Construction of a molecular imprinting catalyst using target analoguetemplate and its application for an amperometric fructosylamine

sensor

Koji Sode *, Shigenori Ohta, Yoshitsugu Yanai, Tomohiko Yamazaki

Department of Biotechnology, Tokyo University of Agriculture and Technology, 2-24-16 Nakamachi, Koganei, Tokyo 184-8588, Japan

Received 28 May 2002; received in revised form 13 December 2002; accepted 8 March 2003

Abstract

Molecular imprinting technology is becoming a versatile tool for preparing tailor-made molecular recognition elements. However,

inherent problems of the molecular imprinting technology include the availability and preparation of template molecules. We

recently reported artificial enzyme sensors for fructosylamines constructed by imprinting with fructosyl valine (Fru-val), a model

compound for HbA1c (Anal. Lett., 2003). However, because the availability of Fru-val is limited, we attempted to construct a Fru-

val-oxidizing molecularly imprinted catalyst (MIC) utilizing the analogue molecule methyl valine (m-val) as template molecule. An

electrode employing the m-val-imprinted polymer showed 1.2-fold higher sensitivity toward Fru-val compared with the control

polymer-employing electrode. We also used the positively charged functional monomer allylamine as functional monomer in order

to increase the selectivity of the MIC toward Fru-val. The selectivity of the electrode immobilizing the allylamine-containing

polymer showed 1.7-fold higher response toward Fru-val than toward Fru-o-lys. By combining the use of both allylamine as the

functional monomer and m-val as the template molecule, an even better MIC-immobilized electrode was produced with a Fru-val

selectivity comparable to that constructed by imprinting with Fru-val.

# 2003 Elsevier Science B.V. All rights reserved.

Keywords: Molecularly imprinted polymer; Hemoglobin A1c; Amperometric sensor; Fructosylamine compound; Analogue compound

1. Introduction

Molecular imprinting technology is becoming a

versatile tool for preparing tailor-made molecular re-

cognition elements (Sellergren, 2001). Polymers with

specific affinities can be prepared in the presence of

template ‘‘imprinting’’ molecules followed by removal of

template from the polymer, thus forming a cavity

complementary to the target molecule. Due to the

availability of tailor-made molecular imprinted poly-

mers applied to affinity chromatography and solid

phase extraction have been reported (Kempe and

Mosbach, 1995; Stevenson, 1999; Bjarnason et al.,

1999). Molecularly imprinted polymers with improved

affinity and selectivity were also developed and applied

as substitutes for antibodies in an immunoassay (Haupt

et al., 1998; Surugiu et al., 2001). However, inherent

problems of the molecular imprinting technology in-

cluded the preparation of template molecules and their

availability. Because the template molecules are washed

out after polymerization without recycling, the resulting

polymer will be more expensive than consumed template

molecule. The target (template) molecules are often

unique and expensive, making the availability of the

template molecules the limiting factor for further

utilization of molecular imprinting technology.

We recently reported on the development of a

polymer catalyst that may work as a fructosylamine

dehydrogenase and showed its potential application for

the construction of an amperometric fructosyl valine

sensor (Sode et al., 2001a,b). Polyvinylimidazole func-

tioned as catalyst of the oxidative hydrolytic reaction of

fructosylamine compounds in the presence of electron

acceptor at neutral pH (Sode et al., 2001a,b; Yamazaki

et al., 2003). The presence of imidazole, a general base* Corresponding author. Fax: �/81-42-388-7027.

E-mail address: sode@cc.tuat.ac.jp (K. Sode).

Biosensors and Bioelectronics 18 (2003) 1485�/1490

www.elsevier.com/locate/bios

0956-5663/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved.

doi:10.1016/S0956-5663(03)00125-8

catalyst, and the electron acceptor 1-methoxy-5-methyl-

phenazinium methylsulfate (methoxy-PMS) resulted in

oxidative cleavage of fructosylamine to produce an

imine compound (Fig. 1). The resulting reduced electron

acceptor is then oxidized on the electrode surface. The

imine compound is then hydrolyzed to glucosone and

valine.

