Improved imide receptors by imprinting using pyrimidine-based fluorescent reporter monomers

Article abstract of DOI:10.1021/jo0477906

(Chemical Equation Presented) Optically responsive receptors toward imides based on 6-substituted 2,4-bis(acrylamido)pyrimidines are presented. The monomers were readily prepared in good yield. Solution binding to 1-benzyluracil (BU) monitored by ¹H NMR appeared lower than a previously reported pyridine-based monomer. However, as indicated by ¹H NMR and IR spectral investigations, the association strength was demonstrated to be "masked" by dimerization of the pyrimidine-based monomer units. Thus, from dilution experiments, a dimerization constant of 731 M^-1 was estimated for the pyrimidine-based monomer 2,4-bis(acrylamido)-6- piperidinopyrimidine whereas for the pyridine-based monomer 2,6-bis(acrylamido) pyridine, no self-association was observed. This precluded an accurate determination of the binding constant for BU to the former monomer whereas for the latter a binding constant of 757 M^-1 was measured. Despite the strong self-association, the novel monomer was shown to lead to enhanced imprinting effects when compared to imprinted polymers prepared analogously, but using the pyridine-based monomer as the recognition element. This was attributed to a higher intrinsic binding affinity exhibited by the pyrimidine based host monomer vis a vis the guest and the existence in the former of more than one interaction site for the guest. The monomers exhibited fluorescence emission informative of the mode of monomer incorporation in the polymer and the presence of guest species. Thus, the fluorescence was rapidly and selectively quenched upon template addition, with the degree of quenching correlating with binding affinity and the amount of template bound to the polymer.

Full text of DOI:10.1021/jo0477906

Improved Imide Receptors by Imprinting Using Pyrimidine-Based
Fluorescent Reporter Monomers
Panagiotis Manesiotis, Andrew J. Hall,* and Bo¨rje Sellergren*
INFU, Universita¨t Dortmund, D-44221 Dortmund, Germany
borje@infu.uni-dortmund.de; andy@infu.uni-dortmund.de
Received December 17, 2004
Optically responsive receptors toward imides based on 6-substituted 2,4-bis(acrylamido)pyrimidines
are presented. The monomers were readily prepared in good yield. Solution binding to 1-benzyluracil
1
(BU) monitored by H NMR appeared lower than a previously reported pyridine-based monomer.
1
However, as indicated by H NMR and IR spectral investigations, the association strength was
demonstrated to be “masked” by dimerization of the pyrimidine-based monomer units. Thus, from
dilution experiments, a dimerization constant of 731 M-1 was estimated for the pyrimidine-based
monomer 2,4-bis(acrylamido)-6-piperidinopyrimidine whereas for the pyridine-based monomer 2,6-
bis(acrylamido)pyridine, no self-association was observed. This precluded an accurate determination
of the binding constant for BU to the former monomer whereas for the latter a binding constant of
757 M-1 was measured. Despite the strong self-association, the novel monomer was shown to lead
to enhanced imprinting effects when compared to imprinted polymers prepared analogously, but
using the pyridine-based monomer as the recognition element. This was attributed to a higher
intrinsic binding affinity exhibited by the pyrimidine based host monomer vis a` vis the guest and
the existence in the former of more than one interaction site for the guest. The monomers exhibited
fluorescence emission informative of the mode of monomer incorporation in the polymer and the
presence of guest species. Thus, the fluorescence was rapidly and selectively quenched upon template
addition, with the degree of quenching correlating with binding affinity and the amount of template
bound to the polymer.
Introduction
Thus, polymerization is performed in the presence of a
template molecule and a complementary functional mono-
Molecular imprinting of polymer matrixes has become
a widely used approach to generate robust macromolecu-
lar receptors for one or a group of target molecules.1-5
(1) Molecularly imprinted polymers. Man made mimics of antibodies
and their applications in analytical chemistry; Sellergren, B., Ed.;
Elsevier Science B.V.: Amsterdam, 2001; Vol. 23.
10.1021/jo0477906 CCC: $30.25 © 2005 American Chemical Society
Published on Web 03/10/2005
J. Org. Chem. 2005, 70, 2729-2738
2729

Manesiotis et al.
FIGURE 1. Modes of incorporation of monomer 2 in the imprinting of BU.
mer, such that binding sites complementary in both
shape and functionality to the original template are
formed.
sized monomers exhibiting enhanced solution binding
properties vis-a`-vis imides. We anticipated that the novel
functional monomers 2 and 3 would satisfy these criteria
for the following reasons: (1) The introduction of electron-
releasing groups at the 6-position of the 2,4-bis(amido)
pyrimidine skeleton will increase the basicity of the N3
ring nitrogen, thereby making it a stronger hydrogen-
bond acceptor.10 (2) The choice of lipophilic 6-position
substituents leads to an enhancement in solubility of the
monomers in comparison to 1. This will allow higher
template loads and correspondingly higher saturation
capacities of the resulting polymers to be achieved. (3)
The nonequivalence of the hydrogen bond donating amido
protons may create a better match toward imides with
nonequivalent hydrogen bond accepting carbonyl oxy-
gens.
In this paper, we report on the synthesis, solution-
binding properties, and polymerization of these mono-
mers for the imprinting of BU. We show that 2 leads to
imprinted polymers performing better than 1 for this task
and attribute this to a higher intrinsic binding constant
which is masked by strong self-association in solution
(Figure 1).
Although simple commodity-type monomers have proven
successful for the imprinting of a large variety of small
target molecules, the binding affinity and capacity are
often insufficient for the corresponding application. One
such example is the binding of the imide structural
motif.6-9 Whereas the traditional monomers lead to
polymers performing poorly, monomers presenting a
donor-acceptor-donor hydrogen bonding array, notably
2,6-bis(acrylamido)pyridine (1), have proven to yield
highly discriminating sites for the RNA/DNA bases uracil
and thymine7,8 and for cyclobarbital,9 albeit mostly in
noncompetitive solvents.
