86/88
90/89
92/89
93/91
4′-MeO 7b
3′-MeO 7c
2′-MeO 7d
2′-MeO 7d
2′-MeO 7d
2′-MeO 7d
-30
-30
3
4
1
1
-30 20
-30 16
a All reactions were run on a 1.0 mmol scale at 0.2 M in aldehyde.
1
b Isolated yield after conversion to acetonides. c Determined by H NMR
of crude mixture. d Determined by chiral HPLC analysis.
a Reactions were run on a 0.30 mmol scale (entries 1-4), 0.67 mmol
scale (entry 5), 1.0 mmol scale (entry 6), and 8.0 mmol scale (1.05 g of 8,
entry 7) at 0.2 M in aldehyde. b Isolated yield after conversion to acetonides.
c Determined by 1H NMR of crude mixture. d Determined by chiral HPLC
analysis of diols.
with 3 mol % of ligand 1 (entry 5). Moreover, satisfactory yield
(94%), dr (syn/anti ) 89/11), and ee (syn ) 92%, anti ) 89%)
were achieved after 20 h with as little as 1 mol % of 1 (entry
6), and the reaction proceeded smoothly without any problem
on a gram scale (entry 7).
selective rearrangements.19 By using electron-rich methoxy
substituted acetophenones, conversions such as a Baeyer-
Villiger oxidation would become facile. We investigated the
direct aldol reaction of aldehyde 6a using methoxy substituted
2-hydroxyacetophenones 7b-7d (Table 2). In the case of
2-hydroxy-4′-methoxyacetophenone (7b), the reaction was run
at -20 °C due to the poor solubility of 7b. Although a slightly
higher ee was obtained, dr and yield were lower than in the
case of 7a (entry 2). In the case of 2-hydroxy-3′-methoxyac-
etophenone (7c), the reaction was run at -30 °C. Yield, dr,
and ee were comparable with those of 7a, and the reaction rate
increased (entry 3). Gratifyingly, the reaction rate, yield, dr,
and ee all improved when using 2-hydroxy-2′-methoxy-
acetophenone (7d; entry 4). It is noteworthy that the aldol
reaction of 7d still proceeded smoothly even when the catalyst
amount was reduced. The reaction was completed within 4 h
The optimized reaction conditions were also applicable to
various R-unsubstituted and R-monosubstituted aldehydes (Table
3). In all cases, 1 mol % of 1 was sufficient to complete the
reaction within 24 h. Both normal (entries 1-3) and branched
(entry 4) R-unsubstituted aldehydes afforded good results
(entries 1-4; yield, 84-94%; dr, syn/anti ) 84/16 to 89/11;
ee, syn ) 92-96%, anti ) 87-91%). It should be mentioned
that R-unsubstituted aldehydes gave the corresponding aldol
adducts in good to excellent yield and ee without forming any
self-condensation products. Remarkably, aldehyde 6e (entry 5)
with a methyl ketone unit also afforded the desired aldol adduct
8e with excellent chemoselectivity (yield of 8e: 91%), diaster-
eomeric ratio (syn/anti ) 93/7), and ee (95%). The results
indicate the high chemoselectivity of the present asymmetric
zinc catalysis. In the case of aldehyde 6f (entry 6), synthesis of
the corresponding diol 8f via Sharpless asymmetric dihydroxy-
lation (AD)20 might be difficult due to the chemoselectivity
issue. Aldehydes with oxygen functionalities such as 6g, 6h,
6i, which lead to useful intermediates for the synthesis of
(18) No reaction proceeded with hydroxyacetone as a donor using Zn/linked-
BINOL ) 2/1 complex. L-Proline afforded excellent results with hydroxy-
acetone. See refs 9c and 13a.
(19) Utility of 2′-hydroxy-2-acetylfuran for further conversion via oxidative
cleavage was demonstrated by Trost et al. Trost, B. M.; Vince, Y. S. C.
Org. Lett. 2002, 4, 3513.
9
J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2171

A R T I C L E S
Kumagai et al.
polyoxygenated compounds, were also converted into diols in
excellent ee (syn: 95% in entry 7, 96% in entry 8, and 90% in
entry 9). R-Monosubstituted aldehydes (entries 10-12) showed
good yield (83-95%) and excellent dr (syn/anti ) 96/4 to
97/3) and ee (98-99%).
