A new enzyme model for enantioselective esterases based on molecularly imprinted polymers

Article abstract of DOI:10.1002/chem.200304783

An efficient enzyme model exhibiting enantioselective esterase activity was prepared by using molecular imprinting techniques. The enantiomerically pure phosphonic monoesters 4L and 5L were synthesized as stable transition-state analogues. They were used

Full text of DOI:10.1002/chem.200304783

FULL PAPER
A New Enzyme Model For Enantioselective Esterases Based On Molecularly
Imprinted Polymers
Marco Emgenbroich and G¸nter Wulff*[a]
Dedicated to Professor Kurt Milow on the occasion of his 80th birthday.
Abstract: An efficient enzyme model
exhibiting enantioselective esterase ac-
tivity was prepared by using molecular
imprinting techniques. The enantiomeri-
cally pure phosphonic monoesters 4l
and 5l were synthesized as stable tran-
sition-state analogues. They were used
as templates connected by stoichiomet-
ric noncovalent interactions to two
equivalents of the amidinium binding
site monomer 1. After polymerization
and removal of the template, the poly-
mers were efficient catalysts for the
hydrolysis of certain nonactivated amino
acid phenylesters (2l, 2d, 3l, 3d)
depending on the template used. Im-
printed catalyst IP4 (imprinted with 4l)
enhanced the hydrolysis of the corre-
sponding substrate 2l by a factor of 325
relative to that of a buffered solution.
Relative to a control polymer containing
the same functionalities, prepared with-
out template 4l, the enhancement was
still about 80-fold, showing the highest
imprinting effect up to now. In cross-
selectivity experiments a strong sub-
strate selectivity of higher than three
was found despite small differences in
the structure of the substrate and tem-
plate. Plots of initial velocities of the
hydrolysis versus substrate concentra-
tion showed typical Michaelis Menten
kinetics with saturation behavior. From
these curves, the Michaelis constant KM
and the catalytic constant kcat can be
calculated. The enantioselectivity shown
in these values is most interesting. The
ratio of the catalytic efficiency kcat/KM,
between the hydrolysis of 2l- and 2d-
substrate with IP4, is 1.65. This enantio-
selectivity derives from both selective
binding of the substrate (KMl/KMd ˆ
0.82), and from selective formation of
the transition state (kcatl/kcatd ˆ 1.36).
Thus, these catalysts give good catalysis
as well as high imprinting and substrate
selectivity. Strong competitive inhibition
is caused by the template used in
imprinting. This behavior is also quite
similar to the behavior of natural en-
zymes, for which these catalysts are good
models.
Keywords: enzyme catalysis
¥
enantioselectivity ¥ enzyme models
¥ molecular imprinting ¥ transition-
state analogue
contains functional monomers that can interact with the
template through covalent or non covalent interactions. After
removal of the template, an imprint containing functional
groups in a certain orientation remains in the polymer. The
shape of the formed imprint, and the arrangement of the
functional groups are complementary to the structure of the
template (for reviews see ref. [3 5]).
For the preparation of imprinted polymeric catalysts, the
functional groups to be introduced should act as binding and
catalytically active sites. Furthermore, it is necessary to create
a shape of the cavity that will support the catalysis, for
example, by stabilizing the transition state of the reaction.
This approach was inspired by the results of Schulz, Lerner,
and others, by generating antibodies against stable transition-
state analogues of a reaction, and obtaining catalytically
active antibodies.[6, 7] In addition to the special shape of the
active site in the antibodies, catalytic antibodies possess,
similar to that in enzymes, special catalytically-active func-
tional groups like guanidine groups, that considerably en-
hance the catalytic reaction. Since most of the research in
antibody chemistry is concerned with ester hydrolysis, these
reactions were the first to be investigated in molecular
Introduction
In order to prepare a catalyst that works similarly to an
enzyme, a cavity that functions as an active site has to be
generated with a shape corresponding to the shape of the
substrate or, even better, to the shape of the transition state of
the reaction. Furthermore, functional groups have to be
introduced into this cavity in an accurate three-dimensional
orientation in order to be effective as binding sites, coenzyme
analogues, or catalytic sites.
Previously, we have introduced a novel method for obtain-
ing such structures in synthetic polymers (™enzyme-analogue
built polymers∫).[1, 2] For this, a highly cross-linked copolymer
is formed around a template molecule. The monomer mixture
[a] Prof. Dr G. Wulff, Dr. M. Emgenbroich
Institute of Organic Chemistry
and Macromolecular Chemistry
Heinrich-Heine-University
D¸sseldorf, Universit‰tsstrasse 1
40225 D¸sseldorf (Germany)
Fax : (‡49)211-811-5840
E-mail: wulffg@uni-duesseldorf.de
4106
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200304783
Chem. Eur. J. 2003, 9, 4106 4117

4106 4117
12]
imprinting.[8
(for a recent review see ref [13]). The main
chelated hydrogen bonds com-
bined with electrostatic interac-
tion. It was therefore possible
to fix a transition-state ana-
logue template during imprint-
ing and, at the same time,
position the amidine groups in
the correct orientation for cat-
alysis. To investigate a similar but non-racemic chiral system,
the substrate should be based on a-amino acid derivatives that
are easily available as both enantiomers. The ester residue
should also contain a nonactivated 3,5-dimethylphenylester to
create a significant imprint, and a 4-carboxybenzoyl group at
the N-terminal part should be introduced.
As a result, the free carboxylic acid can act as a binding
motif during catalysis, and the hydrolysis products, phenol and
dicarboxylic acid, can easily be detected by UV absorption
during HPLC separations. Therefore, the N-(4-carboxyben-
zoyl)-a-amino acid (3,5-dimethylphenyl) esters 2 and 3,
derived from l- and d-leucine and l- and d-valine were
prepared as substrates.
problem to generate an efficient enzyme model by molecular
imprinting is to find a suitable template array that contains the
stable transition-state analogues, the binding sites, and
catalytically-active groups in the desired position.
One possible way to solve this problem is to use non-
covalent interactions between templates and binding-site
monomers with high association constants (so-called stoichio-
metric noncovalent interactions), like the binding of amidine
groups with carboxylic acid, phosphonic acid, or phosphoric
acid groups.[14, 15] In this respect, it is interesting to note that
the most active esterase species among catalytic antibodies
contain a guanidinium group (of the amino acid l-arginine),
which plays an important role in catalyzing the basic
hydrolysis of esters.[16] Similarly, in the investigation of
alkaline hydrolysis of esters, carbonates, and carbamates, we
used amidine groups in polymeric catalysts for binding and
catalysis.[9, 13, 17] We were able to obtain catalysts by molecular
imprinting that showed enhancements in ester hydrolysis,
102 235 times faster relative to the solution reactions. In the
case of carbonates and carbamates, the hydrolysis catalyzed
by an imprinted polymer was 588 and 1400 3860 times faster,
respectively. It could be shown that these catalysts exhibit
Michaelis Menten kinetics similar to enzymes. Further
examples by using amidines in a similar way have recently
been published.[18]
In the next step we now want to form a model for an
enantioselective esterase, and consequently combine the well-
known enantioselectivity in separation with the catalytic
properties of the imprinted polymers.
Previously, Sellergren and Shea presented catalytically-
active imprinted polymers for the hydrolysis of a chiral amino
acid ester.[10, 11] The imprint was obtained by a combination of
covalent and noncovalent interactions. A chiral a-amino
phosphonate was used as the template. In order to mimic the
catalytic triad of the active site in chymotrypsin, imidazole,
phenolic hydroxy, and carboxyl groups were used as catalyti-
cally-active groups. Although the rate enhancements were not
very high (2.5-fold vs control polymer, and 10-fold vs
solution), the enantioselectivity for the hydrolysis was quite
high at kD/kL ˆ 1.85.
The synthesis started from commercially available carbo-
benzoxy protected a-amino acids, which were transformed
into 3,5-dimethylphenyl esters 6 and 7 by coupling in the
presence of (benzotriazol-1-yloxy)tris(dimethylamino)phos-
phonium hexafluorophosphate (BOP)[19] and N-methyl mor-
pholine (see Scheme 1). After deprotection of the N-terminal
group with 33% HBr in acetic acid, the hydrobromides 8 and
9 (as l and d forms) were obtained. Hydrobromides 8 and 9
were coupled in the presence of chlorotripyrrolidinophos-
phonium hexafluorophosphate (PyCloP)[20] with monobenzyl
terephthalic acid (10) to obtain the amides 11 and 12. The
substrates 2 and 3 (in both enantiomeric forms) were obtained
by hydrogenolytic fission of the benzyl esters.
This paper describes the synthesis of enantiomerically pure
template molecules, their corresponding substrates, and the
preparation and catalytic properties of the molecularly-
imprinted polymers.
The synthesis of the optically active templates 4 and 5 (see
Scheme 2) was somewhat more laborious. The two most
common ways to build up a chiral a-amino phosphonic acid
are either by fractionated crystallization of diastereomeric
salts,[21] or synthesis by a diastereoselective reaction.[22] Both
were tested successfully in our laboratory. We chose the
synthesis of chiral a-l-amino phosphonic acid diethyl esters
starting from O-methyl-d-phenylglycinol, as described by
Smith et al.[22] By varying the reaction conditions, we were
able to improve the yield and diastereomeric excess of the
reaction, after column chromatography, for 13 and 14. To
build up our molecules, the N-terminal in 13 and 14 was
Results and Discussion
Synthesis of the substrates and templates: Firstly we inves-
tigated the hydrolysis of an aromatic dicarboxylic monoester
(4-carboxyphenylacetic(3,5-dimethylphenyl) ester), and used
the corresponding phosphonic ester (4-carboxylbenzylphos-
phonic-mono-(3,5-dimethylphenyl) ester) as the stable tran-
sition-state analogue acting as the template. This template was
converted into a bisamidinium salt with N,N'-diethyl(4-vinyl-
phenyl)amidine (1) acting as the binding site monomer.[9] The
amidinium groups are attached to the carboxyl group, as well
as to the phosphonic monoester group of the template, by two
Chem. Eur. J. 2003, 9, 4106 4117
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4107