In our previous work, we used 1 as the functional
monomer in the preparation of molecularly imprinted
polymers targeted toward 1-benzyluracil (BU).8 The
solution association constant for binding between 1 and
BU was found to be of moderate strength (Ka ) 757 M-1
in CDCl3) and the performance of the subsequently
prepared MIPs was better than those prepared using
other designed monomers, based on aminopurines, which
exhibited lower association constants with BU in solu-
tion. To further enhance the monomer-template associa-
tion we have embarked on the design of readily synthe-
In addition, a second interaction mode between 2 and
BU appears at higher concentrations. Binding of BU to
imprinted polymers made from either 1 or 2 is ac-
companied by quenching of the polymer fluorescence,
making the new polymers potentially interesting for the
design of new receptor layers for chemical sensors.
(2) Molecularly Imprinted Materials - Sensors and other Devices;
Shea, K. J., Roberts, M. J., Yan, M., Eds.; Materials Research Society
Symposium Proceedings: San Fransisco, 2002; Vol. 723, M1.3.
(3) Molecular Imprinting - From Fundamentals to Applications;
Komiyama, M., Takeuchi, T., Mukawa, T., Asanuma, H., Eds.; Wiley-
VCH: Weinheim, 2002.
Experimental Section
(4) Wulff, G. Chem. Rev. 2002, 102, 1-27.
3′-Azido-3′-deoxythymidine (AZT) was from Sigma (St.
Loius, MO), and 2,2′-azobis(2,4-dimethylvaleronitrile) (ABDV)
was purchased from Wako (Japan). These reagents, to-
gether with 2,4-diamino-6-chloropyrimidine, 2,6-diaminopy-
ridine, acryloyl chloride, piperidine and sodium, uracil, and
thymine were used as received.
(5) Haupt, K.; Mosbach, K. Chem. Rev. 2000, 100, 2495-2504.
(6) Lanza, F.; Hall, A. J.; Sellergren, B.; Bereczki, A.; Horvai, G.;
S.; Cormack, P. A. G.; Sherrington, D. C. Anal. Chim. Acta 2001, 435,
91-106.
(7) Kugimiya, A.; Mukawa, T.; Takeuchi, T. Analyst 2001, 126, 772-
774.
(8) Hall, A. J.; Manesiotis, P.; Mossing, J. T.; Sellergren, B. Mater.
Res. Soc. Symp. Proc. 2002, 723, M1.3.1-5.
(9) Tanabe, K.; Takeuchi, T.; Matsui, J.; Ikebukuro, K.; Yano, K.;
Karube, I. J. Chem. Soc., Chem. Commun. 1995, 22, 2303-2304.
(10) Abraham, M. H.; Duce, P. P.; Prior, D. V.; Barrat, D. G.; Morris,
J. J.; Taylor, P. J. J. Chem. Soc., Perkin Trans. 2 1989, 1355-1375.
2730 J. Org. Chem., Vol. 70, No. 7, 2005

Improved Imide Receptors by Imprinting
Ethylene glycol dimethacrylate (EDMA) was purified by the
following procedure prior to use: the received material was
washed consecutively with 10% aqueous NaOH, water, brine,
and finally water. After drying over MgSO4, pure, dry EDMA
was obtained by distillation under reduced pressure.
(m, 3H), 9.95 (s, 1H), 10.25 (s, 1H); 13C NMR (DMSO-d6) δ
9.2, 29.1, 108.9, 109.1, 127.4, 131.3, 139.7, 149.9, 150.2, 163.4,
172.6; HRMS calcd 219.1008, found 219.0984.
Synthesis of 2,4-Bis(propylamido)pyridine (5). This
compound was prepared from 2,6-diaminopyridine and propi-
onic acid anhydride following the method of Feibush et al.16
Recrystallization from ethyl acetate gave the product as white
crystals in 40% yield: mp 149 °C; 1H NMR (DMSO-d6) δ 1.00-
1.04 (t, 6H), 2.23-2.39 (q, 4H), 7.67-7.70 (m, 3H), 9.923 (s,
2H); 13C NMR (DMSO-d6) δ 9.8, 29.7, 109.2, 140.2, 150.7, 173.2;
HRMS calcd 221.1164, found 221.1188.
2,6-Bis(acrylamido)pyridine (1),11 1-benzyluracil (BU), and
1,3-dibenzyluracil (DBU)12 were prepared as previously de-
scribed. Deuterated solvents were purchased from Deuterio
GmbH (Kastellaun, Germany). Anhydrous tetrahydrofuran
was stored over appropriate molecular sieves. All other
solvents were reagent grade or higher.
NMR spectra were recorded at 400 MHz. High-resolution
mass spectra were obtained using a JEOL JMS-SX-102A mass
spectrometer. All melting points are uncorrected.
Synthesis of 2,4-Bis(propylamido)-6-piperidinopyri-
midine (6). The same procedure as described above for 5 was
1
followed: mp 223-224 °C; H NMR (DMSO-d6) δ 0.97-1.00
Synthesis of 2,4-Bis(acrylamido)-6-piperidinopyrimi-
dine (2). 2,4-Diamino-6-piperidinopyrimidine13 (20 mmol, 3.87
g) and triethylamine (50 mmol, 5.06 g) were dissolved in THF
(50 mL), and the solution was cooled to 0 °C under N2. To the
cooled, stirred solution was added, dropwise, a solution of
acryloyl chloride (45 mmol, 4.07 g) in THF (20 mL). After
addition, stirring was continued overnight at ambient tem-
perature under N2. After this time the reaction mixture was
partitioned between chloroform (100 mL) and water (40 mL),
and the layers were separated. The organic layer was washed,
consecutively, with saturated aqueous NaHCO3 (2 × 100 mL)
and water (100 mL) before being dried over MgSO4. After
filtration and evaporation of solvent, ethanol (100 mL) was
added to the obtained solid and the mixture stirred for 2 h.
