Ester hydrolysis by a catalytic cyclodextrin dimer enzyme mimic with a metallobipyridyl linking group

Article abstract of DOI:10.1021/ja963769d

A β-cyclodextrin dimer with a linking bipyridyl group is synthesized as a catalyst precursor, a holoenzyme mimic. It binds both ends of potential substrates into the two different cyclodextrin cavities, holding the substrate ester carbonyl group directly above a metal ion bound to the bipyridyl unit. The result is very effective ester hydrolysis with good turnover catalysis. For example, a Cu(II) complex accelerates the rate of hydrolysis of several nitrophenyl esters by a factor of 10⁴-10⁵, with at least 50 turnovers and no sign of product inhibition. In the best case, with an added nucleophile that also binds to the metal ion, a rate acceleration of 1.45 x 10⁷ over the background reaction rate was observed. Hydrolysis by a catalyst with only one cyclodextrin binding group is significantly slower than in the bidentate binding cases. As expected, the binding of a transition state analogue to these catalysts is stronger with the metal ion present than without. This and kinetic evidence point to a mechanism in which the metal ion plays a bifunctional acid-base role, enforced by the binding geometry that holds the substrate functionality right on top of the catalytic metal ion.

Full text of DOI:10.1021/ja963769d

1676
J. Am. Chem. Soc. 1997, 119, 1676-1681
Ester Hydrolysis by a Catalytic Cyclodextrin Dimer Enzyme
Mimic with a Metallobipyridyl Linking Group
Biliang Zhang and Ronald Breslow*
Contribution from the Department of Chemistry, Columbia UniVersity,
New York, New York 10027
ReceiVed October 29, 1996X
Abstract: A â-cyclodextrin dimer with a linking bipyridyl group is synthesized as a catalyst precursor, a holoenzyme
mimic. It binds both ends of potential substrates into the two different cyclodextrin cavities, holding the substrate
ester carbonyl group directly above a metal ion bound to the bipyridyl unit. The result is very effective ester hydrolysis
with good turnover catalysis. For example, a Cu(II) complex accelerates the rate of hydrolysis of several nitrophenyl
esters by a factor of 104-105, with at least 50 turnovers and no sign of product inhibition. In the best case, with an
added nucleophile that also binds to the metal ion, a rate acceleration of 1.45 × 107 over the background reaction
rate was observed. Hydrolysis by a catalyst with only one cyclodextrin binding group is significantly slower than
in the bidentate binding cases. As expected, the binding of a transition state analogue to these catalysts is stronger
with the metal ion present than without. This and kinetic evidence point to a mechanism in which the metal ion
plays a bifunctional acid-base role, enforced by the binding geometry that holds the substrate functionality right on
top of the catalytic metal ion.
Introduction
Computer and solid models indicated that compound 1 would
fit the requirements, its metal complex being able to place the
metal ion right on top of the ester group of various substrates
of the general form 2. Thus we synthesized 1 and examined
Since metal ions are often very effective catalysts, and many
enzymes take advantage of this, it has been attractive to
synthesize enzyme mimics that combine metal ion catalysis with
substrate binding. In fact, the first compound called an
“artificial enzyme” in the literature was such an example, with
a metal ion bound to a ligand that was attached as an ester to
â-cyclodextrin.1 This catalyst was able to use the metal ion to
catalyze the hydrolysis of substrates that would bind into the
cyclodextrin cavity but would not otherwise bind to the metal
ion. However, the rate accelerations seen were not high in this
rather flexible case.
In true enzymes a substrate is rigidly held in the active site,
by multiple binding interactions with the enzyme. The catalytic
groups of the enzyme are held next to the substrate, so that
little entropy must be expended in approaching the transition
state and little conformational enthalpy. In order to achieve
this in an enzyme mimic, the catalyst-substrate complex must
be reasonably stable, and pre-organized into the reactive
geometry.
the binding and catalytic properties of some of its metal ion
complexes. The complexes indeed proved to be outstanding
catalysts, showing large rate accelerations with high catalytic
turnover and no evidence of inhibition by the product fragments.
Detailed studies indicate the mechanism of the catalytic process.5
Results and Discussion
It seemed to us that such requirements could be achieved if
the catalyst was a cyclodextrin dimer, for double binding of
the substrate, in which the linker in the dimer carried an effective
rigidly held catalytic group. The substrate would have to have
hydrophobic groups at both ends, to bind into the two cyclo-
dextrin groups, with an ester or other potentially reactive group
in the middle so it would end up next to the catalytic group.
