Screening of 5-HT1A receptor antagonists using molecularly imprinted polymers

Article abstract of DOI:10.1021/ja067276c

Molecular imprinting produces network polymers with recognition sites for imprint molecules. The high binding affinity and selectivity in conjunction with the polymers' physical robustness positions molecular imprinted polymers (MIPs) as candidates for us

Full text of DOI:10.1021/ja067276c

NIH Public Access
Author Manuscript
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.
Published in final edited form as:
J Am Chem Soc. 2007 February 14; 129(6): 1680–1689. doi:10.1021/ja067276c.
Screening of 5-HT1A Receptor Antagonists using Molecularly
Imprinted Polymers
†
†
‡
*,†
Naphtali A. O’Connor , David A. Paisner , Donna Huryn , and Kenneth J. Shea
†Department of Chemistry, University of California, Irvine, 92717
‡Department of Chemistry, University of Pennsylvania, 19104-6323
Abstract
Molecular imprinting produces network polymers with recognition sites for imprint molecules.
The high binding affinity and selectivity in conjunction with the polymers’ physical robustness
positions molecular imprinted polymers (MIPs) as candidates for use as preliminary screens in
drug discovery. As such, MIPs can serve as crude mimics of native receptors. In an effort to
evaluate the relationship between MIPs and native receptors, imprinted polymers for
WAY-100635, an antagonist of the serotonin (5-HT) receptor subtype 5-HT were prepared. The
1A
resulting MIP P(WAY) was evaluated as an affinity matrix in the screening of serotonin receptor
antagonists with known affinities for the native receptor. Rough correlations in affinity between
the synthetic P(WAY) and native receptor 5-HT were found. These findings provide some
1A
support for the analogy between MIPs and native receptors and their possible use as surrogates.
Introduction
Molecular Imprinting is a general method for preparing synthetic receptors. Molecularly
imprinted polymers (MIPs) can have substrate affinities comparable to those of antibodies
1–4
(Scheme 1) and have been evaluated for a wide variety of applications including
5, 6
7
8, 9
chromatographic stationary phases , solid phase extraction materials , catalysts and as
chemical and biomimetic sensors.
10–13
The high binding affinity in conjunction with physical robustness of MIPs also positions
14–21
them as candidates for use as screens in drug discovery.
Indeed it has often been stated
that MIPs created from imprint molecules with high affinity for biological receptors may
serve as a crude mimic of the native receptor. Implicit in the use of MIP’s as a screen for
biologically active compounds are parallels in ligand binding between the MIP and the
native receptor. Is there justification for this assumption? What evidence is there that the
binding site of a MIP made from a high affinity ligand will resemble the native binding site?
Protein receptors have domains created by polypeptides. The binding domains are
complementary in shape to the substrate and possess an array of strategically positioned
functional groups that enter into numerous weak binding interactions of the substrate.
Molecular imprinting creates domains in crosslinked network polymers that are
complementary in shape to the template molecule. The imprinting process is also believed to
position a limited range of complementary functional groups in three-dimensional space that
22–24
provides for the recognition of analyte molecules.
Upon closer examination however
the similarities wane. Protein receptors (i.e. antibodies) have been refined through eons of
kjshea@uci.edu.
Supporting Information Available: Complete Ref. 28. Data for binding isotherm of P(WAY), P(BA) and P(BL). This material is
available free of charge via the Internet at http://pubs.acs.org.

O’Connor et al.
Page 2
evolutionary fine-tuning. They employ a wide assortment of complementary functionality
for recognition. Once perfected, biological receptors are easily replicated. MIPs on the other
hand have binding sites produced under kinetic control. There is no simple method to correct
or refine them. They are fabricated from organic monomers and a limited range of
complimentary functional groups so there is no correspondence at the molecular level with
native receptors. Furthermore, MIP binding sites are polyclonal.
