Iron- and bismuth-catalyzed asymmetric mukaiyama aldol reactions in aqueous media

Article abstract of DOI:10.1002/asia.201301149

We have developed asymmetric Mukaiyama aldol reactions of silicon enolates with aldehydes catalyzed by chiral Fe^II and Bi^III complexes. Although previous reactions often required relatively harsh conditions, such as strictly anhydrous conditions, very low temperatures (-78 °C), etc., the reactions reported herein proceeded in the presence of water at 0 °C. To find appropriate chiral water-compatible Lewis acids for the Mukaiyama aldol reaction, many Lewis acids were screened in combination with chiral bipyridine L1, which had previously been found to be a suitable chiral ligand in aqueous media. Three types of chiral catalysts that consisted of a Fe^II or Bi^III metal salt, a chiral ligand (L1), and an additive have been discovered and a wide variety of substrates (silicon enolates and aldehydes) reacted to afford the desired aldol products in high yields with high diastereo- and enantioselectivities through an appropriate selection of one of the three catalytic systems. Mechanistic studies elucidated the coordination environments around the Fe^II and Bi^III centers and the effect of additives on the chiral catalysis. Notably, both Bronsted acids and bases worked as efficient additives in the Fe ^II-catalyzed reactions. The assumed catalytic cycle and transition states indicated important roles of water in these efficient asymmetric Mukaiyama aldol reactions in aqueous media with the broadly applicable and versatile catalytic systems. Copyright

Full text of DOI:10.1002/asia.201301149

FULL PAPER
DOI: 10.1002/asia.201301149
Iron- and Bismuth-Catalyzed Asymmetric Mukaiyama Aldol Reactions in
Aqueous Media
Taku Kitanosono,[a] Thierry Ollevier,[b] and Shu Kobayashi*[a]
¯
Dedicated to Professor Teruaki Mukaiyama on the occasion of the 40th anniversary of the Mukaiyama aldol reaction
Abstract: We have developed asym-
metric Mukaiyama aldol reactions of
silicon enolates with aldehydes cata-
lyzed by chiral FeII and BiIII complexes.
Although previous reactions often re-
quired relatively harsh conditions, such
as strictly anhydrous conditions, very
low temperatures (À788C), etc., the re-
actions reported herein proceeded in
the presence of water at 08C. To find
appropriate chiral water-compatible
Lewis acids for the Mukaiyama aldol
reaction, many Lewis acids were
screened in combination with chiral bi-
pyridine L1, which had previously been
found to be a suitable chiral ligand in
aqueous media. Three types of chiral
catalysts that consisted of a FeII or BiIII
metal salt, a chiral ligand (L1), and an
additive have been discovered and
a wide variety of substrates (silicon
enolates and aldehydes) reacted to
afford the desired aldol products in
high yields with high diastereo- and
enantioselectivities through an appro-
priate selection of one of the three cat-
alytic systems. Mechanistic studies elu-
cidated the coordination environments
around the FeII and BiIII centers and
the effect of additives on the chiral cat-
alysis. Notably, both Brønsted acids
and bases worked as efficient additives
in the FeII-catalyzed reactions. The as-
sumed catalytic cycle and transition
states indicated important roles of
water in these efficient asymmetric
Mukaiyama aldol reactions in aqueous
media with the broadly applicable and
versatile catalytic systems.
Keywords: aldol reaction · asym-
À
metric catalysis · bismuth · C C
coupling · iron
Introduction
many natural products, pharmaceuticals, agricultural chemi-
cals, functionalized materials, etc.
In 1973, Mukaiyama et al. reported a TiCl4-mediated aldol
reaction of silicon enolates (silyl enol ethers) with alde-
hydes.[1] In contrast to classical aldol reactions, which were
carried out under basic conditions and led to various side re-
actions, such as isomerization (racemization), retro-reac-
tions, over-reactions, and self-condensation of the alde-
hyde,[2] this aldol reaction provided the desired cross-aldol
adducts in high yields and has been regarded as one of the
Catalytic asymmetric variants on the Mukaiyama aldol re-
action were initiated in 1990 by using a chiral SnII Lewis
acid catalyst[3] and the application of chiral boron,[4] titani-
um,[5] and zirconium[6] Lewis acids soon followed thereafter.
Since then, many chiral Lewis acids have been developed
for asymmetric Mukaiyama aldol reactions.[7] On the other
hand, “direct-type aldol reactions” have been reported, in
which the reactions did not proceed through a silicon eno-
late but rather through a ketone or a synthetic equivalent,
which reacted as a nucleophile directly with the aldehyde.[8]
Thus, whereas silicon enolates have to be prepared (isolat-
ed) from ketones by using stoichiometric amounts of the
base and silicon source in Mukaiyama aldol reactions,
direct-type reactions give the aldol adducts directly.[9] There-
fore, from an atom-economical point of view, the latter pro-
cess has an advantage. However, at this stage, nucleophiles
that can be used in direct-type aldol reactions are limited to
ketones and aldehydes; furthermore, there are few examples
of such reactions with aromatic ketones or acyclic ke-
tones.[10] Thus, the applicable substrates for direct-type aldol
reactions are limited,[11] whereas silicon enolates that can be
applied to the Mukaiyama aldol reaction are not only de-
rived from ketones and aldehydes, but also from esters,
amides, lactones, lactams, enones, unsaturated esters, unsatu-
rated amides, and others. Therefore, from the standpoint of
À
most-reliable carbon carbon bond-forming reactions in or-
ganic synthesis, namely the “Mukaiyama aldol reaction.”
Indeed, this reaction has been applied to the syntheses of
[a] T. Kitanosono, Prof. Dr. S. Kobayashi
Department of Chemistry, School of Science
The University of Tokyo
Hongo, Bunkyo-ku, Tokyo 113-0033 (Japan)
Fax : (+81)3-5684-0634
E-mail: shu_kobayashi@chem.s.u-tokyo.ac.jp
[b] Prof. Dr. T. Ollevier
Dꢀpartement de chimie
Universitꢀ Laval
1045 avenue de la Mꢀdecine
Quꢀbec, G1V 0A6 (Canada)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/asia.201301149.
1
&
&
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers! ÞÞ