Our previous work indicated that various cyclodextrin dimers
could show very strong binding of suitable ditopic substrates
that could occupy both cyclodextrin cavities, and that, once the
substrate was cut in half by the catalytic group, the products
would bind much more weakly.2-4 Thus product inhibition was
not expected to be a problem.
Syntheses. The synthesis of the cyclodextrin bipyridyl dimer
1 is outlined in Scheme 1. The reduction of 2-chloro-5-
nitropyridine by iron and ammonium chloride in a water-
methanol mixed solvent gave 5-amino-2-chloropyridine (3),6
whose amino group was protected with benzaldehyde to form
N-benzylidene-5-amino-2-chloropyridine (4). The coupling of
4 with Ni(II)-Zn7 gave the protected bipyridine 5. The
deprotection of 5 formed the desired product 6. After conver-
sion of 5,5′-diamino-2,2′-bipyridine (6) into the bis-diazonium
compound, and reaction of the bis-diazonium salt of 5,5′-
diamino-2,2′-bipyridine with potassium ethyl xanthate following
(3) Breslow, R.; Chung, S. J. Am. Chem. Soc. 1990, 112, 9659-9660.
(4) Breslow, R.; Halfon, S. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 6916-
6918.
(5) For a preliminary report of some of this work see the following:
Breslow, R.; Zhang, B. J. Am. Chem. Soc. 1992, 114, 5882-5883.
(6) Ramadas, K.; Srinivasan, N. Syn. Commun. 1992, 22, 3189-3195.
(7) Iyoda, M.; Otsuka, H.; Sato, K.; Nisato, N.; Oda, M. Bull. Chem.
Soc. Jpn. 1990, 63, 80-87.
X Abstract published in AdVance ACS Abstracts, February 1, 1997.
(1) Breslow, R.; Overman, L. E. J. Am. Chem. Soc. 1970, 92, 1075-
1077.
(2) Breslow, R.; Greenspoon, N.; Guo, T.; Zarzycki, R. J. Am. Chem.
Soc. 1989, 111, 8296-8297.
S0002-7863(96)03769-9 CCC: $14.00 © 1997 American Chemical Society

Catalytic Cyclodextrin Dimer Enzyme Mimic
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1677
Scheme 1
Scheme 3
Scheme 4
Scheme 2
Ester 21 was synthesized as shown in Scheme 3. 1-Bromo-
adamantane (17) was treated with vinyl bromide; heating the
product with potassium hydroxide yielded 1-adamantylacetylene
(19). Then 19 was converted to 1-adamantylpropiolic acid (20)
by treatment with base, trapping the anion with carbon dioxide,
then acidifying with HCl. Subsequently, carboxylic acid 20 was
coupled with p-nitrophenol, to afford ester 21 in 67% yield.
The synthesis of p-(4-tert-butylphenyl)-o-nitrophenyl 1-ada-
mantylpropiolate (26) is shown in Scheme 4. 4-Phenylphenol
(22) was employed as the starting material, which was nitrated
in the ortho position to give compound 23; subsequent Friedel-
Crafts alkylation furnished p-(4-tert-butylphenyl)-o-nitrophenol
(24) in good yield. 1-Adamantylpropiolic acid (20) was
converted to the acyl chloride 25, which was then reacted with
nitrophenol 24 to afford ester 26.
Katz’s procedure,8 2,2′-bipyridine-5,5′-bis(ethyl xanthate) (7)
was obtained in 35% yield. The bis(ethyl xanthate) 7 was
refluxed with 20% ethanolic potassium hydroxide, then treated
with acetyl chloride to give 2,2′-bipyridyl-5,5′-bisthioacetate 8.
Compound 8 was deprotected by ammonia saturated methanol
solution into dianion 9, which was treated in situ with 6-deoxy-
6-iodo-â-cyclodextrin9 in DMF to afford the cyclodextrin dimer
1. The dimer 1 was purified by reverse-phase chromatography,
eluted with a gradient ranging from water to 40% methanol-
water.
We also synthesized a bipyridyl cyclodextrin monomer 10,
for comparison with dimer 1 in binding and catalytic studies.
The synthesis of monomer 10 is described in Scheme 2. The
key step is the coupling of two pyridine rings by Pd(II) catalyst.