Although the molecular composition of MIP binding sites may not mimic that of the native
protein, they may however respond similarly to variations in ligand structure and thus prove
useful for identifying candidates with similar affinities. To evaluate the correspondence
between MIPs and native receptors, we describe the synthesis and evaluation of MIPs as an
affinity assay for antagonists of the serotonin receptor subtype 5-HT . We chose
1A
WAY-100635, an antagonist of the serotonin receptor subtype 5-HT1A with nanomolar
25
affinity, as our imprint molecule. We then investigated the affinity of the WAY-100635
MIPs (P(WAY)) for WAY-100635 and its analogues. The analogues of WAY-100635 have
known affinities for the receptor, so comparisons between P(WAY) and the native receptor
could be made.
Serotonin is a neurotransmitter and with its receptors have been shown to be involved in
several neurological and psychiatric diseases (e.g. anxiety, depression, feeding disorders,
26
schizophrenia, Alzheimer’s disease and sexual disorders). The mechanism of regulation by
these receptors remains unclear. WAY-100635 was developed as the first potent, “silent”
25, 26
and selective 5HT receptor antagonist.
This compound has an IC value of 2.2 nM
1A
50
at the rat hippocampal 5-HT receptor and selectivity in binding of over 100-fold relative
1A
to binding at other 5-HT receptor subtypes and major neurotransmitter receptor, reuptake
and ion channel sites.
The 5-HT receptors are G protein-coupled receptors (GPCRs) consisting of a single
polypeptide chain that is integrated into the plasma membrane of cells via seven
transmembrane domains. The transmembrane and extracellular portion of the receptor form
the ligand recognition domain and the intracellular portion forms the coupling domain that
interacts with guanine-nucleotide binding proteins (G-proteins). The design of selective
ligands for 5-HT receptors represents a considerable challenge. Far less is known about the
structure of membrane bound proteins than soluble proteins, as X-ray crystal structures have
27
proven difficult to obtain. Not knowing the structure of the 5-HT receptor makes it
difficult to design molecules that target specific protein domains.
15–21
5-HT receptor mimics present an ideal opportunity for MIPs.
The 5-HT receptor is
1A
1A
an important medicinal target and thus synthetic mimics could prove to be a useful tool. We
chose WAY-100635 as the imprint molecule as it is one of the tightest binding with an IC
of 1.35 nM. The molecule is also highly selective with > 100-fold selectivity for the 5-
50
28
HT site relative to a range of other CNS receptors. The implication being that this will
1A
produce a MIP that most closely approximates the native receptor. To test this hypothesis we
evaluated the MIP P(WAY) affinity for a range of WAY-100635 analogues and compared
these to the native receptor (5-HT ).
1A
Zhuang et al have shown in a series of benzamido analogues of WAY-100635 that there is a
29
limited tolerance for variations at the amino pyridine position, in the native receptor. As
the hydrocarbon substituents at the para position becomes more sterically demanding, a
decrease in affinity for the native receptors is observed. We set out to investigate if similar
size constraints are found with the use of the MIP P(WAY) and if any correlation can be
drawn between 5-HT binding affinites and the MIP P(WAY).
1A
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 3
Results and Discussion
Affinity ligand Synthesis
Structural analogues of WAY-100635 (5–12) of known nanomolar affinity to 5-HT were
1A
29
prepared from piperazinyl 3.
MIP synthesis
WAY–100635 (4) was used as the imprint molecule for preparation of crosslinked ethylene
glycol dimethacrylate (EDGMA) polymers with methacrylic acid (MAA) as the functional
monomer (Scheme 3). Solutions of template (1 mol%), MAA (11 mol%), EGDMA (87 mol
%), and 2,2-dimethoxy-2-phenyl acetophenone (DMPA) (1 mol%) in chloroform were
polymerized photochemically by UV irradiation (a Hanovia medium-pressure 500 W
mercury arc lamp) for 24 h at 5 °C to yield polymers imprinted with WAY–100635,
P(WAY). Control polymers containing the same molar ratio of monomers but incorporating
a “generic” template, benzylamine, P(BA), or without template P(BL) were also prepared.