www.chemasianj.org
Shu¯ Kobayashi et al.
Table 1. Evaluation of Lewis acid catalysts in the asymmetric Mukaiya-
ma aldol reaction.
substrate scope, the Mukaiyama aldol reaction still has an
advantage over the direct-type aldol reaction.
As mentioned above, although catalytic asymmetric Mu-
kaiyama aldol reactions have been known since 1990,[3] rela-
tively harsh conditions, such as strictly anhydrous conditions,
very low temperatures (À788C), etc., have been required in
many cases, the substrates that are tolerated have been lim-
ited, and there remains great room for improvement as
a practical asymmetric synthesis. To address these issues, we
have developed catalytic asymmetric reactions that can be
carried out in the presence of water under mild conditions.
Herein, we describe catalytic asymmetric Mukaiyama aldol
reactions that can be applied to a wide range of substrates
in aqueous solutions at 08C.
Entry System Metal salt x/y
Yield
[%][a]
d.r. (syn/
anti)[b]
ee
[%][c]
Results and Discussion
1
2
3
4
5
6
7
8
A
B
A
B
A
B
A
B
A
B
A
B
A
B
A
B
A
B
A
B
A
B
Al
Sc
Fe
N
4
5
61:39
73:27
64:36
80:20
--
À51/3
À7/1
24/1
59/4
--
75/43
À41/4
À57/0
3/13
Evaluation of the Lewis Acids with a Chiral 2,2’-Bipyridine
Ligand
ACHTUNGTRENNUNG
To perform catalytic asymmetric Mukaiyama aldol reactions
in aqueous solution,[12] the development of water-compatible
Lewis acids[13] is required. We have previously reported
a series of water-compatible Lewis acids and determined
suitable factors, including the hydrolysis constant and water-
exchange rate constant,[14] for estimating their catalytic ac-
tivities, by which these Lewis acids were systematically clas-
sified.[15] We also clarified that, whereas conventional mono-
dentate, bidentate, and sometimes tridentate ligands have
been used in asymmetric catalysis in organic solvents, tetra-
dentate or higher-multidentate ligands are effective for cre-
ating chiral water-compatible Lewis acids.[16,17] Among them,
optically active bipyridine ligand L1,[18] as developed by
Bolm et al., was found to be a particularly excellent
ligand.[19] The first use of ligand L1, combined with ScIII as
a chiral water-compatible Lewis acid, in aqueous solvents
was reported in 2004 in the asymmetric hydroxymethylation
of silicon enolates.[20] Since that report, BiIII/L1,[21,22] CuII/
L1,[22,23] and ZnII/L1[,22b,24] systems were shown to be effec-
tive chiral Lewis acids for enantioselective hydroxymethyla-
tion reactions,[20,21,25] ring-opening reactions of meso-epox-
ides,[22,23,26,27] Nazarov cyclization reactions,[28] and a-alkyl-
or a-chloroallylation reactions of aldehydes[24] in aqueous
solvents.
In this context, to find the best chiral Lewis acid catalyst
for the Mukaiyama aldol reaction, we first examined the re-
action of benzaldehyde 1 with propiophenone-derived sili-
con enolate 2a in the presence of 10 mol% of the Lewis
acid and 12 mol% of compound L1 (system A) or in the
presence of 3 mol% of the Lewis acid and 9 mol% of com-
pound L1 (system B) in aqueous media. The data are sum-
marized in the Supporting Information, Tables S1a,b (see
the Supporting Information). Previous studies on organic re-
actions in water showed that the generation of naked metal
cations was crucial for attaining high Lewis acidity in an
aqueous environment. Therefore, metal trifluoromethanesul-
fonates (triflates) or metal perchlorates were chosen as
A
3/9
70
8
5
97:3
Cu
Cd
G
82:18
88:12
89:11
88:12
84:16
85:15
79:21
76:24
79:21
81:19
98:2
9
ACHTUNGTRENNUNG
10
11
12
13
14
15
16
17
18
19
20
21
22
À32/66
Sn
Yb
Lu
Hg
Pb
Bi
1/22
5
75/6
E
60/23
38/28
62/6
AHCTUNGTRENNUNG
53/9
R
66/77
À5/4
À32/78
À1/25
69/17
91/31
2
49:51
89:11
90:10
95:5
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
>99:<1
1
[a] Yield of the isolated product. [b] Determined by H NMR spectrosco-
py. [c] Determined by HPLC analysis; minus values indicate opposite
enantioselectivity. NR=no reaction.
Lewis acids. Selected examples are shown in Table 1. Rela-
tively high enantioselectivities were obtained by using AlIII,
ScIII, FeII, CuII, CdII, SnII, YbIII, LuIII, HgII, PbII, and BiIII.[29]
However, the yields were low by using AlIII and CuII in both
systems A and B. For CdII and PbII, relatively high enantio-
selectivities were only observed for the minor diastereomers.
Whereas system A gave higher enantioselectivities for YbIII
and HgII, system B showed higher enantioselectivities for
ScIII, FeII, and SnII. Both system A and B showed high enan-
tioselectivities for LuIII and BiIII. Notably, BiIII showed very
high selectivities (syn/anti= >99:<1, 91% ee (syn)) in sys-
tem B. Moreover, whereas no reaction occurred by using
FeII in system A, outstanding yield and promising selectivity
(70% yield, d.r. 97:3, 75% ee (syn)) were obtained with sys-
tem B. Regarding the ionic radii of various six-coordinate
metal cations (see the Supporting Information, Table S2),
&
&
2
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

www.chemasianj.org
Shu¯ Kobayashi et al.
these results showed that chiral 2,2’-bipyridine L1 formed
complexes with a wide variety of metal cations; the ionic
radii of the cations were within the range 53.5 (AlIII) to
119 pm (PbII).
enantioselectivity were brought about in the presence of
less-hindered base, especially a-substituted pyridines
a
(Table 2, entries 9 and 10). Coordinative substituents should
also be avoided to obtain higher selectivities (Table 2, en-
tries 11 and 12). The use of 2,6-lutidine as an additive af-
forded the aldol adduct with high enantiomeric purity
(Table 2, entry 13). On the other hand, the enantioselectivity
of the desired aldol adduct was unaffected by the use of
strongly hindered bases (Table 2, entries 14–16). An interest-
ing influence on enantioselectivity was observed when using
imidazole and N-methylimidazole, presumably because of
competitive coordination with the chiral ligand onto FeII
(Table 2, entries 17 and 18). Such unprecedented assistance
of these additives in Lewis acid catalysis, whether Brønsted
acids or bases, intrigued us; regardless of their nature, all of
the additives exhibited the best performance when em-
Identification of Optimal Conditions and Catalysts
In addition to the hitherto developed ScIII and BiIII catalysis
for enantioselective hydroxymethylation reactions, FeII cat-
alysis also seemed to be a promising candidate.
Several additives were examined in the presence of
3 mol% of FeACHTUNGTRENNUNG(OTf)2 and 9 mol% of compound L1
(Table 2); the amount of additive was set at 6 mol%. Acetic
acid and benzoic acid were found to be effective additives
ployed at 6 mol%, twice the amount of FeACTHNURGTNEUNG(OTf)2.
Table 2. Effect of additives in chiral-FeII-catalyzed Mukaiyama aldol re-
actions.
Intensive optimization of the reaction conditions (see the
Supporting Information, Table S3) led to the discovery of
two distinct catalytic systems that included a Brønsted base
(pyridine, defined as “conditions A”) or acid (benzoic acid,
defined as “conditions B”;[30] Scheme 1). In spite of their op-
posing natures, these additives assisted the Lewis acid catal-
Entry
Additive
Yield
[%][a]
d.r. (syn/
anti)[b]
ee
[%][c]
1
2
3
4
5
6
7
8
–
AcOH
TfOH
PhOH
TsOH
70
70
21
37
13
71
76
68
27
31
46
97:3
98:2
98:2
98:2
98:2
97:3
97:3
97:3
97:3
97:3
95:5
75/43
89/13
25/6
83/0
15/12
88/16
86/26
91/À48
93/À9
92/À33
7/7
PhCO2H
2,2’-bipyridyl
R=H
9
10
11
R=Me
R=Et
R=
CH2OH
12
13
14
15
16
R=OMe 60
98:2
98:2
97:3
97:3
98:2
77/14
91/19
76/39
77/38
77/28
Scheme 1. Optimal conditions for the FeII-catalyzed Mukaiyama aldol re-
action.
2,6-lutidine
2,6-diphenylpyridine
R=H
18
58
69
67
R=Me
ysis of a chiral FeII/L1 complex to afford the product in
a highly enantioselective manner in an aqueous environ-
ment, which showed that this powerful assistance could not
be explained by the adjustment of the pH value in the reac-
tion solution.
17
18
imidazole
7
85:15
97:3
1
0/0
73/À31
N-Me-imidazole
26
[a] Yield of the isolated product. [b] Determined by H NMR spectrosco-
py. [c] Determined by HPLC analysis. Ac=acetyl, Ts=4-toluenesulfonyl.
The acidity of these ternary catalytic solutions was mani-
fested by the statistically robust measurement of pH value
in these catalytic systems. The Pourbaix diagram, as calculat-
ed from the redox potentials of various iron species and
water by using the Nernst equation and hydrolysis constants
(see the Supporting Information, Figure S1), shows that FeII
is clearly the active species in these catalytic systems.
Indeed, FeIII exhibited inferior performance (see the Sup-
porting Information, Table S1b).
for Lewis acid catalysis (Table 2, entries 2 and 6). However,
strong acids, such as TfOH and p-toluenesulfonic acid
(TsOH), impeded the function of the catalyst in exerting
precise stereocontrol (Table 2, entries 3 and 5). Interestingly,
a significant enhancement in enantioselectivity was also ob-
served in the presence of a Brønsted base (Table 2, entries 7
and 8). To gain more insight into the interactions between
FeII and an additive, several pyridine derivatives were exam-
ined. Slowing of the reaction rate and a higher level of
The superb performance of BiIII catalysis in asymmetric
hydroxymethylation reactions[21] led us to harbor expecta-
tions about the further evolution of this catalytic system to
3
&
&
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers! ÞÞ

