Molecular imprinting of fructose using a polymerizable benzoboroxole: Effective complexation at pH 7.4

Article abstract of DOI:10.1016/j.polymer.2011.04.002

Covalent molecularly imprinted polymers against d-fructose employing 5-methacrylamido-2-hydroxymethylphenylboronic acid as functional monomer and trimethylpropane trimethacrylate (TRIM) as the crosslinking agent were prepared by a conventional radical bul

Full text of DOI:10.1016/j.polymer.2011.04.002

Polymer 52 (2011) 2485e2491
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Molecular imprinting of fructose using a polymerizable benzoboroxole: Effective
complexation at pH 7.4
, ,
,
Soeren Schumacher a 1, Franziska Grüneberger a 1, Martin Katterle a, Cornelia Hettrich a, Dennis G. Hall b,
*
Frieder W. Scheller a, Nenad Gajovic-Eichelmann a
a Fraunhofer Institute for Biomedical Engineering, Am Mühlenberg 13, 14476 Potsdam, Germany
b Department of Chemistry, University of Alberta, Edmonton, Alberta T6G 2G2, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
Covalent molecularly imprinted polymers against D-fructose employing 5-methacrylamido-2-hydrox-
Received 11 January 2011
Received in revised form
30 March 2011
Accepted 1 April 2011
Available online 19 April 2011
ymethylphenylboronic acid as functional monomer and trimethylpropane trimethacrylate (TRIM) as the
crosslinking agent were prepared by a conventional radical bulk polymerization (MIP-BX(Fru)). Batch
binding studies for fructose in aqueous buffers containing 10% methanol revealed that the binding
capability of MIP-BX(Fru) is paramount compared to a MIP prepared with vinylphenylboronic acid MIP-
BA(Fru). Especially, at the biological important pH-value of 7.4 the rebinding of fructose to the MIP-
BX(Fru) is with 60 nmol per mg polymer about 3.2 higher compared to the MIP-BA(Fru). A pinacol
imprinted polymer was also investigated and showed in case of MIP-BX still an imprinting of 1.7 at pH 7.4
whereas MIP-BA did not show a difference. Cross-reactivity studies at pH 7.4 show the shape-selectivity
Keywords:
Molecular imprinting
Benzoboroxole
Fructose
of the MIP-BX(Fru) in the order of
L-fructose, sorbitol, glucose and sucrose.
Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction
monomer complex is formed and copolymerized with a cross-
linking agent. Affinity maybe tuned by the choice of the functional
monomer e template interaction whereas selectivity can be obt-
ained by the polymer itself a highly crosslinked and therefore rigid.
After the polymerization, the template is extracted, and, in prin-
ciple, the polymer bearing artificial binding sites is able to (re-) bind
the analyte. Molecularly imprinted polymers are described for
a variety of different analytes ranging from small molecules
including saccharides, proteins to even whole cells [16]. These
artificial receptor units are used for solid phase extraction, chro-
matography or sensor applications.
Molecularly imprinted polymers for monosaccharides and
derivatives thereof were synthesized making use of covalent, non-
covalent or metal-coordinated interactions between the saccharide
and the functional monomer. A problem (to be addressed for
aqueous media) lies in the competition of hydroxyl groups attached
to the carbohydrate scaffold and the most abundant ones from
water. Therefore, in many cases organic media or alkaline condi-
tions had to be employed for saccharide rebinding. For example,
Mosbach et al. described a non-covalent molecularly imprinted
Molecular recognition of saccharides in water at physiological
pH-values is one prominent problem in (bio-)organic chemistry
[1e6]. Proteins and nucleic acids exhibit a variety of different
functional groups combined with different structural conforma-
tions, which makes their molecular recognition comparatively
easier [7e10]. In contrast, unsubstituted saccharides differ just in
their intrinsic geometrical orientation of hydroxyl groups and
especially in aqueous media the competition between the sugar’s
hydroxyl groups and hydroxyl groups from water molecules is
a major challenge [11,12]. One method of producing selective
materials in a robust and easy way is the molecular imprinting
approach [13e15].
In general, molecular imprinting offers the possibility to insert
tailor-made artificial binding sites on a molecular level into
a 3-dimensional polymeric network. For that a polymerization is
carried out in the presence of the later analyte which is called the
‘template’. Due to the employment of functional monomers that
are able to interact with the template molecule in a covalent,
non-covalent or metal-coordinated way, a template-functional
polymer with p-nitrophenyl-
a-D-galactoside in ethylene glycol
dimethacrylate (EGDMA) crosslinked polymers [17]. Recently
sucrose was imprinted with MAA and EGDMA for aqueous batch
rebinding [18]. Metal coordination is also a well-known method to
bind saccharides and to polymerize copper-ligands into a molecu-
larly imprinted polymer as functional monomer [19].
* Corresponding author.
E-mail address: soeren.schumacher@ibmt.fraunhofer.de (S. Schumacher).
1
Contributed equally to this work.