2-Bromopyridine (11) was treated with n-butyllithium and then
with trimethyltin chloride in THF.10 The coupling reaction of
2-trimethylstannylpyridine (12) with 2-chloro-5-nitropyridine
was catalyzed11 by Pd(Ph3P)2Cl2 to give 5-nitro-2,2′-bipyridine
(13) with 76% yield. Compound 13 was reduced by sodium
borohydride with catalysis by 10% palladium on activated
carbon to convert it to 5-amino-2,2′-bipyridine (14). The
following steps to prepare bipyridyl cyclodextrin monomer 10
were by essentially the same method as was described for the
bipyridyl dimer 1. The 5-amino-2,2′-bipyridine (14) was
converted to 2,2′-bipyridine-5-ethyldithiocarbonate (15) through
the diazonium intermediate. The dithiocarbonate 15 was
hydrolyzed and treated with acetyl chloride to give 2,2′-
bipyridyl-5-thioacetate (16), which was deprotected by NH3,
then directly reacted with 6-deoxy-6-iodo-â-cyclodextrin to give
the desired bipyridyl cyclodextrin monomer 10.
Detailed procedures for these syntheses, and data on char-
acterization of the compounds, are given in the supporting
information section.
Binding Energy of Dimer 1 with Substrates and with
Transition State Analogues. We have investigated the binding
of 1 with the guests 27-29 (Scheme 5) with and without Zn(II).
The results for the binding of dimer 1 with guests are listed in
Table 1. Compounds 28 and 29 were selected as transition state
analogues; phosphodiester anions are often used as analogues
of the tetrahedral transition state for ester and amide hydrolysis
in antibodies and protein enzymatic studies. The binding
constants of dimer 1 with carbonate 27 and phosphate 28 in
HEPES buffer at pH 7.0 are very similar in the absence of
Zn(II), 1.43 × 106 M-1 and 1.0 × 106 M-1, respectively. The
binding of dimer 1 and bis(1-adamantylethyl) phosphate (29)
with Zn(II) is 50-fold stronger than without Zn(II), and reaches
a binding constant as high as 109 M-1. A similar binding
enhancement (55-fold) was obtained with phosphate 28 binding
into dimer 1 in the presence of zinc ion, but the constant for
the neutral carbonate 27 binding into dimer 1 is only 5-fold
higher with zinc ion than that without zinc ion.
(8) Katz, L.; Schroeder, W.; Cohen, M. J. Org. Chem. 1954, 19, 710.
(9) Tabushi, I.; Kuroda, Y.; Mochizuki, A. J. Am. Chem. Soc. 1980, 102,
1152-53.
(10) Jutzi, P.; Gilge, U. J. Organomet. Chem. 1983, 246, 163.
(11) Bailey, T. R. Tetrahedron Lett. 1986, 27, 4407.

1678 J. Am. Chem. Soc., Vol. 119, No. 7, 1997
Zhang and Breslow
Table 2. Hydrolysis of p-Nitrophenyl 3-Indolepropionate (31) at
37 °Ca
Scheme 5
c
d
entry
catalyst
M(II) pHb
kobs (s-1
)
krel
1
2
3
4
8.0 1.0 × 10-7
1
3
4
Cu(II) 8.0 3.2 × 10-7
Cu(II) 8.0 3.7 × 10-7
bipyridine
bipyridyl dimer 1 Cu(II) 8.0 (1.04 ( 0.02) × 10-3 10 400
5
6
7
8
9
7.0 3.0 × 10-8
Cu(II) 7.0 2.5 × 10-7
7.0 7.1 × 10-7
1
8
24
25
60
â-cyclodextrin
â-cyclodextrin
bipyridyl dimer 1
Cu(II) 7.0 (7.39 ( 0.09) × 10-7
7.0 (1.81 ( 0.42) × 10-6
10 bipyridyl dimer 1 Cu(II) 7.0 (5.49 ( 0.40) × 10-4 18 300
a Reactions were performed at 37 °C in 10 mM HEPES buffer
solution, in 1.0 × 10-4 M bipyridyl dimer 1, cyclodextrin, and
bipyridine; 5.0 × 10-5 M p-nitrophenyl 3-indolepropionate (31), 5.0
× 10-4 M CuCl2. b The pH was checked at the beginning and the end
of each run to hold to within (0.1 unit. c The rate constants were
obtained by analyzing the data using the Kore program and are averages
of at least two runs. All correlation coefficients were g0.9999. d Rate
constants relative to background (no catalyst).
