Carbon-carbon bond formation using substrate selective catalytic polymers prepared by molecular imprinting: An artificial class II aldolase

Article abstract of DOI:10.1021/jo9516805

A class II aldolase-mimicking synthetic polymer was prepared by the molecular imprinting of a complex between dibenzoylmethane (1) and cobalt(II) ion in a 4-vinylpyridine-styrene-divinylbenzene copolymer. This polymer was capable of selectively catalyzing

Full text of DOI:10.1021/jo9516805

5414
J . Org. Chem. 1996, 61, 5414-5417
Ca r bon -Ca r bon Bon d F or m a tion Usin g Su bstr a te Selective
Ca ta lytic P olym er s P r ep a r ed by Molecu la r Im p r in tin g: An
Ar tificia l Cla ss II Ald ola se
J un Matsui,†,‡,§ Ian A. Nicholls,†,| Isao Karube,‡ and Klaus Mosbach*,†
Department of Pure and Applied Biochemistry, University of Lund, Box 124, S-221 00 Lund, Sweden and
Research Centre for Advanced Science and Technology, University of Tokyo, 4-6-1 Komoba, Meguro-ku,
Tokyo 153, J apan
Received September 13, 1995 (Revised Manuscript Received April 19, 1996X)
A class II aldolase-mimicking synthetic polymer was prepared by the molecular imprinting of a
complex between dibenzoylmethane (1) and cobalt(II) ion in a 4-vinylpyridine-styrene-divinyl-
benzene copolymer. This polymer was capable of selectively catalyzing the reaction of acetophenone
(2) and benzaldehyde (3) to produce chalcone (4). The polymer, which demonstrated substrate
selectivity and turnover, increased reaction rate eight-fold, relative to the solution reaction. The
polymer was able to withstand vigourous reaction conditions, DMF and 100 °C for several weeks,
while retaining most (80-95%) of its initial activity. This is the first reported example of catalytic
carbon-carbon bond formation using the molecular imprinting technique.
In tr od u ction
focusing upon hydrolytic and dehydration-based pro-
cesses, though to date success has been modest. The
catalysis of carbon-carbon bond formation poses one of
the most significant objectives for catalytic antibody
technology and organic synthesis19 and is addressed here.
We report the preparation and evaluation of a vinylpy-
ridine-styrene-divinylbenzene copolymer imprinted with
an aldol condensation reactive intermediate analogue, a
Molecular imprinting,1-3 a technique for producing
ligand selective recognition sites in synthetic polymers,
has found use in the preparation of antibody combining
site mimics with affinities and selectivities comparable
to those of biologically derived antibodies4-6 and in a
range of other application areas.7,8 The principles un-
derlying the design9 and preparation8 of molecularly
imprinted polymers (MIPs) are summarized in Figure 1.
In parallel to the development of catalytic antibody
technology,10 attempts have been made to generate
substrate selective catalytically active MIPs,11-18 mainly
complex between dibenzoylmethane (DBM), 1, and Co2+
.
The imprinted polymer demonstrated substrate selectiv-
ity, turnover, and rate enhancement when used to
catalyze an entropically unfavorable C-C bond forma-
tion, namely, the aldol condensation of acetophenone (2)
and benzaldehyde (3) to produce chalcone (4), Figure 2A.
This class II aldolase-like activity, as found in primitive
cells such as yeast and bacteria,20 was competitively
inhibited by the reactive intermediate analogue 1. This
weak enzymelike activity was achieved employing a
combination of temperature (100 °C), reaction time
(days), and solvent (dimethylformamide), not generally
observed with biologically derived catalysts. This par-
ticular reaction system, which requires several weeks to
reach completion, was selected to permit adequate kinetic
analysis and to illustrate the robustness of these materi-
als.
* Author to whom correspondence should be addressed.
† University of Lund.
‡ University of Tokyo.
§ Present address: Synthetic Biochemistry Laboratory, Faculty of
Information Sciences, Hiroshima City University, 151-5 Ozuka, Nu-
mata-cho, Asaminami-ku, Hiroshima 731-31, J apan.
| Present address: Biomedical Chemistry, School of Natural Sci-
ences, University of Kalmar, Box 905, S-391 29 Kalmar, Sweden.
X Abstract published in Advance ACS Abstracts, J une 1, 1996.
(1) Mosbach, K. Trends Biochem. Sci. 1994, 19, 9-14.
(2) Shea, K. J . Trends Polym. Sci. 1994, 2, 166-173.