After filtration, the yellow solid was recrystallized from ethanol
to give the product as light yellow crystals in 60% yield: mp
(2 × t, 6H), 1.47-1.48 (m, 4H), 1.56-1.59 (m, 2H), 2.32-2.38
(q, 2H), 2.46-2.52 (q, 2H), 3.49-3.51 (t, 4H), 7.11 (s, 1H), 9.54
(s, 1H), 10.03 (s, 1H); 13C NMR (DMSO-d6) δ 8.9, 9.0, 23.9,
24.8, 29.2, 29.3, 44.5, 84.2, 159.3, 158.0, 163.0, 172.6, 173.6;
HRMS calcd 305.1852, found 305.1952.
1H NMR Titrations and Estimation of Association
Constants. All 1H NMR titrations were performed in CDCl3.
Association constants (KSL) for the interactions between hosts
and guests were determined by titrating an increasing amount
of functional monomer into a constant amount of guest (BU).
The concentration of BU was 1 mM, and the amounts of added
monomer (1, 2, or 3) were 0, 0.5, 1, 2, 3, 4, 5, 7.5, and 10 equiv,
respectively. The complexation-induced shift (∆δ) of the guest
imide proton was followed and titration curves were then
constructed of ∆δ versus BU concentration. The raw titration
data were fitted to a 1:1 binding isotherm by nonlinear
regression using Microcal Origin 5.0 from which the associa-
tion constants could be calculated.17 Dilution experiments were
performed with 1 and 2 over the concentration range 0.0836-
10 mM; the changes in chemical shifts of the amido protons
were followed, and the dimerization constant (KS2) was esti-
mated in a similar manner as above. In an attempt to esti-
mate the binding constant between BU and 2, the free
concentration of 2 was calculated by taking dimerization of 2
into account as
1
polymerized before melting at 213 °C; H NMR (DMSO-d6) δ
1.48-1.49 (m, 4H), 1.59-1.60 (m, 2H), 3.536 (t, 4H), 5.67-
5.76 (2 × d, 2H), 6.18-6.27 (2 × d, 2H), 6.60-6.70 (2 × dd,
2H), 7.25 (s, 1H), 9.97 (s, 1H), 10.38 (s, 1H); 13C NMR (DMSO-
d6) δ 24.0, 25.0, 44.6, 85.4, 127.2, 128.1, 131.3, 131.9, 156.4,
158.1, 163.1, 163.2, 164.4; HRMS calcd 301.1539, found
301.1501.
Synthesis of 2,4-Bis(acrylamido)-6-ethoxypyrimidine
(3). Preparation as for 2, starting from 2,4-diamino-6-ethoxy-
pyrimidine.14 Recrystallization from ethanol gave the product
as light yellow crystals in 65% yield: mp polymerized before
1
4KS2
1
1
∆δ
1
[S] ) -
+
-
Lt - St
(1)
melting at 217 °C; H NMR (DMSO-d6) δ 1.27-1.30 (t, 3H),
2
(
)
2KS2 ∆δmax
16K
x
4.29-4.34 (q, 2H), 5.73-5.80 (2 x d, 2H), 6.22-6.31 (2 x d,
2H), 6.61-6.72 (2 x dd, 2H), 7.20 (s, 1H), 10.41 (s, 1H), 10.73
(s, 1H); 13C NMR (DMSO-d6) δ 14.7, 62.7, 90.2, 129.3, 131.3,
156.8, 159.5, 165.0, 171.1; HRMS calcd 262.1066, found
262.1049.
S2
where [S] ) concentration of free 2; ∆δ ) the complexation-
induced shift of the imide proton of BU; ∆δmax ) estimated
complexation induced shift at saturation; Lt ) total concentra-
tion of BU; and St ) total concentration of 2.
Synthesis of 2-(Acrylamido)-6-(propylamido)pyridine
(4). Monoacylation of 2,6-diaminopyridine with propionic acid
anhydride was achieved following the method of Berl et al.15
Column chromatography (acetone/n-hexane 4/6) yielded 2-
amino-6-(propylamido)pyridine as a white solid in 50% yield:
mp 142-143 °C; 1H NMR (DMSO-d6) δ 0.96-1.00 (t, 3H),
2.25-2.31 (q, 2H), 5.65 (s, 2H), 6.09-6.11 (d, 1H), 7.17-7.19
(d, 1H), 7.25-7.29 (t, 1H), 9.72 (s, 1H); 13C NMR (DMSO-d6) δ
9.3, 29.0, 100.5, 102.9, 138.5, 150.2, 158.1, 172.1; HRMS calcd
165.0902, found 165.0875. This monoacylated compound was
then reacted with one equivalent of acryloyl chloride in the
presence of triethylamine. After basic, aqueous workup and
evaporation of the dried organic solution, the compound was
FT-IR Characterization. IR spectra were recorded on an
FT-IR spectrometer equipped with an ATR accessory. Samples
(1 mL) of 5 mM solutions of 1, 2, and BU in chloroform were
evaporated on the surface of an ATR crystal. Spectra (32 scans)
were recorded after disappearance of the solvent signals under
continuous purging with dry nitrogen.
