Functional mimicry of carboxypeptidase A by a combination of transition state stabilization and a defined orientation of catalytic moieties in molecularly imprinted polymers

Article abstract of DOI:10.1021/ja8012648

An artificial model for the natural enzyme carboxypeptidase A has been constructed by molecular imprinting in synthetic polymers. The tetrahedral transition state analogues (TSAs 4 and 5) for the carbonate hydrolysis have been designed as templates to allow incorporation of the main catalytic elements, an amidinium group and a Zn²⁺ or Cu²⁺ center, in a defined orientation in the transition state imprinted active site. The complexation of the functional monomer and the template in presence of Cu²⁺ through stoichiometric noncovalent interaction was established on the basis of ¹H NMR studies and potentiometric titration. The Cu²⁺ center was introduced into the imprinted cavity during polymerization or by substitution of Zn²⁺ in Zn²⁺ imprinted polymers. The direct introduction displayed obvious advantages in promoting catalytic efficiency. With substrates exhibiting a very similar structure to the template, an extraordinarily high enhancement of the rate of catalyzed to uncatalyzed reaction (K_cat/k_uncat) of 10⁵-fold was observed. If two amidinium moieties are introduced in proximity to one Cu ²⁺ center in the imprinted cavity by complexation of the functional monomer 3 with the template 5, the imprinted catalysts exhibited even higher activities and efficiencies for the carbonate hydrolysis with k _cat/K_uncat as high as 410 000. These are by far the highest values obtained for molecularly imprinted catalysts, and they are also considerably higher compared to catalytic antibodies. Our kinetic studies and competitive inhibition experiments with the TSA template showed a clear indication of a very efficient imprinting procedure. In addition, this demonstrates the important role of the transition state stabilization during the catalysis of this reaction.

Full text of DOI:10.1021/ja8012648

Published on Web 05/30/2008
Functional Mimicry of Carboxypeptidase A by a Combination
of Transition State Stabilization and a Defined Orientation of
Catalytic Moieties in Molecularly Imprinted Polymers
Jun-qiu Liu*,‡ and Gu¨nter Wulff*,†
Institute of Organic and Macromolecular Chemistry, Heinrich Heine UniVersity, 40225
Du¨sseldorf, Germany and the State Key Laboratory of Supramolecular Structure and Materials,
Jilin UniVersity, Changchun, 130012, P. R. China
Received February 20, 2008; E-mail: wulffg@uni-duesseldorf.de; junqiuliu@jlu.edu.cn
Abstract: An artificial model for the natural enzyme carboxypeptidase A has been constructed by molecular
imprinting in synthetic polymers. The tetrahedral transition state analogues (TSAs 4 and 5) for the carbonate
hydrolysis have been designed as templates to allow incorporation of the main catalytic elements, an
amidinium group and a Zn2+ or Cu2+ center, in a defined orientation in the transition state imprinted active
site. The complexation of the functional monomer and the template in presence of Cu2+ through
stoichiometric noncovalent interaction was established on the basis of 1H NMR studies and potentiometric
titration. The Cu2+ center was introduced into the imprinted cavity during polymerization or by substitution
of Zn2+ in Zn2+ imprinted polymers. The direct introduction displayed obvious advantages in promoting
catalytic efficiency. With substrates exhibiting a very similar structure to the template, an extraordinarily
high enhancement of the rate of catalyzed to uncatalyzed reaction (kcat/kuncat) of 105-fold was observed. If
two amidinium moieties are introduced in proximity to one Cu2+ center in the imprinted cavity by complexation
of the functional monomer 3 with the template 5, the imprinted catalysts exhibited even higher activities
and efficiencies for the carbonate hydrolysis with kcat/kuncat as high as 410 000. These are by far the highest
values obtained for molecularly imprinted catalysts, and they are also considerably higher compared to
catalytic antibodies. Our kinetic studies and competitive inhibition experiments with the TSA template showed
a clear indication of a very efficient imprinting procedure. In addition, this demonstrates the important role
of the transition state stabilization during the catalysis of this reaction.
1. Introduction
but also for practical applications in chemistry, biology, and
medicine.2
Natural enzymes catalyze reactions with high stereoselectivity
and efficiency. To elucidate the catalytic nature of enzyme
efficiency many practical and fundamental efforts have been
made. Although the major evidence of the catalytic mode of
operation were provided by biochemical and structural studies,1
the design of synthetic catalysts as enzyme mimics has become
an important aspect in understanding the molecular interactions
involved in substrate recognition and catalysis.2 In recent years,
chemists showed an increasing interest in the design of highly
selective catalysts, not only for providing more insight into the
mechanisms of molecular recognition and catalysis in enzymes,
Initial efforts in this regard have focused on the construction
of small host molecules with defined cavities as well as carrying
catalytic groups and cofactors.2 Notable examples include
synthetic model systems consisting of cyclodextrins2c,d and
macrocyclic amines2g as well as other macrocyclic compounds.2
More recently, promising success has been achieved in the
design of macromolecular receptors capable of selectively
binding and catalyzing defined substrates.3–5 Among them are
catalytic antibodies,3 chemically modified naturally occurring
enzymes and receptors,4 as well as synthetic polymers.5,2f Using
the concept of transition state stabilization known from enzyme
catalysis, catalytically quite active antibodies have been raised
against stable transition state analogues (TSAs) of the corre-
† Heinrich Heine University.
‡ Jilin University.
(1) (a) Wolfenden, R. Science 1983, 222, 1087–1093. (b) Warshel, A.
Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 444–448. (c) Benkovic, S. J.;
Hammes-Schiffer, S. Science 2003, 301, 1196–1202. (d) Radkiewicz,
J. L.; Brooks, C. L., III J. Am. Chem. Soc. 2000, 122, 225–231. (e)
Kahn, K.; Bruice, T. C. J. Am. Chem. Soc. 2000, 122, 46–51.
(2) (a) Kirby, A. Angew. Chem., Int. Ed. 1996, 35, 707–724. (b) Kirby,
A. Acc. Chem. Res. 1997, 30, 290–296. (c) Breslow, R.; Dong, S. D.
Chem. ReV 1998, 98, 1997–2011. (d) Breslow, R.; Graff, A. J. Am.
Chem. Soc. 1993, 115, 10988–10989. (e) Chin, J. Acc. Chem. Res.
1991, 24, 145–152. (f) Suh, J. Acc. Chem. Res. 2003, 36, 562–570.
(g) Kimura, E. Acc. Chem. Res. 2001, 34, 171–179. (h) Motherwell,
W. B.; Bingham, M. J.; Six, Y. Tetrahedron 2001, 57, 4663–4686.
(i) Houk, K. N.; Leach, A. G.; Kim, S. P.; Zhang, X. Angew. Chem.
2003, 115, 5020–5046; Angew. Chem., Int. Ed. 2003, 42, 4872-4897.
(3) (a) Lerner, R. A.; Benkovic, S. J.; Schultz, P. G. Science 1991, 252,
659–667. (b) Schultz, P. G.; Lerner, R. A. Science 1995, 269, 1835–
1842. (c) Hilvert, D.; Hill, K. W.; Narel, K. D.; Auditor, M.-T. M.
J. Am. Chem. Soc. 1989, 111, 9261–9262.
(4) (a) Qi, D. F.; Tann, C. M.; Haring, D.; Distefano, M. D. Chem. ReV.
2001, 101, 3081–3111. (b) DeSantis, G.; Jones, J. B. Curr. Opin.
Biotechnol. 1999, 10, 324–330. (c) Penning, T. M. Chem. ReV. 2001,
101, 3027–3046.
(5) (a) For reviews, see: Wulff, G. Chem. ReV. 2002, 101, 1–27. (b)
Ramstro¨m, O.; Mosbach, K. Curr. Opin. Chem. Biol. 1999, 3, 759–
764. (c) Severin, K. In Molecularly Imprinted Materials. Science and
Technology; Yan, M., Ramstro¨m, O., Eds.; Marcel Dekker: New York,
2005; pp 619-683.
9
8044 J. AM. CHEM. SOC. 2008, 130, 8044–8054
10.1021/ja8012648 CCC: $40.75
2008 American Chemical Society

Functional Mimicry of Carboxypeptidase A
A R T I C L E S
sponding reaction.3 Similarly, molecularly imprinted polymers
(MIPs) offer an excellent possibility to mimic the active site of
natural enzymes since not only the shape of the transition state
can be mimicked by imprinting but also, at the same time,
suitable catalytically active groups and binding sites can be
introduced into the active site in a predetermined orientation.5
In molecular imprinting a highly cross-linked copolymer is
formed during polymerization around a template molecule. The
monomer mixture contains functional monomers that can
reversibly interact with the template through covalent or
noncovalent interactions. The resultant complex is subsequently
incorporated into a network polymer by copolymerization in
the presence of an excess of a cross-linking monomer and a
solvent acting as an inert porogen. After removal of the template
from the macroporous polymer, 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.5,6
Such imprinted polymers show high selectivity in rebinding to
their own template molecules. Molecularly imprinted polymers
show antibody-like recognition characteristics and have mean-
while been prepared for a large number of compound classes.