(3) Wulff, G. Trends Biotechnol. 1993, 11, 85-87.
(4) Andersson, L. I.; Nicholls, I. A.; Mosbach, K. Antibody Mimics
Obtained by Non-Covalent Molecular Imprinting. In Immunological
Analysis of Agrochemicals: Emerging Technologies; ACS Symposium
Series 586; Nelson, J . O., Karo, A. E., Wong, R. B., Eds.; American
Chemical Society: Washington, DC, 1995; pp 89-97.
(5) Vlatakis, G.; Andersson, L. I.; Mu¨ller, R.; Mosbach, K. Nature
1993, 361, 645-647.
(6) Andersson, L. I.; Mu¨ller, R.; Vlatakis, G.; Mosbach, K. Proc. Natl.
Acad. Sci. U.S.A. 1995, 92, 4788-4792.
(7) Nicholls, I. A.; Andersson, L. I.; Mosbach, K.; Ekberg, B. Trends
Biotechnol. 1995, 13, 47-51.
(8) Andersson, L. I.; Nicholls, I. A.; Mosbach, K. Molecular Imprint-
ing - a Versatile Technique for the Preparation of Separation Materials
of Predeterminated Selectivity. In Highly Selective Separations in
Biotechnology; Street, G., Ed.; Blackie Academic and Professional:
Glasgow, 1994; pp 207-225.
Resu lts a n d Discu ssion
On the basis of a molecular model study, DBM (1) was
perceived as a reactive intermediate analogue for the
cobalt(II) ion-mediated aldol condensation of 2 and 3.21
The bidentate ligand 1 was expected to accommodate two
of the coordination sites of the tetrahedrally configured
Co2+, Figure 1, with 2 equiv of vinyl pyridine (5) satisfy-
(9) Nicholls, I. A. Chem. Lett. 1995, 1035-1036.
(10) Lerner, R. A.; Benkovic, S. V.; Schultz, P. G. Science 1991, 252,
659-667.
(11) Leonhardt, A.; Mosbach, K. Reactive Polymers 1987, 6, 285-
290.
(16) Heilmann, J .; Maier, W. F. Angew. Chem., Int. Ed. Engl. 1994,
33, 471-473. See also Heilmann, J .; Maier, W. F. Z. Naturforsch. 1995,
50b, 460-468.
(17) Matsuishi, T.; Shimada, T.; Morihara, K. Bull. Chem. Soc. J pn.
1994, 67, 748-756.
(12) Wulff, G.; Vietmeier, J .; Poll, H.-G. Makromol. Chem. 1987, 188,
731-740.
(18) Sellergran, B.; Shea, K. J . Tetrahedron: Asymmetry 1994, 5,
1403-1406.
(13) Robinson, D. K.; Mosbach, K. J . Chem. Soc., Chem. Commun.
1989, 14, 969-970.
(19) Danishefsky, S. Science 1993, 259, 469-470.
(20) Heron, E. J .; Caprioli, R. M. Biochim. Biophys. Acta 1975, 403,
563-572.
(14) Mu¨ller, R.; Andersson, L. I.; Mosbach, K. Makromol. Chem.
Rapid Commun. 1993, 14, 637-641.
(15) Beach, J . V.; Shea, K. J . J . Am. Chem. Soc. 1994, 116, 379-
380.
(21) Watanabe, K.; Imazawa, A. Bull. Chem. Soc. J pn. 1982, 55,
3208-3211.
S0022-3263(95)01680-X CCC: $12.00 © 1996 American Chemical Society

A Class II Aldolase Mimic by Molecular Imprinting
J . Org. Chem., Vol. 61, No. 16, 1996 5415
F igu r e 2. (A) The aldol condensation of acetophenone (2) and
benzaldehyde (3) to yield chalcone (4) with the reactive
intermediate analogue dibenzoylmethane (DBM) (1). (B)
Schematic representation of MIP-mediated catalysis of the
aldol condensation of acetophenone (2) with benzaldehyde (3)
to yield chalcone (4). (i) Polymer incubated with cobalt(II) ion.
(ii) Incubation of the polymer with reactants. (iii) Enolization
and nucleophilic addition yields a polymer-stabilized reactive
intermediate. (iv) Proton transfer to the stabilized reactive
intermediate and subsequent dehydration affords chalcone (4),
which has a lower affinity for the polymer than the stabilized
reactive intermediate structure due to 4 having only a single
point coordination to the metal center, motivating its diffusion
from the recognition site. [Note: It is unclear whether the
dehydration step takes place within the “active site” or after
diffusion therefrom.]
