A molecularly imprinted polymer-based synthetic transaminase

Article abstract of DOI:10.1021/ja039622l

The design, synthesis, and evaluation of a molecularly imprinted polymer transaminase mimic is described. Methacrylic acid-ethylene glycol dimethacrylate copolymers were synthesized using, as a template, a transition state analogue (TSA) for the reaction

Full text of DOI:10.1021/ja039622l

Published on Web 06/19/2004
A Molecularly Imprinted Polymer-Based Synthetic
Transaminase
Johan Svenson, Ning Zheng, and Ian A. Nicholls*
Contribution from the Department of Chemistry & Biomedical Sciences, UniVersity of Kalmar,
SE-391 82 Kalmar, Sweden
Received November 16, 2003; E-mail: ian.nicholls@hik.se
Abstract: The design, synthesis, and evaluation of a molecularly imprinted polymer transaminase mimic
is described. Methacrylic acid-ethylene glycol dimethacrylate copolymers were synthesized using, as a
template, a transition state analogue (TSA) for the reaction of phenylpyruvic acid and pyridoxamine to
1
yield phenylalanine and pyridoxal. Polymer suitability was established on the basis of H NMR studies of
template-functional monomer interactions. Polymer recognition characteristics were examined in a series
of HPLC studies using the polymers as chromatographic stationary phases. Selectivity for the TSA, relative
to substrates and products, was observed in both aqueous and nonpolar media. In the latter case
(chloroform/AcOH, 96:4), an enantioseparation factor (R) of 2.1 was obtained, and frontal chromatographic
studies revealed the presence of 11.9 ( 0.2 µmol g-1 (dry weight) of enantioselective sites. Polymers
imprinted with the L-form of the oxazine-based TSA induced a 15-fold enhancement of the apparent reaction
rate (app. Vmax 2.5 × 10-7 mol s-1; app. Km 8.2 × 10-3 M) and enantioselective production of phenylalanine
(32 ( 4% ee) for reactions conducted in an aqueous buffer system. Substrate selectivity was evident, and
a turnover number (kcat) of 0.1 s-1 was determined. This is the first example of the catalysis of sigmatropic
shifts in aqueous media by molecularly imprinted polymers.
Introduction
conversion of pyruvic acid to alanine using pyridoxamine
derivatives with different basic side chains.3b Furthermore, the
Removal of the R-amino groups is the first step in the
catabolism of most L-amino acids.1 In animal species, the
L-amino acids are deaminated and converted to the correspond-
ing R-keto acids in a process involving the transfer of the amino
group to R-ketoglutarate. This reaction is an example of
transamination, a process of central importance in biology,
which is catalyzed by a family of pyridoxal phosphate-dependent
enzymes known as the transaminases, or aminotransferases.
These cofactor-dependent enzymes facilitate the 1,3-prototropic
shift necessary for the reaction. Accordingly, the reverse
reaction, from pyridoxamine and an R-keto acid, yields an amino
acid and pyridoxal, which is a key step in amino acid
biosynthesis, Figure 1.
Pyridoxal phosphate, the coenzyme form of vitamin B6, is
the prosthetic group responsible for catalysis of the transami-
nation reaction and is also involved in a range of mechanistically
related transformations of amino acids including racemization,
decarboxylation, aldol-type condensations, and retro-condensa-
tions.2 On account of the unique reactivities of pyridoxal
phosphate and pyridoxamine phosphate in biology, many
attempts have been made to develop enzyme mimics using these
cofactors, in particular for transamination.3 The seminal work
on transaminase mimics by Breslow and co-workers led to a
system demonstrating up to an 80-fold rate enhancement of the
inherent chirality and hydrophobic binding pockets of func-
tionalized cyclodextrins have been used to produce trans-
aminases.4 By covalent attachment of pyridoxamine, modified
analogues, and various basic groups to cyclodextrins, rate
enhancements for enantio- and substrate-selective transamina-
tions were obtained. More recently, a water-soluble polyethyl-
eneimine (PEI) polymer with incorporated pyridoxamine moi-
eties has been described. These polymers displayed a large
increase in the observed rate of the transamination reaction
between pyruvic acid and pyridoxamine (∼2300-fold increase
at pH 7.0).5 General acid-base catalysis and a hydrophobic
environment within the polymer matrix were believed to be the
main contributors to the rate enhancement, which also seemed
to be dependent on the hydrophobicity of the substrates. Other
reports of transaminase mimics include catalytic antibodies by
(3) (a) Breslow, R.; Hammond, M.; Lauer, M. J. Am. Chem. Soc. 1980, 102,
421-422. (b) Zimmerman, S. C.; Czarnik, A. W.; Breslow, R. J. Am. Chem.