There has been much focus on fructosylamine com-

pounds as important indicators for diabetic control

(Gabbay et al., 1997). Fructosylamine compounds

include the degraded product of HbA1c, in which the

N-terminal valine residue of the b-globin is glycated,

and glycated albumin, in which the o-position amino

group of the lysine residues are glycated. There is an

increasing demand for a selective measurement method

for HbA1c and its degraded derivatives. We, therefore,

attempted to improve the selectivity of the polymer

catalyst and construct an artificial Fru-val dehydrogen-

ase based on molecular imprinting technology. Molecu-

larly imprinting catalyst (MIC) was synthesized by

imprinting the copolymer of polyvinylimidazole and

polyvinylphenylboronate using fructosyl valine (Fru-

val), the model compound for HbA1c, as the template

molecule (Yamazaki et al., 2003). The amperometric

sensor employing MIC showed a higher response to

Fru-val than to fructosyl-o-lysine (Fru-o-lys), the model

compound of glycated albumin (Yamazaki et al., 2003).

However, the Fru-val is unique and expensive. The

availability of Fru-val is, therefore, the limiting factor

for the further utilization of molecular imprinting

technology for artificial fructosylamine dehydrogenase

construction.

We, therefore, attempted to improve the selectivity of

the polymer catalyst by using: (1) an analogue com-

pound of the target molecule as template, (2) a new

functional monomer showing low affinity toward thecompetitive target molecule, and (3) the combination of

(1) and (2). In this paper, we report the construction of

an artificial Fru-val dehydrogenase by molecular im-

printing technology and its application for ampero-

metric Fru-val sensing. We utilized an analogue

molecule of Fru-val as the template to imprint the

catalytic polymer, the co-polymer of polyvinylimidazole

and polyvinylphenylboronate, as well as allylamine asthe functional monomer to decrease the affinity toward

the competitive target molecule, Fru-o-lys.

2. Materials and methods

2.1. Chemicals

Fructosylamine compounds, Fru-val and Fru-o-lys,

were synthesized as previously reported (Keil et al.,

1995). 1-Vinylimidazole (Tokyo Kasei, Tokyo, Japan),

methoxy-PMS (Dojindo, Kumamoto, Japan), and car-

bon paste (Bioanalytical Systems, West Lafayette, IN,

USA) were purchased. Methylvaline (m-val) and

methyllysine (Z) (m-o-lys) were obtained from BachemAG (Bubendorf, Switzerland). 4-Vinylphenylboronate

and allylamine were purchased from Aldrich (Steinheim,

Germany).

Fig. 1. Oxidative fructosylamine cleavage reaction and detection on MIC-employing electrode.

K. Sode et al. / Biosensors and Bioelectronics 18 (2003) 1485�/14901486

2.2. Polymer preparation

m-val imprinted polymer was prepared by dissolving

2.0 mmol ethylene glycol dimethacrylate (EDMA), 1.6mmol 1-vinylimidazole, 0.8 mmol 4-vinylphenylboro-

nate, 0.2 mmol m-val as the template molecule, and 0.08

mmol polymerization initiator 2,2?-azobis(2,4-dimethyl-

valeronitrile) in 695 ml methanol/ultra pure water (4/1) in

a glass tube. The solution was purged thoroughly with

nitrogen gas for 2 min and then polymerized at 45 8C for

12 h. After polymerization, a white bulk polymer was

obtained. The bulk polymer was crushed, ground in amechanical mortar, wet sieved using acetone through a

38 mm sieve, and precipitated using acetone. To remove

the template molecules, the polymer was washed with 30

ml methanol/acetic acid (7/3) two times for 2 h, washed

with 30 ml methanol (2�/30 min), and then dried in

vacuo . Control polymers were prepared in the same

manner but without the template molecule.

Allylamine copolymer was prepared in the samemanner but with 1.6 mmol allylamine as the functional

monomer without any template molecule.