G. Wulff et al.
FULL PAPER
diesters 18 and 19. The last step was a combined hydrolysis of
the carboxylic methyl ester and diphenyl ester to yield the
corresponding monophenyl ester; this was performed analo-
gously according to reference [24] with sodium hydroxide
activated by [18]crown-6 in dioxane/water 1:1 (v/v). The
products 4 and 5 were purified by column chromatography
under basic conditions, and transformed into the dipotassium
salts by an ion exchange column procedure. This workup
procedure was necessary, since the monoesters were not
stable under acidic conditions.
Synthesis and catalytic properties of imprinted and control
polymers: Earlier investigations showed that the amidinium
monomer 1 forms complexes with carboxylic acids and
phosphonic monoesters in acetonitrile with high association
constants. In less polar solvents, like toluene, the association
constant is even higher.[14, 15] From this data it can be
calculated that in toluene/acetonitrile (1:1), under polymer-
izing conditions to form imprinted polymers, more than 97%
of the acidic groups in 4l and 5l are complexed if a 1:2 molar
mixture of template and 1 is present. This stoichiometric
noncovalent interaction further assures that nearly no un-
bound amidines are present in the polymerizing mixture.
Thus, imprinted polymers are prepared from 4l and 5l as
templates with the addition of two equivalents of 1 (see
Scheme 3). The copolymerization with ethylenedimethacry-
late (EDMA) and methylmethacrylate (MMA) is performed
in the usual way by radical initiation (see Table 1).
Scheme 1. Synthesis of the substrate molecules based on valine and
leucine. All molecules were prepared in both enantiomeric forms, inscribed
with d and l for the corresponding forms.
Scheme 2. Synthesis of the transition-state analogues (template mole-
cules) for valine- and leucine-based substrates (only l forms).
Scheme 3. Model of an imprinted cavity with template 4l, and two
equivalents of the binding site monomer 1.
coupled with monomethyl terephthalic acid (15) to prepare 16
and 17. To modify the P-terminal part of the phosphonic acid
diethyl esters, they were hydrolyzed by bromotrimethylsi-
lane[23] in order to yield the free phosphonic acid, which was
dried, directly chlorinated with oxalylchloride, and esterified
with two equivalents of 3,5-dimethylphenol to obtain the
Control polymers are prepared under identical conditions
with the same amount of the amidine monomer 1, but without
templates 4l and 5l. The polarity and the polymerization
behavior of 1 is different from the complexed monomers, to
4108
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4106 4117

A New Enzyme Model for Enantioselective Esterases
4106 4117
Table 1. Polymer composition and properties[a]
Polymer
Amount [g]
template
Splitting off the template [%]
Inner surface area [m2/g]
complex
by HPLC
by titration
IP4
IP5
CPB
CPF
0.300
0.300
0.160
0.137
4l 0.155
5l 0.155
benzoic acid 0.037
formic acid 0.014
76
92
95
75
82
92
82
217
196
287
276
[a] The composition of the monomer mixture for the preparation of the imprinted polymers, consisted in all examples of EDMA (8.20 g, 7.81 mL), MMA
(1.50 g, 1.59 mL), 1 (0.123 g) or 1-HCl (0.144 g) (IP4 and IP5), AIBN (0.10 g), and as porogene a mixture of toluene (5 mL) and acetonitrile (5 mL). The type
and amount of template varied as shown in the Table. The complex is formed from the template and 1; here the amount is calculated. For the determination of
the percentage of splitting off see the Experimental Section. Inner surface area was measured by BET (N2) one point measurement (particle size 45
125 mm).
overcome this difficulty we used simple carboxylic acids for
complexing, such as formic acid and benzoic acid. To some
extent these acids also behave as templates, especially benzoic
acid; however, they only possess one carboxylic group per
molecule and do not have the required transition-state shape.
Thus, the differences in their performance give good indica-
tions of the imprinting effects.
The polymers are prepared by radical initiation in an
ampoule, and are produced as solid macroporous blocks,
which are then crushed, ground, and sieved to obtain a
particle size of 45 125 mm. The polymers are all macroporous
with high inner-surface areas (196 287 m2 gÀ1), as is typical for
these type of polymers (see Table 1). The template is then split
off by repeated washing with a 1:1 mixture of 0.1m sodium
hydroxide and methanol. From HPLC, it was determined that
75 95% of the templates could be released from the
polymer. Alternatively, the amount of accessible amidine
groups was determined by titration after splitting off the
templates (see Table 1). Both methods gave comparable
results, in the case of IP5 there was some deviation.
reaction rate was in the linear range. The catalytic activity of
IP4 was checked at two different buffer concentrations of
0.1m and 0.15m. The kinetics can be treated as a pseudo-first-
order reaction, since the concentration of hydroxide ions
during the reaction is constant. To follow the reaction, the
phenol product was chosen for detection monitoring, because
it shows no product inhibition and should be released from the
catalyst.
The reaction is described by Equation (1).
RCOOR' ‡ OHÀ À!RCOOÀ ‡ R'OH
(1)
For pseudo-first-order kinetics, Equation (2)[25] can be used.
ꢀ
ꢁ
‰RCOOR'Š
À ‰R'OHŠ
t
t
ˆ0
ln
ˆÀ kt
(2)
‰RCOOR'Š
t
ˆ0
The reactions were repeated 2 3 times, and the given data
consists of average values shown with the standard deviation.
The data for k was obtained by linear regression of at least 6
9
The conditions for the application of imprinted polymer
catalysts with amidine-based binding site monomers have
previously been optimized.[17] The hydrolysis of the esters was
carried out under these optimal conditions in a 1:1 mixture of
data points at different time intervals with correlation
coefficients from 0.980 0.999.
Figure 1a and b show the observed calculated pseudo-first-
order kinetics for the hydrolysis of 2l and 2d under catalysis
of the imprinted polymer IP4, relative to the control experi-
ments. A strong catalytic activity of IP4 for the hydrolysis of
the substrate 2l is observed. The catalysis with the control
polymer CPF and the reaction in solution are much slower.
Similar results are obtained for the second polymer IP5
imprinted with the valine analogue template 5 (see Figure 2a
and b). The values calculated from the slopes in Figures 1 and
2 are given in the Experimental Section. The ratios of these
reaction constants result in relative enhancements in reaction
rate constants, as shown in Tables 2 and 3. Based on these
results, one can point out that the polymer IP4, imprinted with
the l-leucine-analogue amino phosphonic acid 4, accelerated
the hydrolysis of the l-substrate 2l in relation to the solution
by factors from 200 to 325 times depending on the buffer
concentration. It was observed that the buffer concentration
had a strong influence on the reaction rate constants. At a
lower buffer concentration, the relative reaction rate con-
stants increase by a factor of around 1.5 relative to a buffer
solution without a catalyst. This results in higher selectivity.
The enhancements show a strong catalytic effect of the
imprinted polymers relative to the solution. These enhance-
ments might not be solely caused by an imprinting effect.
a
2-[4-(2-hydroxyethyl)-1-piperazino]ethanesulfonic acid
(HEPES) buffer, at pH 7.3, with acetonitrile (208C). For this
purpose, a freshly prepared substrate solution of 1 Â 10À3 m in
acetonitrile was added to a suspension of catalyst containing
2 Â 10À3 m active sites (containing 4 Â 10À3 m accessible ami-
dine groups). The hydrolysis of both enantiomers was always
determined separately to check for enantioselectivity.
In the case of the control experiments, an amount of dry
control polymer with the same amount of accessible amidine
groups was used. The hydrolysis in solution, without any
polymer catalyst, was carried out in the same buffer solution
mixture, and at the same substrate concentrations. Control
experiments with monomeric binding sites N,N'-diethyl-4-
vinylbenzamidine (1) in buffered solutions were not per-
formed, since earlier experiments showed that the amidine
groups hardly exert any catalytic effect in buffered solution,
because most of these groups are protonated at this pH
value.[9]
To follow the reaction kinetics, aliquots were taken at
regular intervals and checked by HPLC. The reaction rate
constants for the hydrolysis were measured during conversion
of the first 5 to 10% of substrate conversion, while the
Chem. Eur. J. 2003, 9, 4106 4117
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4109

G. Wulff et al.
FULL PAPER
Figure 1. a) Comparison of the measured pseudo-first-order kinetics for
the hydrolysis of leucine substrates 2l and 2d, in the presence of IP4, CPF,
and just in solution. (In all cases: actetonitrile/0.1n HEPES buffer pH 7.3,
1:1) b) Pseudo-first-order kinetics for hydrolysis of leucine substrates 2l
and 2d in the presence of IP4.
Figure 2. a) Comparison of the measured pseudo-first-order kinetics for
the hydrolysis of valine substrate 3l and 3d, in the presence of IP5, CPF,
and just in solution. (Solvent same as in Figure 1a). b) Pseudo-first-order
kinetics for the hydrolysis of leucine substrates 3l and 3d, in the presence
of IP5 (Solvent same as in Figure 1a).
Possible effects of the polymer backbone and binding sites,
without a specific imprinting effect, can be elucidated by
investigation of the control polymers. The differences be-
tween the catalytic activity of imprinted and control polymers
correspond to the imprinting effect.[13] Polymer IP4 showed a
high enhancement, 67 to 79 times, relative to the control
polymer CPF; this was dependant on the buffer concentration
during the experiment. The acceleration relative to the second
control polymer CPB was about 14. Thus, the activity of the
control polymers was dependent on the size of the complexing
acid. By enlarging the template in the control polymer (from
formic acid to benzoic acid), a small imprinting effect was
observed.[9]
Table 3. Relative enhancements and enantioselectivities for the hydrolysis
of substrates 3l and 3d in acetonitrile/0.1m HEPES buffer pH 7.3 (1:1) at
208C, by using IP5 and the controls CPF and solution.
Substrate
Relative
enhancements[a]
Enantioselectivity
l versus d
IP5 vs solution
IP5 vs solution
IP5 vs CPF
3l
3d
3l
249
206
59
1.21
CPF vs solution
3l/3d
4.2
[a] Relative reaction rate constants ˆ relative enhancements between IP5,
CPF and solution.
Table 2. Relative enhancements and enantioselectivities for the hydrolysis of substrates 2l and 2d in acetonitrile/HEPES buffer pH 7.3 (1:1) at 208C, by
using IP4 and the controls CPF, CPB, and solution in two different buffer concentrations (0.15 and 0.10m).
HEPES buffer 0.15m
M HEPES buffer 0.10
Relative enhancements Enantioselectivity l versus d
Substrate
Relative enhancements[a]
Enantioselectivity
l versus d
IP4 vs solution
IP4 vs solution
IP4 vs CPF
2L
2d
2l
325
234
79
1.39
200
166
67
1.20
IP4 vs CPB
2l
14
CPB vs solution
CPF vs solution
2l/2d
2l/2d
24
4.1
3.0
[a] Relative reaction rate constants ˆ relative enhancements between IP4, CBF, CPB and solution.
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org
4110
Chem. Eur. J. 2003, 9, 4106 4117

