Enzyme-like catalysis via ternary complex mechanism: Alkoxy-bridged dinuclear cobalt complex mediates chemoselective O-esterification over N-amidation

Article abstract of DOI:10.1021/ja400367h

Hydroxy group-selective acylation in the presence of more nucleophilic amines was achieved using acetates of first-row late transition metals, such as Mn, Fe, Co, Cu, and Zn. Among them, cobalt(II) acetate was the best catalyst in terms of reactivity and selectivity. The combination of an octanuclear cobalt carboxylate cluster [Co₄(OCOR)₆O]₂ (2a: R = CF₃, 2b: R = CH₃, 2c: R = ^tBu) with nitrogen-containing ligands, such as 2,2′-bipyridine, provided an efficient catalytic system for transesterification, in which an alkoxide-bridged dinuclear complex, Co₂(OCO^tBu)₂(bpy) ₂(μ₂-OCH₂-C₆H₄-4- CH₃)₂ (10), was successfully isolated as a key intermediate. Kinetic studies and density functional theory calculations revealed Michaelis-Menten behavior of the complex 10 through an ordered ternary complex mechanism similar to dinuclear metallo-enzymes, suggesting the formation of alkoxides followed by coordination of the ester.

Full text of DOI:10.1021/ja400367h

Article
pubs.acs.org/JACS
Enzyme-Like Catalysis via Ternary Complex Mechanism: Alkoxy-
Bridged Dinuclear Cobalt Complex Mediates Chemoselective
O‑Esterification over N‑Amidation
Yukiko Hayashi,† Stefano Santoro,‡ Yuki Azuma,† Fahmi Himo,‡ Takashi Ohshima,*,§
and Kazushi Mashima*,†
†Department of Chemistry, Graduate School of Engineering Science, Osaka University and CREST, JST, Toyonaka, Osaka 560-8531,
Japan
‡Department of Organic Chemistry, Arrhenius Laboratory, Stockholm University, SE-10691 Stockholm, Sweden
§Graduate School of Pharmaceutical Sciences, Kyusyu University and CREST, JST, Fukuoka 812-8582, Japan
S
* Supporting Information
ABSTRACT: Hydroxy group-selective acylation in the
presence of more nucleophilic amines was achieved using
acetates of first-row late transition metals, such as Mn, Fe, Co,
Cu, and Zn. Among them, cobalt(II) acetate was the best
catalyst in terms of reactivity and selectivity. The combination
of an octanuclear cobalt carboxylate cluster [Co4(OCOR)6O]2
t
(2a: R = CF3, 2b: R = CH3, 2c: R = Bu) with nitrogen-
containing ligands, such as 2,2′-bipyridine, provided an
efficient catalytic system for transesterification, in which an
alkoxide-bridged dinuclear complex, Co2(OCOtBu)2-
(bpy)2(μ2-OCH2-C6H4-4-CH3)2 (10), was successfully isolated as a key intermediate. Kinetic studies and density functional
theory calculations revealed Michaelis−Menten behavior of the complex 10 through an ordered ternary complex mechanism
similar to dinuclear metallo-enzymes, suggesting the formation of alkoxides followed by coordination of the ester.
Scheme 1
INTRODUCTION
■
Ester and amide bonds are ubiquitous chemical bonds
abundant in natural and synthetic organic compounds.1
A
common synthetic protocol is the acylation of alcohols and
amines with carboxylic acid, acid chloride, or acid anhydride.
These processes usually require stoichiometric amounts of a
condensation reagent or base, resulting in the formation of
more than stoichiometric amounts of chemical waste. Catalytic
transesterification2,3 and ester−amide exchange reactions4 are
desirable for the synthesis of diverse esters from the perspective
of atom economy and environmental concerns. Because amino
groups have higher nucleophilicity than hydroxy groups, the use
of these reagents for competitive reactions with hydroxy and
amino functional groups selectively produces the amide. Inverse
chemoselectivity was first reported for lipase-catalyzed O-
acylation of hydroxy groups in serine by trifluoroethyl acetate in
the presence of primary alkyl amino groups.5 Recently, we
found that a trifluoroacetate-bridged μ4-oxo-tetranuclear zinc
cluster, Zn4(OCOCF3)6O (1), served as an efficient catalyst for
chemoselective transesterification of methyl esters with
alcohols,6−8 even in the presence of primary or secondary
amino groups, to afford the corresponding esters as the first
artificial catalyst system for O-selective transesterification over
N-amidation (Scheme 1),7 although O-selective acetylation by
activated enol esters as acylating reagents was reportedly
catalyzed by Y5(OiPr)13O.9 Soon after our report, Ishihara and
co-workers reported that a catalyst mixture of La(OiPr)3 and 2-
(2-methoxyethoxy)ethanol, which might generate dinuclear
La3+ species, facilitated the transesterification of methyl esters
with an equimolar amount of alcohol, resulting in selective
acylation of the hydroxy group, even in the presence of amino
groups.10
Here, we report that a wide variety of first-row transition-
metal carboxylates serve as O-selective catalysts, and a new
catalyst system based on Co(II)-carboxylate clusters that
Received: January 19, 2013
Published: March 27, 2013
© 2013 American Chemical Society
6192
dx.doi.org/10.1021/ja400367h | J. Am. Chem. Soc. 2013, 135, 6192−6199

Journal of the American Chemical Society
Article
Figure 1. Catalytic activity and chemoselectivity of (a) various acetates and (b) cobalt carboxylate compounds for acylation of n-hexanol (4a) and n-
hexylamine (5a) with methyl 3-phenylpropionate (3a). Yields were determined by GC analysis of the crude reaction mixture and based on the mean
of duplicate runs.