0032-3861/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2011.04.002

2486
S. Schumacher et al. / Polymer 52 (2011) 2485e2491
The use of aryl boronic acids for saccharide recognition presents
was stirred for 10 min at 0 ꢀC. Then 1.6 mL (22 mmol) of meth-
acryloylchloride 2 was slowly added via a syringe pump and
vigorous stirring within 1 h maintaining the 0 ꢀC. After 4 h the
solution was acidified very slowly using concentrated HCl avoiding
unwanted polymerization. The pale yellow precipitate which
appeared during the acidification was filtered off yielding 3 as pale
yellow crystals which were pure by 1H NMR. Yield: 79e82% 1H
several advantages [20,21]. They are able to bind cis-diol containing
compounds forming a covalent, reversible cyclic boronic acid ester.
This boronic ester is easily formed in alkaline media because an
additional hydroxyl ion can coordinate to the boron and saturates
its electron deficiency e combined with that the resulting tetra-
hedral boron changes its hybridization state from sp2 to sp3. The
optimal binding pH is dependent on the pKa of the boronic acid
employed and the specific cis-diol compounds. In general, electron
withdrawing groups attached to the boronic acid are able to lower
the binding pH. Another principle firstly described by Wulff and
co-workers is to provide a hemilabile ligand in ortho-position to
the boron [22]. Different hemilabile ligands such as amine- or
carbonyl-containing residues are described since the electron lone
pairs of these groups can coordinate intramolecularly to the boron
[23]. Recently, Hall and co-workers screened a variety of ortho-
substituted aryl boronic acids for their ability to bind glycopyr-
anosides and found that an ortho-hydroxylmethyl group is effective
at promoting the covalent complexation of diol molecules [24,25].
This type of benzoboroxole was described as an effective binding
agent for saccharide recognition at pH 7.4 [26]. Consequently, the
benzoboroxole was used in different approaches as binding agent
for the detection of glycoproteins such as the TF-antigen or the
gp120 of HI virus [27,28]. Benzoboroxole-containing molecularly
imprinted polymers are described for monoalcohols or steroids but
so far not for the recognition of unprotected monosaccharides such
as fructose [29,30].
Herein, we report the synthesis of benzoboroxole-containing
covalently imprinted polymers using fructose as a model template
due to its high binding affinity to aryl boronic acids. As the functional
monomer 5-methacrylamido-2-hydroxymethylphenylboronic acid
3 was synthesized and employed due to its known ability to bind
saccharides at a physiological pH-value. For comparison purposes,
the appropriate pinacol-imprinted (MIP-BX(Pin)) and also the
fructose (MIP-BA(Fru)) and pinacol (MIP-BA(Pin)) imprinted poly-
mers with 3-vinylphenylboronic acid 4 as functional monomer were
synthesized. Accordingly, batch binding experiments were per-
formed at pH 11.4, 8.7 and 7.4 to show the binding behavior of these
molecularly imprinted polymers at different pH-values. Moreover,
the binding of fructose to a control polymer was also tested in order
to assess the unspecific binding. The shape-selectivity of the MIP-
NMR (300 MHz, DMSO):
d
[ppm] ¼ 8.06 (m, 1H, AreH); 7.67 (m, 1H,
AreH); 7.34 (m, 1H, AreH); 5.80 (s, 1H, 13-H); 5.50 (s, 1H, 13-H);
4.95 (s, 2H, 7-H); 1.96 (s, 3H, 12-H). 13C NMR (300 MHz, DMSO):
d
[ppm] ¼ 166.71, 148.97, 140.33, 137.67, 123.52, 122.26, 121.25,
119.78, 69.68, 18.69. MS (ESIþ): m/z ¼ 217.12 (Mþ), 218.10 (M þ Hþ),
240.10 (M þ Naþ).
2.3. Synthesis of fructose ester 5a and 5b
Fructose ester synthesis (5a and 5b) was carried out as
described by Wulff et al. [31]
with the desired boronic acid 3 or 4 (5 mmol) in 80 mL of dioxane
solution in presence of nitrobenzene (10 L) under argon atmo-
D-fructose (2.5 mmol) was esterified
m
sphere. The water generated was removed by azeotropic distillation
(88 ꢀC) from the reaction mixture. The residual solvent was
removed in vacuo. Purification was performed by dissolution of the
ester in DCM.
Analytical data 5a: Yield: 80.5% 1H NMR (300 MHz, CDCl3):
d
[ppm] ¼ 8.31e7.51 (m, 6H, AreH), 5.79 (s, 2H, 19-H, 31-H), 5.45
(s, 2H, 19-H, 31-H), 5.03 (s, 4H, 14-H, 26-H), 2.04 (s, 6H, 18-H, 30-H).