F igu r e 1. Schematic representation of molecularly imprinted
polymer preparation: DBM-Co2+ MIP preparation. (i) The
monomers, vinylpyridine (5), styrene (6), and divinylbenzene
(7), are mixed with the imprint species, dibenzoylmethane
(DBM) (1) and cobalt(II) acetate, and initiator in methanol,
allowing formation of complementary sets of interactions
between the monomers and the print species.26 (ii) Polymer-
ization produces a rigid bulk polymer. (iii) The imprint species
is extracted to render sites topographically cimplementary to
the print species in terms of shape and chemical functionality.
(The schematic polymer structure dipicted represents an
idealized case and does not take into account the accessibility
of the substrate into the recognition site.)
Ta ble 1. Ch r om a togr a p h ic Recogn ition Stu d ya
HPLC column (imprint species)
analyte
DBM-Co2+ Co2+ DBM blank
dibenzoylmethane (DBM) (1)
9.3
(5.4)
0.54
(0.55)
0.47
(0.49)
3.50
7.0
(5.6) (4.7) (5.2)
0.55 0.41 0.64
(0.56) (0.47) (0.61)
0.44 0.39 0.61
(0.48) (0.42) (0.56)
2.01 1.59 2.34
(2.10) (1.71) (2.28)
2.5
3.6
ing the remaining two coordination sites of the Co2+
complex. In addition to metal ion coordination, the
pyridinyl residues would provide the base necessary for
generation of the enolate of acetophenone (2). Styrene
(6)-divinylbenzene (7) was used as the cross-linking
agent; these monomers were expected to interact with 1
through π-π stacking and van der Waals interactions
to aid in defining the recognition site topography in the
resultant polymer.
An empirical evaluation of polymer-substrate selectiv-
ity was carried out using the polymer as an HPLC
stationary phase, see Table 1. Retention characteristics
of the reaction substrates, product, and the imprint
species were compared using a DBM-Co2+ imprinted
polymer and a series of reference polymers imprinted
with either Co2+, DBM (1), or nothing. Recognition was
shown to be dependent upon the presence of Co2+ in the
acetophenone (2)
benzaldehyde (3)
chalcone (4)
(2.09)
a
Data represent capacity factors (k′) determined using the
relationship: k′ ) (t - t0)/t0, where t is the retention time of a
given species and t0 that of the void (determined by injection of
acetone). Polymers were packed in Teflon-lined stainless steel
chromatography columns (100 × 5.4 mm). Flow rate 0.35 mL
min-1 and detection at 260 nm. Figures in parentheses are k′
values in the absence of cobalt(II) acetate (0.05 M) in the buffer.
elution buffer, with a substantial reduction in affinity for
1 observed in the absence of the ion, most notable in the
case of the DBM-Co2+ MIP. Optimal recognition of 1, the
reactive intermediate analogue, was dependent upon both

5416 J . Org. Chem., Vol. 61, No. 16, 1996
Matsui et al.
Ta ble 2. Su bstr a te Selectivity
ketone substrate
product
reactivity ratioa
2
8
9
4
10
11
2.0
1.3
1.4
a
Reactivity ratios were determined from product yields (yield
from DBM-Co2+ MIP/yield from Co2+ MIP) after 75 h of reaction.
Adamantyl methyl ketone (8) and 9-acetylanthracene
(9) were substituted for 2 in a set of parallel experiments
to investigate the degree of MIP substrate selectivity, see
Table 2. Lower relative rate enhancements were ob-
served for reaction of the bulkier substrates 8 and 9 with
benzaldehyde (3) to form their respective aldol condensa-
tion products 10 and 11. The greater nonspecific reactiv-
ity of 9 than that of 8 was attributed to its planar
aromatic structure which allows for easier access into and
greater affinity for the MIP through aromatic stacking.
The lower selectivity for the alternate substrates lend
support to the shape recognition effects previously ob-
served for recognition in styrene-divinylbenzene MIPs.23
Long term experiments, of three weeks duration,
afforded up to 80% conversion of starting materials to
product, necessitating some 138 turnovers per theoretical
active site.
To confirm that the imprint recognition site was also
the catalytic center, 1 was tested for its ability to inhibit
the polymer-catalyzed reaction (Vmax ) 0.61 ( 0.06 µmol
h-1; Michaelis constant Km ) 1.23 ( 0.04 µM), Figure 4.