functioned as O-selective acylation catalysts. Further, we reveal
the ternary complex mechanism of the alkoxy-bridged dinuclear
cobalt complex obeying Michaelis−Menten kinetics and report
density functional theory (DFT) calculations of the reaction
mechanism.
served as an O-selective catalyst, while Ni(OCOCH3)2 had very
low catalytic activity. Notably, in the absence of n-hexylamine,
transesterification was comparably very slow or compressed to
almost no reaction, clearly indicating that n-hexylamine
improved the catalytic activity of carboxylate compounds of
first-row transition metals, such as Mn(II), Fe(II), Co(II),
Cu(II), Cu(I), and Zn(II), by maintaining their unique nature
of O-selective acylation, although we previously reported that
the additive effects of amines and N-heteroaromatics
significantly enhance the catalytic activity of zinc cluster 1 for
transesterification.8
RESULTS AND DISCUSSION
■
Catalyst screening of various carboxylates for chemoselective
acylation of alcohols over amines was initiated by a test reaction
of methyl 3-phenylpropanoate (3a) with a 1:1 mixture of n-
hexanol (4a, 1.2 equiv) and n-hexylamine (5a, 1.2 equiv) in the
presence of various metal acetates, Mn(OCOCH3)m (5 mol %
metal), in refluxed diisopropyl ether for 15 h, and the results are
shown in Figure 1. The acetate salts of group 1 elements (Na,
K) and group 2 elements (Mg, Ca) had almost no catalytic
activity. For early and middle transition metals in an oxidation
state of three, such as Sc(III), Ti(III), V(III), Cr(III), and
Mn(III), amidation prevailed over esterification, despite the low
catalytic activities of these acetates. To our surprise, however,
most acetate complexes of middle and late first-row transition
metals in an oxidation state of two showed moderate catalytic
activity and sufficient chemoselectivity, producing O-acylated
product 6aa over N-acylated product 7aa. Among them,
Mn(OCOCH3)2, Fe(OCOCH3)2, Co(OCOCH3)2, [Cu-
(OCOCH3)2]2, and Zn(OCOCH3)2 had relatively high
catalytic activity. The Cu(I) compound Cu(OCOCH3) also
We selected Co(II) as the best metal to study the
mechanistic details of O-selective acylation due not only to
its high catalytic activity and chemoselectivity but also to its
potential advantages based on UV−vis spectroscopic character-
ization, although Cu(I) and Mn(II) also showed interesting
reactivity.11 In addition, a pivalate-bridged octanuclear cluster,
Co8(OCOtBu)12O2,12 has the dimeric structure of a μ-oxo
tetranuclear unit analogous to zinc clusters, Zn4(OCOR)6O,
which are superior catalysts compared with the corresponding
simple salts Zn(OCOR)2·xH2O.6 Since the introduction of a
more electron-deficient ligand, such as trifluoroacetate, to the
tetranuclear zinc cluster entity further improved catalytic
activity, we next examined the catalytic performance of some
carboxylate salts of cobalt, formulated as Co(OCOR)2 (R =
t
CF3, CH3, and Bu) as well as the corresponding carboxylate-
6193
dx.doi.org/10.1021/ja400367h | J. Am. Chem. Soc. 2013, 135, 6192−6199

Journal of the American Chemical Society
Article
bridged clusters, Co8(OCOR)12O2 (2a: R = CF3; 2b: R = CH3;
2c: R = Bu). We prepared carboxylate-bridged μ-oxo cobalt
We then turned our attention to verifying the reactivity of
the in situ-generated Co(II) species supported by 8a (Co/8a =
1/1) toward esters and alcohols. The addition of 8.0 equiv
(Co/3b = 1/1) of methyl benzoate (3b) to the mixture of 2c
and 8a in toluene did not change the spectrum, indicating that
3b could not interact with any cobalt centers (Scheme 2). In
t
clusters 2ac according to previously reported proce-
dures12−15 and confirmed that the cluster 2c existed in
equilibrium with its half segment, Co4(OCOtBu)6O, in a