2.3. Sensor construction and operation

Sensors were constructed by mixing 20 mg polymers

with 50 mg of carbon paste and 20 ml mineral oil, and

packing the mixture into the electrode assembly (3.0 mm

diameter, Bioanalytical Systems). Amperometric mea-surements and cyclic voltammograms were carried out

using a Hokuto Denko potentiostat HA-151 (Tokyo,

Japan) with the three-electrode system. The working

electrode, a reference electrode (Ag/AgCl electrode,

Bioanalytical Systems), and a platinum counter elec-

trode (0.5 mm diameter, Tanaka Noble Metal, Tokyo,

Japan) were joined in the cell through holes in the

Teflon cover. All measurements were carried out at40 8C in 10 ml of 10 mM potassium phosphate buffer

(pH 7.5) containing 1 mM methoxy-PMS as electron

acceptor, with stirring at 250 r.p.m. The anodic applied

potential for the oxidation of methoxy-PMS was �/100

mV versus the Ag/AgCl (3.0 M NaCl) electrode.

3. Results and discussion

3.1. MIC preparation using m-val and m-o-lys as the

template analogue

Fig. 2 shows the schematic diagram of the preparation

of MICs (P3 and P4) using either m-val or m-o-lys as the

template molecule. The MIC was synthesized using 4-

vinylphenylboronate and 1-vinyl-imidazole as the func-tional monomers, and EDMA as the cross-linking

monomer. As we reported previously, the rate of

fructosylamine oxidation reaction also increased with

increasing amount of polyvinylimidaozle (Sode et al.,

2001a,b). In order to enhance the sensitivity of the

polymer-employing electrode to fructosylamine com-

pounds, an imidazole-based MIC containing a smallamount of cross-linker was prepared. We used m-val as

the analogue (template) molecule in this study. Con-

sidering the interaction between boronate and the cis-

diol of the target molecule, we used 4-vinylphenylbor-

onate to increase the affinity of MIC toward fructosy-

lamine compounds. Although m-val does not possess

cis-diols, the random presence of boronate in the

polymer may be located in the cavity that recognizesthe cis-diol of Fru-val. We expected that m-val would

coordinate in the proper orientation with respect to both

EDMA, by hydrophobic interaction with the valine side

chain, and vinylimidazole, by interaction with Schiff

base to be cleaved. We therefore chose m-val as the

analogue molecule of Fru-val. In order to evaluate this

imprinting effect, we also prepared MIC (P4) using m-o-

lys as the template molecule and the analogue moleculeof Fru-o-lys.

Fig. 3 shows the calibration curves for the measure-

ment of fructosylamine compounds using the electrode

employing m-val imprinted MIC (P3) compared with

control polymer (P1). Calibration curves were obtained

by plotting the steady state current versus the concen-

tration of the injected sample. The fructosylamine

compound measurements were carried out at three timesconsecutively, and the reproducibility of MIC was

within 2% errors. The m-val imprinted polymer (P3)-

employing electrode showed 1.2-fold higher sensitivity

toward Fru-val compared with the control polymer

(P1)-employing electrode. Toward Fru-o-lys, however,

m-val imprinted polymer (P3)-employing electrode

showed 80% of the sensitivity compared with the control

polymer (P1)-employing electrode. These results demon-strated that the analogue template m-val would coordi-

nate the proper orientation with respect to the

functional monomers. The selectivity of the polymer

catalyst was improved by molecular imprinting.

We also checked the stability of polymer. The same

response was obtained, even if the MIC sensor was heat-

treated for 85 8C at 10 min (data was not shown). Since

the measurement of Fru-val is generally carried out atroom temperature or physiological condition (around

37 8C), the stability of MIC is for practical use.