The resulting polymers were crushed and subjected to Soxhlet extraction with methanol. The
extracts were evaporated to dryness and dissolved in methanol (10 mL). WAY–100635 was
recovered in > 89% (i.e. splitting yield) as determined by reverse phase HPLC. Further
release of WAY-100635 was not observed in subsequent washings.
Following removal of the template molecule, the polymers were then ground, and sieved to
a 38–125 µm particle size for rebinding experiments and 25–38 µm particle size for
chromatography experiments. For rebinding experiments, 20 mg of dry polymer was
suspended in 1 mL stock solutions of WAY-100635 for 24 h, and their change in
concentration obtained using reverse phase HPLC. For chromatography experiments,
polymer particles were slurry packed in 10 cm × 0.46 cm i.d. stainless steel chromatography
columns for evaluation by HPLC. Approximately 1 g of polymer was required to pack one
column. Once packed, the columns were washed with the various eluents of choice until a
stable baseline was obtained before analysis.
The resulting polymers have a high degree of chemical and thermal stability. They are
mechanically robust and can endure high pressures as HPLC column packing materials
without fracturing. Their physical characteristics from N adsorption analysis are given in
2
Table 1. The polymers all have low surface areas in line with N adsorption analysis of
2
30
previously made imprinted materials with CHCl as porogen. This also demonstrates that
3
both imprinted and control materials are of similar nature. It should be noted however that
the low surface areas and pore volumes determined in the dry state do not represent the
porosities of these polymers in the presence of solvents.
Batch rebinding experiments were performed to evaluate the affinity of WAY-100635 for
the imprinted and control polymers. Polymers were suspended in stock solutions of
WAY-100635 in CHCl for 24 h and then the concentration of WAY-100635 was measured.
3
The equilibrium concentration of free WAY-100635 is labeled as F. The concentration of
WAY-100635 bound to polymer is designated B and is computed by taking the difference
between the total concentration of WAY-100635 (T) and F (B = T – F).
The binding isotherms of WAY-100635 on polymers P(WAY), P(BL) and P(BA) in CHCl
are shown in Figure 2 below. The plots show greater binding to polymer imprinted with
WAY–100635, P(WAY), compared with polymer imprinted with benzylamine P(BA).
Polymer made without any template, P(BL), exhibits similar affinity for WAY100635 as
polymer imprinted with benzylamine P(BA).
3
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 4
31
MIP binding parameters can be estimated from the binding isotherm by a Scatchard plot,
a
linear display of a binding isotherm where the X-axis represents the concentration of bound
analyte and the Y-axis is the ratio of bound to free analyte concentration. The mean
association constant (K ) and number of binding sites (N ) were calculated from a linear
a
t
regression of the Scatchard plot from the expression B/F = K N − K B, where B is the
a
t
a
concentration of bound analyte and F is the concentration of free analyte. The analysis for
imprinted polymer P(WAY) is displayed in Figure 3 below.
The plot in figure 3 exhibits a bimodal distribution of binding sites with two regions that can
be fitted by linear regression. This analysis reveals the MIPs have two sets of sites with
31
differing affinities. The highest affinity sites represented by a linear regression of the first
−1
five points in figure 3, exhibit an estimated mean K of 3186 M and number of binding
a
−1
sites N of 17 µmol g . The lower affinity sites are represented by the latter five points and
demonstrate an estimated mean K of 318 M and number of binding sites N of 63 µmol
t
−1
a
t
−1
−1
g
. Combining the high and low affinity binding sites gives a total of 80 µmol g . The
theoretical total binding sites computed from the amount of template WAY-100635 per
−1
gram of polymer is 56 µmol g , less than that determined from the Scatchard plot. This is
justified by the presence of excess methacrylic acid in the polymer formulation, which will
contribute to the formation of low selectivity binding sites.
The binding and selectivity of WAY-100635 and its analogues were also evaluated from
their performance in HPLC experiments with MIP and control stationary phases. Analysis
was performed by injecting 5µL of 1mM solutions of WAY-100635 and its analogues onto
the polymeric supports using a mobile phase of 2 % isopropanol in CHCl at 0.5 mL/min.