solution state (see Supporting Information). The effects of
nuclearity and carboxylate ligands on cobalt atoms were
investigated, and the results are summarized at the bottom of
Figure 1. Tuning the electronic character of the carboxylate
ligand, i.e., acetate, pivalate, and trifluoroacetate, led to an
increase in the electron-withdrawing ability, thereby improving
the catalytic activity based on the increased yields (80% of 6aa
along with only 4% of 7aa) upon the use of Co(OCOCF3)2,
compared with those of Co(OCOCH3)2 (6aa: 72%, 7aa: 7%)
and Co(OCOtBu)2 (6aa: 67%, 7aa: 7%); as expected, the
pivalate complex had almost the same catalytic performance as
the acetate complex. The most notable finding was that clusters
2ac had better catalytic activities than the corresponding
carboxylate salts. Among the cobalt clusters 2ac tested for O-
selective acylation, trifluoroacetate cluster 2a was selected as the
best catalyst because it had the highest catalytic activity (86% of
6aa). As a result of ligand screening, we selected 2,2′-bipyridine
(8a) as the best chelating nitrogen ligand for the purpose of
isolating and detecting any cobalt species (see Table S1). The
catalytic activity of zinc cluster 1 was also improved by the
addition of 8a.
Scheme 2
To gain mechanistic insight into the cobalt cluster-catalyzed
transesterification, we performed control experiments with 2c,
the most soluble cobalt cluster compared with 2a and 2b, which
are hardly soluble in toluene, to elucidate the solution behavior
of cobalt clusters by UV−vis spectroscopy and to isolate any
key intermediate species. The addition of 8a to 2c in toluene
dramatically altered its absorption spectrum, as shown in Figure
2. The addition of 8a (up to 20 equiv.) increased absorbance at
sharp contrast, the addition of 4-methylbenzyl alcohol (4c)
(Co/4c = 1/1) led to a clear change in the UV−vis spectrum: a
new dinuclear complex 10 in 31% yield and a known trinuclear
complex 1117 in 24% yield were isolated from the reaction
mixture (yields based on Co), where 11 was a product of the
protonation of the μ-oxo moiety of 2c by the alcohol 4c and 10
was a product comprising the alkoxide fate of 4c. Alternatively,
complex 10 was prepared in 76% yield by treating a 1:2 mixture
of Co(OCOtBu)2 and 8a with 1 equiv of the potassium salt of
4c. Complex 10 was crystallographically characterized (Figure
3). Both cobalt atoms adopt a pseudo octahedral geometry
supported by two oxygen atoms of a chelating carboxylate, two
nitrogen atoms of 2,2'-bipyridine, and two oxygen atoms of the
Figure 2. Electronic absorption spectra of [Co4(OCOtBu)6O]2 (2c) in
toluene solutions of 8a at 25 °C. [Co2+] = 1.46 × 10−3 M. The
loadings of 8a referring to curves 18 are 0, 1, 2, 4, 8, 16, 50, and 100
mol ratio to Co2+ ion, respectively.
around 375 nm and decreased absorbance in the 500−700 nm
wavelengths with an isosbestic point appearing at 511 nm with
up to 16 equiv of 8a. Such a decrease in the band intensities
was attributed to the change in the coordination number at the
cobalt center from 4 to a higher number, probably 6. The
addition of excess amounts of 8a to the solution of 2c induced a
deviation from the isosbestic point, suggesting the irreversible
formation of a mononuclear species, cis-Co(OCOtBu)2(bpy)2
(9),16 which was isolated and characterized by X-ray analysis.
Figure 3. Crystal structure of complex 10. All H-atoms and the solvent
molecule (toluene) were omitted for clarity.
6194
dx.doi.org/10.1021/ja400367h | J. Am. Chem. Soc. 2013, 135, 6192−6199

Journal of the American Chemical Society
Article
two bridging alkoxides. The C2 axis passes through the center of
the ‘Co2(μ-O)2’ core, whose planarity is indicated by the sum
(359.80°) of its internal angles.