Their preparation has also become an efficient approach for the
development of synthetic enzyme models with high selectivity.
By analogy to the preparation of catalytic antibodies,3 numerous
experiments have been undertaken to mimic enzyme behavior
by imprinting with stable transition state analogues (TSAs) in
reactions including dehydrofluorination, hydrolyses, Diels-Alder
reactions, transaminase reactions, and others.7,8
in order to emulate carboxypeptidase A-like activity.9 An
unusually high catalytic activity for carbonate hydrolysis was
obtained by imprinting with a stable transition state analogue
template as well as introduction of an amidinium group and a
Cu(II)9a or Zn(II)9b center into the active site in a defined
orientation to each other.
Carboxypeptidase A (CPA, EC3.4.17.1) is a zinc-containing
metalloprotease that has the function of removing the C-terminal
amino acid residue from a peptide chain.10 As one of the most
intensively studied enzymes, CPA has contributed enormously
to the elucidation of the catalytic mechanism of other metallo-
enzymes and also served as a model of an inhibitor in the design
of new medicines.10a In the active site of CPA, the zinc ion
coordinated tightly to the amino acid residues of His 69, Glu
72, and His 196 is essential for the enzyme reaction.10b The
guanidinium moiety of Arg 127 binds the oxyanion generated
in the rate-limiting step of the formation of the tetrahedral
transition state. Substrate specificity is brought about by a
hydrophobic pocket and another guanidinium moiety of Arg
145.
In this paper, in continuation of work published in two short
communications,9 the preparation of further artificial models
for the natural enzyme carboxypeptidase A is described and
the properties of this type of catalysts are discussed in detail.
For this, new functional monomers, substrates, and imprinted
polymers were prepared. The type and strength of interaction
of functional monomers with the template in the presence or
absence of Cu(II) are characterized more in detail. The catalytic
activity of the imprinted polymers is discussed in relation to
their chemical structure and the type of preparation. The
optimized catalysts possess a remarkably high catalytic activity
surpassing that of the corresponding catalytic antibodies by 2.5
orders of magnitude. These are also by far the highest catalytic
activities found for catalysts based on molecularly imprinted
polymers. The highest activity is obtained with a new type of
dimeric binding site containing only one Cu2+ center.
This type of research provides more information for the
understanding of the catalysis of natural enzymes, since these
models show not only high catalytic activity but also other
typical enzyme properties such as selectivity, Michaelis-Menten
kinetics, high efficiency, high proficiency, competitive inhibition,
and a bell-shaped pH rate profile.
Although earlier attempts to mimic enzyme behavior using
imprinted polymers showed only low or moderate catalytic
efficiency,7 very promising results for catalyzed hydrolyses were
obtained in recent years using MIP catalysts.8,9 It follows from
these results that the ability for TSA binding alone is not
sufficient to obtain a really high rate of acceleration; increasing
the transition state binding by steric and electronic effects in
combination with correctly incorporating and positioning the
functional groups is essential for the construction of an effective
enzyme model.2,9
Previously, MIP-based catalysts with both TSA binding sites
and appropriate catalytic groups have been prepared in our group
(6) (a) Wulff, G. Angew. Chem. 1995, 107, 1958–1979; Angew. Chem.,
Int. Ed. 1995, 34, 1812-1832. (b) Cormack, P. A. G.; Mosbach, K.
React. Funct. Polym. 1999, 41, 115–124. (c) Shea, K. J. Trends Polym.
Sci. 1994, 5, 166–173. (d) Sellergren, B., Ed. Molecularly Imprinted
Polymers. Man-made Mimics of Antibodies and Their Application in
Analytical Chemistry; Elsevier: Amsterdam, 2000. (e) Haupt, K.;
Mosbach, K. Chem. ReV. 2000, 100, 2495–2504. (f) Wulff, G. In
Templated Organic Synthesis; Diederich, F., Stang, P. J., Eds.; Wiley-
VCH: Weinheim, 2000; pp 39-73.
(7) (a) Robinson, D. K.; Mosbach, K. J. Chem. Soc., Chem. Commun.
1989, 969–970. (b) Beach, J. V.; Shea, K. J. J. Am. Chem. Soc. 1994,
116, 379–380. (c) Ohkubo, K.; Urata, Y.; Honda, Y.; Nakashima, Y.;
Yoshinaga, K. Polymer 1994, 35, 5372–5374. (d) Sellergren, B.;
Karmalkar, R. N.; Shea, K. J. J. Org. Chem. 2000, 65, 4009–4027.
(e) Svenson, J.; Zheng, N.; Nicholls, I. A. J. Am. Chem. Soc. 2004,
126, 8554–8560. (f) Maddock, S. C.; Pasetto, P.; Resmini, M. Chem.
Commun. 2004, 536–537.
2. Experimental Section
2.1. General Procedures. All reagents were reagent grade and
were used without further purification, unless otherwise noted.
Solvents were purified by standard methods. Column chromatog-
raphy was carried out on silica gel 60 (70-230 mesh). NMR spectra
were recorded using a Bruker AC250F operating at 500 MHz for
1H NMR and 125 MHz for 13C NMR. IR spectra were recorded
on a Bruker IFS-FT66V infrared spectrometer. Elemental analyses
were determined on a Perkin-Elmer 240 DS elemental analyzer.
2.2. Syntheses of Templates and Substrates. The synthesis of
the template 5 and the substrates phenyl-(2-pyridyl)-carbonate (7)
and di-(2-pyridyl)-carbonate (8) (Chart 1) were described previously
in our recent communication (see Supporting Information in ref
9a).
(8) (a) Wulff, G.; Gross, T.; Scho¨nfeld, R. Angew. Chem. 1997, 109, 2049–
2052; Angew. Chem., Int. Ed. Engl. 1997, 36, 1962-1964. (b)
Strikowsky,; A, G.; Kaspar, D.; Gru¨n, M.; Green, B. S.; Hradil, J.;
Wulff, G. J. Am. Chem. Soc. 2000, 122, 6295–6296. (c) Emgenbroich,
M.; Wulff, G. Chem. Eur. J. 2003, 9, 4106–4117. (d) Wulff, G.; Chong,
B.-O.; Kolb, U. Angew. Chem. 2006, 118, 3021–3024; Angew. Chem.,
Int. Ed. 2006, 45, 2955-2958.
2.3. Syntheses of Monomers. The preparation of the new
monomer N,N-[N′′-(2-aminoethyl)-1,5-(3-azapentylen)]-bis[(N′-eth-
yl)-4-vinylbenzamidine] (3) is performed in a similar manner as
(9) (a) Liu, J.-Q.; Wulff, G. J. Am. Chem. Soc. 2004, 126, 7452–7453.
(b) Liu, J.-Q.; Wulff, G. Angew. Chem. 2004, 116, 1307–1311; Angew.
Chem., Int. Ed. 2004, 43, 1287-1290.
(10) (a) Christianson, D. W.; Lipscomb, W. N. Acc. Chem. Res. 1989, 22,
6269.000. (b) Philips, M. A.; Fletterick, R.; Rutter, W. J. J. Biol. Chem.
1990, 265, 20692–20698.
9
J. AM. CHEM. SOC. VOL. 130, NO. 25, 2008 8045

A R T I C L E S
Liu and Wulff
Chart 1. Structures of Designed Functional Monomers (1-3),
Templates (4 and 5), and Substrates (6-8)
Figure 1. Saturation plot of the complexation of the functional monomer
2 and the template 4 (DPP). Data were obtained by 1H NMR titration; the
changes in chemical shift of the phenyl ring protons at δ7.23 were used as
a reference.
of 0.2 M NaOH and acetonitrile (1:1, v/v) to give ∼75% free
cavities. The free cavities in the imprinted polymers were deter-
mined by measuring the complexation of diphenylphosphate (4) to
the polymer at 25 °C (see determination of amidines with
diphenylphosphate).
As a comparison, a control polymer with formic acid instead of
the template was prepared in an analogous manner. For comparison
purposes the imprinted polymers were also prepared with DMSO
and with DMF as the porogen.
that of the monomer N-{2-[bis(2-aminoethyl)amino]ethyl-}N′-ethyl-
(4-vinylbenzamidine) (2) (for synthesis of 2, see Supporting
Information in ref 9a).
