Chiral catalysts dually functionalized with amino acid and Zn2+ complex components for enantioselective direct aldol reactions inspired by natural aldolases: Design, synthesis, complexation properties, catalytic activities, and mechanistic study

Article abstract of DOI:10.1002/chem.200900733

Aldolases are enzymes that catalyze stereospecific aldol reactions in a reversible manner. Naturally occurring aldolases include class I aldolases, which catalyze aldol reactions via enamine intermediates, and class II aldolases, in which Zn²⁺ enolates of substrates react with acceptor aldehydes. In this work, Zn²⁺ complexes of Lprolyl-pendant[15] aneN₅ (ZnL³), Lprolyl-pendant[12]aneN₄ (ZnL⁴), and L-valyl-pendant[12]aneN₄ (ZnL⁵) were designed and synthesized for use as chiral catalysts for enantioselective aldol reactions. The complexation constants for L³ to L⁵ with Zn²⁺ [logK_s(ZnL)] were determined to be 14.1 (for ZnL³), 7.6 (for ZnL⁴), and 9.6 (for ZnL⁵), indicating that ZnL³ is more stable than ZnL⁴ and ZnL ⁵. The depro-tonation constants of Zn²⁺-bound water [pK_a values] for ZnL³, ZnL⁴, and ZnL ⁵ were calculated to be 9.2 (for ZnL³), 8.2 (for ZnL ⁴), and 8.6 (for ZnL⁵), suggesting that the Zn ²⁺ ions in ZnL³ is a less acidic Lewis acid than in ZnL⁴ and ZnL⁵. These values also indicated that the amino groups on the side chains weakly coordinate to Zn²⁺. We carried out aldol reactions between acetone and 2-chlorobenzaldehyde and other aldehydes in the presence of catalytic amounts of the chiral Zn²⁺ complexes in acetone/H₂O at 25 and 37°C. Whereas ZnL³ yielded the aldol product in 43% yield and 1 % ee (R), ZnL⁴ and ZnL⁵ afforded good chemical yields and high enantioselectivities of up to 89% ee (R). UV titrations of proline and ZnL⁴ with acetylacetone (acac) in DMSO/H₂O (1:2) indicate that ZnL⁴ facilitates the formation of the ZnL⁴.(acac)^- complex (K_app = 2.1X10²M^-1), whereas L-proline forms a Schiff base with acac with a very small equilibrium constant. These results suggest that the amino acid components and the Zn²⁺ ions in ZnL⁴ and ZnL⁵ function in a cooperative manner to generate the Zn ²⁺-enolate of acetone, thus permitting efficient enantioselective C-C bond formation with aldehydes.

Full text of DOI:10.1002/chem.200900733

DOI: 10.1002/chem.200900733
Chiral Catalysts Dually Functionalized with Amino Acid and Zn2+ Complex
Components for Enantioselective Direct Aldol Reactions Inspired by Natural
Aldolases: Design, Synthesis, Complexation Properties, Catalytic Activities,
and Mechanistic Study
Susumu Itoh,[a] Masanori Kitamura,[a, b] Yasuyuki Yamada,[a, b] and Shin Aoki*[a, b]
Abstract: Aldolases are enzymes that
catalyze stereospecific aldol reactions
in a reversible manner. Naturally oc-
curring aldolases include class I aldo-
lases, which catalyze aldol reactions via
enamine intermediates, and class II al-
dolases, in which Zn2+ enolates of sub-
strates react with acceptor aldehydes.
In this work, Zn2+ complexes of l-
prolyl-pendant[15]aneN5 (ZnL3), l-
prolyl-pendant[12]aneN4 (ZnL4), and
l-valyl-pendant[12]aneN4 (ZnL5) were
designed and synthesized for use as
chiral catalysts for enantioselective
aldol reactions. The complexation con-
stants for L3 to L5 with Zn2+ [logKs-
tonation constants of Zn2+-bound
water [pKaA
(ZnL) values] for ZnL3,
aldol product in 43% yield and
1% ee (R), ZnL4 and ZnL5 afforded
good chemical yields and high enantio-
selectivities of up to 89% ee (R). UV
titrations of proline and ZnL4 with ace-
tylacetone (acac) in DMSO/H2O (1:2)
indicate that ZnL4 facilitates the for-
CTHUNGTRENNUNG
ZnL4, and ZnL5 were calculated to be
9.2 (for ZnL3), 8.2 (for ZnL4), and 8.6
(for ZnL5), suggesting that the Zn2+
ions in ZnL3 is a less acidic Lewis acid
than in ZnL4 and ZnL5. These values
also indicated that the amino groups
on the side chains weakly coordinate to
Zn2+. We carried out aldol reactions
between acetone and 2-chlorobenzalde-
hyde and other aldehydes in the pres-
ence of catalytic amounts of the chiral
Zn2+ complexes in acetone/H2O at 25
and 378C. Whereas ZnL3 yielded the
ACTHNUTRGNEUNG
mation of the ZnL4·(acac)ꢀ complex
(Kapp =2.1ꢁ102 mꢀ1), whereas l-proline
forms a Schiff base with acac with a
very small equilibrium constant. These
results suggest that the amino acid
components and the Zn2+ ions in ZnL4
and ZnL5 function in a cooperative
manner to generate the Zn2+-enolate
of acetone, thus permitting efficient
ꢀ
(ZnL)] were determined to be 14.1 (for
enantioselective C C bond formation
with aldehydes.
Keywords:
asymmetric catalysis
models · macrocyclic ligands · zinc
aldol
reaction
·
·
enzyme
Introduction
number of excellent studies relating to stereoselective cata-
lytic aldol reactions have appeared.[1] Representative exam-
ples of chiral catalysts for direct asymmetric aldol reactions
include the heterobimetallic LaLi3tris(binaphthoxide) cata-
lysts developed by Shibasaki,[2] Trostꢀs Zn2+-semi-crown
ether,[3] and others.[4] These reactions are largely conducted
in organic solvents.
In living systems, natural enzymes called aldolases work
as catalysts for stereospecific and reversible aldol reac-
tions.[5] As one example, fructose 1,6-bis(phosphate) aldo-
lase (FBP-aldolase, EC 4.1.2.13) catalyzes the cleavage of d-
fructose 1,6-bis(phosphate) (FBP) to give dihydroxyacetone
phosphate (DHAP) and d-glyceraldehyde 3-phosphate
(G3P), as well as the reverse formation of FBP from DHAP
and G3P (Scheme 1). Rabbit muscle aldolase (RAMA), a
FBP aldolase, has so far been most widely used in organic
The aldol reaction is one of the most important reactions in
ꢀ
use for the formation of C C bonds in organic synthesis. A
[a] S. Itoh, Dr. M. Kitamura, Dr. Y. Yamada, Prof. S. Aoki
Faculty of Pharmaceutical Sciences
Tokyo University of Science, 2641 Yamazaki
Noda 278-8510 (Japan)
Fax : (+81)4-7121-3670
E-mail: shinaoki@rs.noda.tus.ac.jp
[b] Dr. M. Kitamura, Dr. Y. Yamada, Prof. S. Aoki
Center for Drug Delivery Research
Tokyo University of Science
2641 Yamazaki, Noda 278-8510 (Japan)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200900733.
10570
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10570 – 10584

FULL PAPER
Scheme 1.
synthesis.[5c,6] Natural aldolases can be classified into two
groups on the basis of their reaction mechanisms. In the
case of class I aldolases, an enamine intermediate is formed
between the lysine residue of the enzyme and the carbonyl
group of the substrate. In class II aldolases, a zinc(II) ion co-
factor acts as a Lewis acid to generate enolates at the active
site.
Chiral organocatalysts[7–16] have recently emerged as a re-
agent class representing a new methodology for stereoselec-
tive aldol reactions, in that they are capable of mimicking
class I aldolases. In 2000, List, Lerner, and Barbas reported
that l-proline serves as a catalyst for direct aldol reactions
of acetone and aldehydes to give aldol adducts such as 1 in
good chemical yields and with moderate enantioselectivities
(Scheme 2).[7] It was suggested that proline forms enamine
intermediates that react with aldehydes to give aldol prod-
ucts (CTN stands for catalytic turnover number).[8]
In later studies, Yamamoto et al. screened a wide variety
of diamine salts including 2 for use in aldol reactions be-
tween several ketones and 4-nitrobenzaldehyde.[9] Maruoka
and co-workers successfully extended the concept of amino
acid catalysis to novel catalysts containing binaphthyl- or bi-
phenyl-based axial chirality (e.g., 3).[10] Barbas et al. report-
ed on asymmetric aldol reactions between ketones and alde-
hydes in water catalyzed by a combination of lipophilic dia-
mine 4 and an acid such as TFA.[11] Hayashiꢀs group inde-
pendently developed a proline-based catalyst—the trans-4-
silyloxyproline 5—for direct aldol reactions of ketones and
aldehydes in water.[12] Some dipeptide derivatives such as 6,
developed by Cꢃrdovaꢀs group, have emerged as another at-
tractive class of catalysts for asymmetric aldol reactions in
aqueous media.[13] Barbas et al. recently reported on some
syn-selective organocatalytic aldol reactions, catalyzed by
amino acid derivatives, of unmodified hydroxyacetone and
dihydroxyacetone to yield syn-1,2-diols.[14]
In 1995, Lerner and co-workers developed catalytic anti-
bodies that displayed aldolase activity by immunizing mice
with a 1,3-diketone hapten.[17] The aldolase antibodies,
which were obtained after screening of polyclonal antibodies
for binding with the hapten, were found to catalyze aldol re-
actions by the enamine mecha-
nism, like the class I aldolases.
To date, natural and artificial
catalysts that possess both func-
tionalities—Schiff-base-forming
part and Zn2+ site—have
scarcely been reported. Some
Zn2+ complexes of proline de-
rivatives have been demonstrat-
ed to have the ability to cata-
lyze direct asymmetric aldol re-
Scheme 2.
actions in aqueous media, mim-
Chem. Eur. J. 2009, 15, 10570 – 10584
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10571