Soc. 1983, 105, 1694-1695. (c) Winkler, J.; Coutouli-Argyropoulou, E.;
Leppkes, R.; Breslow, R. J. Am. Chem. Soc. 1983, 105, 7198-7199. (d)
Czarnik, A. W.; Breslow, R. Carbohydr. Res. 1984, 128, 133-139. (e)
Zimmerman, S. C.; Breslow, R. J. Am. Chem. Soc. 1984, 106, 1490-1491.
(f) Chmielewski, J.; Breslow, R. Heterocycles 1987, 25, 533-540. (g)
Breslow, R.; Chmielewski, J.; Foley, D.; Johnson, B.; Kumabe, N.; Varney,
M.; Mehra, R. Tetrahedron 1988, 44, 5515-5524. (h) Breslow, R.; Canary,
J. W.; Varney, M.; Waddell, S. T.; Yang, D. J. Am. Chem. Soc. 1990, 112,
5212-5219. (i) Fasella, E.; Dong, S. D.; Breslow, R. Bioorg. Med. Chem.
1999, 7, 709-714.
(1) (a) Martell, A. E. Acc. Chem. Res. 1989, 22, 115-124. (b) Breslow, R.
Acc. Chem. Res. 1995, 28, 146-153.
(2) Liu, L.; Rozenman, M.; Breslow, R. Bioorg. Med. Chem. 2002, 10, 3973-
3979.
(4) Breslow, R.; Dong, S. D. Chem. ReV. 1998, 98, 1997-2012.
(5) (a) Liu, L.; Breslow, R. J. Am. Chem. Soc. 2002, 124, 4978-4979. (b)
Liu, L.; Rozenman, M.; Breslow, R. J. Am. Chem. Soc. 2002, 124, 12660-
12661.
9
8554
J. AM. CHEM. SOC. 2004, 126, 8554-8560
10.1021/ja039622l CCC: $27.50 © 2004 American Chemical Society

A Molecularly Imprinted Synthetic Transamine
A R T I C L E S
Figure 1. R-Amino acid synthesis by transamination: (a) pyridoxamine and R-keto acid, (b) ketimine, (c) conjugated quinone-like intermediate, (d) aldimine,
and (e) pyridoxal and amino acid.
Schultz et al.,6 polypeptide-pyridoxamine chimeras by Imperali
and Roy,7 and a pyridoxamine-modified adipocyte lipid binding
protein (ALBP).8
obtained with the application of molecularly imprinted polymers
to a number of hydrolytic reactions,14 the range and catalytic
efficiencies of other reaction types have been limited. The
inherent thermal and chemical stability of these materials renders
MIPs suitable for use under conditions not conducive to
biological macromolecules.17 The flexibility offered by molec-
ular imprinting in terms of choice of template, together with
the significance of the transamination reaction, suggested the
use of molecular imprinting of TSAs for the development of a
synthetic transaminase. Previously, attempts have been made
to produce transaminase-related activities using substrate-
imprinted MIPs, although with modest results.18 In this paper,
we report the synthesis and evaluation of a MIP prepared using
a transition state analogue (TSA) for the reaction of phenylpyru-
vic acid and pyridoxamine to yield phenylalanine and pyridoxal,
the first example of such an approach to this class of reaction,
and the first example of the catalysis of a sigmatropic shift
reaction in aqueous media.
Molecular imprinting9 is a technique for the synthesis of
polymeric materials with predetermined ligand selectivity and
was perceived as an alternative approach for the development
of a synthetic transaminase. The underlying strategy entails the
use of functionalized monomers which can bind reversibly to a
template structure; the resultant complex is subsequently
incorporated into a network polymer by copolymerization in
the presence of an excess of a cross-linking monomer and an
inert solvent (porogen). The removal of the template leaves sites
of complementary shape and functionality, which are capable
of selectively rebinding the template. Molecularly imprinted
polymers (MIPs) with antibody-like recognition characteristics
have been prepared for a large number of compound classes,
and these have been studied with respect to an increasing range
of application areas.10 One area of particular interest, which to
this point in time has received relatively little attention, is their
use in the development of tools for organic synthesis.11 MIPs
have been employed as noncovalent protecting groups for
stereoselective synthesis,12 and, by analogy to the preparation
of catalytic antibodies, MIPs have been synthesized using
template substances with structures mimicking the transition
states of reactions to produce polymers with catalytic function.
Some of the reactions thus far addressed with the technique
include: dehydrofluorination,13 various hydrolyses,14 and the
Diels-Alder15 and aldol16 reactions. While success has been
Results and Discussion
Design and Synthesis of Transition State Analogues. The
reaction between pyridoxamine and phenylpyruvic acid to yield
phenylalanine and pyridoxal was selected for use in this study.