Table 2. Substrate Scope for the Complex 10-Catalyzed
Transesterification
a
The catalytic activity of the alkoxide-bridged cobalt dinuclear
complex 10 and the trinuclear complex 11 along with the
mononuclear complex 9 are exemplified by the transester-
ification of 3b with 4b (Table 1). Complexes 10 and 11 had
Table 1. Transesterification Catalyzed by Isolated Cobalt
a
Pivalate Complexes
yield of 6bb
b
entry
catalyst
(%)
1
2
cis-Co(OCOtBu)2(bpy)2
9
59
c
Co2(μ2-OCOtBu)2(bpy)2(μ2-OCH2-(4-Me-
10
95 (4)
C6H4))2
d
3
56
97
4
Co3(μ2-OCOtBu)5(μ3-OH)(bpy)2
11
a
b
2.0 mmol scale. Determined by GC analysis. Average of two runs.
c
d
Yield of 4-methylbenzyl benzoate (6bc). Catalyst loading: 0.05 mol
% (TON = 1120, TOF = 224 h−1).
almost the same catalytic activities (99% for 10; 97% for 11),
but 10 was the most efficient catalyst for this reaction, as 10 had
high catalytic activity of TON = 1120 and TOF = 224 h−1 upon
reducing the catalyst loading of 10 to 0.05 mol %. A notable
aspect of complex 10 was its catalytic performance for O-
selective acylation of the competitive reaction of 3a with 4a and
5a, in which only 4% of amide 7aa was detected for an excellent
total yield (94%) of two esters, 6aa (90%) and 6ac (4%). When
the mononuclear complex 9 was used as the catalyst, the
product 6bb was obtained in a moderate yield (59% yield). The
lower catalytic activity of mononuclear complexes for trans-
esterification has been noted for Zn(OCOCF3)2(DMAP)2.8
We next investigated the substrate scope of esters 3 and
alcohols 4 for the 10-catalyzed transesterification, and the
results are summarized in Table 2. Acylations of aminoalcohols
with primary and secondary amino groups were smoothly
promoted with excellent chemoselectivity (entries 1 and 2).
The present method was successfully applied to aromatic esters
bearing electron-withdrawing and -donating groups at the para-
position (entries 3 and 4). The transesterification of highly
congested methyl 1-adamantanecarboxylate successfully pro-
ceeded with the catalyst 10 (entry 5). Aliphatic ester with acid-
sensitive THP ether group was converted to the corresponding
ester in good yield without the loss of protecting group (entries
6). The secondary alcohol 4g was also applicable to this
reaction (entry 7).
a
b
2.0 mmol scale. Isolated yield. Amino esters were isolated after Boc
c
d
protection. Diisopropyl ether was used as a solvent. Not detected.
e
f
Yield of hydroxy amide. Yield of the diacylated product at both of O-
g
h
and N-positions. Reaction time: 72 h. Reaction time: 40 h.
[Co2(OCOtBu)2(bpy)2(OPh)]+ by ESI-MS measurement (see
Figure S2). Thus, during the catalytic transformation, the
alkoxide ligands bridging the two cobalt atoms in 10 were
replaced by phenoxides of the phenyl ester 12 to form
phenoxide-bridged species 13-A or 13-B in Scheme 3.
Scheme 3
To elucidate the origin and mechanism of chemoselective
transesterification, we investigated the stoichiometric reaction
and performed kinetic studies of the reaction of phenyl
benzoate (12) with 4c using 10 as the catalyst. We used 12 as a
substrate because the once-liberated phenol could not react
with the resulting 4-methylbenzyl benzoate (6bc), thus
preventing the reverse reaction. We first conducted the
stoichiometric reaction of 10 with the phenyl ester 12 and
the alcohol 4c in toluene, where transesterification of 12 with
4c proceeded irreversibly to give PhCOOCH2-C6H4-4-CH3
(6bc) and PhOH. In the controlled reaction of 10 with 2 equiv
of 12 in toluene, we detected a new peak due to
Under pseudo first-order conditions of excess amounts of 12
(9.87 × 10−2 M) and 4c (1.31 × 10−1 M), the initial velocities
of the transesterification catalyzed by various amounts of 10
1
(1.12−6.69 × 10−4 M) were estimated based on H NMR
measurements by minimizing the paramagnetic anisotropy due
to Co(II) in complex 10. Consistent with the Michaelis−
Menten models of enzymatic reaction mechanisms, we
6195
dx.doi.org/10.1021/ja400367h | J. Am. Chem. Soc. 2013, 135, 6192−6199

Journal of the American Chemical Society
Article
collected the initial velocity data at 60 °C to produce the 1/[S]
vs 1/v plots shown in Figure 4a,b. Both plots of 1/[12] against
of phenol as 1.59 × 10−3 (M), a smaller value than that of Km,4c
.
The small value of Ki indicates that the phenol strongly bound
to the Co atom of the catalyst 10, consistent with very small
exchange of the phenoxide ligand in 13 with alcohol 4c and the
phenolic substrates remaining intact in any catalytic reaction.