3
When less polar eluent was used, peaks were too broad for an accurate determination of
retention times. Measurements were done in triplicate. Figure 4 below depicts a
chromatogram of analyte WAY-100635 on imprinted polymer P(WAY) and control
polymer P(BL). Retention times were marked as the points of absorbance maxima.
Solutions of WAY-100635 and its analogues were injected onto HPLC columns containing
MIP stationary phases imprinted for WAY-100635. Corrected retention volumes (capacity
32
factors, k´) were calculated against a void volume marker (acetone). Imprinting factors
were calculated from the ratios of k´ values on imprinted and nonimprinted polymers (I = k
´
/ k´ ), where I is the imprinting factor and k´ and k´ are the capacity factors for
imp
con imp con
analytes on the imprinted polymer P(WAY) and the control polymer P(BL), respectively.
The imprinting factor can be viewed as a measure of the selectivity for the analytes for the
imprinted polymer over the control polymer and thus is a measure of MIP affinity and
selectivity.
The template molecule WAY-100635 proved the best retained on the MIP support, all other
derivatives have lower affinities for the MIP and neither WAY-100635 or its analogues
showed any affinity for the nonimprinted P(BL) support (Table 2). Variations of the amide
substituent at the N-(pyridine-2-yl) position (N2) were observed to elicit changes in the
capacity factors. As the benzamido analogues 7, 8, 9 and 10 get larger, a drop-off in the
imprinting factor of approximately 2 fold is observed. The largest substituent change is
found with the 4-heptylphenyl derivative 10. This compound elicits the lowest imprinting
factor of 4.0, while the much smaller 4-propylphenyl derivative 8 has a larger imprinting
factor (7.6). Another interesting observation is that in the absence of an amido substituent as
in analogue 2, there is no affinity for the imprinted polymer (imprinting factor of unity).
MIP chromatography columns have been shown to be successful using a variety of eluents,
33, 34
including aqueous mobile phases.
Their ability to operate under a range of conditions
allows them to be useful for the analysis of a variety of molecules. Analyte binding to MIPs
is a combination of specific and non-specific interactions that can be attenuated by
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 5
33, 35, 36
solvent.
In an aqueous based eluent, non-specific hydrophobic interactions are
increased and electrostatic/hydrogen bonding interactions within the MIP binding site are
diminished. To study the effect of solvent on binding, a chromatographic analysis was
performed in an aqueous based mobile phase. After surveying a variety of conditions, a 7/3:
ACN/10mM KH PO +K HPO (aq., pH 5.8) mobile phase at a 1 mL/min flow rate was
2
4
2
4
found to provide the best results (Table 3). P(WAY) was shown to have greater affinity for
WAY-100635 and its analogues over non-imprinted polymer P(BL) with WAY-100635
proving the most retained on the imprinted polymer with an imprinting factor of 2.9.
Interestingly, higher capacity factors were observed on the imprinted and non-imprinted
polymers in the aqueous mobile phase compared to the organic mobile phase. This effect is
attributed to the polymer’s behavior as a reverse phase support where retention is based on,
among other things, dispersive interactions between polymer and analyte.
The imprinting effect was observed to decrease with an increase in volume of the N-
(pyridine-2-yl) amide substituents. This may be visualized in a plot of the van der Waal’s
volume of the N-(pyridine-2-yl) amide substituents against the imprinting factor (Figure 5a).
van der Waal’s volumes were calculated using Connolly’s molecular surface method to
37
provide the solvent excluded volume which approximates the van der Waal’s volume. The
imprinting factor was found to decrease with an increase in the volume of the substituent.
The 4-heptylphenyl group is the largest substituent and produces the lowest imprinting
factor (1.9). This result demonstrates the imprinted polymer’s ability to discriminate
between WAY-100635 analogues is based at least in part on differences in the steric bulk of
the N-(pyridine-2-yl) amide substituents and in part on non-specific dispersive forces.