www.chemasianj.org
Shu¯ Kobayashi et al.
broaden the substrate scope.[31] Furthermore, the availability
of bismuth, because of its affordability, abundance, and
ubiquity, was comparable to that of iron. The reaction pro-
ceeded smoothly in the presence of a Brønsted base to
afford the product with the same level of enantioselectivity.
The best performance was displayed when 12 mol% of pyri-
dine was added to 3 mol% of a chiral BiIII complex that was
formed with threefold excess of compound L1 (Scheme 2).
zation of the O Si bond of the silicon enolate, thus slowing
the rate of hydrolysis after the nucleophilic attack. Substitu-
tion with an electron-withdrawing group led to an impercep-
tible decrease in the reaction rate. The addition of pyridine
affected the enantioselectivity of the product, irrespective of
the substituents on the aromatic ring (Table 3, entries 1–3).
However, the reactions of 4-fluorobenzaldehyde (1d) and 1-
naphthaldehyde (1 f) proceeded in a highly enantioselective
manner with benzoic acid as an additive (Table 3, entries 4
and 5). The optimal conditions could also be applied to ali-
phatic aldehydes, which are often poor electrophiles for
Lewis-acid-mediated asymmetric processes because of both
electronic and steric difficulties in the enantiofacial discrimi-
nation. Interestingly, the reactions of compounds 1h and 1n
proceeded very smoothly to afford the desired products in
high yields with high selectivities with pyridine as an addi-
tive (Table 3, entries 6 and 7). Moreover, the use of FeII per-
chlorate instead of triflate, changing the water/dimethoxy-
ethane (DME) ratio to 3:7, and the use of benzoic acid in-
stead of pyridine retarded the catalytic turnover. On the
other hand, the use of benzoic acid as an additive showed
superior performance in the reactions of thiophenecarboxal-
dehyde (1i) and a,b-unsaturated aldehydes 1l and 1m
(Table 3, entries 7–9). Based on these tendencies, we contin-
ued to examine the substrate scope of the aldehyde and sili-
con enolate.
À
Scheme 2. Conditions for the BiIII-catalyzed Mukaiyama aldol reaction
(conditions C).
Substrate Scope
The substrate scope was evaluated under the optimized re-
action conditions. The intriguing assistance of two distinct
additives suggested the possibil-
ity that the careful selection of
an additive could broaden the
Table 3. Effect of additives on the substrate scope.
substrate scope. Based on such
a perspective, the correlation
between various substrates and
their effective additives was in-
vestigated (Table 3). We found
that the pyridine additive was
more effective in the Mukaiya-
ma aldol reaction of compound
1a with compound 2a (Table 3,
entry 1). Electronic effects that
resulted from the substituents
on the aromatic ring of com-
pound 2a were surveyed. An
electron-donating group (OMe,
2b) resulted in a definite drop
in the reaction rate. Prolonging
the reaction time allowed the
almost-complete consumption
of an aldehyde to afford the
product (3ab) with excellent di-
astereo- and enantioselectivities
in the presence of pyridine.
However, the reaction hardly
proceeded when benzoic acid
was employed as an additive
(Table 3, entry 2). An electron-
donating group on the aromatic
Entry
R1
Silicon enolate
R
Conditions[a]
Yield [%][b]
d.r. (syn/anti)[c]
ee [%][d]
1
2
Ph
Ph
Ph
1a
1a
1a
1d
1 f
1h
1i
H
OMe
Cl
H
2a
2b
2c
2a
2a
2a
2a
2a
2a
2a
A
90 (3aa)
85
91 (3ab)
6
87 (3ac)
70
76 (3da)
99
55 (3 fa)
quantitative
82 (3ha)
75
1 (3ia)
70
41 (3la)
54
82 (3ma)
94
97:3
97:3
97:3
>99:<1
93:7
95:5
92:8
96:4
98:2
96:4
94:6
94:6
--
98:2
86:14
95:5
88:12
88:12
94:6
91/52
88/48
85/21
49/nd
89/66
65/40
65/33
89/39
88/82
91/40
99/25
96/25
–
91/31
75/28
75/30
84/61
95/73
90/75
66/37
B
A[e]
B
3
A[f]
B
4
4-FC6H4
naphthyl
A
B
A
B
A
B
A
B
5
H
6
PhCH2CH2
2-C4H3S
H
7
H
8
CH2=CH
PhCH2=CH
CH3CH2
1l
H
A[e]
B[e]
A
9
1m
1n
H
B[f]
A
10
H
93 (3na)
71
B
87:13
[a] Conditions A: Fe
Fe(ClO4)2 (3 mol%), L1 (3.6 mol%), PhCO2H (3.6 mol%), water/DME (3:7). [b] Yield of the isolated prod-
uct. [c] Determined by 1H NMR spectroscopy. [d] Determined by HPLC analysis. [e] For 48 h. [f] For 36 h.
ACHTUNGRTEN(NUNG OTf)2 (3 mol%), L1 (3.6 mol%), pyridine (7.2 mol%), water/DME (1:9). Conditions B:
AHCTUNGTRENNUNG
ring might contribute to stabili- nd=not determined.
&
&
4
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

www.chemasianj.org
Shu¯ Kobayashi et al.
Conditions A
We found that pyridine was generally the most effective as
an additive in the Mukaiyama aldol reaction. Thus, condi-
tions A were used as the standard conditions for an initial
substrate screen (Scheme 3). The addition of pyridine im-
Scheme 4. Substrate scope, conditions B. [a] Fe
(9 mol%), and PhCO2H (3.6 mol%).
ACHTUNGTRENN(NUG ClO4)2 (3 mol%), L1
Conditions C
A chiral BiIII catalyst (defined as conditions C) was also
found to be attractive as an alternative system for asymmet-
ric Mukaiyama aldol reactions (Scheme 5). In a previous
report on BiIII-catalyzed asymmetric hydroxymethylation re-
actions,[21] 2,2’-bipyridyl was selected as an effective additive.
In that system, only trace amounts of the product were ob-
tained by using BiACTHNUTRGNE(UNG OTf)3 without compound L1, presumably
because of the rapid decomposition of the silicon enolate as
promoted by TfOH that was readily generated from Bi-
Scheme 3. Substrate scope, conditions A. [a] 2,6-di-tert-butylpyridine
(DTBP, 3 equiv) was added.
proved the enantioselectivity of the product, irrespective of
the substituents on the aromatic ring in the silicon enolate.
Not only aliphatic aldehydes (1h, 1n, 1o, and 1r) but also
aromatic aldehydes that contained electron-donating groups
(1b and 1c) reacted with the silicon enolates in a highly ste-
reoselective manner.
AHCTUNGRTEG(NNUN OTf)3 in water. Water-soluble aldehydes, such as aqueous
formaldehyde (1p), and acrolein (1l), reacted with silicon
enolates to afford their corresponding aldol adducts in high
yields with high enantioselectivities. The aldol reaction of
acetaldehyde (1q) is regarded to be difficult because acetal-
dehyde is highly reactive as both a nucleophile and an elec-
trophile, thus causing a self-aldol reaction. Therefore, until
recently, no asymmetric reactions by using acetaldehyde
have been reported.[33] The chiral BiIII complex catalyzed the
reaction of compound 1q with compound 2a to afford the
desired product in good yield with high diastereo- and enan-
tioselectivities. The BiIII complex was also applicable to the
reactions of aldehydes that contained various functional
groups (1j, 1k, 1s, 1t, and 1u). Notably, asymmetric quater-
nary carbon atoms were constructed with high selectivities.
Silicon enolate 2b, which contained an electron-donating
Conditions B
When benzoic acid was employed as an additive, a superior
effect was obtained in the reactions of electron-deficient ar-
omatic aldehydes (1d, 1e, 1 f), secondary aliphatic aldehyde
1g, thiophenecarboxaldehyde (1i), and a,b-unsaturated al-
dehyde 1m (Scheme 4). The introduction of an electron-
withdrawing (chloro) group onto the aromatic ring in the sil-
icon enolate led to a slightly lower (but satisfactory) level of
selectivity in high yield (3mc). The combination of FeII and
benzoic acid could also be applied to the reaction of com-
pound 1a with the silicon enolate that was derived from 3-
pentanone (2 f).[32]
5
&
&
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers! ÞÞ