Polymer Preparation. Imprinted polymers (P1, P2) were
prepared in the following manner. The template molecule BU
(0.2 mmol), the respective functional monomer (1 or 2) (0.3
mmol), and EDMA (20 mmol) were dissolved in chloroform (5.6
mL). To the solution was added the initiator ABDV (1% w/w
total monomers). The solution was transferred to a glass tube,
cooled to 0 °C, and purged with a flow of dry nitrogen for 10
min. The tubes were then flame-sealed while still under
cooling. The tubes were then placed in a thermostated water
bath (pre-set at 40 °C), thus initiating the polymerization
which was then allowed to continue at this temperature for
24 h. After this time, the tubes were broken and the polymers
lightly crushed. Removal of the template molecule from the
polymers was achieved by extraction with methanol in a
1
obtained as a white solid in 40% yield: mp 158-159 °C; H
NMR (DMSO-d6) δ 1.01-1.04 (t, 3H), 2.34-2.40 (q, 2H), 5.72-
7.75 (d, 1H), 6.23-6.28 (d, 1H), 6.58-6.64 (dd, 1H), 7.70-7.80
(11) Oikawa J. Polym. Sci.: Part A: Polym. Chem. 1993, 31, 457-
465.
(12) Kundu, N. G.; Sidkar, S.; Hertzberg, R. P.; Schitz, S. A.; Khatri,
S. G. J. Chem. Soc., Perkin Trans. 1985, 1, 1295-1300.
(13) Roth, B.; Smith, J. M.; Hulquist, J. M. E. J. Am. Chem. Soc.
1950, 72, 1914-1918.
(14) Roth, B.; Smith, J. M.; Hulquist, J. M. E. J. Am. Chem. Soc.
1951, 73, 3, 2864-2868.
(15) Berl, V.; Schmutz, M.; Krische, M. J.; Khoury, R. G.; Lehn, J.-
M. Chem. Eur. J. 2002, 8, 1227-1244.
(16) B. Feibush, A. Figueroa, R. Charles, K. D. Onan, P. Feibush,
B. L. Karger, J. Am. Chem. Soc. 1986, 108, 3310-3318.
(17) Connors, K. A. Binding constants. The measurement of molec-
ular complex stability; John Wiley & Sons: New York, 1987.
J. Org. Chem, Vol. 70, No. 7, 2005 2731

Manesiotis et al.
FIGURE 2. Synthesis of monomers 2 and 3. Conditions: (i) 2, R ) C5H10N: piperidine, reflux; 3, R ) OEt: NaOEt, EtOH, reflux;
(ii) acryloyl chloride, Et3N, 0 °C to rt.
Soxhlet apparatus for 24 h. Thereafter, the polymers were
crushed and sieved to obtain a particle size of 25-50µm.
Nonimprinted polymers (NIPs) (PN1, PN2) were prepared in
the same way as described above, but with the omission of
template molecule from the pre-polymerization solution.
Anal. Calcd: P1 C, 60.6; H, 7.1; N, 0.34. Found: C, 61.1;
H, 7.2; N, 0.45. PN1 C, 60.6; H, 7.1; N, 0.34. Found: C, 61.5;
H, 7.5; N, 0.5. P2 C, 60.6; H, 7.1; N, 0.28. Found: C, 59.7; H,
7.6; N, 0.3. PN2 C, 60.6; H, 7.1; N, 0.28. Found: C, 60.4; H,
7.4; N, 0.3.
HPLC Evaluation of the Polymers. The 25-50 µm
particle size fraction was repeatedly sedimented (80/20 metha-
nol /water) to remove fine particles and then slurry-packed
into HPLC columns (125 mm × 5 mm, i.d.) using the same
solvent mixture as pushing solvent. Subsequent analyses of
the polymers were performed using an HP1050 system equipped
with a diode array-UV detector and a work-station. The mobile
phase was 100% acetonitrile, the flow rate was 1 mL/min,
analyte injection volume was 10 µL, and analyte concentra-
tions were 1 mM. Analyte detection was performed at 260 nm.
Batch Rebinding. Imprinted or nonimprinted polymer (10
mg of 25-50 µm particles) was weighed into 2 mL HPLC vials.
Solutions of BU in acetonitrile (1 mL) made up to the following
concentrations were then added: 0, 0.01, 0.05, 0.1, 0.2, 0.5,
0.75, 1 mM. After 24 h incubation, the supernatants (200 µL)
were transferred into the wells of a 96-well quartz plate
(HELMA) and the free concentration of BU calculated from
the UV absorbance (λ ) 260 nm) measured using a SAFIRE
96-well plate reader (TECAN). A calibration curve was con-
structed by measuring the absorbance of standard solutions
FIGURE 3. (A) 1H NMR complexation induced shifts (∆δ) of
with the same concentrations as the rebinding solutions and
the amount of bound BU obtained by subtraction.
the BU imide proton versus the total concentration of 1 (filled
triangles), 2 (filled squares), or 3 (open circles) in CDCl3. (B)
Binding isotherms for 1 (filled squares) and 2 (open circles)
corresponding to the data in (A) assuming for 1 a 1:1 binding
Fluorescence Measurements. Imprinted or nonimprinted
polymer (10 mg of 25-36µm particles) was weighed into the
wells of a 96-well quartz plate. Solutions of BU in acetonitrile
(200 µL) made up to the following concentrations were then
added: 0, 0.01, 0.05, 0.1, 0.2, 0.5, 0.75, 1 mM. The fluorescence
emission (λex ) 270 nm, λem ) 340 nm) of the sedimented
particles was monitored in the bottom reading mode until no
further change in intensity was observed (ca. 30 min).
model and for 2 a 1:1 model with dimerization of 2 (KS2
)
731M-1, ∆δmax ) 4.5 ppm).
ometry of the association of both 2 and 3 with BU.