Table 1. Binding Constants of Guests with Dimer 1 in Watera
Table 3. Hydrolysis of Esters Catalyzed by Bipyridyl
Cyclodextrin Dimer 1 in the Presence of Cu(II) at 37 °Ca
guest
Ka with Zn(II) (M-1
)
Ka (M-1
)
Ka(Zn(II))/Ka
carbonate 27 (7.70 ( 0.10) × 106 (1.43 ( 0.20) × 106
phosphate 28 (5.47 ( 0.17) × 107 (1.00 ( 0.34) × 106
phosphate 29 (1.02 ( 0.05) × 109 (2.05 ( 0.24) /x× 107
a Values determined by competition with DiTNS 30 (see refs 2-5)
in 10 mM HEPES buffer solution (pH 7.0) at ambient temperature (25
°C); average of more than three independent runs.
5
55
50
b
c
d
entry substrate
pH
kobs (s-1
)
kun (s-1
)
krel
1
2
3
4
5
6
7
ester 34 8.0 ( 0.1 9.61 × 10-5 1.42 × 10-5
ester 32 8.0 ( 0.1 1.35 × 10-4 1.50 × 10-7
ester 33 8.0 ( 0.1 1.74 × 10-4 1.00 × 10-7
ester 31 7.0 ( 0.1 5.49 × 10-4 3.00 × 10-8
ester 31 8.0 ( 0.1 1.04 × 10-3 1.00 × 10-7
7
900
1 740
18 300
10 400
ester 21 7.0 ( 0.1 6.76 × 10-3 3.00 × 10-8 225 000
These results suggest that the two hydrophobic ends of the
ditopic guests bind strongly into the two cyclodextrin cavities
of dimer 1, and that for phosphodiesters there is an extra
interaction of the phosphate anion with the metal cation. This
artificial metalloenzyme is thus much like a true enzyme, in
that binding with substrate in the transition state is tighter than
in the ground state. This is the first example of an artificial
metalloenzyme effectively binding a tetrahedral, negatively
charged phosphate transition state analogue for hydrolysis of
an ester, and indicates that the bipyridyl dimer 1 should have
catalytic power for ester hydrolysis with metal ions. However,
the difference of the binding constants for the carbonate 27 and
the transition state analogue 28 with the Zn(II) complex of dimer
1 greatly underestimates the magnitude of the rate accelerations
we observed.
Ester Hydrolysis. The first investigation of ester hydrolysis
employed p-nitrophenyl 3-indolepropionate (31) under various
conditions in 10 mM HEPES buffer (pH 7.0 or pH 8.0) at 37
°C; the results are listed in Table 2. Comparison of the catalyzed
and uncatalyzed (entry 5) reaction rates obtained at pH 7.0
indicated enhancements of 18 300 for dimer 1 with Cu(II) (entry
11) and 60 for dimer 1 without Cu(II) (entry 10). When the
reaction catalyzed by 1 and Cu(II) (entry 4) was conducted at
pH 8.0, the rate enhancement was 10 400. With simple
â-cyclodextrin, the rate enhancement observed for the hydrolysis
of substrate 31 was only 25 (entries 8 and 9) with and without
Cu(II) at pH 7.0. With 2,2′-bipyridine (entry 7), the hydrolysis
rate of ester 31 increased by a factor of only 12 with Cu(II).
The large rate acceleration by our enzyme mimic clearly
comes from the combination of the binding sites (two hydro-
phobic cavities) and the catalytic site (metal ion ligand). The
two hydrophobic cavities bind the substrate and juxtapose the
ester functional group and the catalytic metal ion of the artificial
metalloenzyme. The Cu(II) complex of dimer 1 is a successful
catalyst for the hydrolysis of this ester and others (Vide infra)
under physiological conditions of temperature and pH, in
contrast to many enzyme model studies that have employed
extreme conditions.
ester 21 8.0 ( 0.1 1.20 × 10-2 1.80 × 10-7
66 700
a All solutions were 1.0 × 10-4 M bipy-dimer 1, 2.0 × 10-4
M
CuCl2, and 6.0 × 10-5 M substrates. Reactions were conducted in 10
mM HEPES buffer solution for substrates 21, 31, and 34. Reactions
were carried out in 60% 10 mM HEPES buffer and 40% DMSO
solution for substrates 32 and 33. Rate constants were obtained by
averaging two or more kinetic runs. Standard deviations were less than
8% of the rate constants in the table. b Reactions were monitored to
>95% completion. The rate constants were obtained by analyzing the
data using the Kore program; the correlation coefficients were g0.9999.
c kun were obtained by calculating the initial rate constant without
d
1-Cu(II). krel ) kobs/kun.