As discussed in the Results and Discussion part, strong peak broad-
ening was observed. 13C NMR (300 MHz, CDCl3):
d
[ppm] ¼ 167.22,
166.92, 149.80, 140.56, 136.85, 129.25, 126.83, 123.82, 123.45, 120.14,
103.60, 73.25, 72.56, 72.22, 71.12, 70.80, 66.04, 61.58, 18.68. IR: n
3276 cmꢁ1 (s), 2927 cmꢁ1 (m), 1655 cmꢁ1 (m), 1614 cmꢁ1 (m),
1524 cmꢁ1 (m), 1398 cmꢁ1 (s), 1214 cmꢁ1 (w), 1054 cmꢁ1 (s),
980 cmꢁ1 (s), 914 cmꢁ1 (m), 822 cmꢁ1 (s), 754 cmꢁ1 (s) MS (ESIꢁ):
m/z ¼ 577.27 (M þ Hþ)
Analytical data 5b: Yield: 82% 1H NMR (500 MHz, CDCl3):
d
[ppm] ¼ 7.89 (m, 2H, AreH), 7.73 (m, 2H, AreH), 7.57 (m, 2H,
AreH), 7.38 (m, 2H, AreH), 6.74 (dd, 2H, 14-H, 22-H, J ¼ 10.9,
17.6 Hz), 5.81 (ddd, 2H, 15-H, 23-H, J ¼ 0.4, 7.5, 17.6 Hz), 5.28 (dd,
2H, 15-H, 23-H, J ¼ 8.5, 11.1 Hz), 5.12 (dd, 1H,4-H, J ¼ 2.4, 8.4 Hz),
4.85 (d,1H, 3-H, J ¼ 2.4 Hz), 4.71 (dd,1H, 5-H, J ¼ 1.6, 8.4 Hz), 3.98(d,
1H, 6-H, J ¼ 13.7 Hz), 3.85 (dd, 1H, 1-H, J ¼ 5.1, 12.1 Hz), 3.78 (dd, 1H,
6-H, J ¼ 2.1, 13.8 Hz), 3.67 (d, 1H, 1-H, J ¼ 12.1 Hz). 13C NMR
BX(Fru) was investigated by competition between
-fructose, glucose, sucrose or sorbitol.
D-fructose and
L
(500 MHz, CDCl3):
d
[ppm] ¼ 137.23, 137.09, 136.56, 136.38, 134.56,
2. Materials and methods
134.34, 133.11, 132.89, 130.08, 129.52, 128.28, 128.16, 114.46, 114.21,
105.01, 72.69, 72.60, 72.43, 65.99, 61.88. MS (ESIþ): m/z ¼ 405, 40
(M þ Hþ), 427.19 (M þ Naþ), 443, 18 (M þ Kþ).
2.1. Chemicals
3-Vinylphenylboronic acid, pinacol (2,3-dimethyl-2,3-butane-
diol), methacryloylchloride (97%), trimethylolpropane trimethacry-
late (techn.) (TRIM), Glucose, Sucrose and Sorbitol were purchased
from SigmaeAldrich, 5-amino-2-hydroxymethylphenylboronic acid
2.4. Synthesis of pinacol ester 6a and 6b
Pinacol esters (6a and 6b) were synthesized analogous. Pinacol
(590 mg, 5 mmol) was esterified with the equimolar boronic acids 3
as a dehydrated HCl salt, from Combi-Blocks,
-fructose [3H-(G)] from Biotrend and AIBN from Fluka. As scintilla-
tion liquid Rotiszint from Carl Roth was purchased. -fructose was
purchased by TCI Europe. Solvents for polymerization were dried
before use or purchased as anhydrous grades. All other substances
were used without further purification.
D
-fructose from Merck,
or 4 in 180 mL of toluene in presence of nitrobenzene (10 mL) under
D
argon atmosphere. Generated water was removed by azeotropic
distillation (84 ꢀC) and finally solvent was removed in vacuo.
Analytical data 6a: Yield: Pale yellow solid, 90% 1H NMR
L
(300 MHz, CDCl3):
AreH), 5.79 (s,1H, 12-H), 5.46 (s, 1H, 12-H), 4.67 (s, 2H, 7-H), 2.06
d
[ppm] ¼ 8.24 (m, 1H, 8-H), 8.01e7.48 (m, 3H,
(s, 12H, 15-H, 16-H, 17-H, 18-H). 13C NMR (300 MHz, CDCl3):
2.2. Synthesis of 5-methacrylamido-2-hydroxymethylphenylboronic
acid 3
d
[ppm] ¼ 166.44, 143.45, 140.76, 136.72, 129.63, 129.24, 127.66,
122.92, 119.81, 117.74, 84.43, 65.44, 24.79, 18.68. MS (ESIþ):
m/z ¼ 340.20 (M þ Hþ þ Naþ).
1.5 g (10.9 mmol) 5-amino-2-hydroxymethylphenylboronic acid
1 and 1.75 g (43.8 mmol) sodium hydroxide were dissolved in
a small amount water to yield an almost saturated solution which
IR: n 3310 cmꢁ1 (m), 2979 cmꢁ1 (m), 2931 cmꢁ1 (w), 1664 cmꢁ1
(m), 1529 cmꢁ1 (s), 1343 cmꢁ1(s), 1142 cmꢁ1 (s), 1068 cmꢁ1 (m),
967 cmꢁ1 (w), 855 cmꢁ1 (w).