Thus dry tri-(2-aminoethyl)ammonium chloride (4.30 g, 0.021
mol) was added to a solution of N-ethyl-4-vinylbenzocarboximide
acid ethyl ester6f (8 g, 0.044 mol) in dry EtOH (150 mL) under
argon. After stirring for 5 h, 4-tert-butylbenzcatechol (0.45 g) was
added, the stirring was continued for 5 days at room temperature,
and the solvent was removed in vacuo. The residue was treated
with ice cooled 6 M NaOH (200 mL) and immediately extracted
four times with ice cooled chloroform. The combined organic phases
were dried with Na2SO4. The solvent was removed in vacuo, and
the residue was purified by silica gel column chromatography using
a 1:1 CH2Cl2/MeOH (MeOH saturated with NH3) solution as the
eluent to give a yellow liquid of 3 (2.01 g, 21% yield). FT-IR (KBr,
cm-1): ν ) 3336 (NH3), 3269 (NH), 2944 (CH3), 2886, 2848 (CH2),
1623 (CdN amidine), 1605 (arom), 1507 (CdN amidine), 1460,
2.4.2. Substitution of Zn2+ by Cu2+ in a Zn2+-Imprinted
Polymer. The preparation of imprinted polymers containing Zn2+
was performed as described previously (see Supporting Information
in ref 9b). A suspension of freshly prepared Zn2+-imprinted polymer
PZn2,5 (100 mg) in a mixture of water and N,N-dimethylformamide
(DMF) (1:1, v/v)(50 mL) was treated with EDTA ·2Na+ (100 mg)
to remove the Zn2+. The mixture was stirred at 100 rpm and room
temperature for 1 day. The polymers were collected by filtration
and washed with DMF, 0.2 M NaOH, and water. This procedure
was repeated three times. The polymer was resuspended in DMF
(20 mL) containing CuCl2 (20 mg), and the mixture was stirred at
100 rpm at room temperature for 1 day. The polymers exhibiting
a blue color were collected by filtration and washed with DMF
and water and then dried in vacuo (PZn2,5Cu). A control polymer
prepared with formic acid instead of the template was treated
identically.
2.4.3. Potentiometric pH Titrations. Potentiometric pH titra-
tions were performed using a standard electrode system (Orion
Research Ross Combination pH Electrode 8100BN). The test
solutions were prepared with freshly distilled water and were
degassed under ultrasonic agitation. All of the test solutions (50
mL) were kept under an argon atmosphere. The potentiometric pH
titrations of the monomer 2 (1.0 mm) in the absence and the
presence of 1 equivalent of Cu(II), respectively, were carried out
at 25 °C with I ) 0.1 (NaNO3), and at least two independent
1
1444,1305, 1135, 985, 916. H NMR (CDCl3, 25 °C): δ(ppm) )
7.51 (d, J ) 7.8 Hz, 4H), 7.35 (d, J ) 7.8 Hz, 4H), 6.68 (dd, J )
10.8 Hz, 2H), 5.82 (d, J ) 17.6 Hz, 2H), 5.38 (d, J ) 11.0 Hz,
2H), 2.52-3.84 (m, CH2’s, 16H), 1.21 (t, J ) 7.2 Hz, 6H). MS
(FAB + NBA): m/z: 461 [M + H]+. Elemental analysis calculated
(%) for C28H40N6 ·2HCl (460.7 + 72.9): C, 63.04; H, 7.88; N, 15.76.
Found: C, 62.80; H, 8.50; N, 16.01.
2.4. Synthesis and Properties of Imprinted Polymers.
2.4.1. Polymerization in the Presence of Copper(II) Ions. As a
typical example the preparation of the imprinted polymer with
monomer 3 and template 5 is described. Sodium phenyl-2-pyridyl-
phosphate was transferred into the potassium salt with a cation ion-
exchanger column (Amberlite IR-120). Potassium phenyl-2-pyridyl-
phosphate (380 mg, 1.32 mmol) and monomer 3 (303 mg, 0.66
mmol) were dissolved in methanol (20 mL). The potassium salt
was used since NaCl is known to have a relatively high solubility
in methanol. After filtering off the KCl, an equivalent of CuCl2
(88 mg) was dissolved in this solution under stirring for 10 min
and the solvent was removed in vacuo to yield the Cu(II) complex
of functional monomer 3 and template 5. The complex obtained
was dissolved in a mixture of MeCN (5.0 mL), DMSO (5.0 mL),
EDMA (4 g), MMA (0.5 g), and AIBN (50 mg). For solubility
reasons we selected a mixture of solvents MeCN/DMSO (1:1, v/v)
as porogen. The reaction mixture was transferred to a 25 mL thick-
walled ampule and was degassed by three freeze-thaw cycles under
vacuum and then sealed. The polymerization was performed at 60
°C for 3 days. After crushing, the polymer was washed with
methanol and dried under vacuum. The crushed polymers were
ground using a mill and sieved to obtain 45-250 µm particles.
The template was removed by exhaustive washing with a mixture
titrations were made for each system. The value of Kw
)
[H+][OH-] at 25 °C and I ) 0.1 (NaNO3) was determined to be
10-13.79. As a comparison, tris(2-aminoethyl)amine was used as a
control (see Figure 3 in Results and Discussion).
2.4.4. 1H NMR Titration. Mixtures of different molar ratios
(0.2-15) of the phosphate (DPP) 4 to the functional monomer 2
in CDCl3 or MeCN-d3 were prepared in 10 NMR tubes and
analyzed. After equilibration for 20 min, the relative changes in
the chemical shifts of 4 were plotted as a function of the molar
ratio of 2 to 4. The titration experiments yielded a saturation
isotherm (Figure 1). Nonlinear regression analysis led to apparent
complex constants (Ka). The stoichiometry of the complex formation
between 4 and the functional monomer 2 was determined with a
Job plot (Figure 2).
2.4.5. Determination of the Amidinium Groups and Their
Complexing Ability in MIPs. The content of the amidinium groups
in the polymers was determined by the complex formation of
diphenylphosphate (4) with the amidinium groups in MeCN solution
9
8046 J. AM. CHEM. SOC. VOL. 130, NO. 25, 2008

Functional Mimicry of Carboxypeptidase A
A R T I C L E S
hydrolysis of 8 by the imprinted polymer PCu3,5 and by the control
polymer is shown in Figure 5 (Results and Discussion). From that,
Michaelis-Menten kinetics were calculated from the double-
reciprocal (Lineweaver-Burke) plots for the hydrolysis of 8
catalyzed by PCu3,5 or by fitting the Michaelis-Menten kinetics
to a hyperbola using the program Origin 7.0. The competitive
inhibition of the hydrolysis by the template 5 was carried out by
adding different amounts of inhibitor 5. The molar ratios of the
active sites of the MIP PCu2,5 to 5 were 1:0, 2:1, and 1:10 (Figure
6).
3. Results and Discussion
Figure 2. Job plot of the complexation of 2 and 4 by 1H NMR titration as
in Figure 1. XDPP ) mole fraction of diphenylcarbonate 4 (DPP), XDPPd )
mole fraction X multiplied by ∆δ.
3.1. Rational Design and Synthesis. Enzymes are essentially
flexible molecules evolved over a very long time; they are
precisely complementary to the reactants in their activated
transition state geometry.12 Although this notion stated by
Pauling 60 years ago has gradually come to be accepted, debates
still continue about the precise nature of enzyme catalysis.13
To elucidate the important role of transition state stabilization
in enzyme catalysis, two major strategies have been developed
and they actually represent important progress in the present
understanding of enzyme mechanisms. In this regard the groups
of Lerner and Schultz independently showed the first successful
examples by generating antibodies against a stable transition
state analogue (TSA). These antibodies show considerable
catalytic activity.3 The other remarkable development is the
preparation of polymers imprinted by TSAs which display a
significant catalytic capacity.5–9 Enzyme-like activity by transi-
tion state stabilization was thus clearly demonstrated in two
different systems.3,5
(see ref 11). Polymer samples (10 mg each) were suspended in
MeCN solution (1 mL) in screw-cap vials, after addition of different
amounts of 4 (resulting in 0.28, 0.7, 1.4, 2.8, 4.2, and 8.4 mM
solutions), the vials were stirred at room temperature for 24 h, then
the polymer was removed by filtration, and the remaining 4 in the
solution was determined by UV spectroscopy at 260 nm (ε ) 5.79
× 102 M-1 cm-1). The amidinium group content and the
complexation constants of the diphenylphosphate (4) with the
amidinium binding site can be analyzed with nonlinear regression
of the plot of the concentration of occupied binding sites against
the concentration of guest 4. The amidinium group content can also
be determined by N-elemental analysis.