We have synthesized some other esters (Scheme 5) to study
their hydrolysis catalyzed by bipyridyl dimer 1 with Cu(II); their
pseudo-first-order rate constants are shown in Table 3. The
hydrolysis of 34 showed only a 7-fold rate enhancement (entry
1); it can bind into only one cyclodextrin unit of dimer 1. We
expected that ester 21, containing an adamantane group, should
bind more strongly to 1 than 31. In fact, the hydrolysis rate of
ester 21 with 1 and CuCl2 at pH 7.0 and 37 °C was 225 000
times (entry 6) faster than the rate of uncatalyzed hydrolysis.
When an excess of substrate 21 was employed, at least 50
turnovers were seen in the ester hydrolysis catalyzed by dimer
1 and CuCl2.
Esters 32 and 33 contain a tert-butylphenyl group which
should bind strongly with dimer 1. However, the solubility of
substrates 32 and 33 is poor in aqueous solution, so the
hydrolyses of esters 32 and 33 were performed in 40% DMSO-
buffer solution. This solvent system decreases the binding
ability of these esters with bipyridine dimer 1 (Vide infra), and
decreases kobs
.
Ester 26 is long enough for binding into two cavities of dimer
1 (Scheme 6). Since the solubility of ester 26 in water is poor,
the hydrolytic experiments of ester 26 were performed in 70%
methanol-aqueous buffer or DMSO-aqueous buffer solvent
systems. The results are listed in Table A (Supporting Informa-
tion) and plotted in Figure A (Supporting Information). The
data show that the rate constant for the Cu(II) complex of dimer

Catalytic Cyclodextrin Dimer Enzyme Mimic
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1679
Scheme 6
Figure 1. Organic solvent effects on the catalytic hydrolysis of ester
26 with bipy-dimer 1 and copper(II) at 30 °C (data from Table 4).
Table 4. Pseudo-First-Order Rate Constants for Catalytic
Hydrolysis of Ester 26 by Dimer 1 with Cu(II) in Different
Solventsa
ET(30)
log kobs (kcal/mol)c
b
entry
solvent
water
kobs (s-1
)
5.10 × 10-2d
1
2
3
4
5
6
7
-1.2924
63.0
55.9
55.5
55.2
51.9
45.0
43.8
ethylene glycol (2.47 ( 0.05) × 10-3 -2.6073
MeOH
formamide
EtOH
DMSO
DMF
(3.40 ( 0.06) × 10-3 -2.4672
(5.75 ( 0.30) × 10-4 -3.2403
(3.03 ( 0.15) × 10-4 -3.5185
(6.78 ( 1.00) × 10-5 -4.1688
(3.10 ( 1.50) × 10-6e -5.5086
a Reactions were run at 30 ( 1 °C in 80% organic solvent and 20%
20 mM Hepes buffer (pH ∼ 7.0). Product formation was monitored
b
by absorbance at 445 nm. kobs was calculated using the Kore program,
with product g95% and the correlation coefficients g0.9999. c ET(30)
d
values from ref 20. kobs was calculated by extrapolating to 0 % volume
e
of methanol from different % volumes of methanol. kobs was calculated
Figure 2. Catalytic hydrolysis of ester 26 by bipy-dimer 1 (0.05 mM)
with Cu(II) in 80%(v/v) DMSO/20 mM HEPES (pH 7.0) at 25 °C.
Best-fit values of Vmax and Km were obtained by fitting the data to the
Michaelis-Menten equation by the KaleidoGraph program. The
continuous line corresponds to initial rate ) Vmax[ester]/(Km + [ester])
with Vmax ) 2.24 × 10-5 M s-1 and Km ) 4.69 × 10-5 M.
from the initial rate using the Passage fitting program.
1 is ca. 100 times higher than that with monomer 10. Since
the solvents used decrease hydrophobic binding, the difference
could be even larger in water solution.