www.chemasianj.org
Shu¯ Kobayashi et al.
Table 4. Correlation between the properties of the aldehydes and aro-
matic ketone-derived silicon enolates and the most-effective catalytic sys-
tems.
R[a]
H
Aldehyde (R1CHO)
More bulky
Sterically small, chelate-type
Electron-rich Electron-poor
conditions A conditions B
conditions C
conditions C
conditions C
OMe (ED) conditions A conditions C
Cl (EW) condition A conditions B
[a] ED=electron donating, EW=electron withdrawing.
none-derived silicon enolates affected neither the reactivity
nor selectivity, except for conditions B. The reaction with
butyrophenone-derived silicon enolate, which led to a signifi-
cant loss in enantioselectivity in the presence of a chiral gal-
lium catalyst with a Trost-type ligand,[34] also led to an excel-
lent result. The introduction of a methoxy group onto the
aromatic ring of the silicon enolate led to a significant drop
in both reactivity and stereoselectivity under conditions B,
whereas the reaction proceeded quantitatively to afford the
product with excellent selectivities under conditions C (e.g.,
3eb and 3mb). In addition, conditions C were effective for
the reactions of less-bulky aldehydes, including water-solu-
ble aldehydes and aldehydes that could chelate to the metal
atoms. The stereoselective construction of asymmetric qua-
ternary carbons was also observed.
The reaction of a chiral b-alkoxy aldehyde that contained
potential directing ability was surveyed to offer a rationale
for the stereocontrol (Scheme 6). The exposure of aldehyde
1u and silicon enolate 2a to the chiral BiIII complex with
(S,S)-L1 exhibited exclusive diastereofacial preference with
high enantioselectivity. The use of FeII instead of BiIII im-
proved the reactivity, with a high level of stereoselection. In
contrast, the “mismatched” case with (R,R)-L1 resulted in
lower enantioselectivity, albeit the same level of yield and
syn-selectivity because of the chelation.
Scheme 5. Substrate scope, conditions C. [a] Bi
(3 mol%), 2,2’-bipyridyl (5 mol%), in water/DME (1:4). See Ref. [21].
[b] DTBP (3 equiv) was added. [c] Sc(OTf)3 (10 mol%), L1 (12 mol%).
ACHTUNGRTEN(NGNU OTf)3 (1 mol%), L1
AHCTUNGTRENNUNG
[d] Without pyridine. Tf=trifluoromethanesulfonyl.
Then, the Mukaiyama aldol reaction of benzyloxyacetal-
dehyde with compound 2a was carried out by using the opti-
mal catalytic systems. The desired aldol adduct is known to
be a key intermediate in the total synthesis of various natu-
ral compounds. For instance, the total synthesis of bafilomy-
cin A1, an in vitro inhibitor of vacuolar H+-ATPase, entailed
three steps to control the stereogenic centers.[35] In the con-
ventional method (Scheme 7), 2-(benzyloxy)acetic acid was
treated with benzotriazole in the presence of DCC. Then,
the lithium enolate of propiophenone was added to afford
the desired diketone, which was reduced under asymmetric
transfer-hydrogenation conditions with a chiral ruthenium
complex to afford the desired b-hydroxyketone with a d.r.
of 68:32 in a total 62% yield (three steps). The major
isomer was formed with 93% ee. On the other hand, in the
group, was not applicable to conditions B, whereas the chiral
BiIII complex catalyzed the reactions with compound 2b in
high yields with high selectivities.
The effect of the choice of ternary catalytic system and
substituent effects on this reaction are summarized in
Table 4. Notably, the introduction of electron-withdrawing
groups on the aromatic ring of benzaldehyde preferred con-
ditions B. In contrast, electron-donating aromatic aldehydes
preferred conditions A, thus indicating that the difference
between pyridine and benzoic acid as the additive may lie in
electronic effects through coordination to FeII. The substitu-
ents on the aromatic ring in the structure of the propiophe-
&
&
6
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

www.chemasianj.org
Shu¯ Kobayashi et al.
Scheme 8. Bismuth-mediated synthesis of a b-hydroxyketone.
sites. This structure resembles that of the corresponding ScIII
and BiIII complexes (for the X-ray structures, see the Sup-
porting Information, Figure S2). Several examples[37] of
a heptacoordination mode of the FeII ion have been report-
ed and X-ray crystals of three types of FeII/L1 complexes[38]
have been obtained, all of which showed heptacoordination.
Thermogravimetric analysis implied that the chiral complex
could behave in a heptacoordinated fashion under the reac-
tion conditions.[39] Apical binding of an aldehyde is favored
because of its lower basicity and its ability to form stereoe-
lectronically favored hydrogen-bonding interactions with the
alcohol moieties of ligand L1 and the relatively greater
trans-effect of the pyridine ligand.[40] The dependence of the
choice of additive on the electronic nature of the aldehydes
indicates the coordination of an additive to the FeII center.
On the other hand, the behavior of BiIII as a naked cation is
inferred from the fact that 2,6-di-tert-pyridine, a proton scav-
enger, could be an efficient additive. To obtain some funda-
mental insight into the catalyst structure, a model reaction
between compounds 1a with 2a was carried out in the pres-
ence of enantiomerically impure 2,2’-bipyridine L1
(Figure 1). Whereas negative nonlinear effects were ob-
served in the absence of an additive, the nonlinearity disap-
peared in the presence of an additive. Because subtle ero-
sion of the enantioselectivity in product 3aa was observed in
Scheme 6. Match/mismatch effect.
Scheme 7. Conventional synthesis of a b-hydroxyketone.[35] DCC=N,N’-
dicyclohexylcarbodiimide, LDA=lithium diisopropylamide, Bn=benzyl.
Mukaiyama aldol reaction, the employment of a chiral BiIII
catalyst (3 mol%) led to the best result; the reaction pro-
ceeded quantitatively with excellent diastereo- and enantio-
selectivities (Scheme 8). The utility of this aldol process is
obvious, even considering the preparation of the silicon eno-
late and chiral 2,2’-bipyridine L1.[36]
Reaction Mechanism
Coordination Environment around the Metal Centers
The complex between an FeII salt and ligand L1, first report-
ed in 2012,[30] adopts a pentagonal bipyramidal structure, in
which the tetradentate ligand occupies four of the equatorial
Figure 1. Nonlinear experiments.
7
&
&
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers! ÞÞ

www.chemasianj.org
Shu¯ Kobayashi et al.
the absence of an additive, the observations may be ration-
alized by the changes in the catalyst structure from a dimeric
or oligomeric form without the additive into a monomeric
form with the additive, which might be an actual catalytic
species that resulted in high enantioselectivity and reactivity.
by a Lewis base in the FeII-catalyzed mechanism and no or
very little involvement in the BiIII-catalyzed mechanism. In
the latter case, the pyridine additive is assumed to be
swamped with the elimination of in-situ-generated triflic
acid.
To determine the role of the benzoate anion as a Lewis
base, the electronic effect on the additive was investigated
(Table 6). When 4-methoxybenzoic acid was used as the ad-
ditive, the reaction proceeded smoothly to provide almost
Effect of Additives
To gain some insight into the structures the catalysts, the re-
action solution was directly subjected to spectrometric inves-
tigation. The chiral FeII complex with ligand L1 did not give
rise to massive changes in the NMR analysis, unlike the
chiral ScIII and BiIII complexes. Likewise, on using an addi-
tive, no changes were observed in the NMR or UV/Vis anal-
yses. ESI-MS experiments could shed some light on the dif-
ferences (see the Supporting Information, Figure S3). In the
absence of an additive, two types of mono-FeII complexes,
Table 6. Effect of the silicon enolate.
that is, [FeII(L1) (DME)]+ and [FeII(L1)(OTf)]+, were
ACHTUNGTRENNUNG(OTf)ACHNUTRTGENNUNG ACTHNUGTRENNUGN
observed; one was coordinated by a DME molecule, where-
as the other was not. Some amounts of an FeII complex that
was composed of two molecules of the chiral ligand were
also identified. When pyridine was used as the additive, the
triflate anion was not identified as a counteranion. Instead
Entry
R
Additive
Yield [%][a] d.r. (syn/anti)[b] ee [%][c]
1
2
3
4
5
6
7
H
H
H
H
H
–
70
85
NR
97:3
97:3
–
75/43
88/48
–
PhCO2H
PhCO2K
4-MeOC6H4CO2H 80
4-NO2C6H4CO2H
PhCO2H
of
[FeII(L1)(OH)]+, was verified. The major peak corresponded
to
FeII/benzoate complex with ligand L1, [FeII(L1)-
ACHTUNGTRENNUNG
(O2CPh)]+. The environment around BiIII, which is known
a
triflate, the formation of
a
hydroxide complex,
97:3
96:4
97:3
93:7
84/43
73/47
49/nd
65/40
52
6
70
MeO
Cl
a
PhCO2H
to be dependent upon the metal-to-ligand ratio, were also
found to be a naked cation without any ligand, as mentioned
above (see the Supporting Information, Figure S3).[20b]
the same results as those with benzoic acid. In contrast, in
the case of 4-nitrobenzoic acid, both the reaction yield and
enantioselectivity suffered from a significant drop. The
enantioselectivity of the syn isomer was almost the same as
that in the absence of any additive. As mentioned above,
the substituents on the aromatic ring in the structure of sili-
con enolates significantly affected the reactivity and selectiv-
ity. Similarly, Mukaiyama et al. reported that 10 mol% of
a metal carboxylate could function as a Lewis basic catalyst
to mediate the reaction between benzaldehyde and a silyl
ketene acetal at À458C in DMF/water (50:1).[42] They as-
sumed that the role of the carboxylate in the catalytic cycle
was the formation of a lithium aldolate through a hexacoor-
dinated hypervalent silicate, which underwent rapid hydroly-
sis and subsequent neutralization to regenerate the lithium
acetate catalyst. The dependence of reactivity on the elec-
trophilicity of the aldehyde supports their hypothesis, in
accord with the results shown in Table 6.
Because of the ability of silicon atoms to adopt higher co-
ordination states, promoting the Mukaiyama aldol reaction
under basic conditions is known to be possible.[41] In that
sense, an additive can function as a Lewis base in our cata-
lytic systems. Thus, to examine whether a nucleophilic
attack takes place at the silicon atom or not, a Mukaiyama
aldol reaction by using dimethylsilyl enol ether was carried
out (Table 5). Throughout the enhancement of the reaction
rate, an almost-quantitative amount of the product was ob-
tained under conditions A, although it also led to a signifi-
cant drop in selectivity. In the case of the chiral BiIII catalyst,
we found that the enantioselectivity remained at the same
level; however, the reactivity was not enhanced at all. These
results imply some involvement of silicon-enolate activation
Table 5. Effect of the silicon enolate.
Ligand field theory is helpful for deepening our under-
standing of the effect of additives. The donation of electrons
by s-donor ligands to d orbitals and some weak bonding and
antibonding interactions with s and p orbitals of FeII are
plausible. Compared with nonbonding orbitals that consist
of dxz and dyz orbitals, the antibonding orbitals between 4s,
4p, dxy, and dx2Ày2 and between 4s, 4pz, and dz2 become desta-
bilized. When pyridine was used as an additive, the OH
ligand in an equatorial position or pyridine in an apical posi-
tion made the former bonding orbital unstable and the
latter stable. The consequent spin transition in the HOMO
Conditions
Si=SiMe3
Si=SiMe2H
Yield [%] syn/anti ee [%] Yield [%] syn/anti ee [%]
A
C
72
76
93:7
98:2
74 (syn)
90 (syn)
95
70
84:16
96:4
28 (syn)
90 (syn)
&
&
8
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