2.4.6. Determination of Cu2+ Content in the MIPs. The
content of copper ions in the imprinted polymers was determined
by elemental analysis for copper. The copper content can also be
calculated from the triamine content. In this case, the triamine
moiety in the polymer was determined by N-elemental analysis,
and the copper content was calculated to be equimolar to the
triamine content assuming that the triamine moiety and Cu2+ forms
a 1:1 complex in MIPs. Both methods gave equivalent values.
2.4.7. Hydrolyses of Substrates. The hydrolyses of the carbon-
ates 6, 7, and 8 were carried out in a 1:1 solution of 50 mM HEPES
buffer (pH 7.3) and MeCN at 20 °C. The formation of phenol or
pyridone was followed by HPLC (Waters 410 pump, Waters 486
UV-detector, integration software CSW from Apex Data) analysis
on a reversed-phase column (RP-18, Merck). The inherent reac-
tivities of the three carbonate substrates against water at pH 7.3
are quite different. Thus the kuncat of 6, 7, and 8 were found to be
7.20 × 10-7, 2.60 × 10-5, and 2.55 × 10-4 min-1, respectively.
In a typical experiment for the hydrolysis of di-(2-pyridyl)carbonate
(8), 0.1-1 mg of dry polymer was added to 1 mL of a
buffer-MeCN solution in a screw-cap vial, and the mixture was
stirred at 20 °C for several hours. After equilibration, the reaction
was initiated by injecting 20 µL of 0.5 mM 8 in dry acetonitrile
(the final concentration of 8 was 0.01 mM), and 50 µL of the
reaction mixture were taken at regular intervals and quickly filtered
to remove the polymer and analyzed by HPLC. Considering the
effect of the pH on the products, MeCN/potassium phosphate buffer
0.01 M (70:30 v/v) at pH 7.4 was used as the mobile phase. The
flow rate was 1.5 mL/min. The UV detection was performed at
218 nm. Pseudo-first-order rate constants k for the hydrolysis were
determined from the slopes of semilogarithmic plots of the reaction
progress against the time.
In our previous work8 in designing polymer catalysts by
molecular imprinting, the functional amidinium group N,N′-
diethyl-4-vinylphenylamidine (1) was rationally positioned by
using phosphonate or phosphates as templates. In the hydrolysis
of esters, carbonates, and carbamates, rate enhancements of
100-3000-fold were obtained using TSA imprinted polymers.
The amidinium group acted as a transition state binding site
similar to the catalytic action of Arg 127 in carboxypeptidase
A (CPA) catalysis.
On the basis of these positive results for TS stabilization, we
made further efforts to adjust the active site of our molecularly
imprinted polymers (MIPs) to that of carboxypeptidase A (CPA).
For these experiments, carbonate hydrolysis was selected as the
model reaction (Scheme 1). Stable analogues of the structurally
well-characterized tetrahedral transition states for this reaction
have been widely used in the preparation of the corresponding
catalytic antibodies.3a,b In our case the functional monomer 2
was chosen since it is expected to form a stable complex with
the phosphate group of template 5. An additional triamine group
was introduced with a defined proximity to the amidinium group.
This group gives rise to a strong 3-fold coordination of Zn2+
or Cu2+ ion leaving the fourth or fifth coordination site free for
other ligands.9 The monomer 2 was synthesized as shown in
Scheme 2 (see Supporting Information9b). Thus, the main
catalytic factors of CPA, an amidinium (instead of a guani-
dinium) group and a Zn2+center, were introduced in a TSA
imprinted polymer cavity. The imprinted polymers exhibited
strong rate enhancements kcatt/ksoln for carbonate hydrolysis by
2.4.8. Kinetics. For the determination of the Michaelis-Menten
kinetics, plots of initial velocities of the catalyzed hydrolysis versus
substrate concentration were measured. In these investigations the
initial reaction rates were monitored while maintaining a constant
concentration of active sites and increasing substrate concentrations.
The procedure to measure the velocity was analogous to pseudo-
first-order kinetics as described. The saturation plots of the
(12) (a) Pauling, L. Chem. Eng. News 1946, 24, 1375–1377. (b) Pauling,
L. Nature 1948, 161, 707–709. (c) Jencks, W. Catalysis in Chemistry
and Enzymology; McGraw-Hill: New York, 1969.
(13) (a) Warshel, A. J. Biol. Chem. 1998, 273, 27035–27038. (b) Garc´ıa-
Viloca, M.; Gao, J.; Karplus, M.; Truhlar, D. G. Science 2004, 303,
186–195. (c) Gao, J. Curr. Opin. Struct. Biol. 2003, 13, 184–192.
(11) Jang, B.-B.; Lee, K.-P.; Min, D.-H.; Suh, J. J. Am. Chem. Soc. 1998,
120, 12008–12016.
9
J. AM. CHEM. SOC. VOL. 130, NO. 25, 2008 8047

A R T I C L E S
Liu and Wulff
Figure 3. Potentiometric pH titration curves. (A) (a) 1 mM 2·4H+ at I ) 0.10 (NaNO3), and (b) 1 mM 2·4H+ + 1.00 mM CuCl2 at I ) 0.10 (NaNO3).
(B) For comparison the potentiometric pH titration curves are given for (a) 1 mM tris(2-aminoethyl)amine ·4H+ at I ) 0.10 (NaNO3), and (b) 1 mM
tris(2-aminoethyl) amine ·4H+ + 1.00 mM CuCl2 at I ) 0.10 (NaNO3). R(OH-) represents the number of moles of base (0.1 mM NaOH) per mole of ligand.
Figure 4. Plot of the concentration of the occupied binding sites versus
that of the guests. Measured is the complexation of 4 (DPP) to the template-
free imprinted polymer PCu2,5 at 25 °C. The concentration of the complexes
was calculated by spectrophotometrically detecting the uncomplexed 4
(DPP) in MeCN (λ ) 260 nm, ꢀ ) 5.79 × 102 M-1) .
Figure 6. Double-reciprocal plots (Lineweaver-Burke) for the catalyzed
hydrolysis of 8 and the inhibition by the template. Inhibition studies of the
hydrolysis of 8 in presence of PCu2,5 using the original template 5 as an
inhibitor in a 1:1 solution of HEPES buffer (pH 7.3) and MeCN at 20 °C.
The molar ratio of the active sites to the inhibitor was (9) 1:0; (2) 2:1;
(b)1:10. The inhibition constant was calculated to be Ki ) 25 µM.
monomer 2; the monomer 3 could be obtained by reaction of 1
equiv of tris(2-aminoethyl)amine with 2 equiv of N-ethyl-4-
vinylbenzocarboximide acid ethyl ester.
Initially, diphenylphosphate (4) was used as template for
catalysts for the carbonate hydrolysis.8b The new template 5
proved to be much more efficient.9 One phenyl ring in 4 was
replaced by a 2-pyridyl ring in 5. This change gives the
possibility of incorporating the pyridine nitrogen atom as the
fourth ligand in the Cu2+coordination sphere (see Scheme 4).
The additional coordination permits the formation of more stable
complexes between the functional monomer and the template
during polymerization. Furthermore, the amidinium group and
the Cu2+ ion are brought into defined proximity to each other
and to the template molecule in the cross-linking polymerization
during the imprinting process. The template 5 was synthesized
as shown in Scheme 3. The attempt to prepare the new template
di-2-pyridyl phosphate failed owing to its instability.
Figure 5. Michaelis-Menten plot of the hydrolysis of 8 by the imprinted
polymer PCu3,5 measured with PCu3,5 (0.051 mM of available cavities)
in buffer solution (HEPES/MeCN 1:1) (9) and by the nonimprinted polymer
CPCu3 (0.098 mM of available Cu2+ in the polymer) (b), pH ) 7.3 at
20 °C.
a factor of 6900. By the replacement of Zn2+ by Cu2+ in the
imprinting procedure, a further dramatic catalytic rate enhance-
ment of 110 000-fold was observed.9a
For further improvement of the catalysis, the new monomer
3 was designed to introduce one Cu2+ center and two amidine
groups in one imprinted cavity (Scheme 2). This design changed
the binding of the substrates and, more importantly, the type of
coordination of the copper in the active site as well as the
reactivity of the copper bound HO- and H2O. This should
increase the binding and catalysis in the imprinted cavities. The
novel design led to further enhancement of the catalytic activity.