The choice of TSA for use was designed based upon transition
state structures derived from semiempirical molecular orbital
calculations (MINDO/3) performed by Andrews et al., Figure
2a.19 Their study described the mechanism of the pyridoxal-
phosphate-dependent enzyme γ-aminobutyric acid (GABA)
transaminase and led to the development of inhibitors for this
enzyme.20 The calculations indicated that in the transition state
the Schiff base is coplanar with the protonated pyridine ring
and an intramolecular hydrogen bond is present between the
phenolic hydroxyl and the Schiff base nitrogen. Protonation of
the pyridine nitrogen (pKa ∼5), instead of the more basic imine
(pKa ∼11, only suggested to be initially protonated), originates
(6) Cochran, A. G.; Pham, T.; Sugasawara, R.; Schultz, P. G. J. Am. Chem.
Soc. 1991, 113, 6670-6672.
(7) (a) Imperiali, B.; Roy, R. S. J. Am. Chem. Soc. 1994, 116, 12083-12084.
(b) Roy, R. S.; Imperiali, B. Tetrahedron Lett. 1996, 37, 2129-2132. (c)
Shogren-Knaak, M. A.; Imperiali, B. Bioorg. Med. Chem. 1999, 7, 1993-
2002.
(8) (a) Kuang, H.; Brown, M. L.; Davies, R. R.; Young, E. C.; Distefano, M.
D. J. Am. Chem. Soc. 1996, 118, 10702-10706. (b) Kuang, H.; Distefano,
M. D. J. Am. Chem. Soc. 1998, 120, 1072-1073. (c) Kuang, G.; Ha¨ring,
D.; Qi, D.; Mazhary, A.; Distefano, M. D. Bioorg. Med. Chem. Lett. 2000,
10, 2091-2095. (d) Ha¨ring, D.; Distefano, M. D. Bioorg. Med. Chem. 2001,
9, 2461-2466. (e) Ha¨ring, D.; Kuang, H.; Qi, D.; Distefano, M. D. J. Mol.
Catal. B 2001, 11, 967-970.
(9) (a) Sellergren, B., Ed. Molecularly Imprinted Polymers. Man-made Mimics
of Antibodies and Their Application in Analytical Chemistry; Elsevier:
Amsterdam, 2000. (b) Wulff, G. Chem. ReV. 2002, 102, 1-28. (c) Haupt,
K.; Mosbach, K. Chem. ReV. 2000, 100, 2495-2504.
(14) (a) Strikovsky, A. G.; Kasper, D.; Gru¨n, M.; Green, B. S.; Hradil, J.; Wulff,
G. J. Am. Chem. Soc. 2000, 122, 6295-6296. (b) Wulff, G.; Gross, T.;
Scho¨nfeld, R. Angew. Chem., Int. Ed. Engl. 1997, 36, 1961-1964. (c)
Sellergren, B.; Shea, K. J. Tetrahedron: Asymmetry 1994, 5, 1403-1406.
(d) Sellergren, B.; Karmalkar, R. N.; Shea, K. J. J. Org. Chem. 2000, 65,
4009-4027.
(15) Liu, X.-C.; Mosbach, K. Macromol. Rapid. Commun. 1997, 18, 609-615.
(16) Matsui, J.; Nicholls, I. A.; Karube, I.; Mosbach, K. J. Org. Chem. 1996,
61, 5414-5417.
(17) Svenson, J.; Nicholls, I. A. Anal. Chim. Acta 2001, 435, 19-24.
(18) (a) Wulff, G.; Vietmeier, J. Makromol. Chem., Macromol. Chem. Phys.
1989, 190, 1727-1735. (b) Andersson, L. I.; Mosbach, K. Makromol.
Chem., Rapid Commun. 1989, 10, 491-495.
(10) (a) Nicholls, I. A.; Rosengren, J. P. Bioseparation 2001, 10, 301-305. (b)
Sellergren, B. Angew. Chem., Int. Ed. 2000, 39, 1031-1039. (c) Andersson,
L. I. Bioseparation 2001, 10, 353-364. (d) Piletsky, S. A.; Panasyuk, T.
L.; Piletskaya, E. V.; Nicholls, I. A.; Ulbricht, M. J. Membr. Sci. 1999,
157, 263-278.
(11) Alexander, C.; Davidson, L.; Hayes, W. Tetrahedron 2003, 59, 2025-
2057.
(12) (a) Bystro¨m, S. E.; Bo¨rje, A.; A° kermark, B. J. Am. Chem. Soc. 1993, 115,
2081-2083. (b) Alexander, C.; Smith, C. R.; Whitcombe, M. J.; Vulfson,
E. N. J. Am. Chem. Soc. 1999, 121, 6640-6651.
(19) Iskander, M. N.; Andrews, P. R.; Winkler, D. A.; Brinkworth, R. I.; Paola,
C. D.; Cavell, S.; Issa, J. Eur. J. Med. Chem. 1991, 26, 129-136.
(20) (a) Andrews, P. R.; Gulbis, J. M.; Iskander, M. N.; Mackay, M. F. Aust. J.