S. Schumacher et al. / Polymer 52 (2011) 2485e2491
2487
Analytical data 6b: Yield: 70e78%, oil 1H NMR (300 MHz,
CDCl3):
AreH); 7.34 (m, 1H AreH); 6.73 (dd, 1H, 7-H, J ¼ 10.9, 17.6 Hz); 5.79
(dd,1H, 8-H, J ¼ 0.9,17.6 Hz); 5.24 (dd,1H, 8-H, J ¼ 0.9,10.9 Hz); 1.36
(s, 12H, 11-H, 12-H, 13-H, 14-H). 13C NMR (500 MHz, CDCl3):
5-amino-2-hydroxymethylphenylboronic acid 1, which reacted
with methacryloylchloride 2 at alkaline pH-values yielding the
polymerizable benzoboroxole 3 after slow acidification.
d
[ppm] ¼ 7.84 (s, 1H, 2-H); 7.70 (m, 1H AreH); 7.52 (m, 1H
For the covalent imprinting approach it is necessary to synthe-
size the functional monomer e template complex before its poly-
merization. In the case of saccharide imprinting using aryl boronic
acids the particular boronic acid ester has to be synthesized.
Depending on the saccharide under investigation different binding
complexes between the aryl boronic acid and the saccharide are
revealed and proposed in the literature [33,34]. In our case, the
binding of fructose to aryl boronic acids has a high binding strength
due to the syn-periplanar arrangement of its hydroxyl groups
and was investigated by means of NMR spectroscopy. Norrild and
Eggert studied the esterification in solution between fructose and
p-tolylboronic acid varying the ratio between fructose and the
boronic acid derivative. By applying a 2:1 ratio between boronic
acid and fructose in DMSO they found different 1:1 (60%) binding
complexes and one 2:1 complex (33%) [33]. Since a stoichiometric
pure 2:1 binding complex for fructose is envisaged. Therefore,
azeotropic distillation in anhydrous dioxane was chosen for the
synthesis of the fructose-benzoboroxole ester 5a and fructose
vinylphenylboronic acid ester 5b since the removal of the released
water forces the reaction into the direction of the 2:1 boronic acid
ester. Using this ester as functional monomer template complex in
molecular imprinting is believed to result in higher binding
constants and superior selectivity also described by other authors.
A conversion of about 80% in each case was reached. Furthermore
the synthesis was started with the desired 2:1 ratio between the
aryl boronic acid derivative and fructose avoiding the removal of
unreacted boronic acid. The pinacol aryl boronic acid esters 6a and
6b were also prepared by azeotropic distillation.
d
[ppm] ¼ 136.86, 136.76, 134.18, 132.72, 128.85, 127.91, 113.86,
83.83, 24.86. MS (ESIþ): m/z ¼ 230.19 (Mþ), 231.19 (M þ Hþ).
2.5. Synthesis of polymers
For polymer synthesis a modified approach was applied where
2 mmol of the desired boronic acid ester 5a, 5b, 6a or 6b and
thermal initiator AIBN (295 mg, 1.8 mmol) were dissolved in 9 mL
THF for 5a/6a or 9 mL acetonitrile/toluene 1:1 for 5b/6b [32]. After
addition of crosslinker TRIM (9 g, 8.5 mL), mixture was mixed well
and purged with argon for 10 min. Polymerization was initiated and
carried out at 65 ꢀC for 48 h and afterward increased up to 95 ꢀC for
24 h. The synthesized polymer monoliths were crushed and ground
in a ball mill (MM200, Retsch) for 2 min at 30 Hz. Polymer powder
was wet sieved (mesh 25 mm) using acetone. Template molecules
fructose and pinacol were removed from polymers by washing with
methanol and water. The polymer particles were dried in vacuo and
stored at room temperatures.
2.6. Batch rebinding studies
Batch rebinding experiments were accomplished at three
different pH-values: pH 11.4 (0.1 M sodium carbonate solution þ 10%
methanol), pH 8.7 and 7.4 (0.1 M phosphate buffer þ 10% methanol
respectively). Fructose stock solutions of 10.0 mM were prepared
with each of these buffers and doped with 3H-fructose. Finally, to
10 mg of polymer, fructose stock solution and buffer was added to
yield 1 mL fructose solutions of 0.1e5 mM. The mixtures were incu-
The characterization of the products was easy in case of the
phenyl boronic acid esters and the analytical data matched data
already reported in Ref. [32]. In contrast, the characterization of the
5-methacrylamido-2-hydroxymethylphenylboronic acid ester 5a is
more complicated. As reported earlier, the esterification of benzo-
boroxole with different substituted saccharides leads to a peak
broadening in the aromatic region due to either slow exchange
times (which cannot be revealed by standard 1H NMR spectros-
copy) or due to the formation of different possible binding
bated over night at 20 ꢀC and centrifuged (9000 g) for 10 min. 500
mL
of the supernatant were discharged to a separate vessel and 5 mL of
scintillation solution (RotiszintÒ eco plus) were added. Radioactivity
wasmeasuredinScintillationcounter LS 650 (Beckmann Coulter)and
unbound fructose could be calculated with a calibration curve.