The functional monomer 3 was synthesized in a similar way
as 2 (Scheme 2). N-Ethyl-4-vinylphenylamide is treated with
triethyloxoniumtetrafluoroborate to yield N-ethyl-4-vinylben-
zocarboximide acid ethyl ester. Reaction of this compound with
an equimolar amount of tris(2-aminoethyl)amine yielded the
The reactivity of the substrates varied clearly with the
template structure used during imprinting. In our earlier study,9b
the template 5 and the substrate 6 were first used in the catalysis
with the imprinted polymers. In order to design a substrate that
is more similar to the template, the substrate phenyl-(2-
pyridyl)carbonate (7) and di-(2-pyridyl)carbonate) (8) were
synthesized.9a This led to a better binding of the substrate in
the imprinted cavity compared to 6, since the nitrogen of the
pyridyl ring can interact with the Cu2+ center similarly to the
template 5 described above. The synthesis of 8 is readily
achieved by the reaction of triphosgene with 2-hydroxypyridine.
Substrate 8 is inherently more reactive toward water than 7;
substrate 6 shows the lowest reactivity.
9
8048 J. AM. CHEM. SOC. VOL. 130, NO. 25, 2008

Functional Mimicry of Carboxypeptidase A
A R T I C L E S
Scheme 1. Hydrolysis of Carbonate
Scheme 2. Synthesis of Functional Monomers
complexes. The observed large change in the chemical shift on
complexation implied a strong binding in the presence of Cu2+
,
but at higher concentrations, the complexes partially precipitated
from the solution. Therefore, it was not possible to determine
with some accuracy the association constant of 2 and 5 in the
1
presence of Cu2+ by H NMR titration. However, it is to be
expected that the complexation of 2 with 5 in the presence of
Cu2+ is much stronger than that of 2 and 5 alone owing to the
strong binding of the Cu2+ ion with the pyridine ring of template
5.
The complexation of Cu2+ ions with the triamine part of
monomer 2 was characterized by potentiometric titration in
aqueous solution (Figure 3A). From the potentiometric titration
a high stability constant of 1015 M-1 was found, a value that is
in accordance with other observations in similar systems.16 In
the case of the reference substance, tris(2-aminoethyl)amine, a
smooth and clear titration curve in the presence of Cu2+ showed
that four protons were neutralized at R(OH-) ) 4, implying a
tetra coordination of the tetraamine for Cu2+ (Figure 3B).
However, the curve for the complexation of monomer 2 is quite
different (Figure 3A,b). The complexation below R(OH-) ) 3
is similar to the foregoing example, but above R(OH-) ) 3,
the curve is different and indicates partial neutralization of other
kinds of protons. This result reflects a strong tricoordination of
the Cu2+ with the triamine part of monomer 2. In comparison
with similar known copper complexes2e,f a pentacoordinated
copper with two additional oxygen ligands (OH-, H2O) can be
assumed. The exact stereochemistry with regard to the position
of the ligands in the complex is not known. An X-ray
crystallographic investigation should bring more information in
this respect. Other types of coordination of Cu2+ are much
weaker. It is known that only a weak coordination of Zn2+ to
amidinium groups of similar systems is occurring.17 A weak
binding of the copper to the amidinium ion would facilitate the
complexation of the amidinium ion to the phosphate template.
On the basis of these binding studies and considering the
solubility, the copper ion complexes for molecular imprinting
were prepared in MeCN/DMSO (1:1, v/v), although in this
mixed solvent lower association of phosphate with amidinium
was observed compared to pure MeCN.
Scheme 3. Synthesis of the Template 5
3.2. Characterization of the Complexation of the Functional
Monomers with the Template. Stoichiometric noncovalent in-
teraction combines the advantages of covalent and noncovalent
interaction during the molecular imprinting procedure and
generates high association constants between TSA and the
functional monomer.5a As shown in previous studies,5a,14 the
amidinium group can bind strongly to carboxylic acids, phos-
phonic acids, and phosphates under imprinting conditions. This
fact allows fixing a stable transition state analogue template
through multiple hydrogen bonds and electrostatic interactions
and to position the amidine groups in a correct orientation for
catalysis. For the characterization of this type of complexation,
1H NMR spectroscopy has proven to be an excellent tool.14 The
formation of the complex of the functional monomer and the
1
template (for example, between 2 and 4) was studied by H
NMR titration experiments yielding a saturation isotherm (Figure
1). Changes in the chemical shift of the benzene ring protons
(δ 7.23) of the template were used as a reference for electrostatic
and hydrogen interaction between the functional monomer and
the template. Job plot analysis14,15 of the complexation dem-
onstrated a 1:1 complexation between the functional monomer
and the template (Figure 2). The apparent association constant
of 2 and 4 was determined to be 5.3 × 103 M-1 in acetonitrile-
d3 (MeCN) by 1H NMR spectroscopic titration; a similar result
was also observed in CDCl3 (5.6 × 103 M-1). In the case of
the complexation of 1 and 4 the association constant was
determined to be 4.6 × 103 M-1 in MeCN at 25 °C.14 The
complexation of 2 and 5 showed an association constant of 1.1
× 104 M-1 in acetonitrile; it is somewhat higher than that
between 2 and 4. The binding of the functional monomer 2 and
5 in the presence of Cu2+ can also be observed in acetonitrile
3.3. Synthesis and Characterization of Imprinted Polymers.
Highly cross-linked molecularly imprinted polymers were
prepared in bulk by standard procedures using the TSA 5 as a
template, 2 or 3 as functional monomers, and ethylene
dimethacrylate (EDMA) as a cross-linker in the presence of
Cu2+ (Scheme 4a and c). Some methyl methacrylate is added
in order to improve the flexibility of the resulting polymer. The
copper monomer-template complex is prepared by dissolving
the potassium salt of the template 5, the monomer 2 or 3, and
an equimolar amount of CuCl2 in methanol. Potassium(I) in 5
is replaced by copper(II), and KCl precipitates. For solubility
reasons the porogen has to contain a certain amount of a polar
1
by H NMR at a relatively low concentration of the copper
(14) Wulff, G.; Knorr, K. Bioseparation 2002, 10, 257–276.
(15) (a) Takeuchi, T.; Dobashi, A.; Kimura, K. Anal. Chem. 2000, 72, 2418–
2422. (b) Striegler, S. Bioseparation 2001, 10, 307–314.
(16) Zompa, L. J. Inorg. Chem. 1978, 17, 2531–2536.
(17) Kimura, E.; Shiota, T.; Koike, T.; Shiro, M.; Kodama, M. J. Am. Chem.
Soc. 1990, 112, 5805–5811.
9
J. AM. CHEM. SOC. VOL. 130, NO. 25, 2008 8049

A R T I C L E S
Liu and Wulff
Scheme 4. Schematic Representation of the Preparation and the Catalytic Centers of Imprinted Polymer Catalystsa
a (a) Molecular imprinting with functional monomer 2 and the template 5 in presence of Cu2+ for the preparation of the imprinted polymer catalyst
PCu2,5. (b) The enzyme-like active site of PCu2,5 prepared after removal of the template 5. (c) Molecular imprinting with functional monomer 3 and two
molecules of the template 5 in presence of Cu2+ for the preparation of the imprinted polymer catalyst PCu3,5. (d) The enzyme-like active site of PCu3,5
after removal of the template 5.
solvent such as DMSO or DMF. Therefore, the solvent acting
as a porogen consisted of a mixture of acetonitrile and DMSO
(1:1, v/v); in other experiments either pure DMSO or DMF was
used. It is interesting that the copper content in the polymeri-
zation mixture does not significantly inhibit the radical polym-
erization. After polymerization and grinding, the templates were
removed from the polymer particles (45-125 µm) to give ∼75%
free cavities (Scheme 4b). The control polymers with identical
composition to that of the imprinted polymer but with formic
acid instead of the template were synthesized in an analogous
manner. The method mentioned above was demonstrated to be
an efficient one for constructing Cu(II)-containing active cavities
(Scheme 4).
In order to compare the catalytic activity of different metal
ions such as Cu(II) and Zn(II), imprinted catalysts were prepared
from both metals. The catalytic activity proved to be very
different, and the question arose whether this is due to special
imprinting effects of the metal ions or if this is mainly dependent
on the inherent catalytic property of the metal ions themselves.