Disregarding self-association, nonlinear curve fitting of
the isotherm17 gave apparent association constants for
both 2 (Ka ) 596 ( 85 M-1) and 3 (Ka ) 561 ( 37 M-1
)
Results and Discussion
that were lower than that obtained for 1 (Ka ) 757 ( 28
M-1). We noted, however, that the imide proton of BU
shifted downfield more than twice when interacting with
1 than with 2 or 3. Although this could in principle be
explained by a different local magnetic environment, the
magnitude of the difference suggested other causes. It
occurred to us that strong self-association of 2 and 3
would lead to a similar behavior, resulting in a “false”
saturation curve. Indeed, compounds such as 2 are known
to self-associate via the structure (S2) shown in Figure
1, where the amide at position 2 adopts an energetically
unfavorable cis-conformation due to repulsion of the
Functional Monomer Synthesis, Solution Con-
formation, and Binding Studies. The two-step syn-
thesis of the new functional monomers reported herein
is illustrated in Figure 2.
Nucleophilic substitution of the 6-chloro group of 2,4-
diamino-6-chloropyrimidine is achieved by reaction with
piperidine or sodium ethoxide, leading to the intermedi-
ate 2,4-diamino-6-substituted pyrimidines. Reaction of
these with 2 equiv of acryloyl chloride in the presence of
triethylamine yields the novel monomers 2 and 3 as
crystalline solids in good yield.
1H NMR titrations were performed with the monomers
2 and 3 using the same procedure previously used for 1,
with BU serving as the guest compound. The complex-
ation induced shift of the imide proton of BU was
followed in both cases and is shown in Figure 3. Based
on the previous, extensive documentation of analogous
host-guest systems18-21 we have assumed a 1:1 stoichi-
(18) Sherrington, D. C.; Taskinen, K. A. Chem. Soc. Rev. 2001, 30,
83-93.
(19) Yu, L.; Schneider, H.-J. Eur. J. Org. Chem. 1999, 1619-1625.
(20) Beijer, F. H.; Kooijman, H.; Speck, A. L.; Sijbesma, R. P.; Meijer,
E. W. Angew. Chem., Int. Ed. 1998, 37, 75-78.
(21) Beijer, F. H.; Sijbesma, R. P.; J. A. Vekemans, J. M.; Meijer, E.
W. J. Org. Chem. 1996, 61, 6371-6380.
2732 J. Org. Chem., Vol. 70, No. 7, 2005

Improved Imide Receptors by Imprinting
FIGURE 4. 1H NMR spectra of an equimolar mixture of 2 and BU (A) and 2 alone (B), present as 5 mM solutions in CDCl3. The
shifts are given in Table 2, and the bands correspond to the protons according to the inserted structures.
obtained.17 This value is considerably higher than that
previously reported for hexanoic acid (2-hexanoylami-
nopyrimidine-4-yl)amide (KS2 ) 170 M-1) by Mejier et
al.,20 but expected since the latter lack the above hydro-
gen bond stabilization of the amide cis-conformation and
also the electron-releasing group at the 6-position. Fur-
ther support for the presence of strong self-association
came from IR investigations. In Figure 6 are shown the
FT-IR spectra of films of the pure monomers obtained
by evaporation of chloroform solutions.
The following peak assignments were made. 1: 3476,
3402 cm-1 CONH free, N-H stretch; 3271 cm-1 CONH
associated, N-H stretch. 2: 3316 cm-1 trans CONH
associated, N-H stretch, 3218 cm-1 cis CONH associated,
N-H stretch, 2954 cm-1 piperidine CH2, C-H stretch
FIGURE 5. 1H NMR dilution experiment showing the shifts
(asym), 2857 cm-1 piperidine CH2, C-H stretch (sym).
(δ) of the amide protons of 2 versus the total concentration of
BU: 3166 cm-1 imide, N-H stretch; 3042 cm-1 phenyl.
2 in CDCl3.
The high-frequency region reveals stark differences in
carbonyl oxygen by the N1 ring nitrogen.20 Evidence for
the positions of the bands corresponding to the NH
1
stretching mode of the amide groups.22 Whereas 1
exhibits bands at 3402 and 3476 cm-1, corresponding to
nonassociated NHs, 2 exhibits somewhat more intense
bands, all below 3316 cm-1, indicative of strong hydrogen
bonds. Although we were not able to unambiguously
assign the bands in the low-frequency region, we noted
that 2 showed bands at frequencies expected for cis-
amides, whereas no such bands were found in the spectra
of 1. Furthermore, the bands of 2 were generally sharper
than those of 1. This indicates the presence of a defined
structure showing little conformational ambiguity.
We thereafter attempted to derive an association
constant between BU and 2 taking the strong dimeriza-
tion into account. Assuming a ∆δmax of the imide proton
of BU equal to that obtained in the titration with 1, it is
seen in Figure 3B that the corresponding isotherm only
covers the low concentration interval showing no sign of
this conformation was obtained from the H NMR band
positions of the vinyl protons c and d of 2 (see Figure 4)
which were found as quartets at 6.6-6.7 and 7.2-7.3
ppm in CDCl3 whereas both appeared at 6.6-6.7 ppm in
DMSO-d6. The downfield shift of d is likely due to a weak
CH-N hydrogen bond between the N1 ring nitrogen and
proton d of the corresponding vinyl group. This would
stabilize a cis-conformation of the amide group and place
the vinyl group nearer to the piperidine ring. In support
of this structure, a 2D-NOESY-spectrum of 2 showed
positive NOEs between the aliphatic carbons of the
piperidine ring (i,j,k) and the corresponding vinyl protons
(d,e,f).
The self-association of 1 and 2 was investigated by a
1H NMR dilution experiment. The corresponding spectra
revealed significant upfield chemical shift changes of the
amide protons of 2 upon dilution (Figure 5), whereas no
movements were seen in a similar experiment using 1.