In the reverse phase (aqueous) chromatography mode, capacity factors were observed to
increase with increasing lipophilicity. Figure 5b reveals a continuous increase of capacity
factor with an increase in the calculated octanol-water partition coefficient (logP) of the N-
(pyridine-2-yl) amide substituent. The largest and most lipophilic 4-heptylphenyl derivative
10 showed the highest capacity factor on both imprinted and non-imprinted polymer. The
logP is a measure of the hydrophobicity and hydrophilicity of a substance. The increased
affinity for both imprinted and non-imprinted polymers with increasing lipophilicity may be
attributed to non-specific dispersive interactions, which come into play in the aqueous
mobile system. It is important to note however that the magnitude of this trend for the non-
imprinted control polymer P(BL) is significantly less.
The preceding chromatographic results are indicative of the contribution of amide
substituents at the N-(pyridine-2-yl) position for retention on the imprinted polymer
P(WAY). Importantly, changes at this position also elicit large variations in the binding
affinities for the 5-HT receptor. The ability of the imprinted polymer to discriminate
1A
between these subtle differences adds credibility to employing imprinted polymers in the
screening of libraries of ligands. A point of interest is to discern what other similarities the
imprinted polymer P(WAY) bears with the native receptor 5-HT
.
1A
The binding interactions that take place between WAY-100635 and the imprinted polymer
and 5-HT receptor are due to a combination of intermolecular forces. These include ionic
1A
interactions, hydrogen bonds and/or van der Waals forces. Binding affinity is related to the
Gibbs free energy of binding ΔG which is composed of an enthalpic and entropic
contribution (ΔG = −RT ln Ki = ΔH − TΔS). The enthalpic contribution is comprised of van
der Waals, dipole-dipole, hydrogen bonding and electrostatic interactions between the ligand
and protein. The binding of a ligand to an imprinted polymer relies on similar interactions,
although they do not share the same specific interactions. Nevertheless, since recognition is
a composite of attractive interactions based upon the host and guest, there is no fundamental
reason why parallels may not be observed.
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 6
Affinities are based on structural features of both ligand and receptor. As a consequence,
variations in binding affinities can be related to deviations in ligand structure. A comparison
between the forces important for ligand binding to the 5-HT receptor and imprinted
1A
polymer can be used as a measure of how well a mimic of the receptor the imprinted
polymer provides. A plot of 5-HT receptor binding affinities against the imprinting factors
1A
for WAY-100635 and its analogues reveals such a relationship (Figure 6 and Table 4). The
figures show a decrease in the imprinting factor in both organic and aqueous based eluents
with an increase in the 5-HT binding affinity. Sterically demanding analogues such as 4-
1A
pentylphenyl 9 and 4-heptylphenyl 10 show low affinities for both the 5-HT receptor and
1A
imprinted polymer. WAY-100635 and other similar sized derivatives show similar 5-HT
1A
affinities while having moderate to high affinity for the imprinted polymer. From this we
gather that a size constraint exists in both 5-HT receptor and imprinted polymer P(WAY).
1A
The imprinted polymer is able to discriminate between analogues having a 20-fold
difference in affinity. Regardless of any electrostatic similarities with WAY-100635,
analogues that are too large for the MIP and 5-HT binding sites are poorly retained. From
1A
the available data, changes to the amide substituents at the N-(pyridine-2-yl) position of
WAY-100635 elicits similar binding response from the imprinted polymer P(WAY) and the
5-HT receptor. This leads us to believe that P(WAY) mimics the binding site shape and
1A
volume of the 5-HT receptor.