(B) Mechanistic Studies. What is the mechanism of the direct
aldol reaction? What is the structure of the catalyst? As
mentioned in the Introduction, many asymmetric catalysts for
the direct catalytic asymmetric aldol reaction were recently
developed. Detailed mechanistic investigations, however, es-
pecially trials to characterize the active species in the presence
of a ketone donor, were absent in those studies. Even a little
information about the structure of the active species would help
to further improve the asymmetric catalysis. Thus, we attempted
to clarify the mechanism and structure of the Et2Zn/(S,S)-linked-
BINOL 1 ) 2/1 complex with and without ketone 7d in detail.
Initially, we postulated that the complex prepared from 2
equiv of Et2Zn and 1 equiv of (S,S)-linked-BINOL 1 would be
a bimetallic monomer13d,e on the basis of the related X-ray
structure of a Ga-Li-linked-BINOL complex15a and Zn bimetallic
complexes.21 Unexpectedly, however, X-ray analysis of a crystal
obtained from the Et2Zn/(S,S)-linked-BINOL 1 ) 2/1 solution
in THF revealed that the complex consisted of Zn and 1 in a
ratio of 3:2 [trinuclear Zn3(linked-binol)2thf3] with C2-sym-
metry.22 The revised structure of the Zn/linked-BINOL pre-
formed complex 9 is shown in Figure 2. Three Zn atoms were
aligned in almost a straight line (Zn(2)-Zn(1)-Zn(3) angle )
6.6°), and each Zn center was pentacoordinated.
Figure 2. X-ray structure of the preformed complex Zn3(linked-binol)2-
thf3 9.
The existence of the trinuclear 9 in Et2Zn/1 ) 2/1 solution
was confirmed by NMR and cold spray ionization mass
spectrometry (CSI-MS) analyses,23 and by measuring ethane gas
1
emission. H NMR spectra of (a) free (S,S)-linked-BINOL 1
and (b) the Zn/1 ) 2/1 mixture in THF-d8 are shown in Figure
1
3. The H NMR spectrum of (b) the Zn/1 ) 2/1 mixture had
four different ArCH signals, suggesting a non-C2-symmetric
environment around each of the two linked-BINOL 1 units in
1
9. H NMR spectrum of crystal 9 in THF-d8 was the same as
that of Zn/1 ) 2/1 solution. Free ligand 1 and 9 were mainly
observed in NMR spectra of the Zn/1 ) 1/1 mixture, indicating
that formation of 9 is thermodynamically preferred. The presence
of 9 in major and additional minor free linked-BINOL 1 was
observed in the NMR spectrum of the Zn/1 ) 1.5/1 mixture,
even when freshly titrated Et2Zn (1.0 M in hexanes) was used.24
The necessity of a slight excess of Et2Zn (g2 equiv against 1)
to form a trinuclear Zn/1 ) 3/2 complex was confirmed by
NMR and ethane gas emission experiments. The presence of
additional Et2Zn (0.5 mol equiv with respect to ligand 1) in the
Zn/1 ) 2/1 solution was checked by measuring ethane gas
emission.25 With CSI-MS, a Zn/1 ) 3/2 complex (m/z )
(20) Review: Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem.
ReV. 1994, 94, 2483.
(21) For selected examples of X-ray structures of bimetallic achiral zinc
complexes, see: (a) Kaminskaia, N. V.; Spingler, B.; Lippard, S. J. J. Am.
Chem. Soc. 2000, 122, 6411. (b) Koike, T.; Inoue, M.; Kimura, E.; Shiro,
M. J. Am. Chem. Soc. 1996, 118, 3091. (c) Abe, K.; Izumi, J.; Ohba, M.;
Yokoyama, T.; Okawa, H. Bull. Chem. Soc. Jpn. 2001, 74, 85 and references
therein.
(22) CIF file of crystal 9 was deposited to CCDC (CCDC 186559). See also
Supporting Information.
Figure 3. 1H NMR spectra of (a) (S,S)-linked-BINOL 1 and (b) Et2Zn/
(S,S)-linked-BINOL 1 ) 2/1 in THF-d8.
(23) The coldspray ionization mass spectrometry (CSI-MS) method enables the
direct observation of labile species in solution phase: Sakamoto, S.; Fujita,
M.; Kim, K.; Yamaguchi, K. Tetrahedron 2000, 56, 955 and references
therein.
1418.9)26 was observed as a prominent peak in both the Zn/1
) 1.5/1 and 2/1 solutions.24
(24) For full details of all of the 1H NMR and CSI-MS charts mentioned in
text, see Supporting Information.
9
2172 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

Direct Catalytic Aldol Reaction of Hydroxyketones
A R T I C L E S
Figure 4. CSI-MS spectrum of a ligand 1/Et2Zn/ketone 7d ) 1/2/10
solution.