To investigate this question an alternative strategy for the
preparation of Cu(II)-containing imprinted polymers was de-
veloped by substitution of Zn(II) by Cu(II) in Zn(II)-imprinted
polymers (Scheme 5). The Zn(II)-containing imprinted polymer
PZn2,5 was synthesized as described previously.9b The Zn(II)
was removed by a 10-fold excess of EDTA·2Na+ and reloaded
by Cu(II). The Cu(II)-containing, blue colored polymer
P(Zn)2,5Cu was obtained. Another model system containing
two amidinium groups (corresponding to two guanidinium
groups in carboxypeptidase A) was obtained using a complex
of the functional monomer 3 with 2 equiv of template 5 in the
presence of Cu(II) (Scheme 4c, d). The imprinted polymer
PCu3,5 was prepared analogously as described for PCu2,5. The
active site in the imprinted polymer is shown in Scheme 4d.
The surface areas of the prepared polymers were determined
by BET analysis (N2, one point measurement), and high inner
surface areas typical for macroporous polymers were obtained.
Comparable inner surface areas for the imprinted and the
nonimprinted control polymers were observed, for example, 310
m2/g for PCu2,5 and 290 m2/g for CPCu2,5; 390 m2/g for
PCu3,5 and 320 m2/g for CPCu3,5. This shows that the
polymer structure of imprinted and control polymers are quite
similar, and a comparison of the catalytic properties should
mainly rely on the imprinting effect of the active site. From
comparison of polymers prepared from monomer 2 and mono-
mer 3, the polymers from the latter show a somewhat higher
inner surface area for both imprinted and control polymers. The
reason may be that 3 is also a cross-linker and may influence
the polymer structure to some extent. A higher surface area
results in a better accessibility and may also influence the
catalytic activity. This assumption can be made since from
earlier investigations it is known that the porosity of the
polymers prepared under these conditions allows a free diffusion
within the pores.6a
The content of copper(II) in the imprinted polymers was
estimated by elemental analysis. No loss of Cu(II) ions was
found during splitting off the template from the imprinted
polymers; similar results had already been observed in the case
of Zn(II)-imprinted polymers.9b This observation was further
supported by the fact that the content of triamine groups
(calculated from elemental analysis) is nearly identical with that
of Cu(II) in the imprinted polymer (e.g., 0.126mmol/g triamine
for PCu2,5 and 0.131mmol/g Cu(II) for PCu2,5).
The content of amidinium ions in the polymers was deter-
mined by measuring the extent of complexation of the ami-
dinium ions with diphenylphosphate (DPP, 4) in MeCN solution.
This method was previously used in similar nonimprinted
systems by Suh et al.11 The plot of the complexation of 4 to
the imprinted polymer PCu2,5 is shown in Figure 4. From the
analysis of the data with a nonlinear regression, the content of
amidinium ions can be calculated. For example, 0.11 mmol/g
9
8050 J. AM. CHEM. SOC. VOL. 130, NO. 25, 2008

Functional Mimicry of Carboxypeptidase A
A R T I C L E S
Scheme 5. Preparation of Imprinted Polymers by Substitution of Zn(II) by Cu(II) in Zn(II)-Imprinted Polymersa
a (a) Molecular imprinting with functional monomer 2 and the template 5 in presence of Zn(II) for the preparation of Zn(II)-imprinted polymer catalyst
PZn2,5. (b) The imprinted cavity of PZn2,5 after removal of the template; (d) Removal of Zn(II) from PZn2,5 by EDTA in a mixture of DMF and water
(1:1, v/v). (c) Reconstructed metal center with a Cu(II) in an originally Zn(II)-imprinted polymer (PZn2,5Cu).
amidinium groups in polymer PCu2,5 and 0.12 mmol/g
amidinium groups in PCu3,5 were obtained. These values
represent the available amidinium ions in the polymer. The total
content of amidinium groups can also be calculated from the
elemental analysis; for PCu2,5 the content of amidinium groups
is 0.136 mmol/g. Considering ∼75% free cavities in the
imprinted polymer, this analysis is in good agreement with the
results of the complexation of amidine with diphenylphosphate
4. The association constant (Ka) for the complex formed between
4 and the active site of the imprinted polymer PCu2,5 was also
estimated from the above experiments (Ka ) (1.65 ( 0.5) ×
104 M-1) This association is considerably higher compared to
that for the complexation of the monomer 2 with 4 in MeCN
(Ka ) (5.3 ( 0.6) × 103 M-1). The result is not surprising
considering both binding of the amidinium ion and a selective
recognition derived from the imprinting effect (although 4 is
similar to but not identical to the template).
3.4. Catalytic Activity of the Imprinted Polymers. The
catalytic activity of the imprinted polymers was determined by
investigating the rate of hydrolysis in the presence of these
polymers for different carbonates in a 1:1 solution of HEPES
buffer (pH 7.3) and MeCN. In order to follow the reaction
kinetics, aliquots were taken at regular intervals and were
analyzed by HPLC on a C18-reversed-phase column. Phenol and
2-hydroxypyridine, as the reaction products, were determined
quantitatively. The pseudo-first-order rate constants were mea-
sured at low conversion from the slopes of a semilogarithmic
plot of reaction progress against time.9 kimpr is the pseudo-first-
order rate constant in the presence of the imprinted catalyst,
kcontr, in the presence of the control polymer, and ksoln, without
catalyst just in solution of HEPES buffer (pH 7.3)/MeCN 1:1.
As in previous studies9 the activity of the catalyst was expressed
by the ratio of reaction in the presence of polymer catalyst to
that in neat buffer solution (kimpr./ksoln). Effects that are only
connected to the imprinting in the polymer can be elucidated
from the ratio kimpr/kcontr; this is designated as the imprinting
factor.
It had been shown9 that the introduction of metal ions into
the TSA-imprinted active site resulted in a strong increase in
catalytic activity. Thus the value of 455 for kimpr/ksoln for the
hydrolysis of diphenylcarbonate (6) in the presence of P1,48b
increased on introduction of Zn2+ (PZn2,5) by a factor of 7
and kimpr/kcontr by a factor of 6.9b The increase was even stronger
in the case of introduction of Cu2+ (PCu2,5). kimpr/ksoln increased
by a factor of 18. It is interesting that additional “free” Cu2+ or
Zn2+ ions do not have a significant influence on the rate. This
is also true if these metal ions are added to the neat buffer
solution.
Now the catalytic activity of imprinted polymers prepared
with two different methods has been compared for the hydrolysis
of phenyl-2-pyridyl-carbonate (7). This substrate is more similar
to the template used during imprinting, and therefore the
catalytic activities of the imprinted catalysts toward 7 are
considerably higher. In Table 1 the catalytic activity of Cu(II)
and Zn(II) imprinted catalysts are compared. The Zn(II)
imprinted catalyst showed a kimpr/ksoln of 4040, and the Cu(II)
imprinted catalyst, that of 15700. Also the imprinting factors
are quite high with 63.2 and 76.9, respectively. If the polymer
is imprinted with Zn(II), and then the Zn(II) is substituted by
Cu(II), the catalytic activity was lower as in the case of direct
imprinting with Cu(II) (Table 1). This result is consistent with
the complexing ability of the two systems during imprinting.
The Cu(II) complex of the triamine part of monomer 2 showed
a very high association constant of 1015.8 M-1, compared to
109.3 M-1 between triamine of 2 and Zn(II).9a Furthermore,
Cu(II) also displays a stronger capacity to complex the pyridine
9
J. AM. CHEM. SOC. VOL. 130, NO. 25, 2008 8051

A R T I C L E S
Liu and Wulff
-1
Table 1. Pseudo-First-Order Kinetics of the Hydrolyses of
Carbonates 7 and 8 in Presence of Different Catalytically Active
Imprinted Polymers Containing Zn(II) and Cu(II)
Table 3. Comparison of Michaelis-Menten Kinetics and Ktx
(Catalytic Proficiency) of Hydrolyses of Carbonate 8 with Imprinted
Polymers and Control Polymers
-1
imprinted
polymera
kimpr
(min-1 b
kcat
(min-1
Km
(mM)
k
cat/Km
Ktx
(M-1 b
substrate
kimpr/ksoln
kimpr/kcontr
polymera
kcat/kuncat
)
)
(min-1 M-1
)
)
PZn2,5
7
7
7
8
7
8
0.105
0.241
0.41
13.0
2.56
4040
9250
15 700
51 000
98 200
217 000
63.2
57.8
76.9
80.1
50.2
53.8
PCu2,5
CPCu2
PZn2,5Cu
CPZn2Cu
PCu3,5
28.0
0.58
6.10
0.73
6.15
0.36
4.16
110 000
1450
40 100
1380
413 000
4520
48 200
61
14 000
57
292 000
276
1.89 × 108
2.40 × 105
5.50 × 107
2.24 × 105
1.15 × 109
1.08 × 106
PZn2,5Cu
PCu2,5c
PCu2,5c
PCu3,5
PCu3,5
0.37
10.2
0.35
105
1.15
55.3
CPCu3
a The imprinted polymer PZn2,5 was prepared using monomer 2 and
template 5 in the presence of Zn(II); PZn2,5Cu was obtained from
PZn2,5 by substitution of Zn(II) by Cu(II); PCu2,5 or PCu3,5 was
prepared using monomer 2 or 3 and template 5 in the presence of
Cu(II). b Hydrolyses of carbonates 7 or 8 in a solution of 50 mM
HEPES buffer (pH 7.3)/MeCN 1:1 at 20 °C (HEPES ) 2-[4-(2-
hydroxyethyl)-1-piperazine]ethanesulfonic acid). There are 2 mM of
available active sites in relation to 1 mM substrate; all k values are the
average of at least three measurements. c Data from ref 9a.