S. Aoki et al.
icking class II (and class I) aldolases. Darbre and Reymond
In this context, we became interested in the reactivities
and enantioselectivities of novel asymmetric catalysts each
dual-functionalized with an enamine-forming amino group
and a Zn2+ complex of a macrocyclic polyamine such as
et al. showed that the 1:2 complex[18] of Zn2+ and l-proline
(ZnACHTUNGTRENNUNG(l-Pro)2, 7) catalyzes aldol reactions in an acetone/H2O
1,4,7,10,13-pentaazacyclopentadecane
([15]aneN5)
or
1,4,7,10-tetraazacyclododecane ([12]aneN4 or cyclen). In par-
ticular, we envisaged the development of enantioselective
ꢀ
C C bond-formation reactions in water, which is considered
to
have
enormous
potential
as
a
reaction
medium.[11–13,15,19,22,25] In this work, we synthesized the l-
prolyl-pendant[15]aneN5 11 (L3), the l-prolyl-pendant[12]-
solvent system.[19,20] They sug-
gested that in the aldol reac-
tions of these substrates,
7
forms enamine species with ace-
tone, but that with dihydroxy-
acetone (DHA) it forms Zn2+
-
enolate intermediates.[21] Mly-
narskiꢀs group reported on a
Zn2+ complex of a C2-symmet-
ric chiral ligand containing two
proline units (compound 8) for
use in catalytic aldol reactions
of acetone and cyclohexanone
with several aldehydes, giving
good yields and enantioselectiv-
ities.[22]
Meanwhile, it has been estab-
lished that a Zn2+ complex of
cyclen
9
(ZnL1; cyclen=
1,4,7,10-tetraazacyclododecane) is a good model for the nat-
urally occurring Zn2+ enzymes (Scheme 3).[23] A Zn2+ com-
plex of the 4-bromophenacyl-pendant cyclen 10 (ZnL2), for
example, was reported to be a good model for the class II al-
aneN4 12 (L4), and the l-valyl-pendant[12]aneN4 13 (L5), to-
gether with the corresponding Zn2+ complexes, 14 (ZnL3),
15 (ZnL4), and 16 (ZnL5), for stereoselective direct aldol re-
actions. From a molecular model study, we expected that the
nitrogen components of the amino acid moieties in these
Zn2+ complexes would scarcely or weakly bind to Zn2+. As
shown in Scheme 4, we hypothesized that Zn2+ catalysts
would initially form complexes 21 with carbonyl substrates
activated by Zn2+. Our question was, which intermediate,
enamine or Zn2+-enolate, would be preferred by these dual-
functionalized catalysts in reactions with acceptor aldehydes
to yield aldol adducts?
dolases.[24] Potentiometric pH titration of 10a [ZnL2
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
-
enolate form 10c [ZnACTHNUTRGNEUNG
(Hꢀ1L2)] in aqueous solution. Note
that Zn2+ ions in Zn2+-cyclen complexes exhibit strong
Lewis acidity and that the pKa values of the Zn2+-bound
water molecules in 9 and 10 are 7.8 and 8.4, respectively.
Here we wish to report on the synthesis and Zn2+-com-
plexation properties of 11 (L3),
12 (L4), and 13 (L5) and to pres-
ent results for some enantiose-
lective aldol reactions per-
formed in the presence of the
Zn2+ complexes 14 (ZnL3), 15
(ZnL4), and 16 (ZnL5) and the
related complexes 17 (ZnL6),
18 (ZnL7), 19 (ZnL8), and 20
(ZnL9). Mechanistic study sug-
gests that the amino acid parts
Scheme 3.
and the Zn2+ ions of the cata-
10572
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10570 – 10584

Chiral Aldol Reaction Catalysts
FULL PAPER
Scheme 4. Initially hypothesized scheme for aldol reactions catalyzed by
Zn2+ complexes of macrocyclic polyamines with pendant amino acids.
lysts function in a cooperative manner to generate Zn2+
enolates to give aldol adducts.
-
Results and Discussion
Synthesis of chiral Zn2+ complexes: The ligand 11 (L3) was
synthesized from [15]aneN5[26] and l-proline, whereas 12
(L4), 13 (L5), and L6 to L9 were synthesized from 3Boc-
cyclen.[27] The Zn2+ complexes of the ligands were obtained
by treatment of acid-free ligand with Zn2+, either before-
hand or in situ immediately before use. For details, see
Schemes S1–S6 in the Supporting Information.
Figure 1. a) Typical potentiometric pH titration curves for H5L3 [0.5 mm,
prepared in situ from H4L3 (0.5 mm)+HNO3 (0.5 mm)] and H5L3
(0.5 mm)+Zn2+ (0.5 mm) with I=0.1 (NaNO3) at 258C. “Equiv (HOꢀ)”
is the number of equivalents of base (NaOH) added. b) Speciation dia-
gram for a mixture of L3 (25 mm)+Zn2+ (25 mm) as a function of pH
with I=0.1 (NaNO3) at 258C.
Deprotonation constants and Zn2+ complexation character-
istics of 11 (L3), 12 (L4), and 13 (L5), as evidenced by poten-
tiometric pH titrations: Typical potentiometric pH titration
curves for H5L3 [0.5 mm, from 11·4TFA (H4L3) (0.5 mm) +
HNO3 (0.5 mm)] with I=0.1 (NaNO3) against NaOH (0.1m)
at 258C (dashed curve in Figure 1a) were analyzed for acid–
base equilibrium (1), where aH+ denotes the activity of H+.
The deprotonation constants, pKai (i=1–5), of H5L3 were
determined to be <3 (pKa1), 3.5ꢁ0.1 (pKa2), 8.2ꢁ0.1
(pKa3), 9.0ꢁ0.1 (pKa4), and 9.5ꢁ0.1 (pKa5) with the aid of
the “BEST” software program,[28] as summarized in
Scheme 5 and Table 1.
lar to that for cyclen (16.2), which is a tetradentate ligand,[29]
and is much greater than those for [12]aneN3 22 (L10, 8.4),[29]
N-dansylcyclen 24 (L11, 6.5),[30] and 26 (L7, 8.0),[31] which
have been established as tridentate ligands for Zn2+
(Scheme 6 and Table 1). The deprotonation constant for the
Zn2+-bound water of ZnL3 (for 14a Q 14b) [pK
aACTHNUTRGNEU(GN ZnL)], as
H
ð5ꢀiÞL Ð Hð4ꢀiÞL þ Hþ
:
defined by Eq. (3), was calculated to be 9.2, which is greater
than the 7.8 observed for the Zn2+-cyclen complex 9 (ZnL1)
(Table 1). These data suggest that L3 serves as a tetradentate
or pentadentate ligand to Zn2+. Figure 1b displays the dis-
tribution diagram for a mixture of L3 (25 mm) and Zn2+
ð1Þ
Kai ¼ ½Hð4ꢀiÞLꢂaHþ =½Hð5ꢀiÞLꢂ ði ¼ 1 ꢀ 5Þ
L þ Zn2þ Ð ZnL : KsðZnLÞ ¼ ½ZnLꢂ=½Lꢂ½Zn2þꢂ ðin mꢀ1
Þ
ð2Þ
ð3Þ
(25 mm) in water, showing that 14 {ZnL3: 14a [ZnL3
+ 14b [ZnL3(HOꢀ)]} is formed almost quantitatively at
pH>7.0.
ACHTUNGTRENNUNG(H2O)]
ZnLðH2OÞ Ð ZnLðHOꢀÞ þ Hþ
:
AHCTUNGTRENNUNG
KaðZnLÞ ¼ ½ZnLðHOꢀÞꢂaHþ =½ZnLðH2OÞꢂ
Analysis of
a potentiometric
pH titration curve (plain curve
in Figure 1a) for a mixture of
H5L3 [0.5 mm, prepared from
H4L3 (0.5 mm) and HNO3
(0.5 mm)] with ZnSO4 (0.5 mm)
gave a complexation constant
[logKsACHTUNGTRENNUNG
(ZnL3)], as defined by
Eq. (2), of 14.1 (see Scheme 5
and Table 1). This value is simi- Scheme 5.
Chem. Eur. J. 2009, 15, 10570 – 10584
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10573

S. Aoki et al.
Table 1. Complexation constants [logK
cyclen (L1), 26 (L7), 22 (L10), and 24 (L11) with I=0.1 (NaNO3) at 258C.[a]
(ZnL)] and deprotonation constants (pKai) of 11 (L3), 12 (L4), 13 (L5),
sACHTUNGTRENNUNG
played in Figure 2, and shows
ACTHNUTRGNEUNG
that ZnL4 [ZnL4(H2O)+ZnL4-
11 (L3)[b]
12 (L4)[b]
13 (L5)[b]
26 (L7)[b]
cyclen (L1)[c]
22 (L10)[c]
24 (L11)[d]
(HOꢀ)]
is
predominant
pKa1
pKa2
pKa3
pKa4
pKa5
logK
<3
3.5
8.0
9.0
9.5
<3
5.7
8.8
9.7
<3
5.6
7.6
2.6
6.4
10.8
<2
<2
9.9
2.4
7.6
12.6
<2
3.5
6.4
9.7
(>95%) above pH 7.4.
As listed in Table 1, the pKa-
AHCTUNGTERNNUNG -
(ZnL) values for the Zn2+
10.9
11.0
bound water in 14 (ZnL3), 15
(ZnL4), and 16 (ZnL5) were
found to be 9.2, 8.2, and 8.6, re-
spectively, larger than those for
9 (ZnL1) of 7.8, 23 (ZnL10) of
7.3, and 25 (ZnL11) of 7.4.
(ZnL)
14.1
G
7.6
G
9.6
G
8.0
T
16.2
G
8.4
G
6.5
ACHTUNGTRENNUNG
pKaACHTUNGTRENNUNG
(ZnL)[e]
[a] See the text for definitions of logK
(ZnL) and pK
H
ence [29] [d] From reference [30]. [e] Deprotonation constants of Zn2+-bound water in ZnL.
Figure 2. Speciation diagram for a mixture of L4 (25 mm)+Zn2+ (25 mm)
as a function of pH with I=0.1 (NaNO3) at 258C.
These data indicate that the Lewis acidity of Zn2+ in 14, 15,
and 16 is weaker than that in 9, 23, and 25. The fact that the
pKaACTHUNRTGNEUNG(ZnL) values for 15 (8.2) and 16 (8.6) are somewhat
larger than those of 18 (7.3),[31] 23 (7.3), and 25 (7.4), which
are Zn2+ complexes of tridentate ligands, indicates the pos-
sibility that the nitrogen components in the prolyl and valyl
units of 15 and 16 coordinate only weakly to Zn2+ (the
cyclen units of these ligands are tridentate binders for Zn2+).
Enantioselective aldol reactions catalyzed by chiral Zn2+
complexes in aqueous media: On the basis of the data on
the Zn2+ complexation behavior of the chiral ligands de-
scribed above, we initially examined the aldol reaction be-
tween acetone and 2-chlorobenzaldehyde in the presence of
the chiral catalysts in a DMSO/acetone system. The results
are summarized in Table 2, in which the catalyst, equivalents
of catalysts, solvent, reaction conditions, chemical yields of
isolated products, catalytic turnover number (CTN), and
enantioselective excesses (ee) of the aldol adduct 1 are
listed. As described in the In-
Scheme 6.
Typical potentiometric pH titration curves for H4L4 and
H4L5 (0.5 mm, dashed curves in Figure S1 and S2a in the
Supporting Information) and for a mixture (plain curves in
Figure S1 and S2a) of H4L4 (0.5 mm) and ZnSO4 (0.5 mm)
with I=0.1 (NaNO3) against NaOH (0.1m) at 258C were an-
alyzed to determine pKa, logKsACHTNUGRTENNUG(ZnL), and pKaAHCUTNGTREN(NUNG ZnL) values,
as summarized in Scheme 7 and Table 1. A speciation dia-
gram for a mixture of L4 (25 mm) and Zn2+ (25 mm) is dis-
troduction, l-proline (20 mol%
relative to the aldehyde) gave 1
with
a good chemical yield
(85%) and enantioselectivity
(67% ee) in DMSO at 258C
(entry 1). As listed in entries 2
and 3, 14 (ZnL3) gave only
traces of 1, whereas metal-free
Scheme 7.
10574
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10570 – 10584