www.chemasianj.org
Shu¯ Kobayashi et al.
orbital overcame the deactivation by the coordination of
a Lewis base[43] and the lower HOMO destabilized the p*
orbital of aldehyde. Given the intriguing observed equilibri-
um shift in the Sc/L1 complex,[44] stabilization of the pentag-
onal bipyramidal structure, even in an aqueous environment,
was expected. In contrast, the bidentate coordination of
a benzoate anion to the iron atom made the latter unstable.
The resultant HOMO–LUMO gaps between the aldehyde
and the silicon enolate correspond to the experimental
trends as shown in Table 4. The strength of the interactions
between the iron center and both substrates governs the ste-
reoselectivities, on the assumption of some interactions be-
tween the iron atom and the p-orbital of the silicon eno-
late.[45]
because water can disturb a weak interaction between the
metal and the oxygen atom of the silicon enolate. A direct
comparison of the effect of enolate geometry was conducted
by using (E)- and (Z)-silicon enolates that were derived
from 3-pentanone under the FeII-catalyzed conditions
(Scheme 9). We found that both (E)- and (Z)-enolates gave
syn-aldol adducts, which also supported the acyclic open-
Thus, there are two functions for the additive in the FeII
complexes: 1) Coordination to the FeII ion and 2) coordina-
tion to the silicon atom of the silicon enolate for activa-
tion.[46] Two kinds of additives can regulate the balance be-
tween the two interactions to afford the desired aldol ad-
ducts in high yields with high diastereo- and enantioselectiv-
ities. On the other hand, a Lewis base additive is consumed
by scavenging the in situ formed TfOH in the case of the
BiIII complex.
Scheme 9. Correlation between the stereochemical outcome and the eno-
late geometry.
transition-state model. The independence of stereochemistry
from enolate geometry was also confirmed for the chiral
BiIII complex (conditions C).
A plausible catalytic cycle for this Mukaiyama aldol reac-
tion in aqueous media is shown in Scheme 10. First, the al-
dehyde coordinates to the pentagonal bipyramidal FeII com-
Reaction Pathways and Transition States
Three different mechanisms have been proposed for Mu-
kaiyama aldol reactions: 1) The Zimmerman–Traxler six-
membered model;[47] 2) the acyclic open-transition-state
model, which involves an antiperiplanar orientation of the
enolate and carbonyl compounds;[48] and 3) the transmetala-
tion model, which involves transmetalation of the silicon
atom of the silicon enolate to the second metal of the Lewis
acid. An initial mechanistic hypothesis in Mukaiyama’s land-
mark report in 1973 involved the formation of a titanium
enolate through transmetalation.[1] Although the resulting ti-
tanium enolate was believed to react with an aldehyde to
afford a stable titanium aldolate with anti-selectivity, no
structural evidence could be supplied at that time. Since
then, an INEPT (insensitive nuclei enhanced by polarization
transfer) 29Si NMR study was emphatic in disproving the op-
eration of a transmetalation pathway.[49] Nevertheless, the
reactions of pre-isolated titanium enolates with aldehydes
were amazingly syn-selective.[50] On the other hand, the
closed-transition-state hypothesis was underscored by com-
putational chemistry and a six-membered transition state
model that did not involve transmetalation was also pro-
posed.[51] To date, the actual mechanism has not yet been
conclusively determined.
Whereas all of these mechanistic investigations were car-
ried out in organic solvents, the kinetics of a hydroxymethy-
lation reaction between aqueous formaldehyde and com-
pound 2a in the presence of a Sc/L1 complex was reported
in 2011.[52] In consequence, a first-order dependence of both
compound 2a and the chiral catalyst was observed, which
ruled out the possibility of a transmetalation mechanism.
The acyclic open-transition-state model is more suitable
when the reaction is performed in an aqueous environment,
Scheme 10. Plausible catalytic cycle.
plex, which consists of the FeII atom, ligand L1, and an addi-
tive in the axial position. One of the ligands, that is, pyri-
dine, benzoic acid, or water, may be replaced by the alde-
hyde. Water that is coordinated to the FeII center is disposed
to exchange and may facilitate the replacement. One face of
9
&
&
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers! ÞÞ