The concentration-dependent inhibition of chalcone (3)
production by 1 (Ki ) 60 ( 10 µM) implies the presence
of a specific reaction center in the polymer matrix. This
is in accord with the chromatographic data, which
demonstrated a higher affinity for DBM (1) by the DBM-
Co2+-imprinted polymer recognition sites, relative to the
reaction substrates 2 and 3. Substrate (1.42 ( 10 mmol
g (polymer)-1) was required to saturate high affinity
nonspecific binding sites in the bulk polymer before
evidence of reaction onset; this necessitated calculation
of an apparate substrate concentration, S*, for use in
kinetic analyses. We propose that the high Ki and low
turnover result from the template structure (1) being a
better product analogue than true transition state ana-
logue for this reaction.
F igu r e 3. Production of chalcone (4) using DBM-Co2+ and
Co2+ MIPs and solvent containing cobalt(II) acetate. Yields
were calculated from the average of triplicate determinations
of duplicate sets of reactions.
the presence of the coordinating metal and the comple-
mentary imprint site interactions, consistent with the
assumptions made in the polymer design. Significantly,
1 showed a much higher affinity for the DBM-Co2+ MIP
than either of the substrates or the reaction product.
Moreover, the affinity of 1 for the DBM-Co2+ MIP was
much higher than in any of the reference polymer
systems.
To evaluate the role of the DBM-Co2+ recognition sites
in mediating this aldol condensation, reactions were
conducted in the presence of the DBM-Co2+ and the Co2+
MIPs and in solution in the presence of pyridine and
cobalt(II) ion, Figure 2B. The rate of the DBM-Co2+ MIP-
mediated reaction was eight-fold higher than the solution
reaction and twice that of the Co2+ MIP reaction, Figure
3. Product yields were determined by reversed phase
high performance liquid chromatographic analysis of the
reaction mixtures. The ratios between substrate and
product peak areas were compared to standard curves
prepared using authentic product samples. HPLC based
stability studies of reactants and products revealed
negligible breakdown of materials under the reaction
conditions and over the time courses employed in these
studies. It is notable that products from these reactions
were obtained on sufficient scale to permit direct spec-
A number of polymer samples were recovered from
reaction mixtures, exhaustively reextracted and reused
for similar reactions. These materials retained 80-95%
of their previous activity. No doubt, the reaction condi-
tions employed here, 100 °C over several days, led to the
gradual decomposition of the polymeric matrix, a plaus-
able explanation for the reduction in activity.
1
troscopic analysis (mass spectrometry and H-NMR) to
further confirm their identity.
The presence of Co2+, or another suitable cation, is a
necessary requirement for catalysis of this aldol conden-
sation under the conditions employed here.21 As im-
printed polymer recognition sites are of a heterogeneous
nature,22 not all the DBM-Co2+ sites will necessarily be
active, nor are they all accessible to solvent and sub-
strates. The contribution of inextricably bound Co2+ to
the MIP reactivity was examined using reactions run
with the washed DBM-Co2+ polymer only as the cobalt
source. A very low background activity was observed,
some 0.1% of that of the Co2+ incubated DBM-Co2+
polymer. The concentration of Co2+ present in MIP assay
samples was determined by the spectrophotometric analy-
sis of incubation mixture supernatants. This data pro-
vides a theoretical maximum number of “active” sites in
the assayed polymers, i.e. assuming that all available
Co2+ ions present in the polymer are active.
Con clu sion s
The results presented demonstrate the use of molecular
imprinting for preparing synthetic nonbiological macro-
molecules capable of enzymelike catalytic turnover,
substrate selectivity, and rate enhancement. Further-
more, this is the first report of true enzymelike catalysis
of C-C bond formation using MIPs. The simple prepa-
ration, high mechanical, thermal, and chemical stability,
and the lack of need for biologically based protocols make
molecularly imprinted polymer “artificial enzymes” in-
teresting alternatives to their biological counterparts.
(22) Ramstro¨m, O.; Nicholls, I. A.; Mosbach, K. Tetrahedron: Asym-
metry 1994, 5, 679-656.
(23) Shea, K. J .; Sasaki, D. Y. J . Am. Chem. Soc. 1991, 113, 4109-
4120.