Based on the above-described experimental results as well as
DFT calculations (see below), a schematic of the possible
catalytic cycle for the transesterification of 12 and 4c using 10 is
shown in Figure 5. The first step is the coordination of ester 12
to one of the cobalt atoms of complex 10, resulting in the
formation of a ternary complex 14-A with η1-coordination of
the pivalate moiety. Intramolecular nucleophilic attack of the
benzyloxy moiety to the coordinated ester 12 proceeds through
transient species 15-A and 15-B, leading to another ternary
complex 14-B coordinated by a transesterificated product 6bc.
Liberation of 6bc from 14-B proceeds smoothly to generate the
half-phenoxide-bridged dinuclear complex 13-A. Under excess
amounts of 4c, intramolecular exchange of the phenoxy ligand
in 13-A by 4c regenerates 10 along with phenol. For standard
transesterification of esters and primary alcohols other than
phenol, these steps are all assumed to be reversible. Reversible
alkoxy ligand exchange is possible for the labile Co(II) ion,
consistent with the findings that Co(III) compounds in argon
atmosphere conditions and Co(II) compounds under air
conditions had very low catalytic activity due to the inertness
of a closed-shell d6 Co(III) species for any ligand exchange
reaction.
To gain more insight into the mechanism of the reaction, we
performed DFT calculations using the B3LYP-D functional20
(see Supporting Information for details). We began our
computational investigation by studying the binding of phenyl
benzoate 12 to complex 10. The adopted computational
scheme allowed us to locate a ternary complex (14-A). Ternary
complex 14-A was slightly higher in energy (+3.2 kcal/mol)
than the separated 10 and 12, mainly due to a loss of entropy in
the binding. As discussed above, the kinetic measurements
indicated that this ternary complex should exist, but the results
of the calculations were not consistent with the kinetic
measurements. From 14-A, the transition state (TS1, Figure
6) for the nucleophilic attack of the benzyloxy group on the
ester was 17.6 kcal/mol higher than the ternary complex (20.8
kcal/mol higher than the isolated 10 and 12). This TS led to
the formation of tetrahedral intermediate 15-A, which, after
reorganization, gave tetrahedral intermediate 15-B in which the
phenoxy group coordinates to one of the cobalt atoms. Note
that at TS1 one of the acetate ligands became a bidentate
bridging ligand between the two metal ions. Also, there is of
course a transition state for the reorganization step, but it is
expected to be considerably lower in energy compared to the
preceding and following TSs and will thus not influence the
mechanism. From 15-B, dissociation of the phenoxy group can
occur through TS2 (20.5 kcal/mol higher than that of 10 + 12,
see Figure 6), leading to the formation of intermediate 14-B in
which the final product is weakly bound to the complex. To
restart the cycle, it is necessary to exchange the phenoxy group
with a benzyloxy group and release product 6bc. This process is
exergonic by 5.3 kcal/mol and leads to the formation of
complex 10.
Figure 4. 1/[S] vs 1/v plots of initial velocity data of the complex 10-
catalyzed transesterification of 12 with 4c in toluene at 60 °C.
Conditions: (a) [10] = 6.03 × 10−4 (M), [12] = 1.24 ( ), 1.85 ( ),
□
●
−1
2.47 ( ), 3.09 ( ), 9.40 ( ) (× 10 M); (b) [10] = 6.03 × 10−4
◆
○
▲
□
◆
●
▲
○
(M), [4c] = 0.63 ( ), 1.25 ( ), 1.87 ( ), 3.12 ( ), 7.57 ( ) (×
10−1 M); (c) [10] = 4.65 × 10−4 (M), [12] = 8.63 × 10−1 (M),
−2
◆
●
▲
[PhOH] = 0.0 ( ), 1.83 ( ), 3.66 ( ) (× 10 M).
1/v and that of 1/[4c] against 1/v produced a series of straight
lines intersecting at a single point to the left of the ordinate,
consistent with a ternary complex mechanism. Because complex
10 already contains deprotonated 4c as a bridging alkoxide
ligand, a so-called catalyst−substrate complex just like an
enzyme−substrate complex, the sequence of substrate binding
and product release was definitely determined in the order of
the first alcohol 4c, followed by the coordination of 12 to form
a ternary complex as a key intermediate. Michaelis−Menten
constants were estimated by a Lineweaver−Burk plot18 as Km,4c
= 4.51 × 10−2 (M) and Km,12 = 1.09 × 10−1 (M).