With isolated crystal 9, the direct aldol reaction of aldehyde
6l did proceed to afford product 8l in 87% yield and 95% ee;
however, the reaction rate was unexpectedly slower (-30 °C,
5 h, with 5 mol % of 9) than that under standard conditions.
The reaction rate with 5 mol % of crystal 9 (which contained
10 mol % of ligand 1) was Vcrystal ) 2.92 × 10-5 M s-1, while
the rate with in situ prepared Zn/1 ) 2/1 catalyst (which
contained 10 mol % of ligand 1) was Vin situ ) 4.47 × 10-5
M
s-1. Because the ee of the products was similar in both cases
(95% ee with crystal 9, 98% ee with in situ prepared Zn/1 )
2/1 catalyst), the effects of a background racemic reaction with
ligand-free Et2Zn should be negligible. A small excess of
Et2Zn would have some interaction with the preformed complex
9 in the presence of ketone 7d, enhancing the reaction rate.27
To clarify the role of a small excess of Et2Zn, we then turned
our attention to observe the formation of a Zn-ketone active
species directly using CSI-MS analysis. When 1 mol equiv of
ketone 7d (against ligand 1) was added to Zn/1 ) 2/1 solution,
peaks derived from 1:Zn:7d ) 2:3:1 (m/z ) 1585.8) and 2:3:2
(m/z ) 1750.9) appeared.24,26 The addition of 1 mol equiv of
aldehyde 6a to the mixture, however, did not afford any aldol
product 8a, which clearly indicates that neither the crystal 9
nor the 1:Zn:7d ) 2:3:1 complex is a real catalyst but only a
preformed complex. As the amount of ketone 7d increased from
1 to 5 and 10 mol equiv, there was a drastic change in the CSI-
MS spectra. With 5 mol equiv of ketone 7d, the peak
corresponding to 9 diminished, and a peak derived from a novel
heptanuclear Zn complex 10 (1:Zn:7d ) 3:7:4, m/z ) 2951.1)
and 11 (1:Zn:7d ) 4:7:4, m/z ) 3565.4)26 appeared as a major
peak. The peak increased by increasing ketone 7d from 5 to 10
mol equiv. In addition, fragment peaks assigned as 1:Zn:7d )
3:7:3 (m/z ) 2785.8), 3:7:2 (m/z ) 2618.2), and 3:7:1 (m/z )
2454.6) were also observed.26 The CSI-MS chart of Zn/1/7d )
2/1/10 is shown in Figure 4. Other charts are shown in the
Supporting Information. The formation of the heptanuclear
complex 10 and 11 should be more favorable under the reaction
Figure 5. Reaction profile with various catalyst loading.
conditions (Table 3), because as much as 200 mol equiv of
ketone 7d exists against ligand 1. More than 100 mol equiv of
7d remained even at the final stage of the reaction under the
conditions listed in Table 3. As expected, the addition of 1 mol
equiv of aldehyde 6a to Zn/1/7d ) 2/1/10 solution led to the
formation of aldol adduct 8a, suggesting that a complicated
oligomeric Zn-rich complex, such as 10 or 11, may be a putative
actual active species.28 All of the trials to elucidate the exact
structure of oligomeric Zn-rich active species, however, failed.
For example, NMR analysis trials only afforded complex charts,
suggesting the existence of several species in equilibrium in
the presence of excess ketone 7d.
To support the CSI-MS observations, we performed kinetic
experiments as follows. (1) The beneficial effects of excess
ketone 7d were confirmed by measuring the initial rate of the
aldol reaction using 1 mol equiv (100 mol %) of 6a, 2 mol
equiv (200 mol %) of 7d, and varied amounts of ligand 1 and
Et2Zn (Zn/1 ) 2/1). The reaction rate increased as usual when
the catalyst loading (based on ligand) increased from x ) 1
and 3 to 5 mol % (Figure 5A). In contrast, the rate gradually
decreased by increasing the catalyst loading (based on ligand)
(25) Three mol equivalents of ethane gas emitted after addition of 2 mol
equivalents of Et2Zn to 1 mol equivalent of ligand 1. When concentrated
H2SO4 was added to the mixture, an additional 1 mol equivalent of ethane
gas was detected, suggesting that 0.5 mol equivalent of Et2Zn remained in
Et2Zn/1 ) 2/1 solution.
(26) The theoretical ion distribution pattern derived from Zn isotopes matched
nicely with the observed peaks.