Chem. 1987, 40, 1073-1081. (b) Andrews, P. R.; Iskander, M. N.; Jones,
G. P.; Winkler, D. A. Eur. J. Med. Chem. 1988, 23, 125-132.
(13) (a) Beach, J. V.; Shea, K. J. J. Am. Chem. Soc. 1994, 116, 379-380. (b)
Mu¨ller, R.; Andersson, L. I.; Mosbach, K. Makromol. Chem., Rapid
Commun. 1993, 14, 637-641.
9
J. AM. CHEM. SOC. VOL. 126, NO. 27, 2004 8555

A R T I C L E S
Svenson et al.
type of polymer in conjunction with the TSA templates used in
this study. d3-Acetic acid was used as an analogue for the
functional monomer (MAA), and complex formation with
L-TSA in CDCl3 was studied in a titration experiment yielding
a saturation isotherm, Figure 4a. Interestingly, only the two
nonequivalent amide protons displayed any significant changes
in chemical shifts. Apparent dissociation constants (app. Kdiss
)
for complexation of these protons with acetic acid were
determined to be 93 ( 10 and 125 ( 15 mM, respectively.
The similarity of these values is indicative of them being
involved in the same or similar interactions with the functional
monomer analogue. Changes in the chemical shifts of the
methylene and methine protons adjacent to the tertiary amine
and that of the pyridinyl ring were barely discernible (<0.02
ppm), which was attributed to the relative remoteness of these
protons from the potential sites for electrostatic interactions (cf.
direct interaction with the amide protons). Nonetheless, these
spectral changes indicate that interactions between the monomer
analogue and the amines do take place.
Figure 2. General structure of the proposed transition state (a), and the
transition state analogue (b).
Job plot analysis22 revealed a 1:1 complex stoichiometry
between the amide proton(s) of the template and acetic acid,
Figure 4b. It is important to note that this stoichiometry pertains
only to the interaction between the functional monomer analogue
and the protons of the amide nitrogen; interactions at other sites,
for example, the hydroxyl and tertiary amine, were discernible
using the NMR experiments described, although the extent of
the chemical shifts involved precluded the calculation of
meaningful dissociation constants and examination of interaction
stoichiometries. It was anticipated that a combination of shape
complementarity afforded by the polymer matrix in conjunction
with the interaction of the functional monomer with the
heteroatom-bearing functionalities should provide the basis for
selective TSA-polymer interactions.
Polymer Synthesis and Characterization. Highly cross-
linked molecularly imprinted polymers were synthesized using
L-TSA as template (P(L)), MAA as functional monomer, and
EGDMA as cross-linker in a relative ratio of 1:5:20 (template:
MAA:EGDMA). Polymerizations were performed at 65 °C
using azo-bis(isobutyronitrile) (AIBN) as the initiator and
chloroform as the porogen. The concentration of template used
in the polymer synthesis was comparable to that used in the
earlier NMR studies, 92 and 30 mM, respectively. Extrapolation
from the nonlinear regression-fitted saturation curves indicated
that approximately 75% complexation of the template is
achieved at the concentrations used in the polymer synthesis,
in terms of interactions with the amide and bridging (N-CH2-
O) methylene of the oxazine ring. P(L) was obtained as particles
(25-63 µm) in better than 80% yields after manual grinding,
sieving, and repeated sedimentation to remove fine particles.
Three reference polymers were synthesized using the same
procedure as described above, although either D-TSA (P(D))
or the anti-asthma drug theophylline (P(T)) was used as
template, or there was no template (P(B)). Residual template
was removed by exhaustive washing using a protocol previously
Figure 3. L-TSA synthesis: (a) methanol, room temperature, 30 min; (b)
methanol, NaBH4, 0 °C; (c) methanol, paraformaldehyde, cat. KOH, reflux,
2 h.
from the resonance stabilization of the conjugated carbanion
during the tautomerization step. Collectively, these findings
suggested that oxazine-based compounds should serve as TSAs
as they allow for the heteroatom orientation and coplanarity to
be maintained.19 In terms of the reaction chosen for the present
study, 2-(5-hydroxymethyl-8-methyl-3,4-dihydropyrido[4,3-e]-
1,3-oxazin-3-yl)phenylalaninamide was proposed as a suitable
TSA, Figure 2b. The amide was selected for use in this case, in
favor of the corresponding carboxylic acid, to avoid a possible
lactonization.