a The control polymers CPCu2 and CPCu3 were prepared in the same
manner as PCu2,5 and PCu3,5, only the template 5 was substituted by
-1
formic acid. b Ktx ) (kcat/Km)/kuncat (catalytic proficiency).18
as a porogen failed owing to the poor solubility of the Cu(II)
complex of functional monomer and template in MeCN. The
reason for the poor efficiencies of the catalysts prepared in a
polar solvent like DMF and DMSO is a twofold one. First, the
interaction among monomer, template, and metal ion is less
pronounced compared to nonpolar solvents. Second, in these
solventssomewhatdifferentphysicalpropertiesofthemacroporous
polymers are observed; thus the stability of the active sites with
regard to shape and arrangement of functional groups is less
pronounced. Both effects lead to less precisely imprinted active
sites and therefore lower catalytic efficiency.
The highest catalytic enhancements were achieved using the
imprinted polymer PCu3,5 prepared with the newly designed
monomer 3 (see Table 1). In this polymer two amidinium
moieties were introduced in a certain proximity to one Cu(II)
center in the imprinted cavity (see Scheme 4c, d). This imprinted
polymer PCu3,5 shows the strongest catalytic activity with a
rate enhancement for kimpr/ksoln of 98 200-fold for the hydrolysis
of substrate 7; in the case of substrate 8 a further rate
enhancement to an extremely high value of 217 000-fold was
observed.
Table 2. Pseudo-First-Order Kinetics of the Hydrolyses of
Carbonate 7 by Imprinted Polymers Prepared by Using Different
Porogens
imprinted
polymer
kimpr
(min-1 a
porogen
kimpr/ksoln
kimpr/kcontr
)
PCu2,5b
PCu2,5
PCu2,5
PCu2,5
PCu2,5
MeCN/DMSO(1:1)
MeCN/DMSO(1:2)
DMSO
MeCN/DMF (1:1)
DMF
0.41
15700
7451
3220
3630
781
76.9
34.5
15.8
17.8
0.195
0.085
0.095
0.020
3.96
a Hydrolysis of carbonate 7 in a solution of 50 mM HEPES buffer
(pH 7.3) /MeCN 1:1 at 20 °C. There are 2 mM of available active sites
in relation to 1 mM substrate. kimpr is the pseudo first-order rate constant
in presence of the polymeric catalyst, kcontr is the rate constant in
presence of the control polymer and ksoln the rate constant in the HEPES
buffer (pH 7.3)/MeCN 1:1 solution, all values are means of at least
three times. b Data from ref 9a.
moiety of 5 than Zn(II). This result shows that the higher
catalytic activity is dependent not only on the better catalytic
properties of Cu(II) but also on the better imprinting properties
of the copper ion. Thus, the stable complexation between
monomer and template provided a better defined conformation
of the complex and a better binding during the imprinting
procedure. A strong binding of the TSA to the imprinted catalyst
during the reaction would exhibit high catalytic efficiency for
the reaction. This result is in good agreement with the notion
regarding TS stabilization in enzymatic catalysis. Furthermore
it shows that similarly, as with Zn2+, the copper ions do not
appreciably interfere in the radical polymerization.
In order to obtain more information on the effect of the
porogen used during the preparation of the imprinted polymers
on the catalytic activity, different solvents were used as reaction
medium as shown in Table 2. Different porogens led to a
remarkable change of the catalytic activity of the imprinted
polymers. Using polar solvents as the porogen, poorer catalytic
activities and imprinting effects were observed. For example,
in pure DMF an imprinting effect of 3.96 for PCu2,5 and low
catalytic activity were obtained. On addition of MeCN to the
DMF using a mixture of solvents the properties improved
somewhat. DMSO as a solvent seems to be better; a relatively
high imprinting factor of 15.8 was obtained with pure DMSO
as a porogen. This is considerably improved by using a mixture
as a porogen with added MeCN. In the case of a 1:1 mixture of
MeCN and DMSO as the porogen, a high imprinting factor of
76.9 was observed (Table 2). The attempt to use pure MeCN
As with natural enzymes the catalysis of the carbonate
hydrolysisbyimprintedpolymersfollowstypicalMichaelis-Menten
kinetics (Figure 5). A plot of initial velocities of the reaction
versus the substrate concentration shows at first an increase
in the rate of reaction with increasing substrate, and the rate
then levels off. At a higher substrate concentration, when all of
the active sites are occupied, the rate remains constant which
means it is zero order with respect to carbonate concentration
(saturation kinetics). The turnover number kcat and the Michaelis
constant Km can be calculated from the double reciprocal plots
of the initial rate versus the substrate concentration (Lineweaver-
Burke plot) or by fitting the Michaelis-Menten curve to a
hyperbola using the program Origin 7.0. Strong turnover
numbers have been observed for all imprinted polymers with a
catalytic center of Cu(II). The kcat is 2.86 min-1 for hydrolysis
of 7 by PCu2,5 and 28.0 min-1 for 8; in the case of PCu3,5
with two amidinium moieties in proximity to the Cu(II) center,
a remarkable turnover number for the hydrolysis of 8 of kcat
)
105 min-1 was obtained (see Table 3).
kcat/kuncat is usually used to express the catalytic activity of
catalytic antibodies and natural enzymes. PCu2,5 showed a kcat/
kuncat of 110 000 for the hydrolysis of 8, a rather high value for
a synthetic enzyme model. PCu3,5 showed an even higher rate
enhancement of 413 000-fold for the hydrolysis of 8; it is by
far the highest value reported until now for molecularly
(18) (a) Miller, B. G.; Wolfenden, R. Annu. ReV. Biochem. 2002, 71, 847–
885. (b) Zhang, X.; Houk, K. N Acc. Chem. Res. 2005, 38, 379–385.
9
8052 J. AM. CHEM. SOC. VOL. 130, NO. 25, 2008

Functional Mimicry of Carboxypeptidase A
A R T I C L E S
workers.2f Their catalysts correspond to our nonimprinted control
systems such as CPCu2.
A unique bifunctional catalysis (similar to that postulated by
Breslow2d for the enolization of ketones) proceeds in the
imprinted polymers such as PCu2,5 via a binding of the pyridyl
ring of 7 or 8 to the copper ion and an activation of the carbonyl
group by the protonated amidinium ion followed by the attack
of the copper bound hydroxyl group to the carbonyl group. The
reaction is further accelerated by the stronger binding of the
tetrahedral transition state compared to the original substrate.
The mechanism in the imprinted polymer PCu3,5 is more
complicated. Two amidinium moieties are oriented in a certain
proximity to one Cu2+ center (see Scheme 6). The coordination
of the copper ion is therefore quite different compared to
PCu2,5. Whereas in the active site of PCu2,5 (Scheme 4b)
copper undergoes three coordinative bonds to amines, in PCu3,5
there are only two coordinative bonds to amines; this influences
the catalytic activity of the additional groups bound to copper
like H2O and OH-. During binding of substrates, this becomes
even more complicated since the active site can bind either one
or two molecules of substrate. In the case of binding two
molecules of substrate 7, copper will be coordinated to four
nitrogens, and in the case of one molecule of substrate 7, copper
will be coordinated by three nitrogens. Coordination of copper
to the nitrogens of the amidines is not to be expected since there
complexation is much lower as the potentiometric pH titrations
also show. Usually catalysts like enzymes are used under
subsaturation conditions, meaning that mixed mechanisms are
to be expected with one or two substrates per active site. One
possible route for the catalysis with the imprinted polymer
PCu3,5 proceeds via binding of two substrates in one cavity as
shown in Scheme 6a and an activation of the carbonyl group
by the protonated amidinium ion, followed by attack of the
copper bound hydroxyl group. The reaction is further accelerated
by the preferred binding of the tetrahedral transition state
(Scheme 6b). It was shown that the carbonate hydrolysis can
be competitively inhibited by adding the TSA template. The
decease in Vmax and the effective inhibition constant (e.g., Ki )
25 µM, for the hydrolysis of 8 by PCu2,5 and inhibited by
template 5) indicates that cavities with high affinity for TSA
are responsible for rate enhancement (Scheme 6b). This is a
further proof for the effect of the molecular imprinting.