A New Enzyme Model for Enantioselective Esterases
4106 4117
While the hydrolysis reaction is enhanced 325-fold with an
imprinting effect of 79 in a 0.1 m HEPES buffer by IP4, the
enantioselectivity between the l- and d-substrates (defined as
the ratio from the reaction rate constants of the hydrolysis of
the l- versus the d-enantiomer), catalysed by polymer IP4 is
1.39. Thus a significant enantioselectivity is observed. This
clearly shows that there is substrate selectivity in these
imprinted polymers. Table 3 shows that similar, though some-
what lower values are obtained for the imprinted catalyst IP5.
In this case, the smaller substituent in the template seemed to
cause a somewhat diminished imprinting accuracy.
Another way to examine the substrate selectivity is to
measure the cross-selectivity. We checked the cross-selectivity
of both l-imprinted polymers IP4 and IP5 with both
corresponding l-substrates 2l and 3l. For this purpose, the
catalytic effect for the hydrolysis of the l-valine-substrate 3l
by IP4 (imprinted with the l-leucine analogue template 4)
and 2l with IP5 were investigated. As shown in Table 4, the
Figure 3. Michaelis Menten kinetics illustrating the rate of hydrolysis for
2l and 2d in acetonitrile/0.1n HEPES buffer pH 7.3 (1:1), in the presence
of IP4 and CPF.
possible to get information on the extent of the influence of
selective binding and selective catalysis on the enantioselec-
tivity of the catalysis in this way.
Table 4. Cross-selectivities of valine substrate 3l in polymer IP4 (leucine-
analogue polymer), and leucine substrate 2l in IP5 (valine-analogue
polymer) in acetonitrile/0.1m HEPES buffer pH 7.3 (1:1) at 208C.
Catalysis by IP4 is shown in Figure 3, the curves in the
graph resemble typical Michaelis Menten curves (especially
in the hydrolysis of 2l). Therefore one can conclude, that,
similar to enzyme catalysis, the substrate (ester) is bound to
the catalyst in a pre-equilibrium step (see Scheme 4a). The
bound ester is then converted by activation of the carbonyl
group of the ester by one of the amidine groups, and the
attacking water molecule; this occurs via a transition state
[a]
Catalyst
Substrate
kIP/ksol
Cross-selectivity
3.2
IP4
IP4
IP5
IP5
2l[b]
3l[b]
3l
325
103
249
79
3.2
2l
IP4 vs. IP5
IP5 vs. IP4
2l
3l
4.2
2.5
[a] Relative enhancement of the imprinted catalyst versus solution [b] 2l ˆ
l-leucine substrate; 3l ˆ l-valine substrate.
™wrong∫ substrate was clearly hydrolyzed slower in both
cases. The cross-selectivity can be calculated in two different
ways. If the cross-selectivity is defined as the quotient of the
reaction rate constant for the hydrolysis of different sub-
strates, with the same polymer used, this value was about 3 for
both polymers. An alternative way to describe the cross-
selectivity, which can be defined as the ratio of the reaction
rate constant for the hydrolysis of one substrate by different
polymers, gave values of 4.2 (leucine substrate 3L) and 2.5
(valine substrate 4l) (see Table 4). The enantioselectivity and
the cross-selectivity show that the templates formed a specific
catalytically active cavity that is able to distinguish between
very similar substrates. It is surprising to realize that the
difference of an isopropyl group and an isobutyl group results
in 3 times more catalytic activity. In each case, the polymer
imprinted with a leucine template analogue hydrolyzes the
leucine ester considerably faster, and the polymer imprinted
with a valine analogue hydrolyzes the valine ester consider-
ably faster.
Mechanism of the catalytic action: In order to elucidate the
mechanism of catalysis for the hydrolysis, a detailed kinetic
investigation was performed. A plot of initial velocities of the
reaction versus the substrate concentration was carried out for
the hydrolysis of 2l and 2d in the presence of the imprinted
catalysts IP4, as well as in the presence of control polymer
CPF (see Figure 3). It was expected that it should also be
Scheme 4. Schematic picture of the different steps in the reaction pathway
during the catalyzed hydrolysis of substrate 2l, beginning with the
substrate-catalyst complex a and the detailed view of the following
steps b e.
Chem. Eur. J. 2003, 9, 4106 4117
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4111

G. Wulff et al.
FULL PAPER
(Scheme 4b) to an intermediate similar to that in Scheme 4c.
The energy difference between the transition state and the
intermediate is small therefore, according to the Hammond
rule, both structures are similar. This justifies why instead of a
transition-state analogue, an intermediate analogue could be
used as a template for the imprinting (compare Scheme 3a
and Scheme 4b and c). The last step involves the breaking
down of the tetrahedral intermediate via transition state II, to
release the carboxylate and the alcohol according to Equa-
tion (1) (see Scheme 4c and d).
curve in Figure 3, a strong participation of the more active
sites can be concluded. Values given for KM and kcat in the
following paragraphs are therefore ™apparent∫ values. Re-
gardless of these difficulties, important conclusions can be
drawn from the results obtained.
Table 5 shows the calculated data for kcat and Km. The
Michaelis constant KM, is a measure of the stability of the
substrate catalyst complex. Values for the template analogue
Table 5. Results of Michaelis Menten kinetics when the imprinted
catalysts IP4 and IP5 were used (see Figure 3).
If the reaction is performed with increasing amounts of
substrate (Figure 3), the rate of the reaction first increases
with increasing substrate concentration, but then levels off. At
higher substrate concentration, when all active sites are
occupied, it remains constant, that is, it is zero-order with
respect to the ester concentration (saturation kinetics). In
contrast, the nonimprinted control polymer CPF exhibits a
curve of low catalytic activity. From the curves shown in
Figure 3, similarly as with enzymes, the influence of binding
and catalysis can be elucidated by using the well-known
Michaelis Menten Equation (3):[25]
IP4
IP5
catalyst substrate
2l
2d
3l
3d
KM [mm]
0.51
4.74
0.94
0.62
3.49
0.57
0.34
3.26
0.95
0.43
2.68
0.62
kcat [10À4 minÀ1
]
kcat/KM [minÀ1mÀ1
l/d ratio of kcat/KM
]
1.66
1.53
substrate of 0.51 and 0.34mm are found. This shows a
relatively high stability of the substrate catalyst complex,
especially if it is taken into account that this noncovalent
interaction occurs in a medium consisting of a mixture of
aqueous buffer and acetonitrile in a 1:1 ratio. kcat is the first-
order rate constant, and is a measure of the catalytic activity
of the catalyst. It corresponds to the number of reactions per
minute an active site of the catalyst can catalyze (turnover
number). With kcat ˆ 4.74  10À4 minÀ1 and 3.26  10À4 minÀ1,
these values are not very high. However, it should be taken
into account that the hydrolysis of a relatively stable phenyl-
ester was investigated, unlike in many other investigations, in
which the kinetics of the hydrolysis for 4-nitrophenylester was
measured. For both polymers IP4 and IP5, the substrate
analogue template relative to its enantiomer is better bound
by a factor of KMl/KMd ˆ 0.82 and 0.79, respectively. The
catalytic constant shows the same tendency kcatl/kcatd of 1.36
and 1.22, respectively. As can be seen in Figure 3 the
hydrolysis of the control polymer does not show typical
Michaelis Menten kinetics. Therefore, meaningful calcula-
tions of kcat and KM can not be performed.
Often kcat/KM values, a measure of catalytic efficiency, are
given for enzymes or enzyme models. If these values are
compared to both substrate enantiomers, an enantioselectiv-
ity for IP4 of 1.66 and for IP5 of 1.53 is obtained. This shows
remarkable enantioselectivity for this catalytic action. Fur-
thermore, it shows that the rate enhancement of one
enantiomer is caused by the better binding of the template
analogue substrate, as well as by a better stabilization of the
template analogue transition state of the reaction. Thus,
thermodynamic and kinetic influences are observed to a
similar extent.
kcat‰CŠ‰SŠ
v0 ˆ
(3)
Km ‡ ‰SŠ
in which, v0 ˆ initial rate; kcat ˆ rate constant of the
catalyzed reaction, KM ˆ Michaelis constant, [S] ˆ molar con-
centration of substrate, and [C] ˆ total molar concentration of
active sites in the catalyst.
To calculate meaningful values for KM and kcat poses some
problems. Primarily, the kinetic measurements are carried out
in a heterogeneous phase, which limits the accuracy of the
values; this is especially true for kinetics at low substrate
concentration. Therefore, we did not use linearization by a
double reciprocal plot (Lineweaver Burk plot), instead we
used a computer program to fit the experimental values into
Equation (3) representing a hyperbola. This usually gives
values of higher accuracy.[26]
To solve Equation (3) the number of active sites in the
imprinted polymers IP4 and IP5 should be determined. A first
value is obtained by the determination of the amount, if
template molecules were released from the imprinted poly-
mer. Since the possibility of shrinking and nonaccessibility of
cavities exists, this value has been further controlled by
titration of the amidine groups with acid (see Table 1). Both
methods gave similar values, indicating that with this stoi-
chiometric noncovalent imprinting most of the cavities are
available. Earlier work showed that in the case of bisamidi-
nium bound templates, nearly the same amount of the
template is taken up, as has been split off after imprint-
ing.[14, 15]
A further problem in the calculation of data from Equa-
tion (3) exists. Unlike enzymes and monoclonal antibodies,
imprinted catalysts have active sites with different binding
ability, different selectivity in binding, as well as different
catalytic activity. This inhomogeneity (™polyclonality∫) is
especially effective if a strong excess of substrate is present,
since all active sites are then occupied. For this reason mean
values will be obtained; however, it is not exactly known how
these averages are generated. From the appearance of the
A typical property of enzymes is the inhibition by certain
molecules. A transition-state analogue imprinted polymer
should show competitive inhibition by the template. This
would be a further proof for catalysis to occur inside the
imprinted cavity. Figure 4 shows a double-reciprocal Line-
weaver Burk plot for initial reaction rates (v0À1) versus
substrate concentration (cÀ1). The reactions were run in the
4112
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4106 4117