(27) For the beneficial effects of additional achiral base to enhance the reaction
rate of bifunctional asymmetric catalysis: (a) Arai, T.; Yamada, Y. M. A.;
Yamamoto, N.; Sasai, H.; Shibasaki, M. Chem.-Eur. J. 1996, 2, 1368. (b)
Tian, J.; Yamagiwa, N.; Matsunaga, S.; Shibasaki, M. Angew. Chem., Int.
Ed. 2002, 41, 3636.
(28) In general, it is true that the observed major species is not always an active
species. However, it seems at least sure that some Zn-rich complex should
be an actual active species as supported by kinetics in the following
paragraph.
9
J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2173

A R T I C L E S
Kumagai et al.
Figure 6. Reaction profile with (a) crystal 9 alone (1.5 mol %) and (b)
crystal 9 (1.5 mol %) + additional Et2Zn (1.5 mol %).
Figure 8. Stereochemical course of the direct aldol reaction of hydroxy-
ketone 7d.
Figure 7. Protected ketone 7e and 7f.
Scheme 1. Postulated Catalytic Cycle for the Direct Aldol
Reaction
from x ) 10, 20, 40 to 100 mol % (Figure 5B). With too much
catalyst loading (x ) 40 and 100), there was not enough ketone
7d (200 mol %) to smoothly form the putative active hepta-
nuclear complex. (2) Kinetic profiles of the reaction (a) with
only crystal 9 (1.5 mol %) and (b) with 9 (1.5 mol %) and
additional Et2Zn (1.5 mol %) also supported the CSI-MS
observations. The aldol reaction of 7d and 6a proceeded 1.7
times faster with the additional Et2Zn (Figure 6, (a) Va ) 4.06
× 10-6 M s-1, (b) Vb ) 6.95 × 10-6 M s-1). The results support
the idea that Zn-rich species, like heptanuclear complex 10 or
11, would be the actual active species. Additional Et2Zn could
react with hydroxyketone 7d to form Zn-alkoxide (Zn-7d), and
then react smoothly with the preformed trinuclear Zn3(linked-
binol)2 complex 9 to form oligomeric Zn-rich species, increasing
the total amount of putative Zn-rich active species. We also
hypothesized that Zn-alkoxide (Zn-7d) generated from small
excess Et2Zn would accelerate the catalyst turn over step, in
which aldol adducts dissociate from catalyst via ligand exchange
with ketone 7d or with the Zn-alkoxide (Zn-7d) (details are
discussed in Scheme 1, vide infra). In fact, with 2-methoxy-
2′-methoxyacetophenone (Figure 7, 7e) and 2-(tert-butyldi-
methylsilyloxy)-2′-methoxyacetophenone (Figure 7, 7f), no
reaction proceeded after 25 h at -20 °C, and 12 h at 0 °C using
3 mol % of 1 and 6 mol % of Et2Zn, respectively. The results
support our assumption that Zn-alkoxide (Zn-7d) has a crucial
role to enhance the reaction rate.
The absolute and relative configuration of aldol adducts also
provided useful information regarding the mechanism of the
present aldol reaction. Identical absolute configuration (R) was
expressed at the R-position of both the syn- and the anti-aldol
products (Figure 8A, syn-(2R,3S), anti-(2R,3R)). Both syn- and
anti-isomers were obtained in similarly high ee (>84% ee, see
Table 3), suggesting that the present catalyst differentiates the
enantioface of the enolate very well and aldehydes come from
the Re-face of the zinc enolate (Figure 8A). Syn-selectivity can
be explained by assuming the transition state shown in Figure
8B. On the basis of the mechanisms of related bifunctional
asymmetric catalysis3 and related achiral bimetallic zinc cata-
lysts,21 it seems reasonable to assume a synergistic function of
two or more zinc centers in an oligomeric Zn-rich complex.
The extreme positive effects of the ortho-MeO-group on the ee
and dr can be rationalized as follows. The MeO-group would
coordinate to one of the Zn centers in the oligomeric complex
(Figure 8A), increasing the preference of one chelate complex
(shielding Si-face) over the other (shielding Re-face). Thus, the
shielding Si-face would become more effective, resulting in the
higher ee of both the syn- and the anti-products. Enhanced syn-
selectivity is explained by the steric hindrance of the aromatic
ring in the enolate against aldehydes.