L-2-(5-Hydroxymethyl-8-methyl-3,4-dihydropyrido[4,3-e]-
1,3-oxazin-3-yl)phenylalaninamide (L-TSA) and its optical
antipode (D-TSA) were readily obtained in a two-pot procedure,
Figure 3. Condensation of L- or D-phenylalaninamide and
pyridoxal, followed by NaBH4 reduction of the resultant imine,
afforded the corresponding secondary amine in 76% yield after
purification on silica. Oxazine ring formation was achieved by
a Mannich cyclization using paraformaldehyde and furnished
the desired products in acceptable yield (56%) after purification.
1H NMR Analysis of TSA-Functional Monomer Interac-
tions. Many previous studies have illustrated the general utility
of methacrylic acid (MAA)-ethylene glycol dimethacrylate
(EGDMA) copolymers for molecular imprinting using nonco-
valent interactions.9 The relative resilience of these polymers
to extremes of chemical environment and temperature suggests
their suitability for applications in organic synthesis. Spectro-
scopic studies have previously been shown to be useful for the
study of template-functional monomer complexation.21 1H
NMR studies were undertaken to establish the suitability of this
(21) (a) Sellergren, B.; Lepisto¨, M.; Mosbach, K. J. Am. Chem. Soc. 1988, 110,
5853-5860. (b) Andersson, H. S.; Nicholls, I. A. Bioorg. Chem. 1997, 25,
203-211. (c) Whitcombe, M. J.; Martin, L.; Vulfson, E. N. Chro-
matographia 1998, 47, 457-464. (d) Svenson, J.; Karlsson, J. G.; Nicholls,
I. A. J. Chromatogr., A 2004, 1024, 39-44.
(22) (a) Takeuchi, T.; Dobashi, A.; Kimura, K. Anal. Chem. 2000, 72, 2418-
2422. (b) Striegler, S. Bioseparation 2001, 10, 307-314. (c) Wulff, G.;
Knorr, K. Bioseparation 2001, 10, 257-276.
9
8556 J. AM. CHEM. SOC. VOL. 126, NO. 27, 2004

A Molecularly Imprinted Synthetic Transamine
A R T I C L E S
1
Figure 4. (a) Saturation plot for the H NMR titration: -∆- amide proton (shift from δ 5.64), -]- amide proton (δ 6.37), -O- oxazine N-CH2-O
methylene (δ 4.93). (b) Job plot of the interaction between amide proton δ 5.64 and acetic acid.
Table 1. Polymer-Ligand Recognition in Aqueous and Organic Media
capacity factors k′ a
polymer
mobile phase
pyridoxamine
pyridoxal
L-TSA
D-TSA
phenylpyruvic acid
P(L)
P(L)
P(B)
P(B)
MeOH/bufferb 1:1
CHCl3/AcOHc
2.82
n.a.d
1.66
n.a.d
0.16
n.a.d
0.03
n.a.d
3.51
10.14
3.22
3.29
4.84
3.23
0.75
0.01
0.47
0.05
0.53
MeOH/bufferb 1:1
CHCl3/AcOHc
0.75
a k′ ) (Va - V0)/V0, where Va is the retention volume for the analyte and V0 is the retention volume for a noninteracting void marker (acetone). All values
are the result of a minimum of three experiments with an uncertainty of (0.02. Flow rate (1.0 mL min-1), injection volume (30 µL), sample concentrations
(1.0 mM). b 0.1 M acetate, pH 7.0. c CHCl3/AcOH 96:4, v/v. d Analyte not soluble in the mobile phase.
developed for similar polymer systems.23 BET-surface area
analysis showed comparable surface areas for the TSA-imprinted
and nonimprinted polymers (P(L) 336 m2/g, P(B) 379 m2/g).
FT-IR studies indicated the presence of the expected function-
alities. To confirm that the complexation of template by
functional monomer survives the polymerization process, a
polymerization mixture containing L-TSA was prepared in an
NMR tube and a proton spectrum was acquired. The shifts
induced by the functional monomer reflected those observed
using acetic acid. Polymerization was initiated and spectra
recorded at regular intervals over 60 min. The chemical shifts
) -RT ln R.24 Importantly, binding of the TSAs to the
nonimprinted P(B) was minimal, k′ 0.75, further supporting the
observation that P(L) contains sites selective for the template.
Significantly weaker recognition was observed for P(L) when
using a polar eluent (methanol/0.1 M sodium acetate (aq); 1:1,
v/v; pH 7.0 - the buffer used for subsequent studies of the
influence of polymers on transamination reaction rates) as
reflected in the markedly lower capacity factors, although some
enantioselectivity was still evident, R ) 1.1 (∆G ) 0.2 kJ mol-1
at 293 K). Thus, the more polar environment is detrimental with
respect to the electrostatic interactions between the carboxylate
residues of the polymer and the analyte. Binding to the
nonimprinted reference polymer, P(B), was stronger in the
aqueous buffer than in chloroform/acetic acid. This is indicative
of the prevalence of nonspecific binding arising from hydro-
phobic interactions with the polymer matrix. Collectively, the
recognition behavior supports the presence of sites selective for
the template and that the basis for selectivity lies in interactions
between the carboxyl groups of the polymer and the template,
as implicated by the NMR studies discussed earlier. Importantly,
the D-TSA-imprinted polymer P(D) behaved identically, al-
though with a reversed enantioselectivity.