Solvent Effects on Catalysis. We have carried out kinetic
studies in various organic solvents, employing the Cu(II)
complex of dimer 1 as catalyst and 26 as the substrate in 20
mM HEPES (pH 7.0) at 25 °C. Water is the best solvent for
inducing strong association between apolar binding partners.
The kobs of ester 26 hydrolysis catalyzed by 1-Cu(II) decreases
with increasing methanol content (Figure B, Supporting Infor-
mation). Methanol is less polar than water and interferes with
hydrophobic binding of the substrate into the cyclodextrin
cavities.
Apolar complexation is not limited to water or alcohols but
occurs in solvents of all polarity.12,13 However, a dramatic
solvent dependence was observed for the catalytic hydrolysis
of ester 26 by dimer 1-Cu(II). When the solvent was changed
from water, the most polar solvent, to DMF, the least polar
solvent, the kobs of catalytic hydrolysis of ester 26 by 1-Cu(II)
decreased from 5.10 × 10-2 s-1 to 3.10 × 10-6 s-1 (Table 4),
a factor of over 104.
groups, such a large solvent effect reflects primarily the strength
of hydrophobic binding.
Kinetic Characteristics. The dependence of the initial rate
of 1-Cu(II)-catalyzed hydrolysis of the ester substrate 26 on
the substrate concentration for a typical set of data is shown in
Figure 2 as a V vs [S] plot and in Figure C (Supporting
Information) as a 1/V vs 1/[S] plot. In each case the continuous
line corresponds to the best fit. Values of the parameters Vmax
and Km were obtained by fitting the initial rate vs [S] data to
the Michaelis-Menten equation. The value of Km for the
1-Cu(II)-catalyzed hydrolytic reaction is 47 µM, a value
consistent with strong double binding. The reactions were
carried out in 80% DMSO solvent, and in water the binding
affinity would be much higher. The kcat for the ester hydrolysis
catalyzed by 1-Cu(II) is 0.45 s-1. The rate enhancement (kcat/
k
uncat) for ester 26 is 1.45 × 107.
Mechanism. We have studied ester hydrolysis by dimer 1
A strong linear relationship exists between the kobs and the
solvent polarity parameter ET(30) of the various solvents (Figure
1). The results are in excellent agreement with some binding
data in the literature.14 Strong correlation between ET(30) and
binding free energies in related host-guest systems have also
been found for binary aqueous solvent mixtures.14-16 Although
the added solvents may also affect the behavior of the catalytic
with copper(II) at 37 °C, using ester 31 in 10 mM buffer at
various pH’s. The pH vs rate profiles in Figure 3 are for the
hydrolysis of p-nitrophenyl 3-indolepropionate (31) by bipyridyl
dimer 1 with Cu(II) and by buffer only. The difference in the
hydrolysis rates between the reaction catalyzed by dimer 1 and
Cu(II) and the reaction run with only buffer is the greatest in
the pH range of 6.5 to 8.0. With respect to rate, the buffer
(12) Siegel, B.; Breslow, R. J. Am. Chem. Soc. 1975, 97, 6869.
(13) Smithrud, D. B.; Diederich, F. J. Am. Chem. Soc. 1990, 112, 339.
(14) Diederich, F.; Smithrud, D. B.; Sanford, E. M.; Wyman, T. B.;
Ferguson, S. B.; Carcanague, D. R.; Chao, I.; Houk, K. N. Acta Chem.
Scand. 1992, 46, 205-215.
(15) Schneider, H.; Kramer, R.; Simova, S.; Schneider, U. J. Am. Chem.
Soc. 1988, 110, 6442-6448.
(16) Ferguson, S. B.; Seward, E. M.; Sanford, E. M.; Hester, M.; Uyeki,
M.; Diederich, F. Pure Appl. Chem. 1989, 61, 1523-1528.