Chiral Aldol Reaction Catalysts
FULL PAPER
(ZnL4, entry 7), whereas 12 (L4)
plus TFA (1 equiv) yielded 1
with a low ee (entry 8).[32]
Table 2. The results of an asymmetric aldol reaction between acetone and 2-chlorobenzaldehyde in the pres-
ence of l-proline, 11, 12, 14, 15, 27, and 28.
Darbreꢀs group performed
aldol reactions with the ZnACHTUNGTRENNUNG(l-
Pro)2 complex 7 in Bicine buf-
Entry Catalyst [mm][a]
Mol% Solvent
Conditions
Yield [%][b] CTN[c] ee [%][d]
fer,[19e] so we examined C C
ꢀ
1
2
3
4
5
6
7
8
l-proline
20
10
10
DMSO/acetone 4:1 4 h, 258C
DMSO/acetone 4:1 26 h, 25–508C
DMSO/acetone 4:1 4 h, 258C
DMSO/acetone 4:1 4 h, 258C
DMSO/acetone 4:1 4 h, 258C
DMSO/acetone 4:1 4 h, 258C
DMSO/acetone 4:1 4 h, 258C
DMSO/acetone 4:1 4 h, 258C
85
3
30
5
4
–
3
–
67 (R)
1 (R)
2 (R)
19 (R)
racemic
12 (R)
34 (R)
9 (S)
bond formation in various
buffer systems at pH 5.5–8.7 by
use of buffer solutions. As sum-
marized in Table 3, 14 (ZnL3)
gave 2-chlorobenzoic acid as a
major product in MES and
HEPES buffers at <pH 7 (en-
tries 1–3), presumably due to
the oxidation of 2-chlorobenzal-
dehyde, and negligible amounts
of aldol adduct in Bicine buffer
(entries 4 and 5). The reactivity
of 15 (ZnL4) at pH 7.0–8.7 was
better (entries 6–10, Table 3).
The higher chemical yields and
low enantioselectivities seen
under these conditions can be
explained by the fact that this
aldol reaction proceeds without
14 (ZnL3)[e] (10)
11 (L3)[f] (10)
11 (L3)[f] +TFA (10) 10
27[f] (10)
10
10
10
8
<1
28 (10)
15
12
12
2
1
1
15 (ZnL4)[e] (10)
12 (L4)[f] +TFA (10) 10
[a] Numbers in parentheses are the concentrations of catalysts in solvent (DMSO/acetone). [b] Isolated yield.
[c] Catalytic Turnover Number (=yield/equivalents of catalyst). [d] Determined by HPLC analysis with a
chiral column (reference [32]). [e] ZnL complexes were formed in situ. [f] L and 27 were extracted with CHCl3
from aq. NaOH (pH>12) prior to use.
Table 3. The results of the asymmetric aldol reaction between acetone and 2-chlorobenzaldehyde in the pres-
ence of 14 (ZnL3) and 15 (ZnL4) in buffer systems.
Entry Catalyst [mm][a] Mol% Solvent
Conditions Yield [%][b] CTN[c] ee [%][d]
1
2
3
4
5
6
7
8
14 (ZnL3)[e] (10) 10
14 (ZnL3)[e] (10) 10
14 (ZnL3)[e] (10) 10
14 (ZnL3)[e] (10) 10
14 (ZnL3)[e] (10) 10
15 (ZnL4)[e] (10) 10
15 (ZnL4)[e] (10) 10
15 (ZnL4)[e] (10) 10
15 (ZnL4)[e] (10) 10
15 (ZnL4)[e] (10) 10
MES buffer (0.1m, pH 5.5)[f]
MES buffer (0.1m, pH 6.0)[f]
58 h, 258C trace
24 h, 258C trace
–
–
–
–
–
2
2
2
5
8
–
–
–
–
–
HEPES buffer (0.1m, pH 7.0)[f] 24 h, 258C trace
Bicine buffer (20 mm, pH 7.1)[g] 3 h, 378C
Bicine buffer (20 mm, pH 8.2)[g] 3 h, 378C
5
15
5 (S)
racemic
39 (R)
20 (R)
8 (S)
7 (S)
7 (S)
–
HEPES buffer (0.1m, pH 7.0)[h] 20 h, 258C 17
a catalyst at basic pH (en-
tries 11 and 12).
The results listed in Tables 2
and 3 prompted us to reex-
amine the effect of pH and
buffer molecules on the aldol
Tris buffer (20 mm, pH 8.0)[g]
3 h, 378C
15
17
48
81
Bicine buffer (20 mm, pH 7.5)[g] 3 h, 378C
Bicine buffer (20 mm, pH 8.0)[g] 3 h, 378C
Bicine buffer (20 mm, pH 8.7)[g] 3 h, 378C
9
10
11
12
none
none
–
–
CES buffer (0.1m, pH 10.0)[h]
24 h, 258C <57[i]
CAPS buffer (0.1m, pH 11.0)[h] 23 h, 258C <78[i]
–
reaction catalyzed by ZnACHTUNGTRENNUNG(l-
[a] Numbers in parentheses are the concentrations of catalysts in solvent. [b] Isolated yield. [c] Catalytic Turn-
over Number (=yield/equivalents of catalyst). [d] Determined by HPLC analysis with a chiral column (refer-
ence [32]). [e] ZnL complexes were formed in situ. [f] Buffer/acetone 5:1. [g] Buffer/acetone 1:1. [h] Buffer/
acetone 4:1. [i] Crude product yield.
Pro)2 complex 7, as reported by
Darbre et al.[19e,20,21] As listed in
Table 4, negligible amounts of
11 (L3) afforded 1 in 30% yield with only 2% ee (R). The
concentration of 14 was determined to be 10 mm, at which,
on the basis of the potentiometric pH titration data shown
in Figure 1, 14 (14a and 14b) would be formed quantitative-
ly (the pH value of an aqueous solution of 14 was ca. 8.0).
Barbas et al. reported on direct asymmetric aldol reactions
in the presence of 4 and TFA, as described above.[11] How-
ever, the addition of TFA
(1 equiv) to 11 had no effect on
the chemical yield and the
enantioselectivity (entry 4,
Table 4. The results of an asymmetric aldol reaction between acetone and 2-chlorobenzaldehyde catalyzed by
7 (Zn
ACHTUNGTERNN(UNG l-Pro)2).
Entry Catalyst [mm][a] Mol% Solvent
Conditions Yield [%][b] CTN[c] ee [%][d]
Table 2). The reactivities of
1,4,7,10,13-pentaazacyclopenta-
decane 27 ([15]aneN5) and 28,
in which the secondary amines
of [15]aneN5 were all protected,
were lower than that of 11 (en-
tries 5 and 6). In contrast, 1 was
obtained in 12% yield and with
1
2
3
4
5
6
7
7
7
7
7
7
[e] (10)
[e] (10)
[e] (10)
[e] (10)
[e] (10)
[e] (10)
10
10
10
10
10
10
aq. HNO3 (0.02n, pH 2.2)[f]
H2O (pH 5.2)[f]
20 h, 378C trace
20 h, 378C trace
20 h, 378C 12
–
–
1
–
7
10
–
–
aq. NaOH (0.02n, pH 9.0)[f]
1 (S)
1 (S)
10 (S)
11 (S)
Bicine buffer (20 mm, pH 6.9)[g] 3 h, 378C
Bicine buffer (20 mm, pH 8.5)[g] 3 h, 378C
Bicine buffer (20 mm, pH 9.4)[g] 3 h, 378C
4
71
99
[a] Numbers in parentheses are the concentrations of catalysts in solvent. [b] Isolated yield. [c] Catalytic Turn-
over Number (=yield/equivalents of catalyst). [d] Determined by HPLC analysis with a chiral column (refer-
34% ee in the presence of 15 ence [32]). [e] ZnL complexes were formed in situ. [f] Solvent/acetone 5:1. [g] Solvent/acetone 1:1.
Chem. Eur. J. 2009, 15, 10570 – 10584
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10575

S. Aoki et al.
good yield (73–78%) and with
moderate
enantioselectivity
(67–80% ee) in acetone and the
acetone/H2O system (entries 4
and 5).[33] Metal-free 12 (L4)
gave only the racemic aldol
product (entry 6), whereas the
Cd2+ and Cu2+ complexes of L4
(compounds 29a and 29b) pro-
moted aldol reactions only to a
negligible extent (entries 7 and
8). It was also found that free
Zn2+
afforded
negligible
amounts of adduct (entry 9). It
had previously been reported
adduct were obtained at acidic
pH and 378C (entries 1 and 2),
Table 5. The results of the asymmetric aldol reaction between acetone and 2-chlorobenzaldehyde catalyzed by
l-proline and Zn2+ complexes in an acetone/H2O system.
whereas
1 was produced in
Entry Catalyst [mm][a]
Mol% Solvent
Conditions Yield [%][b] CTN[c] ee [%][d]
12% yield as a racemic mixture
at basic pH at the same temper-
ature (entry 3), suggesting that
the presence of Bicine in the
reaction mixture is an impor-
tant factor for chiral induction.
When Bicine was used as a
buffer, the chemical yield was
improved, but the enantioselec-
tivity was low (note that we
1
2
3
4
5
6
7
8
l-proline (50)
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
acetone/H2O 4:1 20 h, 378C 22
acetone/H2O 9:1 20 h, 258C 31
acetone/H2O 4:1 20 h, 378C 43
4
6
9
16
15
14
1
–
–
17
14
11
1
2
13
–
48 (R)
10 (R)
1 (R)
67 (R)
80 (R)
racemic
50 (R)
–
14 (ZnL3)[e] (25)
14 (ZnL3)[e] (50)
15 (ZnL4)[e,f] (25)
15 (ZnL4)[g] (50)
12 (L4) (50)
acetone
20 h, 258C 78
acetone/H2O 4:1 20 h, 378C 73
acetone/H2O 4:1 20 h, 378C 72
29a (CdL4)[e] (50)
acetone/H2O 4:1 20 h, 378C
5
29b (CuL4)[e] (50)
acetone/H2O 4:1 20 h, 378C trace
acetone/H2O 4:1 20 h, 378C trace
acetone/H2O 4:1 20 h, 378C 87
acetone/H2O 4:1 20 h, 258C 72
acetone/H2O 4:1 20 h, 378C 54
9
Zn
N
–
10
11
12
13
14
15
16
17
16 (ZnL5)[e] (50)
16 (ZnL5)[e] (50)
17 (ZnL6)[e] (50)
18 (ZnL7)[g] (50)
19 (ZnL8)[e] (50)
20 (ZnL9)[e] (50)
9 (ZnL1)[g] (50)
80 (R)
89 (R)
9 (R)
2 (S)
11 (R)
12 (R)
–
used
2-chlorobenzaldehyde,
acetone/H2O 4:1 20 h, 378C
acetone/H2O 4:1 20 h, 378C
5
9
whereas Darbre et al. used 4-ni-
trobenzaldehyde). It is known
that Bicine is intrinsically capa-
ble of complexing with Zn2+
(logKs1 and logKs2 values are
5.4 and 8.6).[20b] Therefore, we
assume that Zn2+ is coordinat-
ed by l-proline and Bicine mol-
ecules to form complex species
for the enantiofacial discrimination of the aldehyde.
We next carried out the same reaction in acetone/H2O as
solvent in the presence of l-proline, 14 (ZnL3), 15 (ZnL4),
and 16 (ZnL5) and the reference complexes, 17–20 (ZnL6 to
ZnL9), 9 (ZnL1), the Cd2+ complex 29a of L4 (CdL4), and
the Cu2+ complex 29b of L4 (CuL4). The results are sum-
marized in Table 5.
acetone/H2O 4:1 20 h, 378C 66
acetone/H2O 4:1 20 h, 378C trace
acetone/H2O 4:1 20 h, 378C trace
l-proline+9 (ZnL1)[g] (50)
–
–
[a] Numbers in parentheses are the concentrations of catalysts in solvent. [b] Isolated yield. [c] Catalytic Turn-
over Number (=yield/equivalents of catalyst). [d] Determined by HPLC analysis with a chiral column (refer-
ence [32]). [e] ZnL complexes, CdL4, and CuL4 were formed in situ. [f] Zn
complexes were used.
ACHTUNGERTN(NUNG OTf)2 was used. [g] Isolated ZnL
by us that the Lewis acidity of the Zn2+-cyclen complex is
higher than that of Cd2+-cyclen.[34] We therefore concluded
that the Lewis acidity of Zn2+ is an important factor for this
enantioselective aldol reaction.
Other Zn2+ complexes were also tested, as listed in en-
tries 10–13, Table 5. The Zn2+ complex 16 (ZnL5), contain-
ing a valine unit, gave results almost the same (entry 10) as
those obtained with 15 (ZnL4). Most of the reactions in
Tables 3–5 were carried out at 378C as enzyme model stud-
ies. It was shown that the ee values can be improved at
lower temperatures (89% ee, entry 11). The Zn2+ complexes
of N-methylprolyl-pendant cyclen 17 (ZnL6) and (S)-a-
In the acetone/H2O system, l-proline gave 1 in 22% yield
and with 48% ee at 378C (entry 1 in Table 5), both of which
were lower than those obtained in the DMSO/acetone
system (entry 1 in Table 2). In
contrast, 14 (ZnL3) gave better
results
10% ee (R)] at 258C (entry 2 in
Table 5) than in DMSO
(entry 2 in Table 2). At 378C, a
nearly racemic adduct was ob-
tained (entry 3 in Table 5). In-
terestingly, 15 (ZnL4) gave 1 in
[31%
yield
and
methylbenzylbenzylaminocarbonyl-pendant
cyclen
18
(ZnL7) were used for comparison. Although it is unlikely
that the tertiary amine on the side chain of 17 (ZnL6) would
form a Schiff base with acetone, 17 gave 1 in moderate yield
but with low ee (entry 12). On the other hand, 18, with no
amino group, scarcely yielded the aldol adduct (entry 13).
We also conducted the aldol reaction between acetone and
10576
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10570 – 10584