To gain insight into the catalytic cycle of the aldol reaction,
initial rate kinetics were surveyed. The reaction was first-order
on ketone, first-order on the Et2Zn/linked-BINOL 1 complex,
and, importantly, zero-order on aldehyde.29 Thus, the possible
rate-determining step would be the product dissociation step to
regenerate the Zn/linked-BINOL 1/ketone 7d oligomeric puta-
tive active species.30 The postulated catalytic cycle for the direct
aldol reaction is shown in Scheme 1. In the presence of ketone
7d, the oligomeric Zn-rich putative active complex (I) would
(29) See Supporting Information for detailed results, including numerical data.
9
2174 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

Direct Catalytic Aldol Reaction of Hydroxyketones
A R T I C L E S
Table 4. Optimization of Direct Aldol Reaction of 7d with the Second Generation Zn/1 ) 4/1 System
a Reaction was run on a 200 mmol scale (53.7 g of 8l was obtained).
be generated, as observed by CSI-MS analysis (I). Zn-binaph-
thoxide (Ar*O-Zn, Zn-1 in Scheme 1) would function as a
Brønsted base to deprotonate the R-proton in 7d to form (II).
Aldehyde comes from the Re-face of the enolate selectively to
be activated by the Lewis acidic zinc center. Initial rate kinetics
suggest that the 1,2-addition step proceeds quickly to form (IV).
Protonation with the phenolic proton of 1 (Ar*OH) followed
by ligand exchange with 7d would regenerate (I). The catalyst
turn over step would be the rate-limiting step. This is consistent
with the rate acceleration effects due to small excess Et2Zn.
The Zn-alkoxide (Zn-7d complex) should react more smoothly
with (IV) through transmetalation than does ketone 7d itself.27
expected, the reaction proceeded to afford product 8a with the
same ee (93%) as in that entry 2, suggesting that no racemic
pathway with ligand-free Et2Zn was involved in entry 3.
Considering the efficiency, we found that it is important that
the same results (93% ee, yield 78% in entry 3) were obtained
with only one-half the amount of chiral ligand; however, the
chemical yield with 1.1 equiv of ketone 7d was still unsatisfac-
tory. To further improve the turn over step, we surveyed various
achiral additives to find activated MS 3A as the best additive.31
MS 3A was used after activation at 160 °C under reduced
pressure (ca. 0.7 kPa) for 3 h prior to use. In the presence of
MS 3A, the Zn/1 ) 4/1 system promoted the direct aldol
reaction of 2a smoothly using 0.5 mol % (entry 4) or 0.25 mol
% (entry 5) of ligand 1 to give 8a in high yield (entry 4, 93%;
entry 5, 90%), good dr (syn/anti ) 89/11), and high ee (entry
4, 94%; entry 5, 96%) after 18 h. The reaction also proceeded
smoothly with 0.1 mol % ligand loading to afford 8a in high
ee (92% ee), although the reaction time became longer (36 h,
84% yield; entry 6). The optimized conditions were applicable
to R-substituted aldehyde 6l to afford 8l in high yield, dr, and
ee (entry 7, yield 92%, dr ) 98/2, 96% ee). To demonstrate
the practical utility of the present reaction, a direct aldol reaction
was performed on a 200 mmol scale with aldehyde 6l (entry
8). Using 0.25 mol % of 1 (0.5 mmol, 307 mg), 1 mol % of
Et2Zn (2 mL, 2 mmol, 1.0 M in hexanes), and 1.1 equiv of
ketone 7d, we found that the reaction proceeded smoothly and
53.7 g of 8l (yield 96%) was obtained in high dr (syn/anti )
98/2) and ee (94% ee) after 12 h. By using the second generation
Et2Zn/(S,S)-linked-BINOL 1 ) 4/1 with MS 3A system, we
successfully improved the substrate/chiral ligand ratio from 100
(1 mol % with prior 2/1 system) to 400 to 1000 (0.25-0.1 mol
%). This is the most efficient, in terms of catalyst loading,
asymmetric catalyst for the direct catalytic asymmetric aldol
reaction. Considering that the standard catalyst loading for the
direct catalytic asymmetric aldol reaction is 5 to 20 mol %,4-10
we find that the exceptionally low catalyst loading in the present
asymmetric zinc catalysis is remarkable.