1A
A variety of aryl piperazines are known to bind to the 5-HT receptor as antagonists and
1A
agonists. They share the similar motif of an aryl piperazine and rigid aryl segment separated
by a flexible linker. We were interested in examining the imprinted polymer P(WAY) for
the ability to distinguish aryl piperazines which were not simple analogues of the template
WAY-100635 but were also potent binders to the 5-HT receptor. Recognition by the
1A
imprinted polymer for such analogues would demonstrate homology between 5-HT
1A
antagonists and its ability to recognize less obvious similarities between WAY-100635 and
its analogues, parallel to the 5-HT receptor. For this study we obtained the imprinting
1A
factors (chromatographic analysis on P(WAY) and P(BL) with a 2% isopropanol in CHCl
3
mobile phase) of Spiperone, Buspirone and NAN-190 (Figure 7). Spiperone and NAN-190
antagonists and Buspirone is an agonist of the 5-HT receptor.
1A
NAN-190 is the most similar to WAY-100635, sharing a common methoxy phenyl
piperazine motif and antagonist behavior. NAN-190 is also the most potent, displaying high
38
5-HT affinity (K = 0.6 nm). A chromatographic analysis of these 5-HT aryl
1A
i
1A
piperazines on imprinted and control polymer supports show that the imprinted polymer is
able to recognize aryl piperazines that are not simple analogues of the template (Table 5),
with NAN-190 having the highest affinity with an imprinting factor of 3.4. Buspirone
proved the least retained by the imprinted polymer, with an imprinting factor of 1.7. Their
affinities for the imprinted polymer however are not as high as those of the WAY-100635
analogues.
The data indicates a relationship between the 5-HT dissociation constants of the
1A
piperazine analogues and their imprinting factors. The most strongly retained NAN-190 is
also the strongest 5-HT antagonist, with an inhibition constant of 0.6 nM. Spiperone has a
1A
larger imprinting factor (2.8) than buspirone (1.7). This does not correlate with the inhibition
constants, as Buspirone shows greater affinity for the 5-HT receptor at 14 nM than
1A
Spiperone, which has an inhibition constant of 101 nM. Buspirone however is an agonist.
Agonist and antagonist character is dependent on differences in the binding interactions with
the receptor. The MIP, which is imprinted for an antagonist may have enhanced interactions
with antagonists over agonists. This offers additional support for the position that imprinted
polymers function as surrogates for native receptors.
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 7
Molecular Modeling Studies
Molecular modeling has been successfully employed to study the interaction between
40–42
functional monomers and imprinting molecules.
Although MIPs are comprised a
43
collection of heterogeneous binding sites differing in affinity, molecular modeling can be
used to study possible arrangements of complementary functionality that contribute
importantly to the high fidelity sites. Functional monomers can be ligated to template/
analogue molecules at possible contact points and their geometry optimized. This can
provide insight into the possible binding interactions within the MIP cavity and aid the
analysis of binding data. This methodology was applied to the template WAY-100635 in an
effort to identify possible MIP binding site interactions. This model is limited to the
hydrogen bonding interaction between analyte and the functional monomer (MAA) and does
not take into consideration the effect the volume of the MIP binding site has on affinity. The
model was used to analyze the binding of 5-HT antagonists NAN-190, Buspirone and
1A
Spiperone which are similar in volume to WAY-100635. They all share the common feature
of a piperazine nitrogen which is believed to be integral for binding. However, other
structural similarities are not as easily distinguished.
An ionic interaction between an aspartate residue of the 5-HT receptor and the protonated
1A
piperazinyl nitrogen (N3) of antagonists has been considered crucial and via a long range
coulombic intermolecular interaction has been proposed as the first step in the binding site
44
recognition process. This interaction is presumed to be followed by the formation of a
charge reinforced hydrogen bond. Analogously, interaction between the piperazinyl nitrogen
(N3) and methacrylic acid in the prepolymer is presumed to result in a hydrogen bond
between the imprinted polymer P(WAY) and WAY-100635.
The electrostatic potential for WAY-100635 was calculated in an effort to establish the
strongest points of contact with the MIP P(WAY) binding site carboxylates. This was done
45, 46
using a semi-empirical PM3 method from the Spartan 2002 software package.