Chiral Aldol Reaction Catalysts
FULL PAPER
2-chlorobenzaldehyde in the presence of 19 (ZnL8) and 20
(ZnL9), containing a chiral pipecolinic acid unit and a nipe-
cotic acid unit, respectively (entries 14 and 15). Interestingly,
those Zn2+ complexes gave 1 in lower chemical and optical
yields than 15. In addition, Zn2+-cyclen 9 (ZnL1) alone, in
which the Zn2+ is coordinated by four nitrogen atoms, and a
combination of l-proline and 9 (ZnL1) showed negligible
catalytic activity (entry 16 and 17). These results suggest
that primary or secondary amino groups on the side chains
are important and that the amino acid portions of 14
(ZnL3), 15 (ZnL4), and 16 (ZnL5) function as bases for the
deprotonation of acetone activated by Zn2+, rather than as
Schiff-base-forming units.[35]
ty in favor of 33 was improved (entries 4–6). When a HA/
CH3CN solvent system was employed, 16 (ZnL5) gave the
aldol product syn-selectively (syn/anti 1.7:1) in 91% yield
and with 45% ee [(3S,4R), entry 7].
Table 8 shows the results for the aldol reaction between
DHA and 4-nitrobenzaldehyde to give the triol aldol
adduct, which was immediately acetylated to afford 34 in
order to determine its stereoselectivity. As listed in entries 1
and 4, 15 (ZnL4) and 16 (ZnL5) poorly facilitated the aldol
reaction in THF/H2O (H2O is required for the solvation of
both substrates). In the DMSO/H2O solvent system, which
gave a homogeneous reaction mixture, 15 (ZnL4) and 16
(ZnL5) afforded 34 in 22% and 27% yields, respectively,
but with poor enantioselectivities (entries 2 and 5). When
the NMP/H2O solvent system[14a] was used, syn selectivity
Table 6 summarizes the results of aldol reactions between
acetone and other aldehydes in the presence of 15 (ZnL4)
and 16 (ZnL5) as catalysts.
When 4-nitrobenzaldehyde was
Table 6. The results of asymmetric aldol reactions between acetone and various aldehydes in the presence of
15 (ZnL4) and 16 (ZnL5).
used as an acceptor in the pres-
ence of 15 or 16 (5 mol%) in
acetone/H2O 4:1 at 378C, 30
was obtained in 63% ee (R)
and 86% ee (R), respectively
(entries 1 and 4). At 258C, 16
gave 30 with 89% ee (R,
entry 5). Although the reactivi-
ties of 4-bromobenzaldehyde
and 2-naphthylaldehyde were
somewhat lower than that of 4-
nitrobenzaldehyde (entries 2, 3,
6 and 7), 31 and 32 were ob-
tained with good enantioselec-
tivity in the presence of 16 (en-
tries 6 and 7). In all cases, R en-
antiomers were produced (en-
tries 1–7).
We also studied aldol reac-
tions of unprotected a-hydroxy-
ketones such as hydroxyacetone
(HA) and dihydroxyacetone
(DHA) (Tables 7 and 8). This
approach would provide effec-
tive methods for the synthesis
of 1,2-diol derivatives and car-
bohydrates. The reaction be-
tween HA and 4-nitrobenzalde-
Entry Catalyst [mm][a] Mol%
R
Product Conditions Yield [%][b] CTN[c] ee [%][d]
1
2
3
4
5
6
7
15 (ZnL4) (50)
5
4-NO2C6H4
4-BrC6H4
C10H7 (2-naphthyl) 32
4-NO2C6H4
4-NO2C6H4
4-BrC6H4
30
31
20 h, 378C
20 h, 378C
120 h, 378C 59
20 h, 378C
20 h, 258C
20 h, 378C
78
70
16
7
6
17
13
5
63 (R)
52 (R)
57 (R)
86 (R)
89 (R)
75 (R)
83 (R)
15 (ZnL4) (50) 10
15 (ZnL4) (50) 10
16 (ZnL5) (50)
16 (ZnL5) (25)
5
5
30
30
31
86
65
49
16 (ZnL5) (50) 10
16 (ZnL5) (50) 10
C10H7 (2-naphthyl) 32
120 h, 378C 55
6
[a] Numbers in parentheses are the concentrations of catalysts in solvent; ZnL complexes were formed in situ.
[b] Isolated yield. [c] Catalytic Turnover Number (=yield/equivalents of catalyst). [d] Determined by HPLC
analysis with a chiral column (reference [32]).
Table 7. The results of an asymmetric aldol reaction between HA and 4-nitrobenzaldehyde in the presence of
15 (ZnL4) or 16 (ZnL5).
Entry Catalyst[a]
Solvent
Conditions Yield[b] CTN[c] dr[d]
ee [%][e]
ACHTUNGTRENNUNG[syn/anti]
[mm]
[%]
N
1
2
15 (ZnL4) (25) THF/HA/H2O 5:4:1
15 (ZnL4) (50) DMSO/HA/H2O 21:1.1:1 20 h, 378C
24 h, 378C
83
80
17
16
1:3
1:1.1
racemic
4 (3S,4R)/
racemic
15 (ZnL4) (25) HA/CH3CN 1:1
16 (ZnL5) (25) THF/HA/H2O 2:1:1
16 (ZnL5) (25) HA/H2O 3:1
24 h, 258C
20 h, 378C
20 h, 378C
24 h, 258C
24 h, 258C
90
89
68
96
91
18
18
14
19
18
1:3
8 (3S,4R)/
1 (3R,4R)
5 (3S,4R)/
7 (3S,4S)
14 (3S,4R)/
9 (3S,4S)
2 (3R,4S)/
18 (3S,4S)
45 (3S,4R)/
8 (3R,4R)
hyde
(in
THF/HA/H2O,
3
4
5
6
7
DMSO/HA/H2O, and HA/
CH3CN) in the presence of 15
(ZnL4) gave 33 in good yields,
but with poor enantioselectivi-
ties (entries 1–3 in Table 7). In
these reactions, the anti stereo-
isomer was obtained as the
major product. On the other
hand, in the presence of 16
(ZnL5), containing a primary
1:1.8
1:1.4
1:1.7
1.7:1
16 (ZnL5) (25) HA/NMP[f] 1:1
16 (ZnL5) (25) HA/CH3CN 1:1
[a] Numbers in parentheses are concentrations of catalysts in the solvent; ZnL complexes were formed in situ.
[b] Isolated yield. [c] Catalytic Turnover Number (=yield/equivalents of catalyst). [d] Determined by H NMR
spectroscopy and HPLC analysis. [e] Determined by HPLC analysis with a chiral column (reference [32]).
1
amine unit, the enantioselectivi- [f] NMP=N-methylpyrrolidone.
Chem. Eur. J. 2009, 15, 10570 – 10584
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10577

S. Aoki et al.
Table 8. The results of an asymmetric aldol reaction between DHA[a] and 4-nitrobenzaldehyde in the presence
In order to examine whether
amino acid groups of l-proline,
15 (ZnL4), and 16 (ZnL5) would
form Schiff bases with sub-
strates, we initially carried out a
UV titration of acetone with l-
proline and ZnL4. However, the
UV spectral change was negligi-
ble.
of 15 (ZnL4) or 16 (ZnL5).
Entry Catalyst[b]
Mol% Solvent
Conditions Yield [%][c]
dr[d]
[syn/anti]
ee [%][e]
ACHTUNGTRENUN[NG syn/anti]
[mm]
[for aldol]
[for two steps]
U
We then carried out UV ti-
trations of acetylacetone (acac)
with l-proline and chiral Zn2+
complexes (Scheme 9). Fig-
ure 3a shows the results of the
UV absorption titration of acac
(0.2 mm) with l-proline in
DMSO/H2O 1:2 at 258C, in
which the absorption maxima
of acac shifted from 276 nm to
316 nm with increasing concen-
1
2
15 (ZnL4) (12)
15 (ZnL4) (6)
5
5
THF/H2O 1:1
DMSO/H2O 3:1 72 h, 378C
96 h, 378C
trace
22
–
–
1.7:1
6 (3R,4S)/
racemic
3
4
5
6
15 (ZnL4) (100) 10
NMP/H2O 33:1
THF/H2O 3:5
24 h, 258C
72 h, 378C
4
2.8:1
1.4:1
1.7:1
1.9:1
2 (3S,4R)/
3 (3R,4R)
9 (3R,4S)/
3 (3S,4S)
6 (3R,4S)/
racemic
16 (ZnL5) (17)
16 (ZnL5) (6)
5
5
4
DMSO/H2O 2:1 72 h, 378C
NMP/H2O 33:1 42 h, 258C
27
26
16 (ZnL5) (100) 10
44 (3R,4S)/
8 (3R,4R)
[a] We used DHA monomer (entries 1, 2, 4, and 5) or DHA dimer (entries 3 and 6). The equivalents of DHA
(as monomer) relative to 4-nitrobenzaldehyde used were 3.4 (entry 1), 3 (entries 2 and 5), 5 (entry 4), and
2 equiv (entries 3 and 6). [b] Numbers in parentheses are concentrations of catalysts in the solvent; ZnL com-
plexes were formed in situ. [c] Isolated yield. [d] Determined by 1H NMR spectroscopy and HPLC analysis.
[e] Determined by HPLC analysis with a chiral column (reference [32]).
trations
of
l-proline
(0–
500 equiv). These results sug-
gest the formation of a Schiff
was observed (syn/anti 2.8:1 and 1.9:1, entries 3 and 6) with
moderate enantioselectivities (44% ee) in favor of syn-34
(entry 6). Improvement of the aldol reactions of HA and
DHA is currently in progress.
UV titrations of acetylacetone (acac) with proline, valine,
ZnACHTUNGTRENNUNG
(OTf)2, and Zn2+ complexes [15 (ZnL4), 16 (ZnL5), 18
(ZnL7), and 9 (ZnL1)] in a mechanistic study: As described
in the Introduction, antibodies with aldolase activity were
developed by Lerner et al.[17] It was suggested that the
mechanism of the aldolase antibodies involves enamine for-
mation as the result of a reaction between the e-amino
group of Lys and the ketone substrate. In this scenario, the
Lys residue of the antibody attacks the b-diketone 35
(Scheme 8) to yield an iminium cation 37 via 36. It was
reported that 37 enters into equilibrium with 38a and
38b, which exhibit a strong UV absorption at 316 nm
(e316 ~15000mꢀ1 cmꢀ1).
Scheme 9.
base between l-proline and acac. Upon addition of ZnL4
(0–50 equiv), on the other hand, the absorption maxima of
acac shifted from 276 nm to 294 nm, as shown in Figure 3b.
To assign the strong UV absorption at 294 nm shown in Fig-
ure 3b, we measured the change in the UV spectra of acac
as a function of pH in DMSO/H2O 1:2. The deprotonation
constant of the carbonyl a-proton of acac was determined to
be 8.9 by potentiometric pH titrations in H2O with I=0.1
(NaNO3) at 258C. The dashed curve in Figure 4 represents
the UV absorption spectra of acac at pH 4.1 and 7.3, with a
lmax at 276 nm, which shifted to 293 nm upon addition of
HOꢀ (pH 9.8), possibly corresponding to the monoanion of
acac [(acac)ꢀ; plain curve in Figure 4].
These results suggest that l-proline forms enamine 39
with acac (lmax =316 nm), whereas 15 (ZnL4) induces the
ACTHNUTRGNEUNG
formation of the 1:1 ZnL4(acac)ꢀ complex 40 (lmax =294 nm;
Scheme 9). Analysis of the titration curve in the inset of Fig-
ure 3a with the aid of the software program “BIND
WORKS” gave an enamine formation constant (Kapp) of ca.
~10mꢀ1 for acac with l-proline. In DMSO/H2O 95:5, the for-
Scheme 8.
10578
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10570 – 10584