www.chemasianj.org
Shu¯ Kobayashi et al.
the aldehyde is shielded by the tert-butyl group of ligand
L1.[53] Then, a silicon enolate attacks the aldehyde through
an antiperiplanar acyclic transition state. Water also plays
a key role in coordinating to the silicon atom of the silicon
enolate to assist the reaction. The additive, that is, pyridine
or benzoic acid, is also though to play the same role (condi-
tions A/B). In the BiIII-catalyzed reaction (conditions C), the
coordination of pyridine to the silicon atom was not signifi-
cant; however, water coordination was assumed to be cru-
lyze the reactions through acyclic transition states; 2) facili-
tating the catalytic turnover with simultaneous desilylation
to provide direct access to the aldol adducts or facile recov-
ery of the active metal complexes; and 3) stabilizing rigid
transition states that were composed of the metal complexes
and the reactants through entropy-driven aggregation that
was derived from its highest cohesive energy density
(ced).[58] 40 Years after the discovery of the Mukaiyama
aldol reaction, we have established a definitive catalytic
asymmetric variant by using FeII and BiIII salts with chiral bi-
pyridine ligand L1 in aqueous media.
À
cial in that case. Then, carbon carbon bond formation
occurs with high diastereo- and enantioselectivities and the
resulting FeII/alkoxide complex reacts with water to afford
the desired aldol adduct, along with regeneration of the FeII
catalyst. The silicon moiety is quenched with water to de-
Experimental Section
À À
crease the Lewis acidity (Me3Si O SiMe3) and prevent un-
Typical Procedure for the Mukaiyama Aldol Reaction between a Silyl
Enol Ether and an Aldehyde (Table 3, entry 1)
desired achiral Si-catalyzed pathways, which often decrease
the diastereo- and enantioselectivities in asymmetric Mu-
kaiyama aldol reactions. Notably, water also plays an impor-
tant role.[54] In our reaction systems, water is essential for
the in situ generation of active species in the presence of
pyridine or benzoic acid. Indeed, the reaction in DME in
the absence of water did not lead to the formation of the de-
sired aldol adduct.[55] The higher the water/DME ratio, the
lower the yield and the higher the observed diastereo- and
enantioselectivities. The amount of water governed the ste-
reochemical outcome, as well as the consequential catalytic
turnover, although water is prone to destabilize silicon eno-
lates through hydrolysis of the metal complexes.[56]
A mixture of FeACHTNUGTRENUNG(OTf)2 (3.2 mg, 0.009 mmol), the chiral 2,2’-bipyridine
ligand (3.5 mg, 0.0108 mmol), and pyridine (1.6 mg, 0.0216 mmol) in de-
gassed DME (0.65 mL) was stirred at RT for 30 min. Then, the catalyst
solution was cooled at 08C for 30 min and water (70 mL) was added. To
the mixture were successively added aldehyde 1a and silyl enol ether 2a.
After stirring for 24 h at RT, the reaction mixture was quenched with a sa-
turated aqueous solution of NaHCO3 and washed with brine. The aque-
ous layer was extracted three times with CH2Cl2 and the combined or-
ganic layer was washed with brine and dried over anhydrous Na2SO4.
After removal of the solvent under reduced pressure, the residue was pu-
rified by preparative TLC (n-hexane/EtOAc, 3:1) to give the correspond-
ing aldol (3aa, 64.0 mg, 90% yield) as a colorless oil.
Acknowledgements
Conclusions
This work was supported by a Grant-in-Aid for Scientific Research from
the Japan Society for the Promotion of Science (JSPS), the Global COE
Program, The University of Tokyo, MEXT (Japan), the Japan Science
Technology Agency (JST), the Natural Sciences and Engineering Re-
search Council of Canada (NSERC), and the Centre in Green Chemistry
and Catalysis (CGCC, Quꢀbec, Canada). T.K. thanks the JSPS for a Re-
search Fellowship for Young Scientists.
We have developed FeII- and BiIII-catalyzed asymmetric Mu-
kaiyama aldol reactions of silicon enolates with aldehydes.
Three kinds of new ternary catalytic systems, which consist-
ed of a FeII or BiIII metal salt, chiral bipyridine ligand L1,
and an additive, have been exploited. Whilst previous reac-
tions often required relatively harsh conditions, such as
strictly anhydrous conditions, very low temperatures
(À788C), etc., the reactions reported herein proceeded in
the presence of water at 08C. A wide variety of silicon eno-
lates and aldehydes reacted under these conditions to afford
the desired aldol products in high yields with high diastereo-
and enantioselectivities. The superiority of this method over
conventional reactions has also been confirmed by the high
catalytic activity, the simplicity of the experimental proce-
dure, and broad substrate range, including aqueous alde-
hydes, for which the stereochemistry of the reaction has
been regarded to be difficult to govern. The coordination
environments around the FeII and BiIII atoms and the effect
of additives on the chiral catalysts have been elucidated by
mechanistic studies. Notably, both Brønsted acids and bases
worked as efficient additives in the FeII-catalyzed reactions.
The assumed catalytic cycle and transition states clarified
the important roles of water in:[57] 1) producing the active
metal complexes with a high water-exchange rate constant
(3.2ꢂ106) to effectively activate the substrates and to cata-
[1] T. Mukaiyama, K. Narasaka, K. Banno, Chem. Lett. 1973, 1011–
1014.
[2] a) Modern Aldol Reactions (Ed.: R. Mahrwald), Wiley-VCH, Wein-
heim, 2004; b) I. Paterson, Compr. Org. Synth. 1991, 2, 301–319;
c) B. M. Kim, S. F. Williams, S. Masamune, Compr. Org. Synth. 1991,
2, 239–275; d) C. H. Heathcock, Compr. Org. Synth. 1991, 2, 133–
238.
[3] a) S. Kobayashi, Y. Fujishita, T. Mukaiyama, Chem. Lett. 1990,
1455–1458; b) T. Mukaiyama, S. Kobayashi, H. Uchiro, I. Shiina,
Chem. Lett. 1990, 129–132.
[4] a) S. Adachi, T. Harada, Org. Lett. 2008, 10, 4999–5001; b) T.
Harada, K. Shiraishi, Synlett 2005, 1999–2002; c) M. Kinugasa, T.
Harada, A. Oku, J. Am. Chem. Soc. 1997, 119, 9067–9068; d) M. Ki-
nugasa, T. Harada, A. Oku, J. Org. Chem. 1996, 61, 6772–6773;
e) K. Ishihara, T. Maruyama, M. Mouri, Q. Gao, K. Furuta, H. Ya-
mamoto, Bull. Chem. Soc. Jpn. 1993, 66, 3483–3491; f) S. Kiyooka,
Y. Kaneko, K. Kume, Tetrahedron Lett. 1992, 33, 4927–4930; g) E. J.
Corey, C. L. Cywin, T. D. Roper, Tetrahedron Lett. 1992, 33, 6907–
6910; h) E. R. Parmee, Y. Hong, O. Tempkin, S. Masamune, Tetrahe-
dron Lett. 1992, 33, 1729–1732.
[5] a) E. M. Carreira in Comprehensive Asymmetric Catalysis (Eds.:
E. N. Jacobsen, A. Pfaltz, H. Yamamoto), Springer, Heidelberg,
&
&
10
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