2.7. Competitive binding
complexes [25]. Thus, NMR experiments with D-fructose and ben-
Competitive binding studies were done with MIP-BX(Fru) at pH
7.4 with different sugars/sugar alcohol: L-fructose, sorbitol, glucose
zoboroxole were performed in deuterated phosphate buffer at pH
7.4 and at equimolar concentration (Fig. 1).
and sucrose. The same solutions as for batch binding experiments
were used, completed with different 50 mM sugar solutions. For
the experiments, fructose stock solution, buffer solution and 50 mL
of the desired sugar solution were added to 10 mg of polymer to
yield a 1 mL reaction mixture. Fructose concentrations varied from
0.1 to 4.5 mM and the competitive sugar/sugar alcohol had constant
concentrations of 2.5 mM. Incubation, centrifugation and detection
were done the same way as for simple binding experiments.
Beside new peaks at 7.2 and 7.4 ppm which can be assigned to
the formed benzoboroxole ester also huge peak broadening could
be observed. In line with these results, the synthesized polymer-
isable benzoboroxole ester 5a also showed a peak broadening
which leads to the assumption that the ester was formed. In
combination with mass spectrometry and IR data it is concluded
that the fructose e benzoboroxole ester was synthesized.
The synthesized template-functional monomer complexes were
then polymerized with trimethylolpropane trimethacrylate (TRIM)
as the crosslinking agent. Four polymers were synthesized, fruc-
tose-imprinted polymers with either esters 5a or 5b and pinacol
imprinted polymers with either 6a or 6b as control polymers
(Fig. 2). It was not possible to synthesize a polymer without the
template molecule since the solubility of the bare aryl boronic acid
derivatives in the pre-polymerization mixture was not sufficient.
Thus, a bare control polymer without any boronic acid content was
synthesized to evaluate the interaction of the polymeric backbone
with the targeted substances.
2.8. Adsorption/desorption measurements
The porosity of the MIPs was determined by nitrogen adsorp-
tion/desorption porosimetry on a Fisons Sorptomat 1900.
3. Results and discussion
3.1. Synthesis
Since benzoboroxole was described as a potential binding agent
for monosaccharides in aqueous media at pH 7.4 a corresponding
polymerizable derivative 3 was synthesized (Scheme 1). The
synthesis was performed starting from a commercially available
After the polymerization in a two temperature process the
template was removed by simple washing in MeOH/water mixtures
until no further fructose was detected in the washing solution using
a colorimetric anthrone assay (data not shown) [35].

2488
S. Schumacher et al. / Polymer 52 (2011) 2485e2491
Scheme 1. Synthesis of functional monomer and functional monomeretemplate complex.
The ground and sieved polymer particles in the size of
25e50 m (which were also used for batch binding studies) were
found in Table 1. For comparison also a control polymer containing
just crosslinker was analyzed.
m
characterized by means of nitrogen sorption measurements. Here,
the specific surface area (BET) and the pore volumes of either
mesopores (BJH) or micropores (HK) were analyzed. The data are
The values obtained for the surface area and pore volumes are
differing between the boronic acid containing polymers and the
control polymer devoid of it. The control polymer shows the
Fig. 1. NMR comparison of neat benzoboroxole derivatives and their fructose ester.

S. Schumacher et al. / Polymer 52 (2011) 2485e2491
2489
Fig. 2. Synthesis of MIP-BX(Fru), MIP-BA(Fru), MIP-BX(Pin) and MIP-BA(Pin).
highest BET surface area of 482 m2/g whereas the surface areas of
the imprinted polymers are between 332 and 436 m2/g. The values
reflect the synthetic procedure since the overall monomer
concentration is slightly lower for the control polymer. The poly-
meric structure in the mesopore regime shows also the same trend.
The control polymer which consists just of crosslinker shows with
0.42 mL/g the highest pore volume which is up to two times higher
compared to the imprinted polymers. In terms of pore size the
imprinted polymers exhibits a pore size around 2 nm whereas the
control polymer shows a value of 4.7 nm. Thus, a mass transfer of
solvent and saccharides can be present in both cases but due to
the size slightly favored in the control polymer. The values for the
HK-measurements displaying the micropores show that there is no
huge difference between the polymers.
MOPS, HEPES and ACES were also chosen. Moreover, to show the
effect of possible nitrogen coordination, a phosphate buffer con-
taining 1% tetramethylethylenediamine (TEMED) was used [20]. In
general, the binding of fructose in sulfonic acid containing buffers is
lower as compared with binding in phosphate buffer. Among the
sulfonic acid buffers the binding in ACES is higher compared to MOPS
and HEPES due to the possible coordination of the primary amide.
The binding of fructose to the imprinted polymers is comparable if
either a neator a TEMED containing phosphate buffer was used. Since
the imprinting factor (IF) e the difference between the MIP-BX(Fru)
vs. MIP-BX(Pin) e is in both cases about 1.9, just the neat phosphate
buffer was used for the following binding studies at pH 7.4.