A Class II Aldolase Mimic by Molecular Imprinting
J . Org. Chem., Vol. 61, No. 16, 1996 5417
then sparged with dry nitrogen (5 min) at 0 °C. Polymeriza-
tion was carried out at 45 °C (24 h). The bulk polymer was
ground in a mechanical mortar and wet sieved (water/ethanol)
through a 25 µm sieve. The fines were removed by repeated
sedimentation from acetonitrile, and the sediment was col-
lected on a 15 µm filter. The print molecule complex (DBM-
Co2+) was removed by packing the polymer in an HPLC column
and washing with methanol/acetic acid (7/3, v/v) for 24 h at
1.0 mL min-1 and then methanol. The reference polymers
were prepared and treated identically, except for the exclusion
of one or both of DBM and cobalt(II) acetate. BET surface
area measurements (Micromeritics Flowsorb II 2300 instru-
ment, 30% N2 in Ar) found areas of 3.67 and 2.30 m2 g-1 for
the DBM-Co2+ and Co2+ MIPs, respectively.
MIP Assa ys. Polymer samples were typically incubated
at room temperature for 16 h with cobalt(II) acetate (10 mg/
gram of polymer (dry weight)) in methanol (5.0 mL g
(polymer)-1). The polymers were collected on a 15 µm filter
and then dried under vacuum. [Co2+ determinations were
conducted in duplicate on triplicate polymer samples. The
Co2+ present in the polymers were determined by the quan-
titative spectrophotometric (520 nm) analysis of the residual
Co2+ in the filtrate: 5.76 (DBM-Co2+ MIP) and 5.44 (Co2+ MIP)
µmol (site) g (polymer)-1.] Co2+-treated MIP samples (200 mg)
were incubated with ketone (200 µmol) and benzaldehyde (200
µmol) in dry N,N′-dimethylformamide (1.0 mL). Solution
reactions were carried out as above with pyridine (0.01 mL)
and cobalt(II) acetate (8 µmol). Sealed reaction mixtures were
heated in a thermostated oil bath at 100 °C. Aliquots (10 µL)
were taken directly from reaction mixtures, and no significant
difference was observed with results obtained by total extrac-
tion of the polymer matrix. Samples were analyzed by RP-
HPLC using a Serva C-18 column (300 × 2 mm). Samples
were serially diluted 100-fold in buffer before filtering and were
run isocratically using methanol/water (75:25) as eluent at 0.35
mL min-1. The production of 4 and 10 were monitored at 280
nm and 11 at 254 nm. Standard curves of concentration
versus peak area and substrate/product peak ratios were
prepared over the concentration ranges used in the studies
for yield calculation. The reaction products 1024 and 1125 were
prepared as described previously.
F igu r e 4. (A) Lineweaver-Burk plot for the production of
chalcone (4) over a range of inhibitor concentrations [active
site concentration 5.76 µmol g (polymer)-1; maximum velocity
Vmax ) 0.61 ( 0.06 µmol h-1; Michaelis constant Km ) 1.23 (
0.04 µM]. (B) Dixon plot (Ki ) 60 ( 10 µM].
Finally, the technique offers potential for developing
tailor made catalysts, perhaps with catalytic functional-
ities not utilized in biology.
Ack n ow led gm en t. The financial support by the
Swedish Natural Science Research Council and the
Swedish Research Council for Engineering Sciences is
gratefully acknowledged.
Exp er im en ta l Section
Gen er a l. All solvents (Labscan) and chemicals (Aldrich)
were of HPLC or analytical grade. Monomers were distilled
immediately prior to use.
J O9516805
MIP P r ep a r a tion . DBM-Co2+ MIP: vinylpyridine (5) (420
mg), styrene (6) (4.20 g), and divinylbenzene (7) (5.20 g) were
mixed with the imprint molecule, dibenzoylmethane (DBM)
(1) (448 mg), 2,2′-azobis(2,4-dimethylvaleronitrile) (100 mg)
and cobalt(II) acetate (498 mg) in anhydrous methanol (2.5
mL) and chloroform (6.7 mL), briefly sonicated under vacuum
(24) Hori, K.; Ando, M.; Takaishi, N.; Inamoto, Y. Tetrahedron Lett.
1987, 28, 5883-5886.
(25) Batt, D. G.; Goodman, R.; J ones, D. G.; Kerr, J . S.; Mantegna,
L. R.; McAllister, C.; Newton, R. C.; Nurnberg, S.; Welch; Covington,
M. B. J . Med. Chem. 1993, 1434-1442.
(26) Sellergren, B.; Lepisto¨, M.; Mosbach, K. J . Am. Chem. Soc. 1988,
110, 5853-5860.