www.chemasianj.org
Shu¯ Kobayashi et al.
1999, 3, 997–1065; b) R. A. Singer, E. M. Carreira, J. Am. Chem.
Soc. 1995, 117, 12360–12361; c) E. M. Carreira, R. A. Singer, W.
Lee, J. Am. Chem. Soc. 1994, 116, 8837–8838.
Am. Chem. Soc. 2003, 125, 2410–2411; c) B. M. Trost, L. R. Terrell,
J. Am. Chem. Soc. 2003, 125, 338–339; d) B. M. Trost, V. S. C. Yeh,
Org. Lett. 2002, 4, 3513–3516; e) B. M. Trost, V. S. C. Yeh, H. Ito, N.
Bremeyer, Org. Lett. 2002, 4, 2621–2623; f) B. M. Trost, V. S. C.
Yeh, Angew. Chem. 2002, 114, 889–891; Angew. Chem. Int. Ed.
2002, 41, 861–863; g) B. M. Trost, H. Ito, E. R. Silcoff, J. Am. Chem.
Soc. 2001, 123, 3367–3368; h) B. I. Trost, H. Ito, J. Am. Chem. Soc.
2000, 122, 12003–12004.
[6] a) H. Ishitani, Y. Yamashita, H. Shimizu, S. Kobayashi, J. Am.
Chem. Soc. 2000, 122, 5403–5404; b) Y. Yamashita, H. Ishitani, H.
Shimizu, S. Kobayashi, J. Am. Chem. Soc. 2002, 124, 3292–3302.
[7] a) For the Mukaiyama aldol reaction, see: E. M. Carreira, In Com-
prehensive Asymmetric Catalysis I–III (Eds.: E. N. Jacobsen, A.
Pfaltz, H. Yamamoto), Springer, Berlin, 1999, pp. 997–1065; b) R.
Mahrwald, Chem. Rev. 1999, 99, 1095–1120; c) T. Mukaiyama, J. I.
Matsuo in Modern Aldol Reactions (Ed.: R. Mahrwald), Wiley-
VCH, Weinheim, 2004, 1, 127–160; d) K. Ishihara, H. Yamamoto in
Modern Aldol Reactions (Ed.: R. Mahrwald), Wiley-VCH, Wein-
heim, 2004, 2, 25–68.
[18] a) C. Bolm, M. Ewald, M. Felder, G. Schlingloff, Chem. Ber. 1992,
125, 1169–1190; b) C. Bolm, M. Zehnder, D. Bur, Angew. Chem.
Int. Ed. Engl. 1990, 29, 205–207.
[19] Given the specific nature of its complexes with almost all metals,
chiral bipyridine ligand L1 is an efficient candidate for a “privileged”
structure of the chiral ligand that can function in an aqueous envi-
ronment. Indeed, owing to its coordinative potential as a N2O2
cavity, crystallographic investigations on its complexes with CuI,[18b]
[8] B. M. Trost, C. S. Brindle, Chem. Soc. Rev. 2010, 39, 1600–1632.
[9] H. O. House, L. J. Czuba, M. Gall, H. D. Olmstead, J. Org. Chem.
1969, 34, 2324–2336.
[20]
[21]
[27]
ScIII
,
BiIII
,
InIII
,
CuII,[23b] ZnII[24] and FeII[30] metal cations have
[10] The desired b-hydroxycarbonyl product is more basic than the start-
ing substrates, thereby resulting in quite-low catalyst turnover. Such
limitations can be circumvented by combining preformed silicon
enolates and Lewis acid catalysis, because the unproductive side re-
actions can be suppressed under mild conditions and the resulting si-
lylated aldol adduct allows for good catalyst turnover. For recent re-
ports on asymmetric direct-type aldol reactions by using organocata-
lysts, see: a) S. Bertelsen, K. A. Jørgensen, Chem. Soc. Rev. 2009, 38,
2178–2189; b) L. M. Geary, P. G. Hultin, Tetrahedron: Asymmetry
2009, 20, 131–173; c) S. G. Zlotin, A. S. Kucherenko, I. P. Belet-
skaya, Russ. Chem. Rev. 2009, 78, 737–784; d) J. N. Moorthy, S.
Saha, Eur. J. Org. Chem. 2009, 739–748; e) D. Almas¸i, D. A.
Alonso, A. N. Balaguer, C. Nꢃjera, Adv. Synth. Catal. 2009, 351,
1123–1131; f) B. Wang, G. Chen, L. Liu, W. Chang, J. Li, Adv.
Synth. Catal. 2009, 351, 2441–2448; g) B. Wang, X. Liu, L. Liu, W.
Chang, J. Li, Eur. J. Org. Chem. 2010, 5951–5954; h) A. L. Fuentes
de Arriba, L. Simꢄn, C. Raposo, V. Alcꢃzar, F. Sanz, F. M. MuÇiz,
J. R. Morꢃn, Org. Biomol. Chem. 2010, 8, 2979–2985; i) N. Mase,
C. F. Barbas, III, Org. Biomol. Chem. 2010, 8, 4043–4050.
[11] a) J. L. Reymond, Y. Chen, J. Org. Chem. 1995, 60, 6970–6979; b) J.
Kofoed, T. Darbre, J. L. Reymond, Org. Biomol. Chem. 2006, 4,
3268–3281; c) D. Font, S. Sayalero, A. Bastero, C. Jimeno, M. A.
Pericꢅs, Org. Lett. 2008, 10, 337–340.
[12] For the first examples of catalytic asymmetric Mukaiyama aldol re-
actions, see: a) S. Kobayashi, S. Nagayama, T. Busujima, Chem. Lett.
1999, 71–72; b) S. Kobayashi, S. Nagayama, T. Busujima, Tetrahe-
dron 1999, 55, 8739–8746.
been performed.
[20] S. Ishikawa, T. Hamada, K. Manabe, S. Kobayashi, J. Am. Chem.
Soc. 2004, 126, 12236–12237.
[21] a) S. Kobayashi, M. Ueno, T. Kitanosono, Top. Curr. Chem. 2012,
311, 1–18; b) S. Kobayashi, T. Ogino, H. Shimizu, S. Ishikawa, T.
Hamada, K. Manabe, Org. Lett. 2005, 7, 4729–4731.
[22] C. Ogawa, M. Kokubo, S. Kobayashi, J. Synth. Org. Chem. Jpn.
2010, 68, 718–728.
[23] a) M. Kokubo, T. Naito, S. Kobayashi, Chem. Lett. 2009, 38, 904–
905; b) M. Kokubo, T. Naito, S. Kobayashi, Tetrahedron 2010, 66,
1111–1118.
[24] S. Kobayashi, T. Endo, M. Ueno, Angew. Chem. 2011, 123, 12470–
12403; Angew. Chem. Int. Ed. 2011, 50, 12262–12265.
[25] M. Kokubo, C. Ogawa, S. Kobayashi, Angew. Chem. 2008, 120,
7015–7017; Angew. Chem. Int. Ed. 2008, 47, 6909–6911.
[26] a) C. Ogawa, N. Wang, M. Boudou, S. Azoulay, K. Manabe, S. Ko-
bayashi, Heterocycles 2007, 72, 589–598; b) M. Boudou, C. Ogawa,
S. Kobayashi, Adv. Synth. Catal. 2006, 348, 2585–2589; c) S. Azou-
lay, K. Manabe, S. Kobayashi, Org. Lett. 2005, 7, 4593–4595.
[27] M. V. Nandakumar, A. Tschçp, H. Krautscheid, C. Schneider,
Chem. Commun. 2007, 2756–2758.
[28] M. Kokubo, S. Kobayashi, Chem. Asian J. 2009, 4, 526–528.
[29] Some lanthanides, such as CeIII, are known to promote the epimeri-
zation between syn-3a and anti-3a through keto-enolization, see:
D. E. Ward, M. Sales, P. K. Sasmal, Org. Lett. 2001, 3, 3671–3673.
[30] T. Ollevier, B. Plancq, Chem. Commun. 2012, 48, 2289–2291.
[31] For the use of Bi compounds in organic synthesis, see: a) T. Ollevier,
Org. Biomol. Chem. 2013, 11, 2740–2755; b) Bismuth-Mediated Or-
ganic Reactions in Topics in Current Chemistry, Vol. 311 (Ed.: T. Ol-
levier), Springer, Berlin, 2012.
[13] a) S. Kobayashi, C. Ogawa, Chem. Eur. J. 2006, 12, 5954–5960; b) S.
Kobayashi, M. Sugiura, H. Kitagawa, W. W. L. Lam, Chem. Rev.
2002, 102, 2227–2302; c) S. Kobayashi, Lanthanide Triflate-Cata-
lyzed Carbon-Carbon Bond-Forming Reactions in Organic Synthesis
in Lanthanides: Chemistry and Use in Organic Synthesis (Ed.: S. Ko-
bayashi), Springer, Heidelberg, 1999; d) S. Kobayashi, Eur. J. Org.
Chem. 1999, 15–27; e) S. Kobayashi, Synlett 1994, 689–701; f) S. Ko-
bayashi, Chem. Lett. 1991, 2187–2190.
[32] A different metal-to-ligand ratio (5 mol%:15 mol%) allowed us to
obtain higher enantioselectivities for some substrates (3da: 98% ee
versus 89% ee, 3 fa: 98% ee versus 91% ee, 3ga: 98% ee versus
95% ee, 3ia: 98% ee versus 91% ee); see Ref. [30].
[33] a) Y. Qiao, Q. Chen, S. Lin, B. Ni, A. D. Headley, J. Org. Chem.
2013, 78, 2693–2697; b) T. Itoh, H. Ishikawa, Y. Hayashi, Org. Lett.
2009, 11, 3854–3857; c) Y. Hayashi, T. Itoh, M. Ohkubo, H. Ishika-
wa, Angew. Chem. 2008, 120, 4800–4802; Angew. Chem. Int. Ed.
2008, 47, 4722–4724; d) P. Garcꢆa-Garcꢆa, A. LadꢀpÞche, R. Halder,
B. List, Angew. Chem. 2008, 120, 4797–4799; Angew. Chem. Int. Ed.
2008, 47, 4719–4721; e) J. W. Yang, C. Chandler, M. Stadler, D.
Kampen, B. List, Nature 2008, 452, 453–455; f) Y. Hayashi, T. Itoh,
S. Aratake, H. Ishikawa, Angew. Chem. 2008, 120, 2112–2114;
Angew. Chem. Int. Ed. 2008, 47, 2082–2084.
[34] a) H. J. Li, H. Y. Tian, Y. J. Chen, D. Wang, C. J. Li, Chem.
Commun. 2002, 2994–2995; b) H. J. Li, H. Y. Tian, Y. C. Wu, Y. J.
Chen, L. Liu, D. Wang, C. J. Li, Adv. Synth. Catal. 2005, 347, 1247–
1256. See also c) J. Jankowska, J. J. Mlynarski, J. Org. Chem. 2006,
71, 1317–1321.
[14] Coordination Chemistry, Vol. 2 (Ed.: A. E. Martell), ACS Mono-
graph 168, American Chemical Society, Washington, DC, 1978.
[15] S. Kobayashi, S. Nagayama, T. Busujima, J. Am. Chem. Soc. 1998,
120, 8287–8288.
[16] a) Y. Mei, D. J. Averill, M. J. Allen, J. Org. Chem. 2012, 77, 5624–
5632; b) Y. Mei, P. Dissanayake, M. J. Allen, J. Am. Chem. Soc.
2010, 132, 12871–12873; c) T. Hamada, K. Manabe, S. Ishikawa, S.
Nagayama, M. Shiro, S. Kobayashi, J. Am. Chem. Soc. 2003, 125,
2989–2996; d) S. Kobayashi, T. Hamada, S. Nagayama, K. Manabe,
Org. Lett. 2001, 3, 165–167; e) S. Kobayashi, T. Hamada, S. Nagaya-
ma, K. Manabe, J. Braz. Chem. Soc. 2001, 12, 627–633; f) S. Na-
gayama, S. Kobayashi, J. Am. Chem. Soc. 2000, 122, 11531–11532.
[17] Trost et al. developed a dinuclear zinc catalyst system with a chiral
semi-crown ether ligand for catalytic asymmetric direct aldol reac-
tions, Henry reactions, Mannich-type reactions, and diol-desymmet-
rization reactions; see: a) B. M. Trost, A. Fettes, B. T. Shireman, J.
Am. Chem. Soc. 2004, 126, 2660–2661; b) B. M. Trost, T. Mino, J.
[35] F. Eustache, P. I. Dalko, J. Cossy, J. Org. Chem. 2003, 68, 9994–
10002.
11
&
&
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers! ÞÞ