Chiral Aldol Reaction Catalysts
FULL PAPER
We also carried out UV absorption titrations of acac with
l-valine (0–500 equiv), ZnACHTNUGRTNEUNG
(OTf)2 (0–50 equiv), 16 (ZnL5, 0–
50 equiv), 18 (ZnL7, 0–20 equiv), and 9 (ZnL1, 0–20 equiv),
and analyzed the titration curves, in order to obtain the ap-
parent formation constants of enamine or ZnL·ACTHNUTRGNEUNG
(acac)ꢀ com-
plexes (Kapp). The results are summarized in Table 9. Upon
addition of l-valine, the absorption maxima shifted from
276 nm to 316 nm, similarly to what was observed when l-
proline was mixed with acac. On the other hand, ZnACHTUNGTRENNUNG(OTf)2
showed an absorption maximum of 289 nm, similarly to
what was observed for a mixture of ZnL4 and acac. The
Zn2+ complexes of L5 and L7 (16 and 18) also showed ab-
sorption maxima at 292–295 nm. Interestingly, the change in
the UV spectra of acac upon the addition of 9 (ZnL1), in
which Zn2+ is 4-coordinated to four nitrogen atoms, was
negligible. These results suggest that Zn2+ complexes of tri-
dentate ligands (16 and 18) have a higher activity for the
recognition of (acac)ꢀ than Zn2+ complexes of tetradentate
ligands.[36]
Moreover, it was found that the reaction between ZnL4
and acac had gone to completion within a few minutes,
whereas the enamine formation between l-proline and acac
had reached equilibrium after 4 h even in pure DMSO, as
shown in Figure 5.
Proposed mechanism for aldol reactions of acetone cata-
lyzed by chiral Zn2+ complexes: The results of enantioselec-
tive aldol reactions in the presence of chiral Zn2+ complexes
can be summarized as follows.
Figure 3. a) Change in UV spectra of acac (0.2 mm) upon addition of l-
proline (0–500 equiv) in DMSO/H2O 1:2 at 258C. b) Change in UV spec-
tra of acac (0.2 mm) upon addition of ZnL4 (0–50 equiv) in DMSO/H2O
1:2 at 258C.
1) From potentiometric pH titration data, 14 (ZnL3) was
found to be more stable than 15 (ZnL4) and 16 (ZnL5).
Comparison of the pKa values for the Zn2+-bound water
in these Zn2+ complexes indicated that the Lewis acidity
of the Zn2+ in 15 and 16 is stronger than that in 14. This
is possibly due to the fact that the [15]aneN5 unit in 14
functions as a tetradentate ligand to Zn2+ whereas the
[12]aneN4 (cyclen) groups in 15 and 16 act as tridentate
ligands.
2) The amino groups of the l-proline and l-valine units in
14, 15, and 16 coordinate weakly to Zn2+
.
3) Zn2+ complexes of [15]aneN5 and [12]aneN4 derivatives
that contain l-proline and l-valine units are able to cata-
lyze direct aldol reactions between acetone and alde-
hydes. Better results with respect to chemical and optical
yields of the aldol adducts were obtained in acetone/H2O
Figure 4. UV spectra of acac (0.2 mm) in DMSO/H2O 1:2 at 258C (pH 4.1
and 7.3; dashed curve) and of acac (0.1 mm) in DMSO/H2O 1:2 at 258C
after addition of HOꢀ (pH 9.8; plain curve).
mation constant for the proline-
acac enamine 39 was increased
to 62mꢀ1, indicating that these
equilibria are solvent-depen-
dent. On the other hand, analy-
sis of the titration curve in the
inset of Figure 3b gave a com-
plexation constant for 40 of
2.1ꢁ102 mꢀ1 in DMSO/H2O 1:2.
Table 9. The apparent enamine or Zn2+-enolate formation constants (Kapp) for l-proline, l-valine, Zn
15 (ZnL4), 16 (ZnL5), 18 (ZnL7), and 9 (ZnL1) with acac in DMSO/H2O 1:2 at 258C.
ACHTUNGTRENNUNG(OTf)2,
l-Proline
l-Valine
Zn(OTf)2
15 (ZnL4)
16 (ZnL5)
18 (ZnL7)
9 (ZnL1)
lmax 316 nm
316 nm
A
289 nm
157
294 nm
292 nm
295 nm
–
A
(enamine) (Zn2+-enolate) (Zn2+-enolate) (Zn2+-enolate) (Zn2+-enolate)
)
~10
212
139
589 (49[b]
)
n.i.[c]
[a] The Kapp value in DMSO/H2O 95:5. [b] The Kapp value obtained by potentiometric pH titration in H2O
with I=0.1 (NaNO3) at 258C (pKa of acac=8.9). [c] Negligible interaction with acac.
Chem. Eur. J. 2009, 15, 10570 – 10584
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10579

S. Aoki et al.
8) The results of the UV titrations of acac with l-proline
and l-valine suggested that these organocatalysts form
enamine intermediates, analogously to class I aldolases.[8]
In contrast, Zn2+ complexes of the tridentate ligands 15
and 16 facilitate the formation of the ZnL·ACTHNUTRGNEUNG
(acac)ꢀ com-
plexes.
9) The formation of ZnL·ACHTNUTGRNEUNG
(acac)ꢀ complexes (e.g., 40 in
Scheme 9) is much more favorable thermodynamically
and kinetically than the formation of enamine.
Figure 5. Change in UV spectra of acac (0.2 mm)+ZnL4 (10 equiv) in
DMSO/H2O 1:2 at 258C (dashed curve) and of acac (0.2 mm)+l-proline
(200 equiv) in DMSO upon addition at 258C (plain curve).
From these results, we propose a reaction cycle, as shown in
Scheme 10. We assume that 15 and 16 form complexes 41
with acetone, in which the carbonyl group is activated by
Zn2+ ions and the a-proton of acetone is abstracted by
amine units to generate the ZnL-enolate complex 42. We
assume that this step is much faster than the enamine for-
mation. This Zn2+-enolate reacts with an acceptor aldehyde
via 43 to yield 44, which then decomposes to give the aldol
product and Zn2+ complexes in the presence of H2O.
solvent systems than in acetone/DMSO systems, indicat-
ing the importance of the presence of water molecules.
4) The catalytic activities of the [12]aneN4-type Zn2+ com-
plexes 15 and 16 were much higher than those of the
[15]aneN5-type 14. This can be attributed to the fact that
the Lewis acidities of the Zn2+ ions in 15 and 16 are
higher than that of the Zn2+ ion in 14.
Furthermore, we compared the reactivities of an enamine
intermediate (from acetone and l-proline) and a Zn2+-eno-
5) Metal-free ligands and Zn2+ alone failed to yield aldol
reactions with high optical yields. The Zn2+-cyclen com-
plex 9 (ZnL1) gave negligible quantities of aldol adducts.
In addition, the Cd2+ and Cu2+ complexes of L4 (29a
and 29b) had low catalytic activities. These data indicate
that Zn2+ complexes of tridentate ligands have the high-
est catalytic activities for enantioselective aldol reactions.
6) Of the Zn2+ complexes tested in this work, 15 (ZnL4)
and 16 (ZnL5), containing primary or secondary amino
groups on their side chains, gave aldol adducts in good
chemical and optical yields. The Zn2+ complex 17
(ZnL6), containing a tertiary amine, gave the aldol prod-
uct in moderate chemical yield and with low ee. In addi-
tion, 18 (ZnL7), with no amine group, showed negligible
reactivity. These data indicate the importance of amino
groups on the side chains, in
late [from acetone and 15 (ZnL4)] by H NMR spectroscopy
1
(Scheme 11). We first mixed l-proline or 15 (50 mm) with
[D6]acetone (200 equiv) in D2O at 378C to obtain equilibria,
and then added 2-chlorobenzaldehyde. In the presence of
15, it was found that 2-chlorobenzaldehyde was converted
quantitatively in 24 h at 378C (Figure 6). On the other hand,
the aldol reaction from the enamine intermediate of acetone
and l-proline had not gone to completion even after 50 h
(Figure 7). The initial reaction rate of the aldol reaction be-
tween acetone and 2-chlorobenzaldehyde via 15-enolate is
approximately 10 times higher than that via the enamine in-
termediate (by comparison of Figures 6b and 7b), indicating
that the Zn2+-enolate is more reactive than enamine species.
The enantioselectivity could be explained in terms of the
Zimmerman–Traxler-type transition state 43 (Scheme 10).[37]
that they function for the
abstraction of hydrogens at
the
carbonyl
a-position
rather than as Schiff-base-
forming sites.
7) The fact that Zn2+ com-
plexes containing pipecolinic
acid in the case of 19 (ZnL8)
and nipecotic acid in that of
20 (ZnL9) gave lower chemi-
cal and optical yields than
15 (ZnL4) and 16 (ZnL5)
suggested that the orienta-
tion of the lone pairs of
amino groups on the side
chains with respect to the
Zn2+-coordination site is im-
portant.
Scheme 10. Proposed mechanism for aldol reactions of acetone catalyzed by chiral Zn2+ complexes (15 and
16).
10580
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10570 – 10584