Table 1 shows a comparison of the selectivities of Fru-

val sensors employing polymers constructed in this

study. Linear correlations were observed between the

increased current of the electrode employing polymers

and the fructosylamine compound concentrations. The

sensitivities toward Fru-val and Fru-o-lys and thecoefficients of determination (r2) were calculated from

the slopes of the calibration curves, In contrast to the m-

val-imprinted polymer (P3)-employing electrode, the

electrode employing m-o-lys imprinted MIC (P4)

K. Sode et al. / Biosensors and Bioelectronics 18 (2003) 1485�/1490 1487

showed 1.3-fold higher sensitivity toward Fru-o-lys

compared with the control polymer (P1)-employing

electrode, and 10% lower sensitivity toward Fru-val.

Molecular imprinting with m-o-lys, therefore produced a

the polymer catalyst with increased selectivity for Fru-o-

lys versus Fru-val. These results demonstrate that MIC

Fig. 2. A schematic diagram showing the preparation of MIC (P3 and P4) and the proposed molecular recognition mechanism for Fru-val using m-

val-imprinted polymer. 1-Vinylimidazole recognizes the substituted amino group of the template molecule (either m-val or m-o-lys) during the

polymerization step. Polymerization of the functional monomer (4-vinylboronate) and cross-linker (EDMA) is carried out in the presence of either

m-val (P3) or m-o-lys (P4). Subsequent removal of the template yields a substrate binding pocket complementary to the template. Chemical

compounds: (a) m-val, (b) 1-vinylimidazole, (c) EDMA, (d) 4-vinylphenylboronate, (e) Fru-val, (f) m-o-lys, (g) Fru-o-lys.

K. Sode et al. / Biosensors and Bioelectronics 18 (2003) 1485�/14901488

with selectivity for specific target fructosylamine mole-cules can be constructed by imprinting with an analogue

of each target molecule.

3.2. Preparation of polymer catalyst using allylamine as

the functional monomer

The application of a cationic functional monomer was

investigated in order to increase the selectivity of MIC

for Fru-val. Since the discrimination of Fru-val from

Fru-o-lys is necessary for the practical use of a MIC-

based sensor, we attempted to increase the selectivity of

the polymer itself for Fru-val versus Fru-o-lys. Con-sidering that Fru-o-lys is a cationic molecule, we

expected that the use of positively charged functional

monomers, such as allylamine, N ,N ?-diethyl aminoethyl

methacrylamide, and N ,N ,N -trimethyl aminoethyl-

methacrylate, may decrease the affinity of catalytic

polymer for Fru-o-lys. This may consequently increase

the selectivity of MIC toward Fru-val. Allylamine has

already been used to utilize as a positively chargedfunctional monomer (McNiven et al., 1998; Piletsky et

al., 1998; Suarez-Rodrıguez and Dıaz-Garcıa, 2001;

Wizeman and Kofinas, 2001). Allylamine is classified

as a primary amine compound and has a weaker positive

charge than N ,N ?-diethyl aminoethyl methacrylamide

and N ,N ,N -trimethyl aminoethylmethacrylate, which

are classified as secondary and tertiary amine com-pounds, respectively. Considering the formation of

stable electrostatic interactions between the carboxyl

acid group of fructosylamine compounds and secondary

or tertiary amine compounds, we used primary amine

compound allylamine as a positively charged functional

monomer.

The polymer catalyst (P5) was prepared by co-

polymerizing allylamine, 1-vinylimidazole, and 4-vinyl-phenylboronate. The polymer (P5)-based sensor showed

a typical amperometric enzyme sensor response curve

with the injection of Fru-val and Fru-o-lys. The

selectivity of the electrode immobilizing polymer (P5)

showed 1.7-fold higher sensitivity for Fru-val than for

Fru-o-lys (Table 1). The selectivity achieved with this

electrode is mainly due to the decrease in the response

toward Fru-o-lys compared with the polymer withoutallylamine (P1). The use of allylamine as the functional

monomer in the preparation of the artificial fructosyl

valine dehydrogenase therefore resulted in a decrease in

activity toward Fru-o-lys.