The
molecular electrostatic potential is the potential energy of a proton at a particular location
near a molecule. It represents the attraction between the molecule and a proton. The
electrostatic potential is useful in rationalizing the interactions between molecules and
molecular recognition processes because electrostatic forces are primarily responsible for
long-range interactions between molecules. A map of the electrostatic potential onto the
molecular surface of WAY-100635 is shown in Figure 8. Regions attracted to a proton are
signified by blue, and areas that repel a proton are red. Attractive forces between
WAY-100635 and methacrylic acid in the MIP will occur at the regions colored in blue.
These comprise regions around the nitrogen and oxygen heteroatoms and the methoxy
phenyl ring.
Models were built ligating three methacrylic acid molecules to WAY-10635 using the
Spartan 2002 software package. Methacrylic acid was first hydrogen bonded to the
piperazine nitrogen (N3) and the N-(pyridine-2-yl) amide oxygen (O1). The third
methacrylic acid was hydrogen bonded to either the pyridinyl nitrogen (N1) as in Figure 9a
or the phenyl methoxy oxygen (O2) as illustrated in Figure 9b. Their geometries were then
optimized with the PM3 semi-empirical method and are illustrated in Figure 13 below. From
this we gathered that the complex where the third methacrylic acid was hydrogen bonded to
the methoxyl phenyl oxygen (O2) was more favored, with a calculated heat of formation of
−296.18 kcal/mol compared to the −292.47 kcal/mol that was obtained for the pyridinyl
nitrogen (N1) hydrogen bonded structure. It is proposed that an idealized MIP binding site
could include three methacrylic acid hydrogen bonding contacts to the piperazinyl nitrogen
(N3), N-(pyridine-2-yl) amide oxygen (O1) and the methoxy phenyl oxygen (O2).
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 8
The methacrylic acid molecules in Figure 9b were fixed into position to mimic the
WAY-100635 MIP binding site. The WAY-100635 molecule was deleted and 5-HT
1A
antagonists NAN-190, Spiperone and agonist Buspirone inserted into the model. When
inserted, all models had the piperazinyl nitrogen hydrogen bonded to a methacrylic acid
molecule. All molecules were oriented in the binding cavity in a similar manner with the
aryl piperazine in the same region as WAY-100635 in Figure 9b. With the methacrylic
molecules in fixed positions, the geometries were then optimized using the PM3 semi-
empirical method, the result of which can be found in Figure 10 below.
All 5-HT antagonists are able to establish multiple contacts with methacrylic acid
1A
functional monomers in the proposed WAY-100635 MIP cavity. Each has electronegative
atoms in close proximity to acidic hydrogen of methacrylic acid. Most notable are NAN-190
and Spiperone which show contact with all three methacrylic acids. In addition to the
piperazine nitrogen, NAN-190 exhibits hydrogen bonding with the phthalimide oxygen and
methoxy phenyl oxygen (Figure 10a). Spiperone is observed having hydrogen bonds with
the methacrylic acids of the WAY-100635 MIP cavity at its piperazine nitrogen, aryl ketone
oxygen and cyclic amide oxygen. Interestingly, in the optimized structure the agonist
Buspirone could have only two hydrogen bonds with the methacrylic acids of the proposed
WAY-100635 MIP cavity, one at the piperazine nitrogen and the other at imide oxygen.
It is interesting to note that both NAN-190 and Spiperone, which were observed to be
capable of three hydrogen bonds in the proposed WAY-100635 MIP cavity, show greater
retention on the MIP chromatography column P(WAY) than Buspirone (Table 5). It is
possible that the increased imprinting factors on MIP P(WAY) for NAN-190 and Spiperone
compared to Buspirone may be due to the additional ability to enjoy hydrogen bonding at
the aryl portion of the molecule.
Similarities between the proposed MIP binding model and 5-HT receptor binding site are
1a
apparent. In recent homology modeling of the 5-HT receptor in addition to ionic
1a
interaction between the carboxylic acid side chain of Asp3.32 and the nitrogen of the
piperazine moiety, other essential interactions were hydrogen bonding between hydrogen
47
bond acceptors of the ligand with Ser5.42 and Asn7.39. Inspection of the 5-HT receptor
1a
binding to NAN-190 (Figure 11) shows similarities with NAN-190 bound to our proposed
MIP binding model (Figure 10a). Both binding models show interactions with the imide
carbonyl, piperazine nitrogen and phenyl methoxy oxygen atoms of NAN-190.