Chiral Aldol Reaction Catalysts
FULL PAPER
Scheme 11.
1
Figure 7. Changes in the H NMR spectrum of a solution of 2-chloroben-
zaldehyde (50 mm) and l-proline (50 mm) in [D6]acetone/D2O 4:1. a) Im-
mediately after addition of aldehyde to a mixture of l-proline and
[D6]acetone, b) after 2 h, c)after 4 h, and d) after 50 h. Reaction tempera-
ture was 378C.
Conclusions
We report on the design and synthesis of asymmetric cata-
lysts consisting of chiral amino acids and Zn2+–cyclen com-
plexes for use in enantioselective aldol reactions, inspired by
two classes of natural aldolases: class I aldolases, which me-
1
Figure 6. Changes in the H NMR spectrum of a solution of 2-chloroben-
zaldehyde (50 mm) and 15 (ZnL4, 50 mm) in [D6]acetone/D2O 4:1. a) Im-
mediately after addition of aldehyde to a mixture of 15 and [D6]acetone,
b) after 2 h, c) after 4 h, and d) after 24 h. Reaction temperature was
378C.
ꢀ
diate C C bond formation through the creation of Schiff
bases between a Lys residue and the substrates, and class II
aldolases, which generate Zn2+-enolates. Analysis of poten-
tiometric pH titrations revealed that 14 (ZnL3) is more
stable than 15 (ZnL4) and 16 (ZnL5) and that, in contrast,
the Lewis acidity of the Zn2+ in 14 is weaker than those of
the Zn2+ components in 15 and 16. We found that 15 and
16, containing primary and secondary amino groups on the
side chains, are efficient catalysts for asymmetric aldol reac-
tions between acetone and aldehydes (up to 89% ee) in ace-
tone/H2O solvent systems. Aldol reactions of unprotected a-
hydroxyketones such as HA and DHA with 4-nitrobenzalde-
hyde were catalyzed by 16 to give adducts with moderate ee
values (up to 45% ee). The results of aldol reactions in the
presence of some analogous Zn2+ complexes suggested that
amino groups on the side chains of 15 and 16 play important
roles in this reaction, possibly through abstraction of hydro-
gen at the carbonyl a-positions rather than through a Schiff
base mechanism. UV titration data for l-proline and ZnL4
with acac in DMSO/H2O 1:2 suggest that ZnL4 induces the
We assume that the Zn2+ in 43 is 5-coordinated by three ni-
trogen atoms of the cyclen, the enolate of the acetone, and
the carbonyl group of the aldehyde, by analogy with the 5-
ACTHNUTRGNEUNG
coordinated Zn2+ in the 15·(acac)ꢀ complex 40 in Scheme 9,
and that the enolate predominantly attacks at the Re face of
the aldehyde. It is also suggested that metal enolates of acac
may serve as a transition-state analogue for aldol reactions
of acetone or its derivatives mediated by metal enolates.
Whereas 15 (ZnL4) afforded the aldol products anti-33,
from HA and 4-nitrobenzaldehyde, and syn-34, from DHA
and 4-nitrobenzaldehyde, with poor enantioselectivities, 16
(ZnL5) gave the syn forms of 33 and 34 with moderate enan-
tioselectivities (Tables 7 and 8). These results suggest that a
primary amine unit on the side chain is required for enantio-
discrimination in aldol reactions between aldehydes and a-
hydroxyketones such as HA and DHA. We assumed that
the enantioselectivity in aldol reactions of DHA is influ-
enced by the two hydroxy groups. The effects of solvent on
the course of the reaction have yet to be examined.
formation of the ZnL4–
(acac)ꢀ complex (Kapp =2.1ꢁ102 mꢀ1),
ACHTUNGTRENNUNG
whereas proline forms a Schiff base with acac with a very
Chem. Eur. J. 2009, 15, 10570 – 10584
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10581

S. Aoki et al.
small equilibrium constant (Kapp ꢃ10mꢀ1), and that the for-
mation of Zn2+-enolate is more favorable than enamine for-
mation both thermodynamically and kinetically. These date
support a scenario in which the amino acid residues and
Zn2+ ions of 15 and 16 function in a cooperative manner to
generate the Zn2+-enolate of acetone, thus permitting effi-
(0.25 mL) in anhydrous CH2Cl2 (5 mL), to which acetic anhydride
(0.15 mL) was added. After stirring overnight, the mixture was diluted
with CH2Cl2 and washed with HCl (1m). The aqueous layer was further
extracted with CH2Cl2. The combined organic layer was washed with sa-
turated NaHCO3 and saturated NaCl, dried over anhydrous Na2SO4, fil-
tered, and concentrated under reduced pressure. The resulting residue
was purified by silica gel column chromatography (hexane/AcOEt 4:1) to
afford the triacetylated product. Optical purities of the aldol products
were determined by HPLC with a chiral HPLC column, as described
below.[32]
4-(2-Chlorophenyl)-4-hydroxybutan-2-one (1):[32] 1H NMR (300 MHz,
CDCl3/TMS): d=2.23 (s, 3H), 2.64–3.01 (m, 2H), 5.51 (m, 1H), 7.18–
7.63 ppm (m, 4H; ArH); HPLC [Daicel Chiralpak AD-H column
(0.46fꢁ25 cm), hexane/EtOH 95:5, flow rate: 1 mLminꢀ1, l=254 nm,
208C]: tR(S)=14.4 min, tR(R)=19.4 min.
ꢀ
cient enantioselective C C bond formation with aldehydes.
Moreover, it has been suggested that the Zn2+-enolate inter-
mediate formed by ZnL and acetone is more reactive than
the enamine species formed from l-proline and acetone.
These results should afford useful information for the
design, synthesis, and mechanistic study of new asymmetric
catalysts for stereoselective organic reactions.
4-Hydroxy-4-(4-nitrophenyl)butan-2-one (30):[32] 1H NMR (300 MHz,
CDCl3/TMS) d=2.25 (s, 3H), 2.86 (m, 2H), 3.70 (d, J=3.7 Hz, 1H), 5.27
(m, 1H), 7.54 (d, J=8.6 Hz, 2H; ArH), 8.20 ppm (d, J=8.6 Hz, 2H;
ArH); HPLC [Daicel Chiralpak AD-H column (0.46fꢁ25 cm), hexane/
EtOH 90:10, flow rate 1.0 mLminꢀ1, l=254 nm, 108C]: tR (S)=33.6 min,
tR (R)=39.9 min.
Experimental Section
General procedure for catalytic aldol reactions between acetone and al-
dehydes (Tables 2–6):
A given chiral ligand (TFA or HCl salt,
4-(4-Bromophenyl)-4-hydroxybutan-2-one (31):[32] 1H NMR (300 MHz,
CDCl3/TMS) d=2.18 (s, 3H), 2.79 (m, 2H), 3.51 (s, 1H), 5.09 (br, 1H),
7.21 (d, J=8.5 Hz, 2H; ArH), 7.45 ppm (d, J=8.5 Hz, 2H; ArH); HPLC
[Daicel Chiralpak AD-H column (0.46fꢁ25 cm), hexane/iPrOH 92.5:7.5,
0.0125 mmol) was extracted from aqueous NaOH (0.2m) with CHCl3.
After the combined organic layers had been dried over anhydrous
Na2SO4, the solution was filtered and concentrated under reduced pres-
sure. The remaining residue was added to a solution of a mixture of Zn-
flow rate 0.3 mLminꢀ1
41.8 min.
, l =254 nm, 88C]: tR(R)=37.6 min, tR(S)=
ACHTUNGTRENNUNG(NO3)2·6H2O (0.0125 mmol) in solvent (DMSO, H2O, or buffer solutions,
0.05 mL) and acetone (0.2 mL), and the mixture was stirred for 10 min.
The substrate aldehyde (0.25 mmol) was added, and the whole reaction
mixture was stirred for 20–120 h at 258C or 378C. The reaction mixture
was diluted with aq. NH4Cl (6%, 2.5 mL) and extracted with ethyl ace-
tate (30 mLꢁ3). The combined organic layer was dried over anhydrous
Na2SO4, filtered, and concentrated under reduced pressure. The resulting
residue was purified by silica gel column chromatography (hexane/
AcOEt) to provide the pure aldol product. The optical purities of the
aldol products were determined by HPLC with a chiral HPLC column,
as described below.[32]
4-Hydroxy-4-(2-naphthyl)butan-2-one (32):[32] 1H NMR (300 MHz,
CDCl3/TMS) d=2.19 (s, 3H), 2.91 (m, 2H), 3.47 (brs, 1H), 5.30 (m, 1H),
7.46 (m, 3H; ArH), 7.81 ppm (m, 4H; ArH); HPLC [Daicel Chiralpak
AD-H column (0.46fꢁ25 cm), hexane/iPrOH 92.5:7.5, flow rate
1.0 mLminꢀ1, l =254 nm, 208C]: tR(R)=16.5 min, tR(S)=21.6 min.
3,4-Dihydroxy-4-(4-nitrophenyl)butan-2-one (33):[32] 1H NMR (300 MHz,
CD3OD/TMS) anti isomer: d=2.15 (s, 3H), 4.21 (d, J=5.7 Hz, 1H), 4.91
(d, J=5.7 Hz, 1H), 7.64 (d, J=7.8 Hz, 2H; ArH), 8.20 ppm (d, J=
7.8 Hz, 2H; ArH); syn isomer: d=2.31 (s, 3H), 4.28 (d, J=2.5 Hz, 1H),
5.24 (d, J=2.5 Hz, 1H), 7.69 (d, J=8.6 Hz, 2H; ArH), 8.21 ppm (d, J=
8.6 Hz, 2H; ArH); HPLC [Daicel Chiralpak AD-H column (0.46fꢁ
25 cm), hexane/iPrOH 94:6, flow rate 1.0 mLminꢀ1, l=275 nm, 208C]: tR-
General procedure for the catalytic asymmetric aldol reaction between
hydroxyacetone (HA) and 4-nitrobenzaldehyde (Table 7): A given chiral
ligand (TFA or HCl salt) (0.0125 mmol) was extracted from aqueous
NaOH (0.2m) with CHCl3. After the combined organic layers had been
dried over anhydrous Na2SO4, the solution was filtered and concentrated
with reduced pressure. The resulting residue was added to a mixture of
A
t
R
tRACHTUGNTRE(UNNNG syn) (3S,4R)=
CTHUNGTRENNUNG
(34):[32]
1H NMR
ZnACHTUNGTRENNUNG(NO3)2·6H2O (0.0125 mmol) and HA (0.0375–0.14 mmol in a solvent),
(300 MHz, CDCl3/TMS) anti isomer: d=2.08 (s, 3H), 2.17 (s, 3H), 2.18
(s, 3H), 4.75 (d, J=17.6 Hz, 1H), 4.93 (d, J=17.6 Hz, 1H), 5.48 (d, J=
5.5 Hz, 1H), 6.20 (d, J=5.5 Hz, 1H), 7.53 (d, J=8.6 Hz, 2H; ArH),
8.22 ppm (d, J=8.6 Hz, 2H; ArH); syn isomer: d=2.08 (s, 3H), 2.17 (s,
3H), 2.18 (s, 3H), 4.68 (d, J=17.5 Hz, 1H), 4.99 (d, J=17.5 Hz, 1H),
5.54 (d, J=3.5 Hz, 1H), 6.34 (d, J=3.5 Hz, 1H), 7.53 (d, J=8.6 Hz, 2H;
ArH), 8.22 ppm (d, J=8.6 Hz, 2H; ArH); HPLC [Daicel Chiralpak AD-
and the mixture was stirred for 10 min. The 4-nitrobenzaldehyde
(0.25 mmol) was added and the whole reaction mixture was stirred for
20–24 h at 258C or 378C. The reaction mixture was diluted with aq.
NH4Cl (6%, 2.5 mL) and extracted with ethyl acetate (30 mLꢁ3), dried
over anhydrous Na2SO4, filtered, and concentrated under reduced pres-
sure. The remaining residue was purified by silica gel column chromatog-
raphy (hexane/AcOEt 2:1) to provide the pure aldol product. The optical
purities of the aldol products were determined by HPLC with a chiral
HPLC column, as described below.[32]
H column (0.46fꢁ25 cm), hexane/EtOH 90:10, flow rate 1.0 mLminꢀ1
l=254 nm, 258C]: tRA(anti) (3S,4S)=26.9 min, tRA
RA(anti) (3R,4R)=42.6 min, tRA(syn) (3S,4R)=49.9 min.
UV spectrophotometric titrations: UV spectra were recorded on
,
C
CHTUNGTNER(NUNG syn) (3R,4S)=35.3 min,
General procedure for the catalytic asymmetric aldol reaction between
dihydroxyacetone (DHA) and 4-nitrobenzaldehyde (Table 8): A given
chiral ligand (TFA or HCl salt) (0.0125 mmol) was extracted from aque-
ous NaOH (0.2m) with CHCl3. After the combined organic layer had
been dried over anhydrous Na2SO4, the solution was filtered and concen-
trated under reduced pressure. The remaining residue was added to a
t
G
CHTUNGTRENNUNG
a
JASCO V-550 UV/VIS spectrophotometer at 25ꢁ0.18C. The data ob-
tained for the UV titrations (increases in e values at a given wavelength)
were analyzed for apparent complexation constants (Kapp) with the aid of
the “BIND WORKS” software program (Calorimetry Sciences Corp).
For the slow equilibration of l-proline and l-valine with acac (0.2 mm),
solutions of acac with these amino acids at several concentrations (0–
500 equiv) were prepared and their UV spectra were measured after stir-
ring at 25ꢁ0.18C for 12 h (batch method).
Supporting Information for this article is available on the WWW under
http://www.chemeurj.org/or from the author [Experimental Section in-
cluding synthesis of ligands, crystallographic study of 18 (ZnL7), and po-
tentiometric pH titrations].
mixture of ZnACHTUNGTRENNUNG(NO3)2·6H2O (0.0125 mmol) and DHA (monomer or
dimer in a solvent), and the mixture was stirred for 10 min. The 4-nitro-
benzaldehyde (0.25 mmol) was added and the whole reaction mixture
was stirred for 24–96 h at 258C or 378C. The reaction mixture was diluted
with aq. NH4Cl (6%, 2.5 mL) and extracted with ethyl acetate (30 mLꢁ
3), dried over anhydrous Na2SO4, filtered, concentrated under reduced
pressure, and dried in vacuo. The resulting residue was purified by silica
gel column chromatography (hexane/AcOEt 1:2) to afford the aldol
product. The aldol product was dissolved in a solution of pyridine
10582
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10570 – 10584