1
induced in the H NMR spectra of L-TSA by the functional
monomer, MAA, were clearly retained throughout the first 30
min of the polymerization reaction. Sample anisotropy resulted
in severe line broadening as the reaction mixture approached
the gel phase, and complete solidification was apparent after 1
h.
Polymer-Ligand Recognition Studies. The recognition
characteristics of the polymers were examined to establish the
presence of sites selective for the template in the imprinted
polymers. Polymers were packed into HPLC columns and used
as stationary phases in a series of studies that employed two
distinctly different eluent systems, Table 1. In a nonpolar eluent
system, chloroform/acetic acid (96:4, v/v), significant differences
in capacity factors (k′) were observed for the enantiomers on
the P(L) with a separation factor (R) of 2.1. When pure
chloroform was used as eluent, the TSAs failed to elute from
the column. The observed enantioselectivity corresponds to a
difference in Gibbs’ free energy of binding between the
enantiomers of 1.8 kJ/mol at 293 K, as calculated using: ∆G
Polymer recognition of the substrates, pyridoxamine and
phenylpyruvic acid, and one of the products, pyridoxal, was
similarly examined. Difficulty in the detection of phenylalanine
prevented its use in this experiment. In studies in the aqueous
buffer, pyridoxamine was the only analyte to show any
significant retention on P(L), which was attributed to its greater
structural similarity to the template and the strong electrostatic
interactions available through the amine moiety. Pyridoxal,
(23) Karlsson, J. G.; Andersson, L. I.; Nicholls, I. A. Anal. Chim. Acta 2001,
435, 57-64.
(24) Weingarten, M. D.; Sekanina, K.; Still, W. C. J. Am. Chem. Soc. 1998,
121, 9112-9113.
9
J. AM. CHEM. SOC. VOL. 126, NO. 27, 2004 8557

A R T I C L E S
Svenson et al.
Figure 5. Comparison of the phenylalanine production. (a) Time course studies: -×- P(L), -∆- P(T), -0- P(B), -]- solvent reaction. (b) Initial
reaction velocities: -×- P(L), -0- P(B), -]- solvent reaction. In all cases, substrate concentrations were as follows: pyridoxamine (22.5 mM) and
phenylpyruvic acid (22.5 mM), polymer 100.0 mg (except solution reaction) in a total volume of 1.0 mL.
Table 2. Kinetic Data for the Production of Phenylalanine
although relatively poorly retained, also displayed preferential
polymer
inhibitor
app. Vmax (mol s-1
)
app. Km (M)
binding to P(L). Neither phenylpyruvic acid nor pyridoxal
demonstrated significant levels of binding to the polymers
studied. The enantioselectivity of P(L) in both nonpolar and
aqueous media was most encouraging, as the function of MIPs
in aqueous media is a significant challenge for their use as
biomimetic systems.
MeOH/buffer
P(L)
P(B)
P(L)
P(L)
1.7 × 10-8
2.5 × 10-7
1.4 × 10-7
1.7 × 10-7
2.0 × 10-7
1.3 × 10-7
8.2 × 10-3
2.3 × 10-2
2.9 × 10-2
2.3 × 10-2
2.3 × 10-2
L-TSAa
D-TSAa
L-TSAa
P(B)
Frontal chromatography25 was used to determine the strength
of the polymer-TSA interactions in the aqueous buffer used
for the reaction studies described later. Enantioselectivity was
evidenced by the differences in the apparent dissociation
constants observed for the interaction of P(L) with the L- and
D-TSAs (app. Kdiss ) 1.5 ( 0.1 and 2.0 ( 0.1 mM, respectively)
and in terms of the binding site populations (Bt) available for
the analytes, 28.2 ( 0.1 and 16.3 ( 0.1 µmol/g (dry weight),
respectively. Thus, it can be concluded that P(L) possesses 11.9
( 0.2 µmol/g (dry weight) of high fidelity sites with selectivity
for the L-TSA. As anticipated, and reflecting the zonal chro-
matographic studies, binding of the L-TSA to P(B) was weaker,
app. Kdiss ) 1.9 ( 0.1 mM (Bt ) 35.9 ( 0.1 µmol/g). It is
noteworthy that the apparent dissociation constants obtained for
L-TSA binding to P(L) are approximately 2 orders of magnitude
lower than those found for complexation of L-TSA by acetic
acid in the same solvent. This reflects the greater energetic
penalty (translational and rotational free energy) that must be
paid for formation of a complex comprised of a greater number
of monomer/template components.26
Evaluation of Polymer Catalysis of Transamination. The
influence of the polymers on the transamination reaction of
phenylpyruvic acid and pyridoxamine, to afford phenylalanine
and pyridoxal, was studied in an aqueous buffer (methanol/0.1
M sodium acetate, pH 7.0; 1:1) using an HPLC-based assay.