3.3. Binding studies
3.2. Medium optimization
The batch binding studies of fructose were performed in 0.1 M
carbonate solution pH 11.4 with 10% methanol according to earlier
works in our group [32]. Furthermore, the binding pH was lowered
down to pH 8.7 and pH 7.4 to show the ability of the MIP-BX(Fru) to
bind at neutral pH-values as well.
Before starting batch binding experiments the medium was
especially optimized for the recognition of fructose at pH 7.4
(Fig. 3).
A medium optimization was conducted and the MIP-BX(Fru)
compared to the MIP-BX(Pin)was tested with different buffer
compositions at pH 7.4. In a first step the type of co-solvent was
compared.10% of either methanol or acetonitrile were added to 0.1 M
phosphate buffer. The experiments were performed with 2 mM
initial fructose concentration. The rebinding capacity of 2 mM fruc-
tose was 23 nmol/mg polymer and thus, inferior in the case of
acetonitrile compared to methanol with almost 30 nmol. Two
hypotheses can explain this behavior: (i) acetonitrile is able to
interact with the polymer and could be responsible for a morphology
change of the polymer; (ii) the bindingof one fructose releases in case
of a 2:1 (boronic acid:fructose) binding four methanol molecules
which favors the reaction entropically. Therefore, methanol was used
as co-solvent in the subsequent buffer optimization. Beside phos-
phate buffers consisting of substituted sulfonic acids e namely,
The batch binding studies at pH 11.4 revealed high capacities
of the synthesized MIP-BX(Fru) and MIP-BA(Fru) of 135.6 and
114.7 nmol/mg polymer, respectively (Fig. 4A). The theoretical
binding capacities of both fructose-imprinted MIP polymers are
about 200 nmol/mg polymer. Therefore, about 60% of the possible
binding pockets are accessible at an initial fructose concentration of
5 mM. The binding behavior of both MIP-BX(Fru) and MIP-BA(Fru) is
comparable and can be explained by the overwhelming concen-
tration of hydroxyl ions in solution at this alkaline environment
Table 1
Summarized pore volumes obtained by nitrogen sorption measurements (BET).
Polymer
BET surface BJH-mesopore Poresize HK-micropore Poresize
area/m2/g
volume/mL/g BJH/nm volume/mL/g HK/nm
MIP-BX(Fru) 436
MIP-BX(Pin) 360
MIP-BA(Fru) 332
MIP-BA(Pin) 360
0.21
0.24
0.29
0.28
0.42
1.9
2
1.8
1.9
4.7
0.21
0.18
0.16
0.18
0.23
1
1
1.2
0.9
0.9
Fig. 3. Media optimization for 2 mM fructose at pH 7.4 with 10% MeOH; MIP-BX(Fru)
(dark bars); MIP-BX(Pin) (light bars).
CP
482

2490
S. Schumacher et al. / Polymer 52 (2011) 2485e2491
Fig. 4. Batch binding experiments for different fructose binding MIPs at different pH-values. (AeC): Concentration dependency for fructose binding to MIP-BX(Fru) (:), MIP-
BA(Fru) (-), MIP-BX(Pin) (A) and MIP-BA(Pin) (C); (A): carbonate solution, pH 11.4, 10% MeOH; (B): phosphate buffer, pH 8.7, 10% MeOH; (C): phosphate buffer, pH 7.4, 10% MeOH.
which strongly favors fructose complexation regardless of the boron
monomer structure. Consequently, the methylhydroxylgroup in the
benzoboroxole derivative plays a negligible role in terms of binding
strength. Nevertheless, there is a pronounced difference between
the MIP-BX(Pin) and MIP-BA(Pin). In case of the MIPs with vinyl-
phenylboronic acid the difference is (with 15 nmol/mg polymer)
comparably small. In contrast, the difference in capacity of the
MIP-BX(Fru)and the MIP-BX(Pin) is four times higher (about
60 nmol/mg) showing that imprinting efficiency is higher. One
possibility could be that the methylhydroxylgroup increases the
affinity due to steric hindrances within the smaller binding pocket
obtained for the MIP-BX(Pin).
The binding to the control polymer synthesized just with TRIM
as the crosslinker shows negligible binding to fructose (at different
pH-values) at all concentrations studied. This is in line with the
literature since TRIM is a favored crosslinking agent for the recog-
nition of saccharides. The two times higher pore size of the control
polymer shows that the binding of fructose is not dominated by the
polymer morphology. Taking into account the lower binding of
fructose to the MIP-BX(Pin) it can rather be explained by the
presence of boronic acid entities and their right arrangement
showing the desired imprinting effect.
The rebinding of fructose at pH 8.7 shows (Fig. 4B) a clear
difference between MIP-BX(Fru) and MIP-BA(Fru). The capacity at
an initial concentration of 5 mM is about 70 nmol/mg polymer for
MIP-BX(Fru) whereas for the MIP-BA(Fru) a binding of only
40 nmol/mg polymer is reached. Here, most likely a mixture
between intra- and inter-molecular coordination of the boron is
present. Two findings support this assumption. Firstly, the rebound
concentration of fructose to the MIP-BX(Fru) is decreasing with
decreasing pH-value. This could be due to a lower coordination of
hydroxyl ions from the solution to the boron. Secondly, MIP-
BX(Fru) exhibits a higher binding capacity compared to MIP-
BA(Fru) which shows that an intramolecular coordination of the
hydroxyl group to the boron occurs.