(C) Second Generation Et2Zn/Linked-BINOL 1 ) 4/1
System. Although the exact structure of putative active oligo-
meric species was not completely clarified, important informa-
tion to improve the present asymmetric zinc catalysis was
obtained through mechanistic studies in section B: (1) The rate-
determining step of the present direct aldol reaction was
speculated to be the product dissociation step. (2) NMR, CSI-
MS, and ethane gas emission analysis revealed that small excess
Et2Zn remained in the Et2Zn/1 ) 2/1 solution. (3) Kinetic
profiles using preformed trinuclear Zn complex 9 with and
without additional Et2Zn suggested that a small excess of Et2-
Zn accelerates the reaction rate. Thus, we hypothesized that the
addition of more Et2Zn would lead to better reactivity, maintain-
ing high selectivity. In addition, other additives to accelerate
the ligand exchange were also expected to improve the reaction
rate and catalyst loading.
With the first generation Zn/1 ) 2/1 system, the aldol reaction
proceeded smoothly in the presence of 2 equiv of ketone 7d;
however, the chemical yield decreased to 77% with 1.1 equiv
of 7d (Table 4, entry 1 vs entry 2). We assumed that the amount
of ketone loading could be reduced by enhancing the catalyst
turn over step. As the first trial, the reaction with a reduced
amount of ligand 1 (0.5 mol %) was examined to increase the
Zn/1 ratio (entry 3, Zn/1 ) 4/1, vs entry 2, Zn/1 ) 2/1). As
(30) Judging from initial rate kinetic data alone, we cannot completely rule out
the possibility that the enolization step would be the rate-determining step.
Considering the drastic additive effects of MS 3A to accelerate the reaction
(see section C: Second Generation System with MS 3A), we believe it
more reasonable to assume that the rate-limiting step would be the product
dissociation step.
(31) For precedent examples where activated molecular sieves had a key role
to accelerate the catalyst turn over step (the product dissociation step),
see: Iida, T.; Yamamoto, N.; Sasai, H.; Shibasaki, M. J. Am. Chem. Soc.
1997, 119, 4783 and references therein.
9
J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2175

A R T I C L E S
Kumagai et al.
Table 5. Optimization of Direct Aldol Reaction of 2-Hydroxy-2′-methoxypropiophenone (12)
ee (%)
Et2Zn
(x mol %)
ligand 1
(y mol %)
ketone 12
(equiv)
temp
(°C)
time
(h)
yield
(%)
dr
ketone 12
recovered
entry
additive
(syn/anti)
syn
anti
1
2
3
4
5
6
7
8
10
15
20
30
20
20
20
20
5
5
5
5
10
5
2.2
2.2
2.2
2.2
2.2
5
none
none
none
none
none
none
MS 3A
MS 3A
-10
-10
-10
-10
-10
-10
-10
-20
24
24
24
24
24
24
6
39
51
65
64
52
78
94
97
62/38
62/38
64/36
65/35
67/33
64/36
67/33
62/38
88
88
84
76
88
85
75
87
92
93
93
87
92
91
85
95
26
36
61
60
41
25
31
27
5
5
5
5
16
2.2 equiv of 12, the aldol reaction of aldehyde 6a was examined
using Et2Zn/(S,S)-linked-BINOL 1 ) 2/1, 3/1, 4/1, and 6/1 ratio
(entries 1-4). The chemical yield increased by increasing the
ratio from 2/1 (entry 1, 39%) to 4/1 (entry 3, 65%) to afford
13a in high ee (entry 3, syn 84% ee, anti 93% ee), albeit with
modest diastereoselectivity. With a 6/1 ratio (entry 4), the
enantiomeric excess decreased (syn 76% ee, anti 87% ee),
probably due to the racemic pathway being promoted by too
much excess Et2Zn. Interestingly, with the same amount of Et2-
Zn (20 mol %), less (S,S)-linked-BINOL 1 afforded a better
chemical yield (entry 3, 5 mol % of 1, yield 65%; entry 5, 10
mol % of 1, yield 52%), suggesting the effectiveness of the
new 4/1 ratio to improve the reactivity. With the optimized Et2-
Zn/(S,S)-linked-BINOL 1 ) 4/1 ratio, the chemical yield was
improved by using 5 equiv of ketone 12 to afford 13 in 78%
yield (entry 6). The addition of activated MS 3A (entries 7-8)
further improved reactivity, and the product was obtained in
excellent yield (entry 8, 97%) and high ee (entry 8, syn 87%
ee, anti 95% ee) at -20 °C.
Figure 9. Concept for the construction of a tetrasubstituted carbon
stereocenter via direct aldol reaction of hydroxyketone 12.