Studies of the influence of the polymers P(L), P(B), and P(T)
on reaction rates were performed using the solution reaction
(in the absence of polymer) as reference, Figure 5a. The latter
polymer, P(T), was included in the study to provide an
assessment of the extent of any possible influence of polymer
morphology or cavity functionalization (localization of func-
a Inhibitor concentration 0.50 mM.
tional monomer residues in the resultant polymer) on reaction
rates due to the imprinting process. The choice of reaction
medium was based upon a compromise between the feasibility
of analyzing reaction kinetics with the assay employed, substrate
solubility, and a desire to illustrate the utility of these polymers
in aqueous media. The reaction solvent influences the position
of equilibrium, which is rapidly shifted in favor of Schiff base
formation in a less polar solvent, or at higher pH. Indeed,
reactions run in less polar media, pure methanol, were some
15 times faster than those described here in the buffer.
Michaelis-Menten-like kinetics were observed for the reac-
tions performed in the presence of polymers. A 5-fold enhance-
ment of the initial reaction rate was observed with the
nonimprinted reference polymer, P(B), as compared to the
solution reaction, and a 15-fold increase was achieved with P(L),
Figure 5b and Table 2. The general effect of a polymer on
reaction rate is often explained as resulting from local increases
in the concentration of substrate in the vicinity of the polymer
surface due to nonspecific adsorption. However, when compar-
ing the reaction rates obtained with P(B) and P(L), the difference
cannot be explained by differences in surface areas, for, as
determined by BET analysis, the nonimprinted polymer, P(B),
has a greater (12%) surface area than that of P(L). The
enhancement of reaction rate observed with P(L) can therefore
be interpreted as a result of the imprinting of the L-TSA, that
is, template-induced localization of monomer functionalities. To
elucidate the possible influence of a general localization of
functionalities around cavities within the polymer, a polymer
of the same composition but imprinted with an unrelated
template, the anti-asthma drug theophylline, P(T), was exam-
ined. The rate enhancement induced by P(T) was slightly higher
(6-fold relative to the solution reaction) than that of P(B),
(25) Kasai, K.; Oda, Y. J. Chromatogr. 1986, 376, 33-47.
(26) Nicholls, I. A. Chem. Lett. 1995, 1034-1035.
9
8558 J. AM. CHEM. SOC. VOL. 126, NO. 27, 2004

A Molecularly Imprinted Synthetic Transamine
A R T I C L E S
Figure 6. Lineweaver-Burk plots: (a): -∆- P(L), -O- P(B), substrate concentrations: pyridoxamine and phenylpyruvic acid (each 7.5-22.5 mM). (b)
Inhibition studies using P(L): inhibitor concentrations -]- 0.50 mM, -0- 0.25 mM, -∆- 0.00 mM. Substrate concentrations in the presence of
inhibitors: 6.25-13.5 mM.
although significantly lower than for P(L). This result provides
further support for the unique nature of P(L); that is, the
imprinting effect-derived selective recognition of the transition
state analogue is the basis for the observed reaction rate
enhancement.
Moreover, no product enantiomeric excess was observed in
reactions with P(B) or buffer in the presence of the L-TSA.
These results highlight the enantioselectivity of the polymer and
the role the recognition sites play in the observed chiral
induction.
The apparent reaction velocities (Vmax) and Michaelis con-
stants (Km) were determined using substrate concentrations
ranging from 7.5 to 22.5 mM, Figure 6 and Table 2. Unfortu-
nately, higher concentrations were impossible on account of
substrate solubilities, and lower concentrations were prohibited
due to assay detection limits. A 1:1 pyridoxamine/phenylpyruvic
acid stoichiometry was employed in the data presented to
optimize the width of the concentration range studied. Accord-
ingly, the kinetic parameters reflect the performance of the
reaction system as a whole.