From the curve in Figure 5 a dimerization constant
(KS2 as defined in Figure 1) of 731 M-1 ((51) was
(22) Colthup, N. B.; Daly, L. H.; Wiberly, S. E. Introduction to IR
and Raman Spectroscopy, 3rd ed.; Academic Press: New York, 1990.
J. Org. Chem, Vol. 70, No. 7, 2005 2733

Manesiotis et al.
TABLE 1. Retention Factors (k) and Imprinting Factors
(IF) for Uracil- and Thymine-Based Solutes Retained on
Polymers P1 and P2a
P1
P2
solute
kMIP
23
3.82
1.04
0.31
3.46
kNIP
IF
25
5.5
3.1
1.1
5.3
kMIP
41
7.00
1.95
0.37
6.62
kNIP
IF
30
6.1
2.9
1.0
5.4
BU
U
FU
DBU
AZT
0.93
0.69
0.34
0.28
0.66
1.37
1.13
0.66
0.35
1.23
a k ) (tR - t0)/t0, where tR is the retention time of the analyte
and t0 is the retention time of the nonretained void marker
(acetone). IF ) kMIP/kNIP. The mobile phase was 100% ACN, the
injection volume was 10 µL, analyte concentration was 1 mM, flow
rate was 1 mL/min, and detection was performed at 260 nm.
are little different from that of the template molecule,
the imprinted polymer shows a high degree of selectivity
for its template.
The HPLC retention data, although demonstrating a
clear improvement using 2, are not sufficient for deter-
mining its origin. Since only one point on the isotherm
is tested, the higher k and IF obtained on P2 may be
due to a larger number of accessible sites, a higher
intrinsic binding affinity of these sites or a combination
of these effects. To discriminate between these possibili-
ties, complete equilibrium binding isotherms were mea-
sured (Figure 7A,B). Small samples (10 mg) of the
polymers were incubated with solutions (1 mL) of the
template (BU) in acetonitrile at different concentrations
and the amount of free and bound solute estimated from
the supernatant concentrations of BU after overnight
incubation.
When considered individually, the isotherms described
close to straight lines and none of the polymers exhibited
visible saturation within the measured concentration
interval. This shows that the nonspecific adsorption was
weak in acetonitrile although, as indicated by the slope
of these curves, a significant number of such weak sites
was present. When comparing the isotherms of the
imprinted with the nonimprinted polymers, the former
were steeper showing that the MIPs had adsorbed more
BU at a given concentration of free BU. This difference
was larger for P2 than for P1, which appears more
clearly from a plot of the differential adsorptions (Figure
7C). From this we conclude that P2 contains additional
binding sites, which appear to be stronger and to exist
in a greater abundance than those of P1. These results
were supported by fluorescence emission measurements
of the polymers.
As seen from the fluorescence spectra in Figure 8 and
the Stern-Volmer plots in Figure 9, binding of BU was
associated with a quenching of the fluorescence emission
(λem ) 340 nm) of the polymers. This occurred rapidly
after addition of the analyte (Figure 10) and contrasts
with the recently reported fluorescence enhancement
observed for cyclohexylbarbital interacting with an im-
printed polymer made using 1.23 Possible factors contrib-
uting to this conflicting result are the different UV
absorptions exhibited by the guests at the excitation
wavelength of the fluorophore and the fact that barbi-
turates are intrinsically fluorescent.
FIGURE 6. ATR FT-IR spectra of films of 1, 2, and BU and
corresponding equimolar mixtures of 1+BU and 2+BU in the
high-frequency region, obtained by evaporation of correspond-
ing chloroform solutions on the surface of an ATR crystal.
saturation. Thus, the intrinsic association constant be-
tween BU and 2 cannot be determined by this method.
Binding Properties of Imprinted Polymers. We
thereafter prepared imprinted (MIP) and nonimprinted
(NIP) polymers as shown in Figure 1 and packed them
into HPLC columns for chromatographic assessment of
their ability to retain the template molecule and other
analytes of related structure and functionality. This latter
group comprises molecules where there exist either
additional sites for interaction with the pendant func-
tionalities within the polymer matrixes (uracil (U) and
5-fluorouracil (FU)), a molecule where a hydrogen-
bonding site has been blocked (1,3-dibenzyl uracil (DBU))
and a molecule containing a different substituent at the
1-position (AZT). The results of the chromatographic
evaluation are shown in Table 1.
Opposite to the expectations we had based on the
solution-binding results, the polymers prepared using 2
(P2) exhibited stronger retention and higher imprinting
factors (IF ) kMIP/kNIP) than those prepared using 1 (P1).
As seen in Table 1, the template molecule was nearly two
times more retained on P2 than on P1. On injection of
the other, similar analytes onto the polymers, the dif-
ference between the MIP and NIP becomes even more
stark. While the retention of these analytes on the NIP
(23) Kubo, H.; Nariai, H.; Takeuchi, T. Chem. Commun. 2003, 2792-
2793.
2734 J. Org. Chem., Vol. 70, No. 7, 2005

Improved Imide Receptors by Imprinting
Mode of Monomer Incorporation. Although the
elemental analysis data indicated that the monomers had
been stoichiometrically incorporated into the polymers
the question remained whether they were different
concerning the conversion of the second vinyl group. The
fluorescence spectral properties of 1 and 2 provided
unique structural information concerning the mode of
monomer incorporation in the polymer. We found that
monomers 1 and 2 exhibited fluorescence emission at 435
nm (on excitation at 270 nm) (Figure 11A) which we
anticipated could serve as a direct measure of the
accessibility and microenvironment of the imprinted
binding sites in the cross-linked polymers.24
As seen in Figure 8, the fluorescence spectra corre-
sponding to polymers P1 and P2 prepared from mono-
mers 1 and 2, respectively, displayed emission maxima
at ca.340 nm, corresponding to a blue shift of nearly 100
nm compared to the free monomers. Furthermore, the
emission intensity was observed to be at least an order
of magnitude stronger than for the free monomers.