Ultimately, we conducted an inhibition experiment with
phenol in the transesterification of a fixed concentration of the
phenyl ester 12 and variable concentrations of alcohol 4c, and
the obtained initial velocity data are plotted in Figure 4c. The
observed alteration of the Km,4c with no change in Vmax is
characteristic of the competitive inhibition between 4c and
phenol, and a Dixon plot19 afforded the inhibition constant (Ki)
Before initiating the catalytic cycle, the formation of an
alkoxide species such as 10 is a key step, because cobalt
precursors have no alkoxide ligands. In this context, the
superiority of the cobalt cluster 2c over the mononuclear
compound 9 as well as the simple salt of cobalt pivalate,
6196
dx.doi.org/10.1021/ja400367h | J. Am. Chem. Soc. 2013, 135, 6192−6199

Journal of the American Chemical Society
Article
Figure 5. Possible reaction mechanism for transesterification of phenyl ester 12 with alcohol 4c catalyzed by cobalt cluster 2c, calculated using the
B3LYP-D method. Energies are in kcal/mol.
alcohol was first incorporated into the metal center through the
formation of μ-alkoxy ligation that was stabilized by two cobalt
centers similar to the active site of metallo-enzymes, such as
aminopeptidase and phosphoesterase.23 On the other hand, no
amide complex was detected when amines were added to the
cobalt clusters, whereas amines coordinated to metals to
generate the corresponding cobalt mononuclear complexes. In
other words, the cobalt cluster system is able to deprotonate
alcohols; however, amines are not deprotonated and instead
coordinate to the metal center because the pKa value of amine
is naturally higher than that of the corresponding alcohol. The
chemoselectivity of the catalyst 10 remained high even when 10
was treated with amine before the addition of ester and alcohol;
thus, the possible formation of cobalt-amide species could be
excluded (see Supporting Information). Such a clear difference
between alcohols and amines in the deprotonation step for
generating the corresponding alkoxides and amides plays an
important role in determining the hydroxy group selectivity in
transesterification in the presence of amines. Based on the fact
that early transition-metal carboxylates favorably assisted the
acylation of the amino group, it is likely that the chemo-
selectivity depends on the Lewis acidity of the metal centers.
Thus, cobalt ions of the cobalt cluster behave as a Lewis acid
and its μ-oxo moiety acts as a Brønsted base to deprotonate
alcohols.
Figure 6. Optimized transition states TS1 and TS2 based on DFT
calculations. Distances are in Å.
Co(OCOtBu)2, combined with 8a, is due to the basic μ-oxo
moiety that deprotonated 4c to form the catalytically active
alkoxide species 10. In fact, the function of basic oxo species
has been noted in tin catalyst systems: Otera et al. reported an
efficient catalyst system of distannoxane,21 which has a μ-oxo
moiety and therefore might facilitate the exchange of its original
ligands to alkoxide, and Houghton and Mulvaney proposed
such a process in the stannoxane-catalyzed urethane syn-
thesis.22
Kinetic studies revealed that complex 10 mediated trans-
esterification through the ternary complex mechanism, where
6197
dx.doi.org/10.1021/ja400367h | J. Am. Chem. Soc. 2013, 135, 6192−6199

Journal of the American Chemical Society
Article
(b) Chen, C.-T.; Kuo, J.-H.; Ku, C.-H.; Weng, S.-S.; Liu, C.-Y. J. Org.
Chem. 2005, 70, 1328. (c) Ishihara, K.; Niwa, M.; Kosugi, Y. Org. Lett.
2008, 10, 2187. (d) Kondaiah, G. C. M.; Reddy, L. A.; Babu, K. S.;
Gurav, V. M.; Huge, K. G.; Bandichhor, R.; Reddy, P. P.; Bhattacharya,
A.; Anand, R. V. Tetrahedron Lett. 2008, 49, 106. (e) Magens, S.;
Ertelt, M.; Jatsch, A.; Plietker, B. Org. Lett. 2008, 10, 53. (f) Magens,
S.; Plietker, B. J. Org. Chem. 2010, 75, 3715. (g) Weng, S.-S.; Ke, C.-S.;
Chen, F.-K.; Lyu, Y.-F.; Lin, G.-Y. Tetrahedron 2011, 67, 1640.
(4) For selected examples, see: (a) Kuroki, Y.; Ishihara, K.; Hanaki,
N.; Ohara, S.; Yamamoto, H. Bull. Chem. Soc. Jpn. 1998, 71, 1221.
(b) Han, C.; Lee, J. P.; Lobkovsky, E.; Porco, J. A., Jr. J. Am. Chem. Soc.