The observed rate enhancements were competitively inhibited
by L-TSA. It induced a decrease in Vmax (∼30%) and an almost
4-fold increase in Km for P(L), indicative of competitive
inhibition by the template for the imprinted cavities. Most
importantly, D-TSA had a less pronounced effect on the reaction
rate, a 3-fold decrease in Km and a 20% reduction in Vmax, which
reflects the stereochemical integrity of the high affinity sites
and implies that sites with affinity for L-TSA are responsible
for the observed rate enhancements. The influence of inhibitor
chirality on the performance of the polymer was further
underscored by the effect on P(B), with which no significant
inhibition was observed at the concentrations employed. The
catalytic performance of these materials is reflected in the
turnover number (kcat), which for P(L) is 0.1 s-1, based upon
28 µmol of sites per gram of polymer. While a number of
entropically favorable hydrolytic reactions have been catalyzed
by MIPs,11 in some cases with high catalytic efficiencies, the
catalysis of the sigmatropic transamination reaction described
here is unique not only by virtue of the observed catalytic
efficiencies, but also the fact that the polymer system is
functional in an aqueous medium.
Together with the enantioselectivity of P(L) demonstrated
in the frontal and zonal chromatographic studies of polymer-
ligand recognition, these data support the existence of recogni-
tion sites selective for the L-TSA, which are due to its presence
as a template during the polymerization process. This concurs
with the evidence obtained from the NMR studies, which
indicate specific interactions between functional monomer
(analogue) and the template - the basis for the observed ligand
selectivity in the resultant polymers. Furthermore, the enantio-
selectivity observed in product formation provides the strongest
evidence in support for the origin of the observed catalytic effect
being the result of the presence of sites selective for the specific
enantiomer of the transition state employed in the polymerization
process. The catalytic effect observed using this transition state
analogue as a template is superior to those previously obtained
by the molecular imprinting of substrate-based structures.18 This
supports the use of mimics of transition state structures in the
synthesis of catalytic MIPs.
To further examine the selectivity of the polymers, reactions
were run using an alternative substrate, pyruvic acid. The initial
solution reaction rate (1.8 × 10-8 mol s-1) was similar to that
observed using phenylpyruvic acid (1.3 × 10-8 mol s-1). Both
P(B) and P(L) induced increases in initial reaction rate for
pyruvic acid, 7.3 × 10-8 and 9.8 × 10-8 mol s-1, respectively.
The fact that P(L) has only a marginal effect on rate in this
case, as compared to P(B), can be attributed to the substrate
being less well recognized by the sites selective for the transition
state for the reaction due to its smaller size. To establish the
robustness of the polymeric material, polymer used in assays
was washed using the protocol used for polymer preparation
and evaluated in subsequent assays where the polymer retained
at least 90 ( 5% of the initial activity. This highlights the
potential of these types of polymers for use as substitutes for
enzymes in industrial applications.
The enantioselectivity of P(L) was further reinforced upon
analysis of the product, phenylalanine, which was obtained in
favor of the L-form, 32 ( 4% ee. Importantly, both reference
polymer systems, P(B) and P(T), furnished a racemic product.
Reactions performed in the presence of the L-TSA (0.50 mM)
resulted, in addition to a lower product yield, in a significantly
lower product enantiomeric excess, 9 ( 4% ee. However, no
change in the enantiomeric excess of the reaction product was
observed when run in the presence of the D-TSA (0.50 mM).
Conclusion
The results presented demonstrate the use of molecular
imprinting for preparing a synthetic polymer with enzyme-like
activity (catalytic turnover, substrate selectivity, enantioselec-
9
J. AM. CHEM. SOC. VOL. 126, NO. 27, 2004 8559

A R T I C L E S
Svenson et al.
tivity, and rate enhancement). Furthermore, this is the first report
of a TSA-imprinted polymer capable of mediating a trans-
amination reaction and is the first example of catalysis of a
sigmatropic shift in aqueous media by a molecularly imprinted
polymer. The versatility of the technique with respect to choice
of template suggests the use of this system for the catalytic
asymmetric synthesis of nonnatural amino acids and related
structures. The relatively simple preparation and high mechan-
ical, chemical, and thermal stability make MIP “artificial
enzymes” interesting alternatives to their biological counterparts.
Ha˚kan S. Andersson and Dr. Susanne Wikman are thanked for
many helpful discussions.
Supporting Information Available: Synthesis and charac-
terization of transition state analogues, including 1H NMR and
13C NMR spectra of the L-TSA. 1H NMR studies of template-
monomer analogue complexation. Polymer syntheses and
polymer characterization. Zonal and frontal chromatographic
procedures and data from frontal chromatographic studies. Assay
protocol and reagent preparation. Data from kinetics studies on
reactions with pyruvate, and reactions of phenylpyruvic acid in
methanol. This material is available free of charge via the
Internet at http://pubs.acs.org.
Acknowledgment. The financial support of the Swedish
Science Council (VR), Graninge Foundation, Carl Trygger’s
Foundation, Swedish Knowledge (KK) Foundation, and the
University of Kalmar is most gratefully acknowledged. Dr.
JA039622L
9
8560 J. AM. CHEM. SOC. VOL. 126, NO. 27, 2004