Scheme 2. Direct Aldol Reaction of R,R-Disubstituted Aldehyde
6m
The optimized reaction conditions, Et2Zn 20 mol %, (S,S)-
linked-BINOL 1 5 mol %, and MS 3A, were applied to various
R-unsubstituted aldehydes. As shown in Table 6, both linear
(entries 1-3) and branched (entry 4) R-unsubstituted aldehydes
afforded products in moderate to high yield (72-97%) and in
moderate to excellent ee (68-96% ee). Aldehydes with oxygen
functionality were also applicable substrates (entries 5-7), and
products were obtained in good yield (80-92%) and high ee
(85-97% ee). In all cases, diastereoselectivity was moderate
(syn/anti ) 59/41-71/29). With R-substituted aldehyde 6l, no
reaction proceeded, probably because the aldol adduct 13l was
thermodynamically unfavorable under the reaction conditions.
Our investigation to improve diastereoselectivity of the direct
aldol reaction of 12 revealed that the heteroatom in the linker
part of linked-BINOL affected the diastereoselectivity. With a
new (S,S)-sulfur-linked-BINOL 2, the aldol reaction of 12
proceeded anti-selectively (Table 7). The synthetic scheme for
the new ligand 2 is shown in Scheme 3. Reactivity with 2 was
With the new 4/1 with the MS 3A system, aldol adduct 8m
from the sterically hindered R,R-disubstituted aldehyde 6m was
also obtained in excellent dr (>98/2) and high ee (92% ee,
Scheme 2). In the case of 6m, 5 mol % of ligand 1 and 2 equiv
of ketone 7d were necessary to achieve 8m in moderate
chemical yield (75%). Because 8m was unstable, the aldol
adduct was isolated after conversion into acetonide.
(D) Construction of Chiral Tetrasubstituted Carbon
Stereocenter. Catalytic asymmetric construction of a chiral
tetrasubstituted carbon stereocenter,32 which is not accessible
via asymmetric hydrogenation reactions,33 is one of the most
important topics in recent synthetic organic chemistry. We
envisioned that the present asymmetric zinc catalysis would also
differentiate the enantioface of tetrasubstituted enolate derived
from 2-hydroxy-2′-methoxypropiophenone (Figure 9, 12). The
aldol reaction of 12 would then afford products with a chiral
tetrasubstituted carbon stereocenter (Figure 9).34 As shown in
Table 5, the ratio of Et2Zn/(S,S)-linked-BINOL 1 and the
addition of MS 3A were important to achieve high yield. With
(34) For the construction of a chiral tert-OH group at the R-position to the
carbonyl group via the diastereoselective aldol reaction using chiral
auxiliary, see: (a) Kamino, T.; Murata, Y.; Kawai, N.; Hosokawa, S.;
Kobayashi, S. Tetrahedron Lett. 2001, 42, 5249. For the construction of a
chiral tert-OH group at the â-position to the carbonyl group via kinetic
resolution (enantioselective retroaldol reaction) using catalytic antibody,
see: (b) List, B.; Shabat, D.; Zhong, G.; Turner, J. M.; Li, A.; Bui, T.;
Anderson, J.; Lerner, R. A.; Barbas, C. F., III. J. Am. Chem. Soc. 1999,
121, 7283.
(32) For a review of catalytic enantioselective synthesis of chiral quaternary
centers, see: Corey, E. J.; Guzman-Perez, A. Angew. Chem., Int. Ed. 1998,
37, 388.
(33) For excellent achievements in catalytic asymmetric hydrogenation of
ketones, see review: Noyori, R.; Ohkuma, T. Angew. Chem., Int. Ed. 2001,
40, 40.
9
2176 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003

Direct Catalytic Aldol Reaction of Hydroxyketones
A R T I C L E S
Scheme 3. Synthesis of a New (S,S)-Sulfur-linked-BINOL 2a
Table 6. Direct Aldol Reaction of Various Aldehydes with
2-Hydroxy-2′-methoxypropiophenone (12)
a (a) AcSK, DMF, 0 °C, 10 min, 97%; (b) NaOMe, MeOH, THF 0 °C,
5 min; (c) 14, NaH, THF, 0 °C, 30 min, 75% (two steps); (d) TsOH‚H2O,
CH2Cl2, MeOH, 40 °C, 12 h, 88%.
Scheme 4. Transformations of Aldol Adducts via Regioselective
Rearrangementa
a Isolated yield. b Determined by 1H NMR of mixture (entries 1-3, 5-7),
determined after isolation (entry 4).
Table 7. Direct Aldol Reaction with
2-Hydroxy-2′-methoxypropiophenone (12) Using
(S,S)-Sulfur-linked-BINOL 2
a (i) mCPBA, NaH2PO4, ClCH2CH2Cl, 50 °C, 2 h; (ii) O-mesitylene-
sulfonylhydroxylamime, CH2Cl2, room temperature, 4 h; (iii) DIBAL, -78
°C to room temperature, 2 h.