A New Enzyme Model for Enantioselective Esterases
4106 4117
points were measured by using a B¸chi Melting Point B-545. The inner
surfaces (one point BET-isotherm) were measured by JUWE Laborger‰te
GmbH, Viersen (JUWE BET-A-MAT).
Materials: The solvents for HPLC (acetonitrile: isocratic grade), and all
solvents for the coupling reactions and polymerisations: methylene
chloride (peptide grade), chloroform (dry), methanol (extra dry), toluene
(dry), acetonitrile (dry), and ethanol (dry) were supplied from Biosolve. All
amino acid reactants (puris.), and the coupling reagent BOP and oxalyl
chloride were purchased from Fluka. N-Methyl morpholine, terephthalic
acid monomethyl ester, and bromotrimethylsilane, trifluoro acetic acid
were obtained from Aldrich, and 1,4-dioxane (puriss p.a.), [18]crown-6, and
30% hydrogen bromide in acetic acid were obtained from Merck.
3,5-Dimethyl phenol from Aldrich was further purified by sublimation.
Azobisisobutyronitrile from Aldrich was further purified by recrystalliza-
tion from dry methanol. Ethylene glycol dimethycrylate and methyl
methacrylate were purchased from Aldrich, and purified by drying over
calcium hydride and distillation. All other solvents were used in distilled
grade.
Figure 4. Lineweaver Burk plot of the kinetics for the hydrolysis of 3l in
acetonitrile/0.1n HEPES buffer pH 7.3 (1:1), in the presence of polymer
IP5 with different inhibitor 5 concentrations.
HPLC: All HPLC measurements were performed with a set-up consisting
of a Waters 410 pump, a Waters 486 UV-detector, and spectral recording
and integration software CSW Chromatography Station for Windows,
presence of catalyst IP5 under standard conditions, firstly
without addition of an inhibitor, and secondly with two
different concentrations of template 5l. Evidently, 5l is a
very effective inhibitor that binds much better than the
substrate 3l by a factor of 24 [Ki(competitive) ˆ 0.016mm].
From the plot in Figure 4, this does not suggest to be pure
competitive inhibition. Since the straight lines intersect in the
negative part of the plot (and not on the ordinate), a mixed
competitive (noncompetitive) inhibition seems to be present
(i.e., not only the catalyst, but also the catalyst substrate
complex can bind to the inhibitor). Since the accuracy of the
kinetic measurements is not high enough, we have not
calculated the different inhibitor constants.
Version 1.7, 2000, Apex Data Ltd. As columns,
a RP-18 ACE-EPS
(Bischoff) or a RP-18 (Merck) were used. Acetophenone was used as an
internal standard, which was distilled before usage.
General procedure for esterification; example 7d: N-Cbz-d-valine (6.68 g,
0.0266 mol) in dry methylene chloride (160 mL) were treated at room
temperature with 3,5-dimethyl phenol (3.25 g, 0.0266 mol), BOP (11.76 g,
0.0266 mol), and N-methyl morpholine (1.88 mL, 0.0266 mol). After being
stirring overnight, the solvent was removed, and the product was purified
by column chromatography (chloroform/hexane/acetone 4:10:1). The
product was obtained in a yield of 57.5% (5.43g).
N-Benzyloxycarbonyl-l-leucine 3,5-dimethylphenyl ester (6l): Yield:
1
39.5%; colorless oil; [a]2D5 ˆ À15.98 (c ˆ 2.0 in CHCl3); H NMR: d ˆ 1.01
(d, 3J(H,H) ˆ 5.4 Hz, 3H; CHCH3), 1.02 (d, 3J(H,H) ˆ 4.8 Hz, 3H;
CHCH3), 1.67 (m, 1H; CH2CH(CH3)2), 1.82 (m, 2H; CH2CH(CH3)2),
2.31 (s, 6H; CCH3), 4.61 (dt, 3J(H,H) ˆ 4.4, 3J(H,H) ˆ 8.7 Hz, 1H;
NHCHCH2), 5.14 (s, 2H; OCH2), 5.21 (d, 3J(H,H) ˆ 8.5 Hz, 1H; NH),
6.69 (s, 2H; ArH), 6.87 (s, 1H; ArH), 7.33 ppm (m, 5H; PhH); 13C NMR:
d ˆ 19.8, 20.5, 21.5, 23.5 (CH3), 40.3 (CH2), 51.3 (CH), 65.7 (CH2), 117.4,
126.4, 126.7, 126.8, 127.1 (ArCH), 134.8, 137.9, 148.9 (ArC), 154.6,
Conclusion
Very active, molecularly-imprinted catalysts with enzyme-like
properties were prepared. They showed pronounced sub-
strate- and enantioselectivity. The origin of the enantioselec-
tivity was investigated by determining the plots of initial
velocities versus substrate concentration. As a result, the
Michaelis constants Km and the catalytic constants kcat were
determined for both enantiomers. The enantioselectivity is
caused, to a similar degree, by selective-substrate binding
(thermodynamic control), and selective formation of the
transition state of the reaction (kinetic control). Strong
competitive inhibition by the template (a stable transition-
state analogue of the reaction) was observed in the catalysis.
ˆ
170.6 ppm (C O); FT-IR: nÄ ˆ 3339 (NH), 3034 (Ar-H), 2969 (CH3), 2871
ˆ
ˆ
(CH2), 1764 (C O ester), 1723 (C O carbamate), 1618 (carbamate),
1529 cmÀ1 (NH); MS (EI): m/z: 369 [M ]; elemental analysis calcd (%) for
C22H27NO4 (369.5): C 71.52, H 7.37, N 3.79; found: C 70.95, H 7.34, N 3.91.
‡
N-Benzyloxycarbonyl-d-leucine 3,5-dimethylphenyl ester (6d): Yield:
60.8%; colorless oil; [a]2D5 ˆ ‡16.08 (c ˆ 2.0 in CHCl3); spectra analogue
to 6l; elemental analysis calcd (%) for C22H27NO4 (369.5): C 71.52, H 7.37,
N 3.79; found: C 70.87, H 7.13, N 3.91.
N-Benzyloxycarbonyl-l-valine 3,5-dimethylphenyl ester (7l): Yield:
98.0%; m.p. 528C; [a]2D5 ˆ À11.68 (c ˆ 2.0 in CHCl3); 1H NMR: d ˆ 1.04
(d, 3J(H,H) ˆ 6.7 Hz, 3H; CHCH3), 1.08 (d, 3J(H,H) ˆ 6.6 Hz, 3H;
CHCH3), 2.30 (s, 6H; CCH3), 2.34 (m, 1H; CH(CH3)2), 4.53 (dd,
3J(H,H) ˆ 4.7, 3J(H,H) ˆ 9.2 Hz, 1H; NHCHCH), 5.14 (s, 2H, OCH2),
5.35 (d, 3J(H,H) ˆ 9.2 Hz, 1H; NH), 6.68 (s, 2H; ArH), 6.87 (s, 1H; ArH),
7.36 ppm (m, 5H; PhH); 13C NMR: d ˆ 18.0, 19.5, 21.4 (CH3), 31.8, 59.6
(CH), 67.5 (CH2), 119.3, 128.3, 128.6, 129.0, 136.7 (ArCH), 139.8, 150.7,
ˆ
150.7, 156.7 (ArC), 171.3, 171.5 ppm (C O); FT-IR (KBr): nÄ ˆ 3352 (NH),
Experimental Section
ˆ
ˆ
3035 (Ar-H), 2986 (CH3), 2963 (CH3), 1772 (C O ester), 1693 (C O
carbamate), 1618 (carbamate), 1528 cmÀ1 (NH); MS (EI): m/z: 355 [M ];
‡
Instrumentation: Elemental analysis was performed in the microanalytical
laboratories of the Faculty of Natural Sciences of the Heinrich-Heine
University in D¸sseldorf (Perkin Elmer 2400). 1H (500 MHz) and 13C{1H}
NMR (125.8 MHz) spectra were recorded on a Bruker DRX 500, and the
31PNMR (81 MHz) on a Bruker AC 200 with TMS as an internal standard.
NMR spectra were measured in CDCl3 at 258C if not otherwise stated.
Optical rotations were measured with a Perkin Elmer 241MC polar-
imeter (accuracy Æ 0.0028). The mass spectra were measured on a Varian
MAT311A (EI), a Finnigan Mat 8200 (EI), a Finnigan Mat8200 (FAB with
3-nitrobenzylalcohol (NBA) as matrix, and if required with addition of
sodium chloride), or a Finnigan INCOS 50 (CI). All infrared spectra were
recorded with a Bruker Vector 22 FT-IR-spectrophotometer. The melting
elemental analysis calcd (%) for C21H25NO4 (355.4): C 70.96, H 7.09, N 3.94;
found: C 70.98, H 7.28, N 3.89.
N-Benzyloxycarbonyl-d-valine 3,5-dimethylphenyl ester (7d): Yield:
57.5%; m.p. 548C; [a]2D5 ˆ ‡11.68 (c ˆ 2.0 in CHCl3); spectra analogue to
7l; elemental analysis calcd (%) for C21H25NO4 (355.4): C 70.96, H 7.09, N
3.94; found: C 70.75, H 7.10, N 3.96.
General procedure for deprotection; example for 9l: Ester 7l (1.50 g,
4.220 mmol) in 30% hydrogen bromide in acetic acid (20 mL) was stirred
for 3 h at room temperature. The excess hydrogen bromide in acetic acid
was removed by evaporation, and the remaining solid was dried. The
product was taken up in dry diethyl ether, filtered off, and washed with a
Chem. Eur. J. 2003, 9, 4106 4117
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4113