1680 J. Am. Chem. Soc., Vol. 119, No. 7, 1997
Zhang and Breslow
Table 5. Catalytic Hydrolysis of Esters by Dimer 1 in pH 7.0
Buffer at 37 °Ca
entry
ligandb
metal ion
kobs (s-1) × 103 c
krel
Ester 31
1
2
3
4
5
6
7
8
9
0.000030d
0.30
0.55
0.14
0.27
0.55
0.62
1.77
8.89
1
Ni(II)
Cu(II)
Co(II)
Cu(II)
Tb(III)
Eu(III)
Ni(II)
Zn(II)
10 000
18 000
4 700
PAO 35
PAO 35
PAO 35
PAO 35
PAO 35
PAO 35
0 900
18 000
21 000
59 000
300 000
Ester 21
10
11
12
13
14
0.000030d
1.2
6.8
11.2
51.2
1
40 000
220 000
371 000
1700 000
Zn(II)
Cu(II)
Ni(II)
Zn(II)
Figure 3. pH-rate (log kobs) profile for the hydrolysis of p-nitrophenyl
3-indolepropionate (31) by bipyridyl cyclodextrin dimer 1 (9) and
buffer only (b). Reactions were run at 37 °C in buffer solutions using
10 mM MES for pHs between 5.5 and 6.5, 10 mM HEPES for pHs
between 6.70 and 8.25, 10 mM CHES for pHs between 8.50 and 9.50,
and 10 mM CAPES for pHs between 10.00 and 10.50.
PAO 35
PAO 35
a All solutions were 1.0 × 10-4 M in bipy-dimer 1, 1.0 × 10-3
M
in metal ion and 6.0 × 10-5 M in substrate. Reactions were carried
out in 10 mM HEPES buffer solution (pH 7.0) at 37 °C. b PAO 35 is
2-pyridinecarbaldehyde oxime. c Reactions were monitored to >95%
completion. The rate constants were obtained by analyzing the data
using the Kore program; the correlation coefficient s were g0.9999,
and are the average of two or more kinetic experiments. The errors in
kobs were <8%. d Uncatalyzed rate constants of esters.
Scheme 7
carbaldehyde oxime (PAO, 35) are remarkably active catalysts
for the hydrolysis of 8-acetoxyquinoline-5-sulfonate.18 Thus
we have examined the cleavage of ester 31 with added PAO
35 bound to the dimer 1-metal ion complex (Figure E,
Supporting Information). The cleavage of ester 21 by dimer 1
with various metal ions in the presence of PAO 35 was
examined in 20 mM HEPES buffer (pH 7.0 ( 0.1) at 37.0 (
0.2 °C and followed by monitoring the formation of p-
nitrophenoxide by UV at 400 nm. The results are listed in Table
5 (and displayed in Figure F of Supporting Information). A
rate enhancement of 1.7 × 106-fold over the background reaction
rate was observed for cleavage of ester 21 by 1-Zn(II) with PAO
35 (entry 14). In line with our previous work, we suggest the
mechanism shown in Scheme 8. As can be seen from Table 5,
the relative catalytic activities of metal ions in the presence of
PAO 35 ligand were different from those in the absence of PAO.
The order of catalytic activity with PAO 35 is Zn(II) > Ni(II)
> Eu(III) > Tb(III) > Cu(II) > Co(II). The order of catalytic
activity without PAO 35 is Cu(II) > Ni(II), Zn(II). As Scheme
8 shows, we propose that the Zn(II) becomes pentacoordinate
in the transition state, which is less likely for Cu(II). Without
PAO, four-coordination by the metal ion is enough. In our
previous work,18 we showed that the acylated catalyst that is
formed by the mechanism of Scheme 8, with an acyl group on
the oxygen of PAO, is then hydrolyzed by the metal ion. Thus
turnover catalysis was seen with such systems, but we have
not examined this question in the present case.
hydrolysis of ester 31 shows a linear pH dependence, whereas
the 1-Cu(II)-catalyzed ester hydrolysis shows a different pH
dependence.
The pH profile in Figure D (Supporting Information) is
expanded from Figure 3 at pH 6.5 to 9.0 for the dimer catalyzed
hydrolysis reaction. This pH profile shows a pH-independent
plateau and a pH-dependent range. The experimental data fit
the theoretical curve described by eq 1, where Kf is the binding
constant of 1-Cu(II) with ester 31, Ka is the acid dissociation
constant for the coordinated water molecule, and k1 is the rate
constant of the rate-determining step. The best fitting of the
data gave pKa ) 7.15, k1 ) 2.05 × 10-4 s-1, and Kf ) 7.0 ×
104 M-1
.
Summary. The catalysts that can bind a substrate at both
ends so as to hold a substrate ester group right above a catalytic
functionality are the most effective. They perform ester
hydrolysis with turnover and very high rate accelerations. When
the catalytic group is only a metal ion, Cu(II) is more effective
than Zn(II) or Ni(II). However, when the catalyst also carries
a nucleophilic bound oxime ligand, the best metal ion is Zn(II).