(E) Transformation of Aldol Adducts. As mentioned
previously, the usefulness of the aldol adducts becomes much
higher by assuming the 2-methoxyphenyl moiety as a place-
holder for further conversions. As shown in Scheme 4, the
Baeyer-Villiger oxidation proceeded smoothly by treating the
ketones 17a and 18a with mCPBA, probably with the aid of
neighboring oxygen atoms. Interestingly, benzoate 19a was
obtained in 89% yield when acetonide 17a was used. On the
other hand, phenyl ester 20a was obtained in 93% yield when
carbonate analogue 18a was subjected to the same conditions.
In both cases, no regioisomer was observed. Furthermore, 18a
afforded amide 21a exclusively in 97% yield in one step via
Beckmann rearrangement with O-mesitylenesulfonylhydroxyl-
amine (MSH).35 The amide 21a was readily transformed into
22a by reduction with DIBAL. The ortho-MeO-phenyl group
of 22a can be regarded as an oxidatively removable protecting
group of amine.36
1
a Isolated yield. b Determined by H NMR of mixture.
somewhat lower than that with 1, and aldol adducts were
obtained in moderate to good yield (56-82%) by using 10 equiv
of 12. Major anti-isomers were obtained in high ee (81-93%
ee), although the ee of minor syn-isomers was rather low (48-
60% ee). We believe that heteroatoms in the linker would
coordinate to the Zn atom, affecting the structure of the putative
active oligomeric complex. It is difficult to explain precisely
the effect of heteroatoms in the linker, however, because the
exact structure of the putative oligomeric active species is not
clarified yet.
(35) Tamura, Y.; Fujiwara, H.; Sumoto, K.; Ikeda, M.; Kita, Y. Synthesis 1973,
215.
9
J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003 2177

A R T I C L E S
Kumagai et al.
Summary
ligand ) 1000). The new system was also applicable to the
direct aldol reaction of 2-hydroxy-2′-methoxypropiophenone
(12) to afford products with a tetrasubstituted carbon stereocenter
in good yield (up to 97%) and ee (up to 97% ee), although
diastereoselectivity was still unsatisfactory. The reaction pro-
vided a new method to synthesize chiral tert-alcohol with a small
molecular chiral catalyst. Application of the present efficient
asymmetric zinc catalysis to other asymmetric reactions such
as direct aldol and Mannich reactions of unmodified esters is
currently underway in our group.
We successfully developed new highly efficient direct
catalytic asymmetric aldol reactions of hydroxyketones by
catalysis with asymmetric zinc complexes as mimics of zinc
metalloenzymes (class II aldolases). The first generation Et2-
Zn/linked-BINOL 1 ) 2/1 system was initially developed, which
promoted the direct aldol reaction of 2-hydroxy-2′-methoxy-
acetophenone (7d) with 1 mol % ligand loading. Through
mechanistic studies using X-ray, NMR, CSI-MS, and kinetic
experiments, the role of a small excess of Et2Zn to accelerate
the catalyst turn over step was revealed. On the basis of
mechanistic studies, the second generation Et2Zn/linked-BINOL
1 ) 4/1 with MS 3A was developed as a much more efficient
system. With the new system, as little as 0.1 mol % of ligand
and 0.4 mol % of Et2Zn efficiently promoted the aldol reaction
in the presence of only 1.1 equiv of ketone 7d. This is the most
efficient, in terms of catalyst loading, asymmetric catalyst for
direct catalytic asymmetric aldol reactions (substrate/chiral
Acknowledgment. We thank financial support by RFTF and
Grant-in-Aid for Encouragements for Young Scientists (B) (for
S.M.) from JSPS and MEXT. We thank Dr. T. Ohshima, Dr.
T. Suzuki, and Dr. N. Yoshikawa for contributions at the
beginning stage of this project.
Supporting Information Available: Experimental procedures,
characterization of the products, X-ray data (CIF), and detailed
data for kinetic studies (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
(36) (a) Porter, J. R.; Traverse, J. F.; Hoveyda, A. H.; Snapper, M. C. J. Am.
Chem. Soc. 2001, 123, 10409. (b) Ishitani, H.; Ueno, M.; Kobayashi, S. J.
Am. Chem. Soc. 2000, 122, 8180 and references therein.
JA028926P
9
2178 J. AM. CHEM. SOC. VOL. 125, NO. 8, 2003