G. Wulff et al.
FULL PAPER
small amount of dry diethyl ether. The crude product was pure and could
further be used directly, and was obtained in a yield of 90.2% (1.15 g).
analysis calcd (%) for C28H29NO5 (459.5): C 73.18, H 6.36, N 3.05; found: C
72.84, H 6.38, N 3.22.
l-Leucine 3,5-dimethylphenyl ester hydrobromide (8l): Yield: 75.2%;
m.p. 2138C; [a]2D5 ˆ ‡ 26.38 (c ˆ 2.0 in MeOH); 1H NMR ([D6]DMSO):
d ˆ 0.97 (d, 3J(H,H) ˆ 4.8 Hz, 3H; CHCH3), 0.98 (d, 3J(H,H) ˆ 4.7 Hz, 3H;
CHCH3), 1.76(m, 1H; CH2CH(CH3)2), 1.85 (m, 2H; CH2CH(CH3)2), 2.30
(s, 6H; CCH3), 4.26 (t, 3J(H,H) ˆ 6.9 Hz, 1H; NHCHCH2), 6.80 (s, 2H;
ArH), 6.97 (s, 1H; ArH), 8.49 ppm (brs, 3H; NH3); 13C NMR
([D6]DMSO): d ˆ 21.1, 22.4, 22.5 (CH3), 24.3 (CH), 39.5 (CH2), 51.1
N-(O-Benzylterephthaloyl)-d-valine 3,5-dimethylphenyl ester (12d):
Yield: 78.0%; m.p. 1068C; [a]2D5 ˆ À10.98 (c ˆ 2.0 in CHCl3); spectra
analogue to 12l; elemental analysis calcd (%) for C28H29NO5 (459.5):
C 73.18, H 6.36, N 3.05; found: C 72.96, H 6.34, N 3.09.
General procedure for the hydrogenation, example for 2l: Palladium
hydroxide on carbon (0.50 g) was added to 11l (1.20 g, 2.53 mmol) in
ethanol (60 mL), and treated with hydrogen gas (1 bar). After 2 to 3 days
the catalyst was filtered off over celite (reaction was followed by TLC). The
solvent was removed, and the obtained crude product was purified by
column chromatography (ethyl acetate with 1% acetic acid/methanol 3:1).
To remove the acetic acid completely, the product was dried over potassium
hydroxide. The product was recrystallized from cyclohexane and dry
ethanol, and a yield of 91.1% (0.88 g) was obtained.
ˆ
(CH), 119.0, 128.3 (ArCH), 139.6, 150.0 (ArC), 169.2 ppm (C O); FT-IR
(KBr): nÄ ˆ 2952 (CH3), 1771 (C O ester), 1591 (NH3 ), 1495 cmÀ1 (NH3 );
‡
‡
ˆ
‡
À
MS (FAB, NBA): m/z: 236 [M Br ]; elemental analysis calcd (%) for
C14H22BrNO2 (316.2): C 53.17, H 7.01, N 4.43; found: C 53.03, H 6.75, N 4.49.
d-Leucine-3,5-dimethylphenyl ester hydrobromide (8d): Yield: 71.9%;
m.p. 2178C; [a]2D5 ˆ À26.28 [c ˆ 2.0 in MeOH]; spectra analogue to 8l;
elemental analysis calcd (%) for C14H22BrNO2 (316.2): C 53.17, H 7.01, N
4.43; found: C 53.04, H 6.94, N 4.37.
N-Terephthaloyl-l-leucin 3,5-dimethylphenylester (2l): Yield: 91.1%;
m.p. 1628C; [a]2D5 ˆ À6.68 (c ˆ 2.0 in CHCl3); 1H NMR: d ˆ 1,07 (d,
3J(H,H) ˆ 1.6 Hz, 3H; CHCH3), 1.08 (d, 3J(H,H) ˆ 1.9 Hz, 3H; CHCH3),
1.87 (m, 2H; CH2CH(CH3)2), 1.73 (m, 1H; CH2CH(CH3)2), 2.32 (s, 6H;
CCH3), 5.08 (m, 1H; NHCHCH2), 6.74 (s, 2H; ArH), 6,74 (d, 3J(H,H) ˆ
7.3 Hz, 1H; NH), 6.89 (s, 1H; ArH), 7.88 (d, 3J(H,H) ˆ 8.5 Hz, 2H; ArH),
8.12 ppm (d, 3J(H,H) ˆ 8.2 Hz, 2H; ArH); 13C NMR: d ˆ 21.2, 22.1 (CH3),
22.9 (CH), 25.2, 25.2 (CH3), 41.7 (CH2), 51.6 (CH), 118.8, 127.3, 128.0, 130.5
l-Valine 3,5-dimethylphenyl ester hydrobromide (9l): Yield: 90.2%; m.p.
2048C; [a]2D5 ˆ ‡ 19.58 (c ˆ 2.0 in MeOH); 1H NMR ([D6]DMSO): d ˆ 1.08
(d, 3J(H,H) ˆ 7.0 Hz, 3H; CHCH3), 1.11 (d, 3J(H,H) ˆ 7.0 Hz, 3H;
CHCH3), 2.30 (s, 6H; CCH3), 2.35 (m, 1H; CH(CH3)2), 4.20(d,
3J(H,H) ˆ 4.8 Hz, 1H; NHCHCH), 6.79 (s, 2H; ArH), 6.97 (s, 1H; ArH),
8.51 ppm (brs, 3H; NH3); 13C NMR ([D6]DMSO): d ˆ 18.1, 18.7, 21.1
(CH3), 29.9 (CH2), 57.7 (CH), 119.1, 128.4 (ArCH), 139.7, 149.9 (ArC),
ˆ
(ArCH), 132.1, 138.5, 139.5, 150.3 (ArC), 166.6, 169.9, 172.3 ppm (C O);
ˆ
ˆ
ˆ
168.1 ppm (C O); FT-IR (KBr): nÄ ˆ 2970 (CH3), 1753 (C O ester), 1583
FT-IR (KBr): nÄ ˆ 3356 (NH), 2960 (CH3), 1752 (C O ester), 1701
(NH3 ), 1501 cmÀ1 (NH3 ); MS (FAB, NBA): m/z 222 [M Br ]; elemental
analysis calcd (%) for C13H20BrNO2 (302.2): C 51.67, H 6.67, N 4.63; found:
C 51.48, H 6.63, N 4.53.
(COOH), 1641 (C O amide), 1528 cm (amide); MS (FAB, NBA): m/z
‡
‡
‡
À1
À
ˆ
‡
‡
406 ([M ‡Na]), 384 ([M ‡H]); elemental analysis calcd (%) for
C22H25NO5 ¥ 0.5H2O (383.4 ‡ 9.0): C 67.33, H 6.68, N 3.57; found: C 67.04,
H 6.66, N 3.48.
d-Valine 3,5-dimethylphenyl ester hydrobromide (9d): Yield: 61.3%; m.p.
2028C; [a]2D5 ˆ À19.98 (c ˆ 2.0 in MeOH); spectra analogue to 9l;
elemental analysis calcd (%) for C13H20BrNO2 (302.2): C 51.67, H 6.67, N
4.63; found: C 51.34, H 6.46, N 4.67.
N-Terephthaloyl-d-leucine 3,5-dimethylphenyl ester (2d): Yield: 94.8%;
m.p. 1628C; [a]2D5 ˆ ‡ 6.68 (c ˆ 2.0 in CHCl3); spectra analogue to 2l;
elemental analysis calcd (%) for C22H25NO5 (383.4): C 68.91, H 6.57, N 3.65;
found: C 68.75, H 6.71, N 3.76.
General procedure for the coupling to the amide, example for 11l: Ester
8l (0.80 g, 2.530 mmol) in dry methylene chloride (60 mL) under argon
were mixed with terephthalic acid monomethyl ester (0.49 g, 2.530 mmol),
PyCloP (1.17 g, 2.783 mmol), and N-methyl morpholine (0.77 g,
7.583 mmol). The mixture was stirred overnight at room temperature.
The product was purified by column chromatography (chloroform/hexane/
acetone 4:2:1 for 12l, and for 12d the eluent was chloroform/hexane/
acetone 10:4:1), and a yield of 94.5% (1.13 g) was obtained.
N-Terephthaloyl-l-valine 3,5-dimethylphenyl ester (3l): Yield: 87.8%;
m.p. 1538C; [a]2D5 ˆ ‡ 6.78 (c ˆ 2.0 in CHCl3); 1H NMR: d ˆ 1.18 (d,
3J(H,H) ˆ 2.5 Hz, 3H; CHCH3), 1.22 (3J(H,H) ˆ 2.5 Hz, 3H; CHCH3), 2.37
(s, 6H; CCH3), 2.55 (m, 1H; CH(CH3)2), 5.07 (dd, 3J(H,H) ˆ 4.8,
3J(H,H) ˆ 8.8 Hz, 1H; NHCHCH), 6.77 (s, 2H, ArH), 6,88 (d, 3J(H,H) ˆ
8.8 Hz, 1H; NH), 6.94 (s, 1H; ArH), 7.93 (d, 3J(H,H) ˆ 8.5 Hz, 2H; ArH),
8.19 ppm (d, 3J(H,H) ˆ 8.5 Hz, 2H; ArH); 13C NMR: d ˆ 18.1, 19.2, 21.2
(CH3), 31.7, 57.7 (CH), 118.8, 127.3, 128.1, 130.5 (ArCH), 132.1, 138.7, 139.5,
N-(O-Benzylterephthaloyl)-l-leucine-3,5-dimethylphenyl ester (11l):
Yield: 94.5%; m.p. 1088C; [a]2D5 ˆ ‡2.58 (c ˆ 2.0 in CHCl3); 1H NMR:
d ˆ 1.06 (d, 3J(H,H) ˆ 4.1 Hz, 3H; CHCH3), 1.07 (d, 3J(H,H) ˆ 4.1 Hz, 3H;
CHCH3), 1.85 (m, 2H; CH2CH(CH3)2), 1.96 (m, 1H; CH2CH(CH3)2), 2.31
(s, 6H; CCH3), 5.06 (m, 1H; NHCHCH2), 5.38 (s, 2H; OCH2), 6.62 (d,
3J(H,H) ˆ 8.2 Hz, 1H, NH), 6.72 (s, 2H; ArH), 6.88 (s, 1H; ArH), 7.40 (m,
5H; PhH), 7.87 (d, 3J(H,H) ˆ 8.5 Hz, 2H; ArH), 8.14 ppm (d, 3J(H,H) ˆ
8.4 Hz, 2H; ArH); 13C NMR: d ˆ 21.2, 22.1, 22.9 (CH3), 25.2 (CH), 41.8
(CH2), 51.5 (CH), 67.1 (CH2), 118.8, 127.2, 128.0, 128.3, 128.4, 128.7, 130.0
ˆ
150.2 (ArC), 166.8, 170.0, 171.2 ppm (C O); FT-IR (KBr): nÄ ˆ 3345 (NH),
ˆ
ˆ
2968 (CH3), 1756 (C O ester), 1681 (COOH), 1644 (C O amide),
1531 cmÀ1 (amide); MS (FAB, NBA): m/z 392 ([M ‡Na]), 370 ([M ‡H]);
elemental analysis calcd (%) for C21H23NO5 ¥ 0.25H2O (369.4 ‡ 4.5):
C 67.46, H 6.33, N 3.75; found: C 67.38, H 6.51, N 3.61.
‡
‡
N-Terephthaloyl-d-valine 3,5-dimethylphenyl ester (3d): Yield: 93.3%;
m.p. 1538C; [a]2D5 ˆ À6.68 (c ˆ 2.0 in CHCl3); spectra analogue to 3l;
elemental analysis calcd (%) for C21H23NO5 ¥ 0.5H2O (369.4 ‡ 9,0): C 66.65,
H 6.39, N 3.70; found: C 66.51, H 6.47, N 3.75.
(ArCH), 133.0, 135.7, 137.9, 139.5, 150.3 (ArC), 165.6, 166.4, 171.9 ppm
À1
ˆ
ˆ
(C O); FT-IR (KBr, cm ): nÄ ˆ 3260 (NH), 2957 (CH3), 1762 (C O ester),
General procedure for coupling of the amide to 17: The a-l-amino-
phosphonic acid 14[22] (5.19 g, 0.0247 mol) in dry methylene chloride
(180 mL) under an argon atmosphere were treated with PyCloP (10.41 g,
0.0247 mol), N-methyl morpholine (5.00 g, 0.0494 mol), and monomethyl
terephthalate (15) (4.45 g, 0.0247 mol); the mixture was stirred for 48 h at
room temperature. The solvent was then evaporated, and the product was
purified by column chromatography in chloroform/hexane/acetone 4:1:1 to
give a yield of 87.5% (8.03 g).
ˆ
ˆ
1722 (C O ester), 1638 (C O amide), 1538 (amide); MS (EI): m/z: 473
[M ]; elemental analysis calcd (%) C29H31NO5 (473.6): C 73.55, H 6.60, N
‡
2.96; found: C 73.28, H 6.56, N 3.06.
N-(O-Benzylterephthaloyl)-d-leucine-3,5-dimethylphenyl ester (11d):
Yield: 92.3%; m.p. 1088C; [a]2D5 ˆ À2.58 (c ˆ 2.0 in CHCl3]; spectra
analogue to 11l; elemental analysis calcd (%) for C29H31NO5 (473.6): C
73.55, H 6.60, N 2.96; found: C 73.65, H 6.88 , N 3.10.
N-(O-Benzylterephthaloyl)-l-valine-3,5-dimethylphenyl
ester
(12l):
l-N-[1-(Diethoxyphosphoryl)-3-methylbutyl]terephthalamic acid methyl
ester (16): Yield: 90.5%; colorless oil; [a]2D5 ˆ À10,88 [c ˆ 2.0 in CHCl3];
1H NMR: d ˆ 0.96 (s, 3H; CHCH3), 0.99 (s, 3H; CHCH3), 1.22 (t,
Yield: 78.0%; m.p. 1068C; [a]2D5 ˆ ‡10.88 (c ˆ 2.0 in CHCl3); 1H NMR:
d ˆ 1.29 (d, 3J(H,H) ˆ 7.0 Hz, 3H; CHCH3), 1.14 (d, 3J(H,H) ˆ 7.3 Hz, 3H;
CHCH3), 2.32 (s, 6H; CCH3), 2.48 (m, 1H; CH(CH3)2), 5.01 (dd, 3J(H,H) ˆ
4.8, 3J(H,H) ˆ 8.5 Hz, 1H; NHCHCH), 5.39 (s, 2H; OCH2), 6.70 (d,
3J(H,H) ˆ 8.8 Hz, 1H; NH), 6.71 (s, 2H; ArH), 6.89 (s, 1H; ArH), 7.36 (m,
5H; PhH), 7.88 (d, 3J(H,H) ˆ 8.9 Hz, 2H; ArH), 7.88 ppm (d, 3J(H,H) ˆ
8.4 Hz, 2H; ArH); 13C NMR: d ˆ 18.1, 19.1, 21.2 (CH3), 21.8, 57.7 (CH), 67.1
(CH2), 118.8, 127.2, 128.0, 128.3, 128.4, 128.7, 130.1 (ArCH), 133.0, 135.7,
3
3J(H,H) ˆ 7.0 Hz, 3 H ; CH2CH3), 1.36 (t, J(H,H) ˆ 7.2 Hz, 3H; CH2CH3),
1.77 (m, 3H; CH2CH(CH3)2), 3.94 (s, 3H; OCH3), 4.11 (m, 4H; CH2CH3),
4.79 (m, 1H; NHCHP), 7.33 (d, 3J(H,H) ˆ 11.0 Hz, 1H; NH), 7.92 (d,
3J(H,H) ˆ 8.5 Hz, 2H; ArH), 8.10 ppm (d, 3J(H,H) ˆ 8.5 Hz, 2H; ArH);
13
À
C NMR: d ˆ 16.8 (d), 17.0 (d) (Et CH3), 21.7, 22.1 (CH3), 24.5 (d) (CH),
38.6 (CH2), 44.5 (d) (CH2), 52.8 (OCH3), 62.8 (d), 63.2 (d) (OCH2), 127.8,
138.1, 139.5, 150.2 (ArC), 165.6, 166.6, 170.9 ppm (C O); FT-IR (KBr): nÄ ˆ
130.0 (ArCH), 133.1, 138.4 (ArC), 166.6, 166.6 ppm (C O); 31P NMR
ˆ
ˆ
3299 (NH), 2973 (CH3), 1756 (C O ester), 1724 (C O ester), 1642 (C O
(81 MHz, CDCl3): d ˆ 26.3 (s); FT-IR (cmÀ1): nÄ ˆ 3267 (NH), 2957 (CH3),
ˆ
ˆ
ˆ
amide), 1542 cmÀ1 (amide); MS (FAB, NBA): m/z: 460 [M ‡H]; elemental
2871 (CH2), 1728 (C O ester) 1661 (C O amide), 1542 (amide), 1280
‡
ˆ
ˆ
4114
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4106 4117