Furthermore, in the case of the MIP-BX(Fru) and MIP-BX(Pin)
there is still a significant difference between both polymers for
fructose binding. The difference in capacity at 5 mM initial fructose
concentration is about 12 nmol/mg polymer. The binding isotherms
of fructose to the MIP-BA(Fru) and MIP-BA(Pin) at pH 8.7 are almost
similar and show no difference.
At pH 7.4 the advantage of intramolecular coordination provided
by the benzoboroxole units becomes evident (Fig. 4C). The poly-
mers bearing vinylphenylboronic acid exhibit a small binding
ability with just about 18.5 nmol/mg polymer at 5 mM initial
fructose concentration and no differentiation between the fructose
and pinacol imprint was seen. At this concentration the MIP-
BX(Fru) is still able to bind 60 nmol/mg polymer which is a factor
of 3.2 compared to the MIP-BA(Fru). The MIP-BX(Pin) binds
40 nmol/mg polymer at 5 mM fructose resulting in a difference of
about 20 nmol/mg polymer. The decreasing difference in capacity
between the MIP-BX(Fru) and MIP-BX(Pin) with decreasing
pH-values underlines the effect of the possible free hydroxyl group
in ortho-position to the boron which can lead to an additional
hydrogen bonding to the saccharide. Since this effect could also be
present during the polymerization, it is just possible in the MIP-
BX(Fru).
3.4. Cross reactivity
To verify the selectivity, the affinity between boronic acids and
different saccharides in solution has to be considered. In general,
the affinity is dependent on the pKa of the aryl boronic acid
derivative as well as of the structure of the saccharide. Moreover, in
case of aqueous environment the tautomeric form of the saccharide
is also one of the key factors regarding binding strength. The

S. Schumacher et al. / Polymer 52 (2011) 2485e2491
2491
4. Conclusion
For the first time a molecularly imprinted polymer with
5-methacrylamido-2-hydroxymethylphenylboronic acid as func-
tional monomer for the binding of fructose was prepared. In
comparison to earlier reports for fructose recognition, the binding
pH could be lowered and a favored rebinding in comparison to the
MIP-BX(Pin) analog could be observed. It was shown that the MIP-
BX can be used for saccharide recognition at pH 7.4 in aqueous
environment which was so far not reported for unprotected
saccharides such as fructose.
Appendix. Supporting information
Supplementary data related to this article can be found online at
doi:10.1016/j.polymer.2011.04.002
Fig. 5. D-fructose binding in presence of competitors at equimolar concentration.
References
binding strength of fructose is comparably high due to the high
[1] Jelinek R, Kolusheva S. Chemical Reviews 2004;104(12):5987e6016.
[2] Jin S, Cheng YF, Reid S, Li MY, Wang BH. Medicinal Research Reviews 2010;
30(2):171e257.
[3] Mazik M. Chemical Society Reviews 2009;38(4):935e56.
[4] Sharon N. Biochemical Society Transactions 2008;36:1457e60.
[5] Wu AM, Lisowska E, Duk M, Yang ZG. Glycoconjugate Journal 2009;26(8):
899e913.
[6] Zhang HL, Ma Y, Sun XL. Medicinal Research Reviews 2010;30(2):270e89.
[7] Matile S. Chemical Society Reviews 2001;30(3):158e67.
[8] Nygren PA, Skerra A. Journal of Immunological Methods 2004;290(1e2):
3e28.
[9] Giovannoli C, Baggiani C, Anfossi L, Giraudi G. Electrophoresis 2008;29(16):
3349e65.
presence of about 25% of the b-D-fructose furanose species (water,
31 ꢀC). The binding to this form is preferred due to the presence of
a syn-periplanar hydroxyl pair (C2eC3). Compared to glucose the
binding strength is much higher because the binding of the
pyranose-isomer (
deuterated water at 27 ꢀC. Taking these limitations into account, the
cross reactivity was evaluated with -fructose since the discrimi-
nation of enantiomers is described for many different molecular
imprints. Beside -fructose, -glucose, sucrose and sorbitol were
a-D-glucofuranose) is just present with 0.14% in
L
L
D
used as competitors (Supporting information; at equimolar
concentration shown in Fig. 5). Therefore, the difference of the
binding with and without competitor reflects the displacement of
fructose showing the degree of cross reactivity.
[10] Chen Y, Liu Y. Chemical Society Reviews 2010;39(2):495e505.
[11] Lindhorst T. Essentials of carbohydrate chemistry and Biochemistry. 3rd. rev.
ed. Weinheim: Wiley-VCH; 2007.
[12] Walker DB, Joshi G, Davis AP. Cellular and Molecular Life Sciences 2009;
66(19):3177e91.