The differences are reasonable in terms of the mechanisms
involved and the coordinating abilities of the different metal
ions.
Ka
kobs ) k1Kf
(1)
Ka + [H+]
We propose the mechanism shown in Scheme 7, in which
the substrate binds to the dimer and is attacked by hydroxide
bound to copper, from a bound water with pKa ) 7.15.17 There
are other kinetically equivalent mechanisms that seem less likely.
Effects of Metal Ions and Ligand. In 1965 we reported
that the zinc or nickel ion complexes of (E)-2-pyridine-
(17) The pKa of the Cu(bipy)(OH2)2 complex is 7.8: Rosch, M. A.,
Trogler, W. C. Inorg. Chem. 1988, 29, 2409-2416.
(18) Breslow, R.; Chipman, D. J. Am. Chem. Soc. 1965, 87, 4195-4196.

Catalytic Cyclodextrin Dimer Enzyme Mimic
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1681
Kinetics. Aqueous buffers were prepared with deionized water and
pH value were adjusted with 1 N NaOH or 1 N HCl solution. Stock
solutions of metal ions, substrates, and catalysts were prepared in the
appropriate buffer solution, except for water insoluble substrates which
were prepared in CH3OH or DMF. The pH was checked at the
beginning and the end of each run to hold to within (0.1 unit.
All reactions for p-nitrophenol product were monitored by following
absorbance change against time at 400 nm. A kinetic run was initiated
by adding 1.00 mL of 100 µM bipy-dimer 1 dissolved in 10 mM
HEPES solution and 10 µL of metal ion stock solution to make the
solution 500 µM in metal ion and brought to 37 ( 0.1 °C in the
spectrophotometer chamber. A 6 µL sample of substrate solution was
injected (to make the solution 60 µM in substrate), shaken vigorously,
and the absorbance at 400 nm monitored as a function of time.
For the catalytic reactions (fast reactions), the pseudo-first order rate
constants were obtained by analyzing the data using the Kore program.19
For the uncatalytic reactions (slow reactions), the rate constants were
obtained by an initial rate treatment. The final product absorbance
was determined by making a solution with the same concentrations of
products and catalyst, and with the same buffer solution used as the
kinetic runs.
Scheme 8
Experimental Section
General. 1H-NMR spectra were recorded on a Varian VXR 200 or
400 MHz spectrometers. 13C-NMR spectra were recorded on a 75 MHz
spectrometer. All spectra were with the residual solvent peaks as
reference signals. Mass spectra were recorded on a Nermag R-1010
instrument (for chemical ionization(CI) or electron impact ionization(EI)
spectra) or a Jeol JMS-DX-303 HF instrument (for FAB spectra). pH
values were measured on an Orion 701A ionalyzer with a glass electrode
after calibration to standard buffer solutions. Solvents, drying agents,
and inorganic salts were obtained from Fisher Scientific Co., Amend
Drug and Chemical Co., Sigma Chemical Co., or Aldrich Chemical
Co. unless otherwise indicated. â-Cyclodextrin was obtained from the
American Maize-Product Co. Diethyl ether (Et2O) was dried by
distillation under argon from K0-Na0 amalgam/benzophenone, tetra-
hydrofuran (THF) was dried by distillation under argon from K0/
benzophenone, and benzene and methylene chloride (CH2Cl2) were
dried by distillation from calcium hydride. Anhydrous N,N-dimethyl-
formamide, N,N-dimethylacetamide, pyridine, and acetonitrile were
obtained in Sure/SealTM bottles from Aldrich Chemical Co. Most
starting reagents were obtained from Aldrich.
Syntheses. The synthetic sequences are outlined in the paper. The
detailed procedures are available in Supporting Information.
Acknowledgment. This paper is dedicated to the memory
of Paul Dowd. This work has been supported by grants from
the National Institutes of Health and the Office of Naval
Research.
Supporting Information Available: Full experimental
details on the syntheses, product characterization, and kinetic
studies, including one table and six figures (18 pages). See
any current masthead page for ordering and Internet access
instructions.
JA963769D
(19) Swain, C. G.; Swain, S. M.; Berg, F. L. J. Chem. Inf. Comput. Sci.
1980, 20, 47.
(20) Reichardt, C. Angew. Chem., Int. Ed. Engl. 1988, 27, 490.