A New Enzyme Model for Enantioselective Esterases
4106 4117
‡
ˆ
À À
(P O), 1031 (P O C); MS (EI): m/z 385 (M ); elemental analysis calcd
(%) for C18H28NO6P ¥ 0.25H2O (385.4 ‡ 4.5): C 55.45, H 7.37, N 3.59; found:
C 55.56, H 7.50, N 3.69.
General procedure for the combined hydrolysis to produce 5l: Compound
19 (1.00 g, 1.91 mmol) suspended in 1,4-dioxane (10 mL) and water
(10 mL), was treated with [18]crown-6 (30 mg), and a 1n sodium hydroxide
solution (3.82 mL, 3.82 mmol). After the mixtures was stirred for 12 h at
room temperature, the solution was evaporated and dried. The product was
purified by means of column chromatography with a solvent mixture of
chloroform and ammoniacal methanol (10% NH3 in MeOH) in the ratio
3:1, and was then transformed in the dipotassium salt by an ion exchange
column loaded with potassium ions. This produced a yield of 76.2%
(0.59 g).
l-N-[1-(Diethoxyphosphoryl)-2-methylpropyl]terephthalamic acid methyl
ester (17): Yield: 87.5%; m.p. 818C; [a]2D5 ˆ À2.58 (c ˆ 2.0 in CHCl3);
1H NMR: d ˆ 1.11 (dd, J(H,H) ˆ 6.8, 4J(P,H) ˆ 1.3 Hz, 3H; CHCH3), 1.12
3
(d, 3J(H,H) ˆ 6.8 Hz, 3H; CHCH3), 1.27 (t, 3J(H,H) ˆ 7.5 Hz, 3H;
CH2CH3), 1.37(t, 3J(H,H) ˆ 7.0 Hz, 3H; CH2CH3), 2.35 (m, 4H;
CH(CH3)2), 3.98 (s, 3H; OCH3), 4.14 (m, 3H; CH2CH3), 4.62 (ddd,
3J(H,H) ˆ 4.8 Hz, 3J(H,H) ˆ 10.4 Hz, 2J(P,H) ˆ 18.1 Hz, 1H; NHCHP),
6.72 (dd, 3J(H,H) ˆ 10.4 Hz, 3J(P,H) ˆ 2.6 Hz, 1H; NH), 7.90 (d,
3J(H,H) ˆ 8.8 Hz, 2H; ArH), 8.14 ppm (d, 3J(H,H) ˆ 8.8 Hz, 2H; ArH);
Dipotassium salt of N-(4-carboxybenzoyl)-1-l-amino-3-methylbutylphos-
phonic acid-3,5-dimethylphenyl ester (4l): Yield: 72.5%; m.p. decompo-
sition >2808C; [a]2D5 ˆ ‡ 7.28 (c ˆ 2.0 in MeOH); 1H NMR ([D6]DMSO):
d ˆ 0.88 (d, 3J(H,H) ˆ 4.0 Hz, 6H; CHCH3), 1.60 (m, 3H; CH2CH(CH3)2),
2.12 (s, 6H; CCH3), 4.22 (m, 1H; NHCHP), 6.51 (s, 2H; ArH), 6.67 (s, 1H;
ArH), 7.59 (dd, 3J(H,H) ˆ 9.5 Hz, 4J(P,H) ˆ 2.8, 1H; NH), 7.77 (d,
3J(H,H) ˆ 8.3 Hz, ArH), 7.84 ppm (d, 3J(H,H) ˆ 8.0 Hz, 2H; ArH);
13C NMR: d ˆ 21.7 (d), 22.3 (CH3), 24.5 (CH), 26.5 (d) (CH2), 41.1 (CH),
120.0 (d), 126.2, 128.2, 130.6 (ArCH), 137.6, 140.2, 142.5, 149.3, 155.1 (ArC),
13
1
À
C{ H} NMR: d ˆ 16.8 (d), 18.2 (d) (Et CH3), 20.6 (d) (CH3), 29.3 (d), 50.8
(d) (CH), 52.8 (OCH3), 62.5 (d), 62.6 (d) (OCH2), 127.1 (d), 129.9 (d),
(ArCH), 133.0, 138.0 (ArC), 166.2, 166.6 ppm (d) (C O); 31P NMR: d ˆ
ˆ
ˆ
25.2 (s); FT-IR (KBr): nÄ ˆ 3261 (NH), 2981 (CH3), 1725 (C O ester) 1659
À1
ˆ
ˆ
‡
À
À
(C O amide), 1544 (amide), 1278 (P O), 1031 cm (P O C); MS (CI,
‡
NH3): m/z 370 ([M‡H] ), 371 (M ); elemental analysis calcd (%) for
C17H26NO6P (371.4): C 54.98, H 7.06, N 3.77; found: C 54.82, H 7.19, N
3.81.
169.5, 174.6 (C O); 31P NMR ([D6]DMSO)): d ˆ 15.6 (s); FT-IR (KBr):
ˆ
À
ˆ
nÄ ˆ 3384 (NH), 2957 (CH3), 1626 (C O amide), 1595 (COO ), 1559
(amide), 1420 (COOÀ), 1208 cmÀ1 (P O);; MS (FAB, NBA): m/z 464
General procedure for the hydrolysis, chlorination, and esterification of the
P-terminal part to produce 19: Bromo trimethylsilane (4.27 g, 0.0337 mol)
was added in batches to 17 (0.50 g, 6.73 mmol) in dry methylene chloride
(60 mL) under an argon atmosphere. After the mixture was stirred for 14 h
at room temperature, the solution was evaporated, and THF (20 mL) and
water (2 mL) were added. After an additional 5 h, the solution was again
evaporated and dried by oil pump vacuum (0.05 mbar). The obtained crude
product was used without further purification. Some drops of DMF were
added to the hydrolysis product in dry methylene chloride (60 mL) under
argon oxalyl chloride (5.13 g, 0.0404 mol) for activation. After the mixture
was stirred for 11 h, the solvent and the excess oxalyl chloride were
removed, and the product was dried under an oil pump vacuum
(0.05 mbar). The crude product was used without further purification.
After chlorination, the residue in dry methylene chloride (60 mL) under an
argon atmosphere was treated with 3,5-dimethylphenol (1.73 g, 0.0141 mol)
and N-methyl morpholine (1.43 g, 0.0141 mol); it was then stirred for 13 h
at room temperature. After evaporation, the product was purified by
column chromatography (hexane/chloroform/acetone 10:4:1). The product
obtained was in a yield of 47.4% (1.67 g) over 3 steps, and was used directly
because of its instability towards oxidation.
ˆ
([M‡H]?), 486 ([M‡Na]?); elemental analysis calcd (%) for C21H24NO6P-
K2 ¥ 1 H2O(463.59 ‡ 18,0): C 49.11, H 5.10, N 2.73; found: C 49.27, H 4.82, N
2.76.
Dipotassium salt N-(4-carboxybenzoyl)-1-l-amino-2-methylpropylphos-
phonic acid 3,5-dimethylphenyl ester (5l): Yield: 76.2%; m.p. (decomp)
>2008C; [a]2D5 ˆ ‡ 19.78 [c ˆ 2.0 in MeOH]; 1H NMR ([D6]DMSO): d ˆ
0.93 (d, 3J(H,H) ˆ 6.8 Hz, 3H; CHCH3), 0.98 (d, 3J(H,H) ˆ 6.8 Hz, 3H;
CHCH3), 1.25 (m, 1H; CH(CH3)2), 2.09 (s, 6H; CCH3), 3.99 (ddd,
3J(H,H) ˆ 4.2, 3J(H,H) ˆ 9.5, 2J(P,H) ˆ 17.2 Hz, 1H; NHCHP), 6.49 (s, 1H;
ArH), 6,62 (s, 2H; ArH), 7.62 (d, 3J(H,H) ˆ 8.3 Hz, 2H; ArH), 7.85
(d,3J(P,H) ˆ 8.3 Hz, 2H; ArH); 13C NMR ([D6]DMSO): d ˆ 19.1, 21.2, 21.6
(d) (CH3), 29.9 (CH), 63.2 (d) (CH), 118.8 (d), 123.6, 126.6, 129.2 (ArCH),
136.4, 137.8, 138.0, 154.1 (ArC), 166.3, 172.0 (C O); 31P NMR
ˆ
([D6]DMSO): d ˆ 14.1 (s); FT-IR (KBr): nÄ ˆ 3428 (NH), 2960 (CH3),
À
À
ˆ
1634 (C O amide), 1594 (COO ), 1554 (amide), 1385 (COO ), 1222
À1
‡
ˆ
À À
(P O), 1034 cm (P O C); MS (FAB, NBA): m/z: 404 ([M À 2 K ‡ H ]);
elemental analysis calcd (%) for C20H22NO6PK2 ¥ 4H2O (449.57 ‡ 72.1):
C 43.39, H 5.46, N 2.53; found: C 43.43, H 5.21, N 2.70.
Preparation of the imprinted polymers: The binding site monomer N,N'-
diethyl-4-vinylbenzamidine was used in the form of amidinium chloride salt
1 Â HCl.[28] One equivalent of template 4l or 5l, was dissolved together
with two equivalents of the amidine 1 ¥ HCl in dry methanol to form the
polymerisable 1:2 complex. The remaining potassium chloride precipitates
were removed by membrane filtration. The solution was evaporated, and
the dried complex was dissolved in 82% (w/v) ethylene glycol dimethyl-
acrylate (EDMA), 15% (w/v) methyl methacrylate (MMA) in 10 mL
acetonitrile/toluene 1:1. The mixtures were homogenized in an ultrasonic
bath at a maximum temperature of 408C. Finally, the initiator azobisiso-
butyronitrile (AIBN) was added, and the monomer mixture was degassed
by a ™freeze-and-thaw∫ procedure. The polymerization was carried out in
bulk at 608C for 72 h in an evacuated ampoule. The polymers were crushed
and sieved, and only the fraction from 45 to 125 mm was used for the
measurements.
l-N-{1-[Bis-(3,5-dimethylphenoxy)phosphoryl]-3-methylbutyl}terephtha-
lamic acid methyl ester (18): Yield: 57.3% over 3 steps; colorless oil;
1H NMR: d ˆ 1.03 (d, J(H,H) ˆ 5.8 Hz, 3H; CHCH3), 1.06 (d, J(H,H) ˆ
6.0 Hz, 3H; CHCH3), 1.88 (m, 3H; CH2CH(CH3)2), 2.15 (s, 6H; CCH3),
3.99 (s, 3H; OCH3), 5.17 (m, 1H; NHCHP), 6.62 (brdd, 1H; NH), 6.79 (3s,
6H; ArH), 7.75(d, 3J(H,H) ˆ 8.3 Hz, 2H; ArH), 8.05 ppm (d,3J(H,H) ˆ
8.5 Hz, 2H; ArH); 13C NMR: d ˆ 18.5, 18.6, 20.9 (d), 21.4 (CH3), 26.5 (d)
(CH2), 41.1 (CH), 51.7 (OCH3), 52.9 (CH), 113.5, 118.3, 122.7, 127.4, 127.7
(d), 130.2 (ArCH), 133.4, 137.8, 139.9 (d), 140.2, 150.4 (ArC), 166.7,
3
3
167.2 ppm (d) (C O); 31P NMR: d ˆ 18.4 ppm (s); FT-IR (cmÀ1): nÄ ˆ 3279
ˆ
ˆ
(NH), 3018 (Ar-H), 2962 (CH3), 2921 (CH3), 2875 (CH), 1737 (C O ester),
ˆ
ˆ
À À
1669 (C O amide), 1536 (amide), 1280 (P O), 1030 (P O C); MS (EI):
‡
m/z: 402 ([M À 135 ]?); elemental analysis calcd (%) for C30H36NO6P
(537.6): C 67.03, H 6.75, N 2.61; found: C 67.22 , H 6.72, N 2.57.
To remove the template, the polymer was first swollen in acetonitrile, and
after one hour the template was washed out by stirring three to five times in
a 1:1 mixture of 0.1m sodium hydroxide and methanol in an ice bath, until
no template was present. The amount of released template was controlled
by HPLC (eluent: 0.2% (v/v) trifluoroacetic acid/acetonitrile 65:35 (v/v),
internal standard acetophenone, flow rate 1 mLminÀ1). A calibration curve
for the determination of the template in the presence of the internal
standard acetophenone was established. At the time when no template was
detectable, the polymer was washed with water and methanol. Before use,
all polymers were dried over phosphorous pentoxide in an vacuum oven at
408C.
l-N-{1-[Bis-(3,5-dimethylphenoxy)phosphoryl]-3-methylpropyl}terephtha-
lamic acid methyl ester (19): Yield: 47.4% over 3 steps; colorless oil;
3
1H NMR: d ˆ 1.21 (dd, J(H,H) ˆ 6.8, 4J(P,H) ˆ 1.4 Hz, 3H; CHCH3), 1.23
(d, 3J(H,H) ˆ 6.9 Hz, 3H; CHCH3), 2.11 (s, 6H; CCH3), 2.34 (s, 6H;
CCH3), 2.55 (m, 1H; CH(CH3)2), 4.00 (s, 3H; OCH3), 5.10 (ddd, 3J(H,H) ˆ
3
2
4.8, J(H,H) ˆ 10.4, J(P,H) ˆ 19.4 Hz, 1H; NHCHP), 6.72 (m, 4H; ArH),
6.91 (m, 2H; ArH), 7.77 (d, 3J(H,H) ˆ 8.8 Hz, 2H; ArH), 8.07 ppm (d,
3J(H,H) ˆ 8.8 Hz, 2H; ArH); 13C NMR: d ˆ 18.5, 18.6, 20.9 (d), 21.4 (CH3),
30.1 (d) (CH), 51.7 (OCH3), 52.9 (CH), 113.5, 118.3, 122.7, 127.4, 127.7 (d),
130.2 (ArCH), 133.4, 137.8, 139.9 (d), 140.2, 150.4 (ArC), 166.7, 167.2 ppm
(d, C O); 31P NMR (81 MHz, CHCl3): d ˆ 17.7 ppm (s); FT-IR: nÄ ˆ 3289
ˆ
À
ˆ
(NH), 3018 (Ar H), 2965 (CH3), 2922 (CH3), 2875 (CH), 1729 (C O ester),
The polymers were characterized by measuring the inner surface (BET-
isotherm), and the amount of active amidine groups that were carried out
by acid-base titration. For titration in 11 vials for each, the same amount of
polymer (20 mg) as suspended in a 1:1 mixture of a 1m solution of sodium
chloride and 1,4-dioxane (10 mL) was used. An increasing amount of HCl
À1
ˆ
ˆ
À À
1668 (C O amide), 1538 (amide), 1279 (P O), 1031 cm (P O C); MS
‡
‡
(CI, NH3 , 2008C): m/z: 524 ([M ‡H]), 541 ([M ‡NH4]); elemental
analysis calcd (%) for C29H34NO5P (523.6): C 66.53, H 6.55, N 2.68; found:
C 66.23, H 6.58, N 2.58.
Chem. Eur. J. 2003, 9, 4106 4117
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4115