3.3. Fru-val sensor employing artificial fructosyl valine

dehydrogenase

We then prepared a novel MIC (P6) using allylamine

as the functional monomer and m-val as the template

molecule in order to enhance the selectivity of the m-val-based artificial fructosyl valine dehydrogenase prepared

using m-val as a template. The sensitivity of the sensor

employing MIC (P6) toward Fru-val was about 1.9-fold

higher than those for Fru-o-lys. This selectivity was

higher than that observed in allylamine/vinylimidazole/

vinylphenylboronate copolymer (P5) and also higher

than that observed in m-val imprinted vinylimidazole/

vinylphenylboronate copolymer (P3). The selectivity forFru-val of the sensor employing MIC (P6) using

allylamine as the functional monomer and m-val as

the template molecule is the best analogue imprinted

Fig. 3. Fru-val (circle) and Fru-o-lys (triangle) calibration curves for

the sensor electrodes immobilizing either MIC (P3) prepared by m-val-

imprinting (filled symbols) or control polymer (P1) (open symbols).

Table 1

Sensitivity and selectivity of polymers for fructosylamine compounds

Polymer Template Allylamine Sensitivity (nA/mM) Selectivity (Fru-val/Fru-o-lys)

Fru-val Fru-o-lys

P1 �/ �/ 79 (R2�/0.984) 70 (R2�/0.988) 1.1

P2 Fru-val �/ 135 (R2�/0.992) 75 (R2�/0.982) 1.8

P3 m-val �/ 95 (R2�/0.984) 60 (R2�/0.968) 1.6

P4 m-o-lys �/ 84 (R2�/0.980) 101 (R2�/0.976) 0.8

P5 �/ �/ 91 (R2�/0.996) 54 (R2�/0.998) 1.7

P6 m-val �/ 95 (R2�/0.992) 50 (R2�/0.980 1.9

K. Sode et al. / Biosensors and Bioelectronics 18 (2003) 1485�/1490 1489

catalyst and is similar to those achieved with the sensor

employing MIC (P2) prepared by Fru-val imprinting.

The requirement in the specificity for the fructosyla-

mine detection is the discrimination of Fru-val fromFru-o-lys. Current proposed biochemical methods for

HbA1c measurement require proteolytic digestion of

HbA1c to generate small molecular weight fructosyla-

mine compounds to be subjected to the enzyme fructosyl

amine oxidase (FAOD) (Tsugawa et al., 2000; Ogawa et

al., 2002). This process also generates the proteolytic

products of glycated albumin containing Fru-o-Lys. An

acceptable catalytic component of a diagnostic kit forHbA1c should therefore be selective for Fru-val over

Fru-o-lys. The enzyme FAOD, which oxidatively de-

grades fructosylamine compounds, has been used as

fructosylamine sensor constituent (Tsugawa et al., 2000;

Ogawa et al., 2002). FAOD shows broad substrate

specificity and is generally not specific for fructosyl

valine (Yoshida et al., 1996; Sode et al., 2001b). MIC

(P6) is among the most selective catalysts recognizingfructosyl valine. Because our aim is selective measure-

ment of Fru-val, further optimization of the functional

monomer and operational conditions of MIC-employ-

ing sensors is required to improve the specificity of MIC

for Fru-val.

An artificial fructosyl amine dehydrogenase with

similar or higher selectivity toward Fru-val than MIC

prepared Fru-val imprinting was constructed usingappropriate analogue template and a functional mono-

mer with decreased affinity toward the competitive

molecule.

Acknowledgements

The authors thank Dr Stefano Ferri for kindlyrevising the manuscript.

References

Bjarnason, B., Chimuka, L., Ramstrom, O., 1999. On-line solid-phase

extraction of triazine herbicides using a molecularly imprinted

polymer for selective sample enrichment. Analytical Chemistry 71,

2152�/2156.

Gabbay, K.H., Hasty, K., Breslow, J.L., Ellison, R.C., Bunn, H.F.,

Gallop, P.M., 1997. Glycosylated hemoglobins and long-term

blood glucose control in diabetes mellitus. Journal of Clinical

Endocrinology and Metabolism 44, 859�/864.