By achieving a combination of a preferred binding of the
transition state in the imprinted active sites and a proper
orientation of the catalytic moieties, a high catalytic efficiency
and proficiency of molecularly imprinted polymers have been
obtained. Rate enhancements (kcat/kuncat ) 4.1 × 105) are even
higher by more than 500-fold compared to catalytic antibodies
for which kcat/kuncat ) 810 has been reported for carbonate
hydrolysis.19 The examples reported here are the first examples
where the catalytic activity of molecularly imprinted polymers
obviously surpasses the activity of the catalytic antibodies
reported (generally the catalytic rate enhancements for catalytic
antibodies is on the order of kcat/kuncat ) 102 -104).3 However,
in natural enzymes kcat/kuncat can go up much higher with values
Figure 7. pH Profile of the Rate of Hydrolysis of 6 in Presence of 0.048
mM Active Sites of PCu2,5. Buffer solution/MeCN 1:1. Buffers used were
pH 5.5-6.5 MES; pH 7.3-8.1 HEPES; pH 8.4-9.5 CHES. k (10-1 min-1).
imprinted catalysts. The Michaelis constants Km demonstrate a
much better binding in the imprinted polymers compared to the
control polymers (see Table 3). By combination of substrate
recognition and catalysis the catalytic efficiency can be recog-
nized. The imprinted polymers exhibit by far better catalytic
efficiencies kcat/Km (min-1 M-1) compared to the control
polymers; for example, enhancement factors of 1056 and 790
for the hydrolysis of 8 by PCu3,5 and PCu2,5 were observed
in comparison to the controls CPCu3 and CPCu2, respectively.
However, for the hydrolysis of 8 by PZn2,5Cu which was
prepared by substitution of Zn(II) by Cu(II) from Zn(II)-
imprinted polymer PZn2,5, this value is significantly lower
(245). Obviously, the remarkable differences in the catalytic
efficiency relate to the accuracy of the imprinting procedure.
kcat/Km is an important measure for the efficiency of enzymes
and enzyme analogues. In order to include in the assessment of
enzymes the difference in the difficulty of the tasks that they
perform, it is necessary to know the reaction rate under
uncatalyzed conditions and relate the efficiency to this reaction
rate. Thus (kcat/Km)/kuncat has been introduced by Wolfenden18
and called the catalytic proficiency Ktx-1.The proficiency is
formally the equilibrium constant for the formation of the
complex between the transition state and the enzyme or enzyme
model. It therefore reflects the hypothetical binding affinity of
an enzyme for the transition state.18 Our new imprinted polymers
were shown to possess very high Ktx-1 values (108-109 M-1
)
in catalysis which are, as the efficiencies, significantly higher
by 2 to 3 orders of magnitude compared to the nonimprinted
control polymers (Table 3). These differences are remarkable
since the control polymers also contain the same catalytically
active functional groups. The comparison of Ktx-1 for imprinted
polymers and the control polymers unambiguously demonstrates
the stabilization of the transition state in the imprinted cavities.
-1
Interestingly, Ktx is higher for PCu3,5 compared to PCu2,5
which means the transition state is more firmly bound though
the imprinting factor is somewhat lower (50 instead of 80; see
Table 2).
With regard to the mechanism of catalysis for these imprinted
catalysts it is important that the active site of PCu2,5 contains
a copper aqua hydroxy form (see Scheme 4 b), which is the
catalytically most active form.2e This has been demonstrated
in our case by measuring the pH profile for the carbonate
hydrolysis in the presence of PCu2,5 (see Figure 7). It shows
a typical bell-shaped profile with an optimum at pH 7.2 . The
maximum rate is obtained when the aqua hydroxy form is at
its highest level. Similar catalysts with Cu2+ and guanidinium
ions for peptidase activity were investigated by Suh and co-
-1
of 106-1020. The catalytic proficiency Ktx in our examples
goes up to values of 108-109 M-1. Compared to natural
enzymes with 107-1023 M-1 our catalysts are approaching the
less proficient enzymes.18 The catalytic efficiency for our
catalysts (kcat/Km ) 3 × 105 min-1 M-1) is lower compared to
enzymes.
(19) Jacobs, J. W.; Schultz, P. G.; Sugawara, R.; Powell, M. J. Am. Chem.
Soc. 1987, 109, 2174–2176.
9
J. AM. CHEM. SOC. VOL. 130, NO. 25, 2008 8053

A R T I C L E S
Liu and Wulff
Scheme 6. Schematic Representation of the Binding and Catalysis of Substrates in the Cavities of the Imprinted Polymer PCu3,5a
a (a) Substrate 7 bound in the cavity of the imprinted polymer catalyst PCu3,5. (b) Intermediate in the catalysis by imprinted polymer catalyst PCu3,5
(in both cases with reactions of two substrates at the same time).
In the present enzyme model, the cavities of the molecularly
imprinted polymers are designed toward a single TSA template.
Therefore, the affinity induced by the TSA imprinting is used
preferentially for TS stabilization. This is in contrast to natural
enzymes which are evolved to recognize a series of structures
that connect the substrate and the multistep transition states
along the reaction pathway.20 Furthermore, enzymes are evolved
to have a flexible surface to stabilize not only the TS but also
the intermediate states in a multistep reaction.1,20 In comparison,
the cavities generated by molecular imprinting exhibit a more
rigid structure owing to the use of a high degree of cross-linking
agent in the polymerization process but still have some flexibility
and swellability. The relatively slow mass transfer inside the
molecularly imprinted polymers and the “polyclonality” (the
active sites prepared by imprinting with one template do not
all have exactly the same structure) of imprinted cavities are
also obstacles for further improvement of the catalysis. Although
this approach is intellectually promising, there still is a long
way to go to design molecularly imprinted catalysts that can
compete with naturally occurring enzymes with regard to their
activity and selectivity.
active catalysts prepared by the molecular imprinting strategy.
The catalytic activity, the catalytic efficiency, and the catalytic
proficiency are clearly surpassing those of the catalytic antibod-
ies and approaching those of the natural enzymes. This strategy
provides a promising tool to elucidate the role of transition state
stabilization in enzyme catalysis and shows direct evidence
supporting Pauling’s transition state stabilization theory. The
study demonstrates that molecularly imprinted polymers offer
an excellent possibility to mimic the active site of natural
enzymes not only by mimicking the shape of the transition state
but also for introducing suitable catalytically active groups into
the active site in a predetermined orientation. It is also possible
to introduce catalytic groupings that do not exist in nature and
which cannot be introduced by biochemical methods.
From these results some important information to improve
the catalytic efficiency of molecularly imprinted polymer
catalysts can be summarized: First, the affinity of molecularly
imprinted polymers for their related transition states should be
high. Second, the orientation of the catalytic moieties with
respect to the reactive groups of the substrates has to be
optimized. The high efficiency and selectivity together with the
strong chemical, mechanical, and thermal stability as well as
the relatively simple preparation method give these catalysts a
real advantage compared to catalytic antibodies and also provide
a good alternative compared to natural enzymes.
4. Conclusion
New models for the enzyme carboxypeptidase A were
prepared by molecular imprinting showing high catalytic
efficiency. This was achieved by a combination of molecular
imprinting using a stable transition state analogue as a template
and the introduction of amidinium functions and a Cu2+ binding
site in the active site in a predetermined orientation. These
models show typical enzyme properties such as selectivity,
Michaelis-Menten kinetics, competitive inhibition, and a bell-
shaped curve of the pH rate profile. They represent the most
Acknowledgment. This research was supported by Deutsche
Forschungsgemeinschaft and Fonds der Chemischen Industrie and
by the National Natural Science Foundation of China (20534030,
20725415), the 111 project (B06009). J.-q.L. acknowledges a
fellowship from the Alexander von Humboldt Foundation. Helpful
discussions with Prof. Dr. W. Kla¨ui, Institute of Inorganic Chemistry
of the Heinrich Heine University, are greatly acknowledged.
(20) Fersht, A. Structure and Mechanism in Protein Science: A Guide to
Enzyme Catalysis and Protein Folding, 3rd ed.; Freeman, W. H. and
Company: New York, 2000.
JA8012648
9
8054 J. AM. CHEM. SOC. VOL. 130, NO. 25, 2008