G. Wulff et al.
FULL PAPER
was added until all theoretically present amidine groups were protonated. Table 6. Reaction rate constants for the hydrolysis of substrates 2l and 2d in
After the mixture was stirred overnight, an aliquot of 500 mL was taken, acetonitrile/HEPES buffer pH 7.3 (1:1) at 208C, by using IP4 and the controls
and, after removing the polymer by membrane filtration, the solution was CPF, CPB and solution in two different buffer concentrations.
transferred in a small plastic vial to measure the pH value (pH-micro-
Catalyst
Substrate
0.10m HEPES buffer
k [minÀ1
0.15m HEPES buffer
k [minÀ1
electrode, Schott). The resulting curve from the measured pH-value plotted
versus the HCl concentration had an inflection point, from which, the
amount of accessible amidine-groups could be calculated.
]
]
IP4
IP4
CPF
CPF
average CPF
CPB
CPB
average CPB
solution
solution
average solution
2l[a]
2d
2l
2d
2
2l
2d
2
8.27 Â 10À4 Æ8.5%
5.95 Â 10À4 Æ2.4%
1.00 Â 10À5 Æ1.5%
1.08 Â 10À5 Æ3.4%
1.05 Â 10À5 Æ3.5%
6.11 Â 10À5
6.69 Â 10À4
5.57 Â 10À4
General procedure for the kinetic measurements: The hydrolysis of the
esters were carried out at optimal conditions for amidine based polymers in
an 1:1 mixture of a 2-[4-(2-hydroxyethyl)-1-piperazino]ethanesulfonic acid
(HEPES) buffer, pH 7.3, and acetonitrile at a temperature of 208C.[17] For
this purpose, dry polymer with free cavities (6 Â 10À6 mol; 1.2 Â 10À5 mol
amidine groups), was dispersed in a 3 mL screw-capped, in the buffer
mixture (2.7 mL), which included 30 mL of an 0.05m internal standard
solution (acetophenone in acetonitrile). This gave a 2mm concentration of
free cavities. The polymer was stirred in this solution overnight to swell the
polymer, and equilibrate all amidine groups inside. To start the reaction, a
freshly prepared substrate solution in acetonitrile (30 mL, 0.1m, 3 Â
10À6 mol) was added to get a final substrate concentration of 1mm. The
hydrolysis of both enantiomers was determined separately at all times to
check for enantioselectivity.
1.01 Â 10À5
9.84 Â 10À6
9.97 Â 10À6 Æ1.2%
5.97 Â 10À5
6.04 Â 10À5 Æ 1.2%
2.53 Â 10À6 Æ 3.8%
2.56 Â 10À6 Æ7.3%
2.54 Â 10À6 Æ5.9%
2l
2d
2
3.28 Â 10À6 Æ1.7%
3.42 Â 10À6 Æ9.8%
3.35 Â 10À6 Æ7.5%
[a] 2l ˆ l-leucine substrate; 2d ˆ d-substrate; IP4 ˆ polymer imprinted with
template 4; CPF ˆ control polymer imprinted with formic acid; CPB ˆ control
polymer imprinted with benzoic acid; Solution ˆ acetonitrile/buffer (1:1). When a
measurement error is given, the value is investigated two to three times; errors
denote standard deviation.
In the case of the control experiments, same amounts of dry control
polymer and amidine groups were used (1.2 Â 10À5 mol). Without the
presence of a polymer catalyst, the hydrolysis reactions in solution were
carried out in the same solution mixture at the same substrate concen-
trations.
Table 7. Reaction rate constants for the hydrolysis of the substrates 3l and
3d using IP5 and the controls CPF and solution. All experiments were
performed in acetonitrile/0.1m HEPES buffer pH 7.3 (1:1).
For each condition the reaction was performed two to three times to reduce
measurement errors. During the reactions, six to nine aliquots of 150 mL
were taken, the polymer was filtered off through a membrane, and the
sample was then collected in an Eppendorf vessel and frozen in liquid
nitrogen to stop the reaction. All samples from the reaction with imprinted
polymers were taken within four hours, within three days for the control
polymers, and within two weeks for the control solution (with no polymer
presence). Afterwards each sample was defrosted to room temperature and
measured by injecting 20 mL in the HPLC-system. For the mobile phase,
acetonitrile, and trifluoroacetic acid (0.2% v/v) in water 30:70 (v/v) with a
flow rate of 1 mLminÀ1 were used, and for the stationary phase, a RP-18
column (Merck) was used. The chosen wavelength corresponded to the
absorption of 3,5-dimethyl phenol with lmax 218 nm, which is optimal to
detect the product of the reaction. The system was optimized to control the
product peak and the internal standard. A detection software package was
used to record and integrate the chromatograms.
substrate
k [minÀ1
]
IP5
IP5
CPF
CPF
average CPF
solution
solution
average solution
3l[a]
3d
3l
3d
3
3l
3d
3
5.54 Â 10À4 Æ2.7%
4.53 Â 10À4 Æ6.2%
9.02 Â 10À6 Æ2.2%
9.44 Â 10À6 Æ0.8%
9.23 Â 10À6 Æ2.8%
2.16 Â 10À6 Æ1.1%
2.23 Â 10À6 Æ3.0%
2.20 Â 10À6 Æ2.6%
[a] 3l ˆ l-valine substrate; 3d ˆ d-substrate; IP5 ˆ polymer imprinted
with template 5; CPF ˆ control polymer imprinted with formic acid;
solution ˆ acetonitrile/buffer (1:1). When a measurement error is given,
the value was investigated twice, except for IP5, which was investigated
three times.
A calibration curve for the determination of the product phenol in presence
of the internal standard acetophenone was established. With this calibra-
tion a conversion up to 20% of the investigated hydrolysis could easily be
followed.
Table 8. Reaction rate constants for cross-selectivity of valine substrate 3l
with polymer IP4 (leucine-analogue polymer), and leucine substrate 2l[a]
with IP5 (valine-analogue polymer) in acetonitrile/0.1m HEPES buffer
pH 7.3 (1:1) at 208C.
Due to the experimental circumstances at the beginning of all the
measurements, a small amount of the reaction product was already
present. Thus the product concentration, and the total substrate concen-
tration of all data were corrected by a factor of 0.994; this meant, that on
average, 0.6% of the substrate had already been hydrolyzed at the start of
the reaction.
Catalyst
Substrate
k [Â10À4 minÀ1
]
IP4
IP5
3l
2l
2.26 Æ4.1%
2.00 Æ3.1%
The calculations of the reaction rate constants were performed with a linear
fit by Origin 7.0. (See Tables 6, 7, and 8)
[a] 2l ˆ l-leucine substrate; 3l ˆ l-valine substrate.
Michaelis Menten kinetics: For the
imprinted polymers IP4 and IP5, and
Table 9. Data of Michaelis Menten kinetics with IP4 and CPF.[a]
for the control polymer CPF, plots of
initial velocities of the catalyzed hy-
drolysis versus substrate concentration
were measured. For these investiga-
tions the initial reaction rates were
monitored at constant concentration
of active sites, and increasing substrate
concentration. We chose six different
substrate concentrations in Table 9
and five in Table 10. The procedure
to measure the velocity was analogous
to pseudo-first-order kinetics, as pre-
viously described.
Substrate con-
centration
2l with IP4
2d with IP4
2l with CPF
2d with CPF
v0 [10À7 m minÀ1
]
v0 [10À7 m minÀ1
]
v0 [10À8 m minÀ1
]
v0 [10À8 m minÀ1
]
[mmolLÀ1
]
0.124
0.497
0.994
1.988
9.940
19.88
2.05
3.94
6.99
7.65
8.88
9.19
2.24
5.09
5.62
6.20
6.81
0.61
0.94
1.45
2.04
2.02
0.62
0.10
1.47
2.08
2.06
[a] Reaction velocities obtained by increasing substrate concentration of substrates 2l and 2d in acetonitrile/0.1m
HEPES buffer pH 7.3 (1:1) at 208C.
4116
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4106 4117