Weaker intensities were also observed at higher wave-
lengths with a shoulder at 435 nm. It occurred to us that
these spectral features could be informative about the
mode of monomer incorporation in the polymer. We
therefore compared the fluorescence emission spectra (λex
) 270 nm) of unreacted monomers 1 (Figure 11A) with
the model compounds corresponding to monoreacted (4)
(Figure 11A) and completely reacted (5) monomers
(Figure 11B). Whereas 1 showed a single broad and weak
emission band at 435 nm, the model compound 4 in
addition exhibited a weak band at 340 nm. However,
model compound 5 showed a single and strong emission
peak at 340 nm which coincided with the emission
spectra of the polymers. The same result was also
obtained when comparing the spectra of 2 and model
compound 6.
FIGURE 7. Adsorption isotherms in acetonitrile obtained for
BU adsorbed on (A) P1 (upper symbols) and PN1 (lower
symbols) and (B) P2 (upper symbols) and PN2 (lower sym-
bols) and the corresponding differential plots (C) of P1
(squares) and P2 (circles). The data points are the results of
two independent experiments where the spread is indicated
as error bars. The data have been corrected for bleeding of
UV-active material from the polymers.
These results suggest that the blue shift is predomi-
nantly due to a chemical change in the monomer emission
properties upon polymerization rather than micro-
environmental effects. Based on these observations we
conclude that a significant portion of both monomers (1
and 2) has been doubly incorporated in the polymer. The
large difference in emission intensity between the model
compounds precludes a quantitative evaluation of the
data.
Also the fluorescence data reveal larger imprinting
effects for P2 than for P1, as reflected in the differential
plots (Figure 9C). The correlation between the differential
Stern-Volmer curves and the corresponding binding
isotherms shows that a simple measure of the fluores-
cence quenching reports on how much template is bound
to the polymer.
Why is P2 Better Than P1? To understand the origin
of the better performance of the polymers made using 2
we performed a further characterization of the polymers
and the solution interactions between BU and 2.
Complexation between BU and Monomers Stud-
ied by FT-IR. A first inspection of the FT-IR spectra in
(24) Shea, K. J.; Sasaki, D. Y.; Stoddard, G. J. Macromolecules 1989,
21, 1722.
J. Org. Chem, Vol. 70, No. 7, 2005 2735

Manesiotis et al.
FIGURE 8. Fluorescence emission spectra (λex ) 270 nm) of
P1 in acetonitrile in the presence of, from top to bottom,
increasing amounts of BU as described for the experiment in
Figure 9.
Figure 6 reveals an interesting and opposite behavior of
1 and 2. As concluded above, the bands corresponding to
neat 1 were broader than those of 2, indicating the
presence of a more defined structure in the latter case.
In the presence of a stoichiometric amount of BU, the
bands of 1 become sharper whereas those of 2 become
broader. For 1 the sharpening of the bands is explained
by the formation of a well-defined triple hydrogen bonded
complex. Further support for this structure is the rela-
tively intense NH bands and the apparent absence of
bands corresponding to nonassociated NH. Opposite to
the complex between BU and 1, addition of BU to 2
appears to result in a less defined structure. This is
possibly caused by multiple amide bond conformations
or interaction modes vis-a`-vis BU. Studying more closely
the high-frequency region, the NH stretch vibrations are
found at lower frequencies (even extending below 2800
cm-1) in the complex between BU and 2 than those found
in neat 2 or in the BU-1 complex. These spectral features
indicate the presence of strongly hydrogen-bonded com-
plexes.22 Another important observation concerns the
relative intensities of bands tentatively assigned to cis-
and trans-amide bonds. Addition of BU leads to an
apparent weakening of the former indicating that the
trans-conformation is favored. This would be an expected
result if BU competes for the same interaction sites of 2
as those involved in dimerization.
FIGURE 9. Stern-Volmer plots (λex ) 270 nm, λem ) 340 nm)
showing the fluorescence quenching upon addition of BU to a
suspension of (A) P1 (solid diamonds) and PN1 (open dia-
monds) and (B) P2 (solid diamonds) and PN2 (open diamonds)
in acetonitrile and the corresponding differential plots (C) of
P1 (open diamonds) and P2 (solid diamonds). The data points
are the average results of independent experiments using two
different batches of polymer where the spread is indicated as
error bars.
Complexation between BU and Monomers Stud-
1
ied by H NMR. Further attempts to characterize the
solution complexes formed between BU and 1 and 2,
respectively, were carried out by a closer analysis of the
1H NMR spectra corresponding to the titration of BU
with 1 or 2. First it is of interest to identify which of the
protons of BU experience a complexation induced shift
in addition to the imide protons shown in Figure 3 and
the extent of these shifts (Table 2). Study of the spectra
corresponding to BU titrated with 1 revealed downfield
shifts of all protons (largest for the pyrimidine ring
protons) that saturated at roughly 3 times larger ∆δmax
than those obtained in the titration with 2.
BU. Nevertheless, it does not exclude additional associa-
tion modes occurring at higher concentrations of BU (L)
and 2 (S) such as in the prepolymerization mixture (Lt
) 20 mM, St ) 30 mM). The existence of other interaction
modes was suggested from a comparison of the spectra
of 2 alone with that of an equimolar mixture of BU and
2 (Figure 4). Addition of BU led to significant downfield
shifts of the piperidine i protons whereas the vinyl
protons d and f and amide protons a and b shifted upfield
(Table 3).
Band interference precluded accurate position to be
obtained at concentrations of 1 larger than 0.0075 M.