Haupt, K., Mayes, A.G., Mosbach, K., 1998. Herbicide assay using an

imprinted polymer based system analogous to competitive fluor-

oimmunoassays. Analytical Chemistry 70, 3936�/3939.

Keil, P., Mortensen, H.B., Christophersen, C., 1995. Fructosylvaline: a

simple mode of the N-terminal residue of human hemoglobin A1c.

Acta Chemica Scandina Series B 39, 191�/193.

Kempe, M., Mosbach, K., 1995. Molecular imprinting used for chiral

separations. Journal of Chromatography A 694, 3�/13.

McNiven, S., Kato, M., Levi, R., Yano, K., Karube, I., 1998.

Chloramphenicol sensor based on an in situ imprinted polymer.

Analytica Chimica Acta 365, 69�/74.

Ogawa, K., Stollner, D., Scheller, F., Warsinke, A., Ishimura, F.,

Tsugawa, W., Ferri, S., Sode, K., 2002. Development of a flow-

injection analysis (FIA) enzyme sensor for fructosyl amine

monitoring. Analytical Bioanalytical Chemistry 373, 211�/214.

Piletsky, S.A., Piletskaya, E.V., Panasyuk, T.L., El’skaya, A.V., Levi,

R., Karube, I., Wulff, G., 1998. Imprinted membranes for sensor

technology: opposite behavior of covalently and noncovalently

imprinted membranes. Macromolecules 31, 2137�/2140.

Sellergren, B. (Ed.), Molecularly Imprinted Polymers-Man-Made

Mimics of Antibodies and Their Application in Analytical Chem-

istry. Elsevier, Amsterdam, The Netherlands 2001.

Sode, K., Takahashi, Y., Ohta, S., Tsugawa, W., Yamazaki, T., 2001a.

A new concept for the construction of an artificial dehydrogenase

for fructosylamine compounds and its application for an ampero-

metric fructosylamine sensor. Analytica Chimica Acta 435, 151�/

156.

Sode, K., Ishimura, F., Tsugawa, W., 2001b. Screening and character-

ization of fructosyl-valine utilizing marine microorganisms. Marine

Biotechnology 3, 126�/132.

Stevenson, D., 1999. Molecular imprinted polymers for solid-phase

extraction. Trends in Analytical Chemistry 18, 154�/158.

Suarez-Rodrıguez, J.L., Dıaz-Garcıa, M.E., 2001. Fluorescent compe-

titive flow-through assay for chloramphenicol using molecularly

imprinted polymers. Biosensors and Bioelectronics 16, 955�/961.

Surugiu, I., Danielsson, B., Ye, L., Mosbach, K., Haupt, K., 2001.

Chemiluminescence imaging ELISA using an imprinted polymer as

the recognition element instead of an antibody. Analytical Chem-

istry 73, 487�/491.

Tsugawa, W., Ishimura, F., Ogawa, K., Sode, K., 2000. Development

of an enzyme sensor utilizing a novel fructosylamine oxidase from a

marine yeast. Electrochemistry 68, 869�/871.

Wizeman, W.J., Kofinas, P., 2001. Molecularly imprinted polymer

hydrogels displaying isomerically resolved glucose binding. Bio-

materials 22, 1485�/1491.

Yamazaki, T., Ohta, S., Yanai, Y., Sode, K., 2003. Molecular

imprinting catalyst based artificial enzyme sensor for fructosyla-

mines. Analytical Letters 36, 75�/89.

Yoshida, N., Sakai, Y., Isogai, A., Fukuya, H., Yagi, M., Tani, Y.,

Kato, N., 1996. Primary structures of fungal fructosyl amino acid

oxidases and their application to the measurement of glycated

proteins. European Journal of Biochemistry 242, 499�/505.

K. Sode et al. / Biosensors and Bioelectronics 18 (2003) 1485�/14901490