A New Enzyme Model for Enantioselective Esterases
4106 4117
Table 10. Data of Michaelis Menten kinetics with IP5 without and with two concentration of inhibitor
5L.[a]
[8] X.-C. Liu, K. Mosbach, Macromol. Rapid
Commun. 1997, 18, 609 615.
[9] G. Wulff, T. Gro˚, R. Schˆnfeld, Angew.
Chem. 1997, 109, 2049 2052; Angew.
Chem. Int. Ed. Engl. 1997, 36, 1961 1964.
[10] B. Sellergren, K. J. Shea, Tetrahedron:
Asymmetry 1994, 5, 1403 1406.
[11] B. Sellergren, R. N. Karmalkar, K. J. Shea,
J. Org. Chem. 2000, 65, 4009 4027.
[12] K. Ohkubo, Y. Funakoshi, Y. Urata, S.
Hirota, S. Usui, T. Sagawa, J. Chem. Soc.
Chem. Commun. 1995, 2143 2144.
Substrate
concentration
[mmolLÀ1
3l with IP5
3d with IP5
3l with IP5
3l with IP5
[I] ˆ 0.001m
v0 [10Àm minÀ1
]
v0 [10À7 m minÀ1
]
[I ˆ 0.0]001m
]
v0 [10À8 m minÀ1
]
v0 [10À8 m minÀ1
]
0.497
0.994
1.988
9.940
19.88
3.67
5.10
5.58
6.09
6.51
2.95
3.71
4.34
5.05
5.38
9.25
1.58
2.43
3.25
3.80
0.38
0.65
1.12
2.09
2.43
[a] Reaction velocities obtained by increasing the substrate concentration of substrates 3l and 3d in
acetonitrile/0.1m HEPES buffer pH 7.3 (1:1) at 208C. [I] ˆ concentration of inhibitor 5l.
[13] G. Wulff, Chem. Rev. 2002, 102, 1 28.
[14] G. Wulff, R. Schˆnfeld, Adv. Mater. 1998,
10, 957 959.
G. Wulff, K. Knorr, B[i1o5s]eparation 2002, 10, 257 276.
[16] J. D. Stewart, L. J. Liotta, S. J. Benkovic, Acc. Chem. Res. 1993, 26,
396 404.
[17] A. G. Strikovsky, D. Kasper, M. Gr¸n, B. S. Green, J. Hradil, G. Wulff,
J. Am. Chem. Soc. 2000, 122, 6295 6296.
In addition, the influence of the inhibiting molecule template 5, was
monitored with polymer IP5. Two different inhibitor concentrations were
added to the hydrolysis media to reach the concentration of 1mm and
0.1mm. The velocities were monitored in the same way. The obtained
results are listed in Tables 9 and 10.
[18] a) W. Chen, D. K. Han, K. D. Ahn, J. M. Kim, Macromol. Res. 2002,
10, 122 126b) J. M. Kim, K. D. Ahn, G. Wulff, Macromol. Chem.
Phys. 2001, 1105 1108.
The calculations of the Michaelis Menten kinetics were performed by
fitting them to a hyperbola by the program Origin 7.0.
[19] a) B. Castro, G. Evin, C. Selve, R. Seyer, Synthesis 1977, 413; b) B.
Castro, J. R. Dormoy, G. Evin, C. Selve, Tetrahedron Lett. 1975, 1219
1222.
Acknowledgement
¬
[20] J. Coste, E. Frerot, P. Jouin, J. Org. Chem. 1994, 59, 2437 2446.
[21] P. Kafarski, B. Lejczak, J. Szewczyk, Can. J. Chem. 1983, 61, 2425
2430.
[22] A. B. Smith, K. M. Yager, C. M. Taylor, J. Am. Chem. Soc. 1995, 117,
10879 10888.
This work was supported by Deutsche Forschungsgemeinschaft, Fonds der
Chemischen Industrie, and by the Graduierten Kolleg for Biology and
Chemistry ™Molecular Physiology∫, Heinrich-Heine-University of D¸ssel-
dorf.
[23] V. A. Solodenko, T. N. Kasheva, V. P. Kukhar, E. V. Kozlova, D. A.
Mironenko, V. K. Svedas, Tetrahedron 1991, 47, 3989 3998.
[24] J. Szewczyk, B. Lejczak, P. Kafarski, Synthesis 1982, 5, 409 412.
[25] H. Bisswanger, Enzyme Kinetics–Principles and Methods, Wiley-
VCH, Weinheim, 2002.
[26] a) A. Cornish-Bowden, Fundamentals of Enzyme Kinetics, Portland
Press, London, 1996; b) P. J. F. Hernderson, in Enzyme Assays (Eds.:
R. Eisenthal, M. J. Danson), IRL, Oxford, 1992.
[1] G. Wulff, A. Sarhan, K. Zabrocki, Tetrahedron Lett. 1973, 44, 4329
4332.
[2] G. Wulff, A. Sarhan, German patent, Offenlegungsschrift DE
A 2242796, 1974, [Chem. Abstr. 1975, 83, P 60300w]; US Patent,
continuation in part US-A 4127730, 1978.
[3] G. Wulff, Angew. Chem. 1995, 107, 1958–1979; Angew. Chem. Int. Ed.
Engl. 1995, 34, 1812 1832.
[27] G. Wulff, A. Biffis, in Molecularly Imprinted Polymers–Man-Made
Mimics of Antibodies and their Application in Analytical Chemistry
(Ed.: B. Sellergren), Elsevier, Amsterdam, 2001, pp. 71 111.
[28] a) G. Wulff, in Templated Organic Synthesis (Eds. F. Diederich, P. J.
Stang), Wiley-VCH, Weinheim, 1999, pp. 39 73; b) G. Wulff, R.
Schˆnfeld, M. Gr¸n, R. Baumstark, G. Wildburg, L. H‰ussling,
German Patent, Offenlegungsschrift DE A 19720345 A1, 1998[Chem.
Abstr. 1998, 128 P 49155].
[4] ™Molecular and Ionic Recognition with Imprinted Polymers∫ (Eds.;
R. A. Bartsch, M. Maeda), ACS Symp. Ser. Vol 703, Washington, 1998.
[5] Molecularly Imprinted Polymers–Man-Made Mimics of Antibodies
and their Application in Analytical Chemistry (Ed.: B. Sellergren),
Elsevier, Amsterdam, 2001.
[6] R. A. Lerner, S. J. Benkovic, P. G. Schultz, Science 1991, 252, 659 667.
[7] P. G. Schultz, Angew. Chem. 1989, 101, 1336 1348; Angew. Chem. Int.
Ed. Engl. 1989, 28, 1283 1295.
Received: January 28, 2003 [F4783]
Chem. Eur. J. 2003, 9, 4106 4117
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4117