Chiral Aldol Reaction Catalysts
FULL PAPER
K. N. Houk, H. J. Martin, B. List, J. Am. Chem. Soc. 2003, 125,
2475–2479; f) B. List, L. Hoang, H. J. Martin, Proc. Natl. Acad. Sci.
USA 2004, 101, 5839–5842; g) C. Allemann, R. Gordillo, F. R.
Clemente, P. H.-Y. Cheong, K. N. Houk, Acc. Chem. Res. 2004, 37,
558–569; h) A. Bassan, W. Zou, E. Reyes, F. Himo, A. Cꢃrdova,
Angew. Chem. 2005, 117, 7190–7194; Angew. Chem. Int. Ed. 2005,
44, 7028–7032.
Acknowledgements
This work was supported by the grants-in-aid from the Ministry of Edu-
cation, Culture, Sports, Science, and Technology (MEXT) of Japan (Nos.
18390009, and 19659026) and the High-Tech Research Center Project for
Private Universities (matching fund subsidy from MEXT). S.A. is grate-
ful to the grants from the Sumitomo Foundation (Tokyo), the Naito
Foundation (Tokyo), and the Tokuyama Science Foundation (Tokyo).
M.K. is thankful for a grant-in-aid from MEXT of Japan (No. 20750081)
and a Sasakawa Scientific Research Grant from the Japan Science Soci-
ety (Tokyo).
[9] a) S. Saito, M. Nakadai, H. Yamamoto, Synlett 2001, 1245–1248;
b) M. Nakadai, S. Saito, H. Yamamoto, Tetrahedron 2002, 58, 8167–
8177; c) S. Saito, H. Yamamoto, Acc. Chem. Res. 2004, 37, 570–579.
[10] a) T. Kano, J. Takai, O. Tokuda, K. Maruoka, Angew. Chem. 2005,
117, 3115–3117; Angew. Chem. Int. Ed. 2005, 44, 3055–3057; b) T.
Kano, O. Tokuda, J. Takai, K. Maruoka, Chem. Asian J. 2006, 1,
210–216; c) T. Kano, O. Tokuda, K. Maruoka, Tetrahedron Lett.
2006, 47, 7423–7426.
[11] a) N. Mase, Y. Nakai, N. Ohara, H. Yoda, K. Takabe, F. Tanaka,
C. F. Barbas III, J. Am. Chem. Soc. 2006, 128, 734–735; b) N. Mase,
F. Tanaka, C. F. Barbas III, Angew. Chem. 2004, 116, 2474–2477;
Angew. Chem. Int. Ed. 2004, 43, 2420–2423.
[12] a) Y. Hayashi, T. Sumiya, J. Takahashi, H. Gotoh, T. Urushima, M.
Shoji, Angew. Chem. 2006, 118, 972–975; Angew. Chem. Int. Ed.
2006, 45, 958–961; b) Y. Hayashi, S. Aratake, T. Okano, J. Takaha-
shi, T. Sumiya, M. Shoji, Angew. Chem. 2006, 118, 5653–5655;
Angew. Chem. Int. Ed. 2006, 45, 5527–5529; c) S. Aratake, T. Itoh,
T. Okano, T. Usui, M. Shoji, Y. Hayashi, Chem. Commun. 2007,
2524–2526; d) S. Aratake, T. Itoh, T. Okano, N. Nagae, T. Sumiya,
M. Shoji, Y. Hayashi, Chem. Eur. J. 2007, 13, 10246–10256; e) Y.
Hayashi, S. Samanta, T. Itoh, H. Ishikawa, Org. Lett. 2008, 10, 5581–
5583; f) Y. Hayashi, T. Itoh, S. Aratake, H. Ishikawa, Angew. Chem.
2008, 120, 2112–2114; Angew. Chem. Int. Ed. 2008, 47, 2082–2084.
[13] a) P. Dziedzic, W. Zou, J. Hafren, A. Cꢃrdova, Org. Biomol. Chem.
2006, 4, 38–40; b) A. Cꢃrdova, W. Zou, P. Dziedzic, I. Ibrahem, E.
Reyes, Y. Xu, Chem. Eur. J. 2006, 12, 5383–5397.
[14] a) S. S. V. Ramasastry, H. Zhang, F. Tanaka, C. F. Barbas III, J. Am.
Chem. Soc. 2007, 129, 288–289; b) S. S. V. Ramasastry, K. Al-
bertshofer, N. Utsumi, F. Tanaka, C. F. Barbas III, Angew. Chem.
2007, 119, 5668–5671; Angew. Chem. Int. Ed. 2007, 46, 5572–5575;
c) S. S. V. Ramasastry, K. Albertshofer, N. Utsumi, C. F. Barbas III,
Org. Lett. 2008, 10, 1621–1624.
[15] For reviews: a) P. I. Dalko, L. Moisan, Angew. Chem. 2004, 116,
5248–5286; Angew. Chem. Int. Ed. 2004, 43, 5138–5175; b) D.
Enders, M. Voith, A. Lenzen, Angew. Chem. 2005, 117, 1330–1351;
Angew. Chem. Int. Ed. 2005, 44, 1304–1325; c) J. Mlynarski, Eur. J.
Org. Chem. 2006, 4779–4786; d) S. Mukherjee, J. W. Yang, S. Hoff-
mann, B. List, Chem. Rev. 2007, 107, 5471–5569; e) H. Kotsuki, H.
Ikishima, A. Okuyama, Heterocycles 2008, 75, 493–529; f) H. Kotsu-
ki, H. Ikishima, A. Okuyama, Heterocycles 2008, 75, 757–797; g) D.
Enders, A. A. Narine, J. Org. Chem. 2008, 73, 7857–7870; h) J. Mly-
narski, J. Paradowska, Chem. Soc. Rev. 2008, 37, 1502–1511; i) M.
Gruttadauria, F. Giacalone, R. Noto, Adv. Synth. Catal. 2009, 351,
33–57; j) Asymmetric Synthesis with Chemical and Biological Meth-
ods (Eds.: D. Enders, K.-E. Jaeger), Wiley-VCH, Weinheim, 2007.
[16] Selected examples of direct asymmetric aldol reactions: a) N. Yoshi-
kawa, M. Shibasaki, Tetrahedron 2001, 57, 2569–2579; b) T. J. Dick-
erson, K. D. Janda, J. Am. Chem. Soc. 2002, 124, 3220–3221;
c) A. B. Northrup, D. W. C. MacMillan, J. Am. Chem. Soc. 2002, 124,
6798–6799; d) A. Cꢃrdova, W. Notz, C. F. Barbas III, Chem.
Commun. 2002, 3024–3025; e) V. Gnanadesikan, Y. Horiuchi, T.
Ohshima, M. Shibasaki, J. Am. Chem. Soc. 2004, 126, 7782–7783;
f) Z. Tang, Z.-H. Yang, L.-F. Cun, L.-Z. Gong, A.-Q. Mi, Y.-Z.
Jiang, Org. Lett. 2004, 6, 2285–2287; g) D. Enders, C. Grondal,
Angew. Chem. 2005, 117, 1235–1238; Angew. Chem. Int. Ed. 2005,
44, 1210–1212; h) S. Samanta, J. Liu, R. Dodda, C.-G. Zhao, Org.
Lett. 2005, 7, 5321–5323; i) Y. Horiuchi, V. Gnanadesikan, T. Ohshi-
ma, H. Masu, K. Katagiri, Y. Sei, K. Yamaguchi, M. Shibasaki,
Chem. Eur. J. 2005, 11, 5195–5204; j) J. Mlynarski, J. Jankowska, B.
Rakiel, Chem. Commun. 2005, 4854–4856; k) Z. Jiang, Z. Liang, X.
Wu, Y. Lu, Chem. Commun. 2006, 2801–2803; l) P. M. Pihko, K. M.
Laurikainen, A. Usano, A. I. Nyberg, J. A. Kaavi, Tetrahedron 2006,
[1] a) Modern Aldol Reactions (Ed.: R. Mahrwald), Wiley-VCH, Wein-
heim, 2004; b) K. Furuta, T. Maruyama, H. Yamamoto, J. Am.
Chem. Soc. 1991, 113, 1041–1042; c) E. R. Parmee, O. Tempkin, S.
Masamune, A. Abiko, J. Am. Chem. Soc. 1991, 113, 9365–9366;
d) E. J. Corey, C. L. Cywin, T. D. Roper, Tetrahedron Lett. 1992, 33,
6907–6910; e) S. Kobayashi, H. Uchiro, I. Shiina, T. Mukaiyama,
Tetrahedron 1993, 49, 1761–1772; f) R. A. Singer, E. M. Carreira, J.
Am. Chem. Soc. 1995, 117, 12360–12361; g) D. A. Evans, J. A.
Murry, M. C. Kozlowski, J. Am. Chem. Soc. 1996, 118, 5814–5815;
h) S. E. Denmark, S. B. D. Winter, X. Su, K.-T. Wong, J. Am. Chem.
Soc. 1996, 118, 7404–7405; i) S. Nagayama, S. Kobayashi, J. Am.
Chem. Soc. 2000, 122, 11531–11532; j) S. Ishikawa, T. Hamada, K.
Manabe, S. Kobayashi, J. Am. Chem. Soc. 2004, 126, 12236–12237.
[2] a) M. Shibasaki, N. Yoshikawa, Chem. Rev. 2002, 102, 2187–2209;
b) M. Shibasaki, S. Matsunaga, Chem. Soc. Rev. 2006, 35, 269–279;
c) Y. M. A. Yamada, N. Yoshikawa, H. Sasai, M. Shibasaki, Angew.
Chem. 1997, 109, 1942–1944; Angew. Chem. Int. Ed. Engl. 1997, 36,
1871–1873; d) N. Yoshikawa, Y. M. A. Yamada, J. Das, H. Sasai, M.
Shibasaki, J. Am. Chem. Soc. 1999, 121, 4168–4178.
[3] a) B. M. Trost, H. Ito, J. Am. Chem. Soc. 2000, 122, 12003–12004;
b) B. M. Trost, H. Ito, E. R. Silcoff, J. Am. Chem. Soc. 2001, 123,
3367–3368.
[4] a) M. Nakagawa, H. Nakao, K.-i. Watanabe, Chem. Lett. 1985, 391–
394; b) N. Kumagai, S. Matsunaga, N. Yoshikawa, T. Ohshima, M.
Shibasaki, Org. Lett. 2001, 3, 1539–1542; c) G. Lalic, A. D. Aloise,
M. D. Shair, J. Am. Chem. Soc. 2003, 125, 2852–2853; d) M. A.
Calter, R. K. Orr, Tetrahedron Lett. 2003, 44, 5699–5701; e) B. M.
Trost, S. Shin, J. A. Sclafani, J. Am. Chem. Soc. 2005, 127, 8602–
8603.
[5] a) C. Walsh, Enzymatic Reaction Mechanisms, Freeman and Compa-
ny, New York, 1979; b) D. Voet, J. G. Voet, Biochemistry, Wiley,
New York, 1990; c) C.-H. Wong, G. M. Whitesides, Enzymes in Syn-
thetic Organic Chemistry, Pergamon, Oxford, 1994; d) J. McMurry,
Organic Chemistry, A Biological Approach, Thomson Bools/Cole,
Belmont, 2007.
[6] a) H. J. M. Gijsen, L. Qiao, W. Fitz, C.-H. Wong, Chem. Rev. 1996,
96, 443–473; b) S. Takayama, G. J. McGarvey, C.-H. Wong, Chem.
Soc. Rev. 1997, 26, 407–415; c) T. D. Machajewski, C.-H. Wong,
Angew. Chem. 2000, 112, 1406–1430; Angew. Chem. Int. Ed. 2000,
39, 1352–1374; d) C.-H. Wong, M. C. Bryan, P. T. Nyffeler, H. Liu,
E. Chapman, Pure Appl. Chem. 2003, 75, 179–186; e) S. M. Dean,
W. A. Greenberg, C.-H. Wong, Adv. Synth. Catal. 2007, 349, 1308–
1320; f) M. Sugiyama, Z. Hong, P.-H. Liang, S. M. Dean, L. J.
Whalen, W. A. Greenberg, C.-H. Wong, J. Am. Chem. Soc. 2007,
129, 14811–14817.
[7] a) B. List, R. A. Lerner, C. F. Barbas III, J. Am. Chem. Soc. 2000,
122, 2395–2396; b) W. Notz, B. List, J. Am. Chem. Soc. 2000, 122,
7386–7387; c) K. Sakthivel, W. Notz, T. Bui, C. F. Barbas III, J. Am.
Chem. Soc. 2001, 123, 5260–5267; d) B. List, Acc. Chem. Res. 2004,
37, 548–557; e) W. Notz, F. Tanaka, C. F. Barbas III, Acc. Chem.
Res. 2004, 37, 580–591.
[8] a) S. Bahmanyar, K. N. Houk, J. Am. Chem. Soc. 2001, 123, 11273–
11283; b) S. Bahmanyar, K. N. Houk, J. Am. Chem. Soc. 2001, 123,
12911–12912; c) K. N. Rankin, J. W. Gauld, R. J. Boyd, J. Phys.
Chem. A 2002, 106, 5155–5159; d) L. Hoang, S. Bahmanyar, K. N.
Houk, B. List, J. Am. Chem. Soc. 2003, 125, 16–17; e) S. Bahmanyar,
Chem. Eur. J. 2009, 15, 10570 – 10584
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10583