The similar saturation curves and relative ∆δmax values
indicate that the two monomers interact similarly with
The extent of the shifts of protons d and f were larger
but in the same direction as those observed in the dilution
experiment of 2. Of particular interest is the relatively
2736 J. Org. Chem., Vol. 70, No. 7, 2005

Improved Imide Receptors by Imprinting
FIGURE 11. Fluorescence emission spectra of (A) model
compound 4 (dashed line) and monomer 1 (solid line) and (B)
of model compound 5 dissolved in CHCl3 (1 mM). λex ) 270
nm.
FIGURE 10. Time dependence of the fluorescence quenching
(λex ) 270 nm, λem ) 340 nm) after addition of BU (triangles)
and DBU (circles) to (A) P1 (solid symbols) and PN1 (open
symbols) and (B) P2 (solid symbols) and PN2 (open symbols)
in acetonitrile to reach a final concentration of 0.5 mM.
large upfield shift of vinyl proton d. This supports the
hypothesis, based on the FT-IR analysis, that BU desta-
bilizes the amide cis-conformation. The concomitant
downfield shift of the piperidine protons i and the
appearance of two amide bands indicate that an ad-
ditional complex species exists at this concentration.
TABLE 2. Complexation-Induced Shifts Observed in
the Titration of BU with 1 or 2
∆δmax (ppm)
proton
δ start (ppm)
1
2
Ha′
7.99
5.66
7.13
4.90
7.34
+4.19
+0.06a
+0.15
+0.08
+0.05
+1.94
+0.04
+0.05
+0.02
+0.015
Hb′
Hc′
Hd′
He′,f′,g′
Conclusions
a Band interference precluded accurate position to be obtained
Despite the apparently weaker binding of BU dis-
played by 2 compared to 1, imprinted polymers made
using 2 exhibit increased imprinting effects reflected in
higher retention and imprinting factors. We attribute
these differences in the properties of the polymers to the
binding mode presented by 2 and also its ability to
dimerize (Figure 1). Thus, the lower apparent solution
binding constant is likely the result of pronounced
dimerization of 2 masking the inherently stronger hetero-
complex formation 2:BU. These differences are carried
through into the polymer matrix during the polymeriza-
tion step. Template removal exposes the high affinity
DAD hydrogen bonding array of 2 whereas the dimers
are locked up and are thereby poorly accessible. Possibly,
2 also exposes additional interaction sites for BU but the
exact nature of these is still unknown. Monomer 1 binds
more weakly to BU but shows, on the other hand, no
tendency to dimerize. The net effect should be the
creation of weaker imprinted sites and more pronounced
nonspecific binding.
at concentrations of 1 larger than 0.0075 M.
TABLE 3. Complexation-Induced Shifts of 2 Observed
in an Equimolar Mixture of BU and 2 (5 mM each in
CDCl3) Compared to 2 Alone and upon Dilution of 2 from
10 to 0.8 mM
proton
δ (ppm) 2
∆δ (ppm) 2+BU
∆δ (ppm) 2 dilution
Hj
Hk
Hi
Hf,g
He,h
Hc
Hd
Hl
Ha,b
d 1.578
d 1.658
s 3.589
m 5.745
m 6.438
q 6.660
q 7.228
s 7.444
s 10.93
+0.003
-0.008
+0.023
-0.016
+0.002
-0.035
-0.18a
-0.005
-0.25b
0
-0.01
0
+0.02
0
-0.12
-0.05
-0.01
-1.0
a Broadened quartet. b Signal appeared as two peaks at 10.65
and 10.72 ppm.
significant interest. In an effort to prevent dimerization
we synthesized model compounds 7 and 8 exhibiting
more bulky amide substituents which we anticipated
Obviously, molecules incorporating the affinity en-
hancing features of 2 but unable to dimerize would be of
J. Org. Chem, Vol. 70, No. 7, 2005 2737

Manesiotis et al.
would stabilize the amide trans isomer and thus prevent
dimerization.27
This was further supported by the low chromatographic
imprinting factors obtained for polymers prepared against
BU using 8 as functional monomer.
An interesting property of the system is the correlation
between the ligand binding with the extent of fluores-
cence quenching. Although fluorescence reporter groups
have been incorporated in imprinted binding sites previ-
ously,25,26 the use of simple DAD monomers as combined
binding and reporter groups has only recently been
reported.23 The enhanced affinity vis-a`-vis imides using
2, seems promising for the future design of polymers for
the separation and sensing of such guests.
Acknowledgment. Work supported in part by the
European Community’s Training and Mobility Program
under Contract No. FMRX-CT-98-0173 [MICA] and by
Heineken Technical Services (Zoeterwoude, NL).
Supporting Information Available: FTIR spectra and
chromatographic retention data. This material is available free
of charge via the Internet at http://pubs.acs.org.
JO0477906
(25) Turkewitsch, P.; Wandelt, B.; Darling, G. D.; Powell, W. S. Anal.
Chem. 1998, 70, 2025-2030.
Whereas 7 exhibited similar dimerization behavior as
2 judged from IR and NMR data, 8 alone showed IR
spectra more similar to 1. However, addition of BU to 8
did not lead to disappearance of the bands corresponding
to free NH groups (see Supporting Information) indicat-
ing that these interacted weakly compared to 1 and 2.
(26) Matsui, J.; Higashi, M.; Takeuchi, T. J. Am. Chem. Soc. 2000,
122, 5218-5219.
(27) Early attempts to synthesise the methacrylamide version of 2
failed due to its unexpectedly high tendency to polymerize. This may
in itself be an indication that also this molecule prefers the cis isomer
resulting in a less conjugated double bond.
2738 J. Org. Chem., Vol. 70, No. 7, 2005