2005, 127, 10039. (c) Sabot, C.; Kumar, K. A.; Meunier, S.;
Mioskowski, C. Tetrahedron Lett. 2007, 48, 3863. (d) Movassaghi,
CONCLUSION
■
The findings of the present study indicate that some
carboxylate compounds of first-row late transition metals,
such as Mn(II), Fe(II), Co(II), and Cu(I), exhibit unique O-
selectivity with appropriately high catalytic activity in the
transesterification of alcohols in the presence of primary or
secondary amines. In such catalytic systems, the addition of
amines was essential to produce high catalytic activity, and
among them, cobalt was the best metal in terms of not only
catalytic activity and chemoselectivity but also its applicability
for mechanistic studies, leading us to select the system of
octanuclear cobalt clusters 2c and 8a. Transesterification
catalyzed by the combined system of 2c and 8a proceeded
through the formation of the key intermediate alkoxide-bridged
cobalt dinuclear complex 10, which performed enzyme-like
catalysis following the Michaelis−Menten mechanism via a
ternary complex formation. We also confirmed that the
advantages of cluster complexes as catalyst precursors are the
deprotonation ability of the basic μ-oxo moieties in the cobalt
cluster and stabilization of the alkoxy ligand by bridged
dinuclear complexation. Deprotonation of nucleophiles was
thus the most important step not only for achieving high
catalytic activity but also for determining the chemoselectivity,
resulting in the chemical differentiation of alcohols and amines.
́
M.; Schmidt, M. A. Org. Lett. 2005, 7, 2453. (e) Price, K. E.; Larrivee-
Aboussafy, C.; Lillie, B. M.; McLaughlin, R. W.; Mustakis, J.;
Hettenbach, K. W.; Hawkins, J. M.; Vaidyanathan, R. Org. Lett.
2009, 11, 2003. (f) Yang, X.; Birman, V. B. Org. Lett. 2009, 11, 1499.
(g) Ohshima, T.; Hayashi, Y.; Agura, K.; Fujii, Y.; Yoshiyama, A.;
Mashima, K. Chem. Commun. 2012, 48, 5434.
(5) Gardossi, L.; Bianchi, D.; Klibanov, A. M. J. Am. Chem. Soc. 1991,
113, 6328.
(6) (a) Iwasaki, T.; Maegawa, Y.; Hayashi, Y.; Ohshima, T.; Mashima,
K. J. Org. Chem. 2008, 73, 5147. (b) Iwasaki, T.; Maegawa, Y.; Hayashi,
Y.; Ohshima, T.; Mashima, K. Synlett 2009, 10, 1659. (c) Iwasaki, T.;
Agura, K.; Maegawa, Y.; Hayashi, Y.; Ohshima, T.; Mashima, K.
Chem.Eur. J. 2010, 16, 11567. (d) Hayashi, Y.; Ohshima, T.; Fujii,
Y.; Matsushima, Y.; Mashima, K. Catal. Sci. Technol. 2011, 1, 230.
(e) Maegawa, Y.; Agura, K.; Hayashi, Y.; Ohshima, T.; Mashima, K.
Synlett 2012, 23, 137.
ASSOCIATED CONTENT
* Supporting Information
Characterization data for all new compounds, synthetic
procedures, detailed experimental data, crystallographic data
(CIF) for 9−11, and Computational details and Cartesian
coordinates of all stationary points. These materials are
available free of charge via the Internet at http://pubs.acs.org.
■
S
(7) Ohshima, T.; Iwasaki, T.; Maegawa, Y.; Yoshiyama, A.; Mashima,
K. J. Am. Chem. Soc. 2008, 130, 2944.
(8) Maegawa, Y.; Ohshima, T.; Hayashi, Y.; Agura, K.; Iwasaki, T.;
Mashima, K. ACS Catal. 2011, 1, 1178.
(9) Lin, M. -H.; RajanBabu, T. V. Org. Lett. 2000, 2, 997.
(10) Hatano, M.; Furuya, Y.; Shimmura, T.; Moriyama, K.; Kamiya,
S.; Maki, T.; Ishihara, K. Org. Lett. 2011, 13, 426.
(11) Further studies of Cu(I) and Mn(II) systems are in progress.
(12) Sidorov, A. A.; Fomina, I. G.; Ponina, M. O.; Aleksandrov, G.
G.; Nefedov, S. E.; Eremenko, I. L.; Moiseev, I. I. Russ. Chem. Bull., Int.
Ed. 2000, 49, 958.
AUTHOR INFORMATION
Corresponding Author
mashima@chem.es.osaka-u.ac.jp; ohshima@phar.kyushu-u.ac.jp
■
(13) Blake, A. B. Chem. Commun. (London) 1966, 569.
(14) Fomina, I. G.; Aleksandrov, G. G.; Dobrokhotova, Zh. V.;
Proshenkina, O. Yu.; Kiskin, M. A.; Velikodnyi, Yu. A.; Ikorskii, V. N.;
Novotortsev, V. M.; Eremenko, I. L. Russ. Chem. Bull., Int. Ed 2006, 55,
1909.
(15) Dobrokhotova, Zh. V.; Fomina, I. G.; Kiskin, M. A.; Sidorov, A.
A.; Novotortsev, V. M.; Eremenko, I. L. Russ. Chem. Bull., Int. Ed.
2006, 55, 256.