www.chemasianj.org
Shu¯ Kobayashi et al.
[36] S. Ishikawa, T. Hamada, K. Manabe, S. Kobayashi, Synthesis 2005,
2176–2182.
[49] a) T. Chan, M. Brook, Tetrahedron Lett. 1985, 26, 2943–2946; b) R.
Helmer, R. West, Organometallics 1982, 1, 877–879; c) G. Morris,
R. Freeman, J. Am. Chem. Soc. 1979, 101, 760–762.
[50] I. Kuwajima, E. Nakamura, Acc. Chem. Res. 1985, 18, 181–187.
[51] C. T. Wong, M. W. Wong, J. Org. Chem. 2007, 72, 1425–1430.
[52] C. Mukherjee, T. Kitanosono, S. Kobayashi, Chem. Asian J. 2011, 6,
2308–2311.
[53] The enantiofacial differentiation of an aldehyde carbonyl group was
trivial, owing to the relative preference of the carbonyl group that
was coordinated by a Lewis acid for an (E)-configuration. One face
of the aldehyde was shielded by the tert-butyl group of ligand L1
and the orientation of the aldehyde might be fixed by a hydrogen
bond between the aldehyde and the hydroxy moieties of ligand L1.
[37] For example, see: a) M. G. B. Drew, A. H. B. Othman, S. M. Nelson,
J. Chem. Soc. Dalton 1976, 1394–1399; b) M. M. Bishop, J. Lewis,
T. D. OꢇDonoghue, P. R. Raithby, J. N. Ramsden, J. Chem. Soc.
Chem. Commun. 1978, 828–829; c) S. B. Larson, S. H. Simonsen,
J. N. Ramsden, J. Lagowski, Acta Crystallogr. Sect. C 1990, 46, 1930–
1932; d) P. Lainꢀ, A. Gourdon, J. P. Launay, J. P. Tuchagues, Inorg.
Chem. 1995, 34, 5150–5155.
[38] CCDC 828098 ([FeII(L1)
([FeII(L1) (H2O)]2+·2ClO4À), and CCDC 864123 ([FeII(L1)-
(MeCN)2A
(THF)2A
(H2O)]2+·2BrÀ) contain the supplementary crystallographic
(H2O)]2+·2ClO4À), CCDC 850236
ACHUTGTNRENNUG(DME)ACHTUTGNRENNUGN
A
CHTUNGTRENNUNG
A
CHTUNGTRENNUNG
data for this paper. These data can be obtained free of charge from
The Cambridge Crystallographic Data Centre via www.ccdc.cam.a-
c.uk/data_request/cif.
À
The formyl C H···O hydrogen bond is known to be important for
fixing an aldehyde in chiral-Lewis-acid-catalyzed reactions, including
the Mukaiyama aldol reaction; see: E. J. Corey, D. Barnes-Seeman,
T. W. Lee, Tetrahedron Lett. 1997, 38, 4351–4354.
[39] B. Plancq, T. Ollevier, Aust. J. Chem. 2012, 65, 1564–1572.
[40] a) B. J. Coe, S. Glenwright, Coord. Chem. Rev. 2000, 203, 5–80;
b) T. G. Appleton, H. C. Clark, L. E. Manzer, Coord. Chem. Rev.
1973, 10, 335–422.
[41] a) S. E. Denmark, G. L. Beutner, Angew. Chem. 2008, 120, 1584–
1663; Angew. Chem. Int. Ed. 2008, 47, 1560–1638; b) H. Fujisawa,
T. Mukaiyama, Chem. Lett. 2002, 182–183.
[42] T. Nakagawa, H. Fujisawa, Y. Nagata, T. Mukaiyama, Bull. Chem.
Soc. Jpn. 2004, 77, 1555–1567.
[43] For example, see: I. Iwakura, T. Ikeno, T. Yamada, Angew. Chem.
2005, 117, 2580–2583; Angew. Chem. Int. Ed. 2005, 44, 2524–2527.
[44] T. Kitanosono, M. Sakai, M. Ueno, S. Kobayashi, Org. Biomol.
Chem. 2012, 10, 7134–7147.
[54] Lubineau et al. demonstrated that the Mukaiyama aldol reaction in
water without any catalyst was feasible and could be accompanied
by significant rate acceleration compared to that in organic solvents.
The stereoselective preference of the products was reversed in com-
parison with that in conventional TiCl4-catalyzed reactions, see:[1]
a) A. Lubineau, E. Meyer, Tetrahedron 1988, 44, 6065–6070; b) A.
Lubineau, J. Org. Chem. 1986, 51, 2142–2144. Furthermore, the rate
acceleration and the stereochemical outcome in water bore a striking
resemblance to that under high pressure. A curious coincidence be-
tween the reaction in water and that under high pressure has been
observed in many reactions since, which is nowadays understood as
a result of the high dielectric constant and the highest cohesive
energy density (ced) of water; see: c) Y. Yamamoto, K. Maruyama,
K. Matsumoto, J. Am. Chem. Soc. 1983, 105, 6963–6965; d) A. Lubi-
neau, J. Augꢀ, Top. Curr. Chem. 1999, 206, 1–39; e) K. Hori, T. Ya-
maguchi, M. Uezu, J. Comput. Chem. 2011, 32, 778–786; f) Y. Jung,
R. A. Marcus, J. Am. Chem. Soc. 2007, 129, 5492–5502; g) J. F.
Blake, W. L. Jorgensen, J. Am. Chem. Soc. 1991, 113, 7430–7432.
[55] Full recovery of the starting materials was observed.
[45] Taking the low reactivity of silyl ketene acetals into consideration,
À
silyl enol ethers should be somewhat activated through d p interac-
tions with the FeII ion.
[46] We confirmed that the reaction did not proceed at all in the pres-
ence of an additive alone.
[47] a) H. E. Zimmerman, M. D. Traxler, J. Am. Chem. Soc. 1957, 79,
1920–1923; b) J. E. Dubois, M. Dubois, Tetrahedron Lett. 1967, 8,
4215–4219; c) J. E. Dubois, M. Dubois, Bull. Soc. Chim. Fr. 1969,
3553; d) J. E. Dubois, P. Fellmann, C. R. Acad. Sci. Ser. C 1972, 274,
1307; e) D. A. Evans, L. R. McGee, Tetrahedron Lett. 1980, 21,
3975–3977; f) C. H. Heathcock, Science 1981, 214, 395–400;
g) D. A. Evans, J. V. Nelson, T. R. Taber, Top. Stereochem. 1982, 13,
1–115; h) C. H. Heathcock in Comrehensive Carbanion Chemistry,
Part B (Eds.: E. Buncel, T. Dust), Elsevier, Amsterdam, 1984, Chap-
ter 4; i) C. Gennari, L. Colombo, C. Scolastico, R. Todeschini, Tetra-
hedron 1984, 40, 4051–4052; j) R. W. Hoffmann, K. Ditrich, Tetrahe-
dron Lett. 1984, 25, 1781–1784; k) C. H. Heathcock in Asymmetric
Synthesis, Vol. 3, Part B (Ed.: J. D. Morrison), Academic Press, New
York, 1984, Chapter 2.
[56] Profound correlation between the results and the amount of water
was observed in initial reports on the catalysis of YbACTHUNRGTNEUNG(OTf)3 in aque-
ous media; see: a) S. Kobayashi, I. Hachiya, J. Org. Chem. 1994, 59,
3590–3596; b) S. Kobayashi, I. Hachiya, Tetrahedron Lett. 1992, 33,
1625–1628.
[57] After submitting this manuscript, the following paper on the role of
water in lanthanide-catalyzed Mukaiyama aldol reactions has been
published: M. Hatanaka, K. Morokuma, J. Am. Chem. Soc. 2013,
135, 13972–13979.
[58] The ced is often used as an indicator of entropy-driven aggregation
and, thus, is experimentally proportional to “hydrophobicity”; see:
G. Graziano, J. Chem. Phys. 2004, 121, 1878–1882.
[48] a) S. Murata, M. Suzuki, R. Noyori, J. Am. Chem. Soc. 1980, 102,
3248–3249; b) Y. Yamamoto, K. Maruyama, Tetrahedron Lett. 1980,
21, 4607–4610.
Received: August 25, 2013
Published online: && &&, 0000
&
&
12
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPER
BiIIIFeII Clyro: Ternary catalytic sys-
tems that comprise a chiral ligand,
a (FeII or BiIII) metal salt, and an addi-
tive broaden the scope for the asym-
metric Mukaiyama aldol reaction. The
facile synthesis of the ligand under-
scores its versatility.
Aldol Reaction
Taku Kitanosono, Thierry Ollevier,
Shu¯ Kobayashi*
&&&& — &&&&
Iron- and Bismuth-Catalyzed Asym-
metric Mukaiyama Aldol Reactions in
Aqueous Media
13
&
&
Chem. Asian J. 2013, 00, 0 – 0
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers! ÞÞ