S. Aoki et al.
62, 317–328; m) C. Grondal, D. Enders, Tetrahedron 2006, 62, 329–
337; n) D. Enders, J. Palecek, C. Grondal, Chem. Commun. 2006,
655–657; o) C. Grondal, D. Enders, Synlett 2006, 3507–3509; p) S.
Luo, H. Xu, J. Li, L. Zhang, J.-P. Cheng, J. Am. Chem. Soc. 2007,
129, 3074–3075; q) V. Maya, M. Raj, V. K. Singh, Org. Lett. 2007, 9,
2593–2595; r) S. Luo, H. Xu, L. Zhang, L. Jiuyuan, J.-P. Cheng, Org.
Lett. 2008, 10, 653–656; s) J. Liu, Z. Yang, Z. Wang, F. Wang, X.
Chen, X. Liu, X. Feng, Z. Su, C. Hu, J. Am. Chem. Soc. 2008, 130,
5654–5655; t) K. Nakayama, K. Maruoka, J. Am. Chem. Soc. 2008,
130, 17666–17667.
[27] E. Kimura, S. Aoki, T. Koike, M. Shiro, J. Am. Chem. Soc. 1997,
119, 3068–3076.
[28] A. E. Martell, R. J. Motekaitis, Determination and Use of Stability
Constants, 2nd ed., VCH, Weinheim, 1992.
[29] a) E. Kimura, T. Shiota, T. Koike, M. Shiro, M. Kodama, J. Am.
Chem. Soc. 1990, 112, 5805–5811; b) X. Zhang, R. Eldik, T. Koike,
E. Kimura, Inorg. Chem. 1993, 32, 5749–5755.
[30] S. Aoki, Y. Tomiyama, Y. Kageyama, Y. Yamada, M. Shiro, E.
Kimura, Chem. Asian J. 2009, 4, 561–573.
[31] It was confirmed that L7 serves as a tridentate Zn2+ ligand by X-ray
crystal structure analysis of 18 (ZnL7), as shown in Figure S3 in the
Supproting Information. CCDC 72468 contains the supplementary
crystallographic data for this paper. These data can be obtained free
of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[32] The absolute configuration of the aldol products 1, and 30–34 were
determined by comparison with the following references: a) K. Sak-
thivel, W. Notz, T. Bui, C. F. Barbas III, J. Am. Chem. Soc. 2001,
123, 5260–5267; b) W. Miao, T. H. Chan, Adv. Synth. Catal. 2006,
348, 1711–1718; c) S. S. V. Ramasastry, H. Zhang, F. Tanaka, C. F.
Barbas III, J. Am. Chem. Soc. 2007, 129, 288–289; d) S. S. V. Rama-
sastry, K. Albertshofer, N. Utsumi, F. Tanaka, C. F. Barbas III,
Angew. Chem. 2007, 119, 5668–5671; Angew. Chem. Int. Ed. 2007,
46, 5572–5575.
[17] a) J. Wagner, R. A. Lerner, C. F. Barbas III, Science 1995, 270,
1797–1800; b) C. F. Barbas III, A. Heine, G. Zhong, T. Hoffmann, S.
Gramatikova, R. Bjçrnestedt, B. List, J. Anderson, E. A. Stura, I. A.
Wilson, R. A. Lerner, Science 1997, 278, 2085–2092.
[18] a) C.-H. Ng, H.-K. Fun, S.-B. Teo, S.-G. Teoh, K. Chinnakali, Acta
Crystallogr. Sect. C 1995, 51, 244–245; b) Y. Inomata, T. Takeuchi,
T. Moriwaki, Inorg. Chim. Acta 1983, 68, 187–193.
[19] a) T. Darbre, M. Machuqueiro, Chem. Commun. 2003, 1090–1091;
b) J. Kofoed, M. Machuqueiro, J.-L. Reymond, T. Darbre, Chem.
Commun. 2004, 1540–1541; c) J. Kofoed, J.-L. Reymond, T. Darbre,
Org. Biomol. Chem. 2005, 3, 1850–1855; d) R. Fernandez-Lopez, J.
Kofoed, M. Machuqueiro, T. Darbre, Eur. J. Org. Chem. 2005, 5268–
5276; e) J. Kofoed, T. Darbre, J.-L. Reymond, Chem. Commun.
2006, 1482–1484.
[20] For the reported pKa values and Zn2+ complexation properties of l-
proline, see: a) K. K. Girdhan, K. K. Vaid, R. S. Relan, J. Indian
Chem. Soc. 1970, 47, 715–716; b) Critical Stability Constants, Vol. 1
(Eds.: A. C. Martell, R. M. Smith), Plenum Press, New York, 1974;
c) C. W. Childs, D. D. Perrin, J. Chem. Soc. A 1969, 1039–1044.
[33] The [a]D values of 5 mm 15 (ZnL4) in H2O hardly changed after stir-
ring at 378C for 1 day (+10.0ꢁ2.0 (c=0.23 in H2O)), indicating
negligible racemization at the asymmetric carbon center of 15.
[34] a) S. Aoki, K. Sakurama, N. Matsuo, Y. Yamada, R. Takasawa, S.-i.
Tanuma, M. Shiro, K. Takeda, E. Kimura, Chem. Eur. J. 2006, 12,
9066–9080; b) S. Aoki, K. Sakurama, R. Ohshima, N. Matsuo, Y.
Yamada, R. Takasawa, S.-i. Tanuma, K. Takeda, E. Kimura, Inorg.
Chem. 2008, 47, 2747–2754.
[35] Note that retro aldol reaction of 1 to acetone and 2-chlorobenzalde-
hyde was hardly observed in the presence of 15 (ZnL4) (data not
shown).
[36] The aldol reaction of acac and 2-chlorobenzaldehyde in DMSO/H2O
4:1 at room temperature for 20 h catalyzed by 16 (ZnL5) afforded
45 in 15% yield.
[21] They performed potentiometric pH titrations of 10 mm Zn2+
+
20 mm l-proline against 1m NaOH and they assumed that 7 decom-
poses between pH 6–8 to form Schiff-base species with acetone, al-
though distribution diagram for this system was not reported.
[22] a) J. Paradowska, M. Stodulski, J. Mlynarski, Adv. Synth. Catal.
2007, 349, 1041–1046; b) J. Jankowska, J. Mlynarski, J. Org. Chem.
2006, 71, 1317–1321; c) J. Jankowska, J. Paradowska, B. Rakiel, J.
Mlynarski, J. Org. Chem. 2007, 72, 2228–2231; d) J. Paradowska, M.
´
Rogozinska, J. Mlynarski, Tetrahedron Lett. 2009, 50, 1639–1641.
[23] a) S. Aoki, E. Kimura in Comprehensive Coordination Chemistry II,
Vol. 8 (Eds.: L. Que, Jr., W. B. Tolman), Elsevier, Amsterdam, 2004,
pp. 601–640; b) E. Kimura, T. Shiota, T. Koike, M. Shiro, M.
Kodama, J. Am. Chem. Soc. 1990, 112, 5805–5811; c) T. Koike, E.
Kimura, I. Nakamura, Y. Hashimoto, M. Shiro, J. Am. Chem. Soc.
1992, 114, 7338–7345; d) E. Kimura, T. Koike, M. Shionoya, M.
Shiro, Chem. Lett. 1992, 787–790; e) T. Koike, S. Kajitani, I. Naka-
mura, E. Kimura, M. Shiro, J. Am. Chem. Soc. 1995, 117, 1210–
1219; f) E. Kimura, Y. Kodama, T. Koike, M. Shiro, J. Am. Chem.
Soc. 1995, 117, 8304–8311.
[24] E. Kimura, T. Gotoh, T. Koike, M. Shiro, J. Am. Chem. Soc. 1999,
121, 1267–1274.
[25] a) C.-J. Li T.-H. Chan, Organic Reactions in Aqueous Media, Wiley,
New York, 1997; b) Organic Synthesis in Water (Eds.: P. A. Grieco),
Blacky Academic and Professional, London, 1998.
[37] H. E. Zimmerman, M. D. Traxler, J. Am. Chem. Soc. 1957, 79, 1920–
1923.
[26] a) T. J. Atkins, J. E. Richman, W. F. Oettle, Org. Synth. 1978, 58, 86–
97; b) Z. Kovacs, E. A. Archer, M. K. Russell, A. D. Sherry, Synth.
Commun. 1999, 29, 2817–2822.
Received: March 21, 2009
Published online: September 11, 2009
10584
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10570 – 10584