(16) Mikhailova, T. B.; Malkov, A. E.; Sidorov, A. A.; Fomina, I. G.;
Aleksandrov, G. G.; Golovaneva, I. F.; Demyanovich, V. M.;
Novotortsev, V. M.; Ikorskii, Eremenko, I. L. Russ. J. Inorg. Chem.
2002, 47, 1829.
(17) Aleksandrov, G. G.; Fomina, I. G.; Sidorov, A. A.; Mikhailova, T.
B.; Zhilov, V. I.; Ikorskii, V. N.; Novotortsev, V. M.; Eremenko, I. L.;
Moiseev, I. I. Russ. Chem. Bull., Int. Ed. 2004, 53, 1200.
(18) Lineweaver, H.; Burk, D. J. Am. Chem. Soc. 1934, 56, 658.
(19) Dixon, M. Biochem. J. 1953, 55, 170.
(20) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.;
Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (c) Grimme, S. J.
Comput. Chem. 2006, 27, 1787.
(21) (a) Orita, A.; Mitsutome, A.; Otera, J. J. Org. Chem. 1998, 63,
2420. (b) Otera, J.; Danoh, N.; Nozaki, H. J. Org. Chem. 1991, 56,
5307. (c) Otera, J.; Yano, T.; Kawabata, A.; Nozaki, H. Tetrahedron
Lett. 1986, 27, 2383. (d) Otera, J. Acc. Chem. Res. 2004, 37, 288.
(e) Xiang, J.; Orita, A.; Otera, J. Adv. Synth. Catal. 2002, 344, 84.
(f) Xiang, J.; Toyoshima, S.; Orita, A.; Otera, J. Angew. Chem., Int. Ed.
2001, 40, 3670.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the Core Research for Evolutional
Science and Technology (CREST) program of the Japan
Science and Technology Agency (JST), Japan. Y.H. is the
grateful recipient of a scholarship from JSPS (DC1; 2010
2012). S.S and F. H. acknowledge financial support from the
Swedish Research Council and the Goran Gustafsson and Knut
̈
and Alice Wallenberg Foundations. Computer time was
generously provided by the Swedish National Infrastructure
for Computing.
REFERENCES
■
(1) (a) Larock, R. In Comprehensive Organic Transformations, 2nd ed.;
Wiley-VCH: New York, 1999. (b) Mulzer, J. In Comprehensive Organic
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: New York,
1992; Vol 6. (c) Otera, J. In Esterification; Wiley-VCH: Weinheim,
2003.
(2) For reviews, see: (a) Otera, J. Chem. Rev. 1993, 93, 1449.
(b) Hoydonckx, H. E.; De Vos, D. E.; Chavan, S. A.; Jacobs, P. A. Top.
Catal 2004, 27, 83. (c) Grasa, G. A.; Singh, R.; Nolan, S. P. Synthesis
2004, 971. (d) Enders, D.; Niemeier, O.; Henseler, A. Chem. Rev.
2007, 107, 5606.
(3) For selected recent examples, see: (a) Ramalinga, K.;
(22) Houghton, R. P.; Mulvaney, A. W. J. Organomet. Chem. 1996,
Vijayalakshmi, P.; Kaimal, T. N. B. Tetrahedron Lett. 2002, 43, 879.
517, 107.
6198
dx.doi.org/10.1021/ja400367h | J. Am. Chem. Soc. 2013, 135, 6192−6199

Journal of the American Chemical Society
Article
(23) For reviews, see: (a) Cleland, W. W.; Hengge, A. C. Chem. Rev.
2006, 106, 3252. (b) Sreedhara, A.; Cowan, J. A. J. Biol. Inorg. Chem.
2001, 6, 337. (c) Hegg, E. L.; Burstyn, J. N. Coord. Chem. Rev. 1998,
173, 133. (d) Kramer, R. Coord. Chem. Rev. 1999, 182, 243.
̈
(e) Weston, J. Chem. Rev. 2005, 105, 2151. (f) Jedrzejas, M. J.;
Setlow, P. Chem. Rev. 2001, 101, 607. (g) Kimura, E. Curr. Opin. Chem.
Biol. 2000, 4, 207. (h) Kobayashi, M.; Shimizu, S. Eur. J. Biochem.
1999, 261, 1. (i) Vitale, L. Food Technol. Biotechnol. 1999, 37, 29.
(j) Liu, C.; Wang, L. Dalton Trans. 2009, 227. (k) Gahan, L. R.; Smith,
S. J.; Neves, A.; Schenk, G. Eur. J. Inorg. Chem. 2009, 2745. (l) Liu, C.;
Wang, M.; Zhang, T.; Sun, H. Coord. Chem. Rev. 2004, 248, 147.
6199
dx.doi.org/10.1021/ja400367h | J. Am. Chem. Soc. 2013, 135, 6192−6199