As expected, the degree of competition is dependent on strength
[13] Schumacher S. Molecularly imprinted polymers: science goes market? e A
market analysis based on the patent situation. In: Lee S-W, Kunitake T, editors.
Handbook of molecular imprinting: advanced sensor Application; 2011.
[14] Allender C, Mosbach K. Biosensors & Bioelectronics 2009;25(3):539e42.
[15] Lieberzeit PA, Gazda-Miarecka S, Halikias K, Schirk C, Kauling J, Dickert FL.
Sensors and Actuators B-Chemical 2005;111:259e63.
[16] Alexander C, Andersson HS, Andersson LI, Ansell RJ, Kirsch N, Nicholls IA, et al.
Journal of Molecular Recognition 2006;19(2):106e80.
[17] Mayes AG, Andersson LI, Mosbach K. Analytical Biochemistry 1994;222(2):
483e8.
[18] Kirk C, Jensen M, Kjaer CN, Smedskjaer MM, Larsen KL, Wimmer R, Yu DH.
Biosensors & Bioelectronics 2009;25(3):623e8.
[19] Striegler S. Analytica Chimica Acta 2005;539(1e2):91e5.
[20] Hall DG. Boronic acids: preparation and applications in organic synthesis and
medicine. 1. Edition. Weinheim: Wiley-VCH; 2005.
[21] James TD, Phillips MD, Shinkai S. Boronic acids in saccharide recognition. 1 ed.
Cambridge: The Royal Society of Chemistry; 2006.
[22] Wulff G. Pure and Applied Chemistry 1982;54(11):2093e102.
[23] Yang XP, Lee MC, Sartain F, Pan XH, Lowe CR. Chemistry-a European Journal
2006;12(33):8491e7.
[24] Hall DG, Dowlut M. Abstracts of Papers of the American Chemical Society
2006;231.
[25] Berube M, Dowlut M, Hall DG. Journal of Organic Chemistry 2008;73(17):
6471e9.
of the saccharide e boronic acid interaction (Fig. 5).
sucrose influence the binding of -fructose to the MIP-BX(Fru)
slightly whereas -fructose and sorbitol have a higher impact. It
is noteworthy that the extent of competition between sorbitol and
-fructose are comparable even their binding constants to aryl
boronic acids vary by a factor of two. More than 90% of -fructose is
still bound in the presence of glucose or sucrose (more than 95%) as
competitors. In contrast, sorbitol and -fructose act as competitors
since less -fructose could bind to the polymer. 36 nmol fructose
D-glucose and
D
L
L
D
L
D
per mg polymer are bound without competitor, whereas around
24 nmol fructose are bound with equimolar addition of either
L
-fructose or sorbitol. Consequently, 2/3 of the possible binding
cavities are occupied by -fructose and just 1/3 by either -fructose
or sorbitol at equimolar concentrations. Moreover, at smaller
-fructose concentrations than the concentration of the competitor
also high amounts of -fructose are bound to the MIP-BX(Fru).
The separation factor which describes the interaction of
targeted molecule with stationary phase compared to
a competitor gives a quantitative value (equation: 2). The values for
-fructose and sorbitol are 2 and 2.5 at an equimolar concentration
of -fructose to the competitor, respectively. These values show that
- and - fructose can be distinguished in a normal batch binding
experiment, which is equal to one plate in terms of chroma-
tography. The -values for -glucose and sucrose are 10.7 and
D
L
D
D
a
a
a
[26] Schumacher, S, Katterle, M, Hettrich, C, Paulke, B-R, Pal, A and Hall, DG, et al.,
Chemical Sensors 1 (1), in press.
[27] Pal A, Berube M, Hall DG. Angewandte Chemie-International Edition 2010;
49(8):1492e5.
[28] Jay JI, Lai BE, Myszka DG, Mahalingam A, Langheinrich K, Katz DF, Kiser PF.
Molecular Pharmaceutics 2010;7(1):116e29.
[29] Whitcombe MJ, Rodriguez ME, Villar P, Vulfson EN. Journal of the American
L
D
L
D
a
D
Chemical Society 1995;117(27):7105e11.
71 underlining the small degree of competition with these
saccharides.
[30] Alexander C, Smith CR, Whitcombe MJ, Vulfson EN. Journal of the American
Chemical Society 1999;121(28):6640e51.
[31] Wulff G, Schauhoff S. Journal of Organic Chemistry 1991;56(1):395e400.
[32] Rajkumar R, Warsinke A, Mohwald H, Scheller FW, Katterle M. Talanta 2008;
76(5):1119e23.
[33] Norrild JC, Eggert H. Journal of the Chemical Society-Perkin Transactions 2
1996;(12):2583e8.
[34] Norrild JC, Eggert H. Journal of the American Chemical Society 1995;117(5):
1479e84.
[35] Somani BL, Khanade J, Sinha R. Analytical Biochemistry 1987;167(2):327e30.
2
a
¼ (nbound_fructose ꢂ nfree_saccharide)/(nfree_fructose ꢂ nbound_saccharide).