Summary
In summary we have developed a synthetic MIP receptor for the recognition of
WAY-100635 and its analogues. The binding sites are heterogeneous in nature, as expected
for a noncovalent imprinted polymer. A Scatchard analysis shows selective binding sites
−1
with a mean association constant of 3186 M and non-selective binding sites with a mean
−1
association constant of 318 M in chloroform. In analogy with the native receptor there is a
steric constraint on the analogues within the imprinting cavity, and as N-(pyridine-2-yl)
amide substituents became larger a drop-off in affinity for the imprinted polymer was
observed. We have also shown that imprinted polymer shows affinity for aryl piperazines
which are not direct analogues of the imprinting template. A relationship was observed to
exist between the 5-HT receptor and MIP affinities, with both showing limited tolerance
1A
for variations at the amino pyridine position in a series of benzamido analogues of
WAY-100635. These results support the viability of imprinting polymers as surrogates for
native receptors in drug discovery.
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 9
Experimental Section
General Procedures
29, 48
Compounds 1–12 were prepared in a similar manner documented procedures.
High
pressure liquid chromatography (HPLC) experiments were performed on a HP 1100 dual
solvent delivery system. Molecular modeling was done using the Spartan 2002 software
package.
General Procedure for Molecularly Imprinted Polymer Synthesis
A solution of ethylene glycol dimethacrylate (6.78 g, 34.21 mmol), methacrylic acid (0.41 g,
4.77 mmol), and dimethoxy-2-phenyl acetophenone (0.103 g, 0.4 mmol) in chloroform (9.3
mL) was deoxygenated using nitrogen gas for 5 min. Additional deoxygenated solvent was
added to bring the volume of the solution to its initial level. The appropriate template (0.4
mmol) was added and solution further deoxygenated. The polymerizations were initiated
photochemically by a mercury arc UV light source at 5 ± 2 °C and allowed to proceed for 24
h under constant irradiation. After crushing, the polymers were extracted with methanol for
24 h, then dried under vacuum. The polymers were further crushed using a mortar and pestle
and sieved to isolate 38 – 125 µm particles for use in batch rebinding studies.
General Procedure for Batch Rebinding Experiments with Molecularly Imprinted Polymers
A stock solution of 5 mM WAY–100635 in solvent was prepared. This solution was diluted
with solvent 0.1, 0.2, 0.4, 0.6, 0.8, 1, 1.4, 1.8, and 2 mM solutions. One mL of each WAY–
100635 solution was then added to 2.0 mL screwcap HPLC vials containing either no
polymer or 20 mg of polymer P(WAY), P(BA) or P(BL) (38–125 µm size range). The vials
were then shaken for 24 hours and the unbound WAY-100635 concentrations quantified
using HPLC analysis. This concentration is labeled as F and the concentration of bound
WAY-100635 to polymer is designated B and is computed by taking the difference between
the total concentration of WAY-100635 (T) and F (i.e. B = T – F).
General Procedure for Chromatographic Experiments
Polymer particles (25–38 µm) were slurry packed into stainless steel columns (length: 100
mm, internal diameter: 4.6mm). Once packed, the columns were washed using an eluent of
70/30 ACN/H O until a stable baseline was observed. Chromatographic experiments were
2
then performed using these columns with the Agilent 1100 system.
Supplementary Material
Refer to Web version on PubMed Central for supplementary material.
Acknowledgments
We are grateful to Wyeth Pharmaceuticals, the NIH and University of California, Irvine for financial support.
References
1. Yan, M.; Ramstrom, O., editors. Molecularly Imprinted Materials: Science and Technology. New
York: Marcel Dekker; 2005. p. 734
2. Komiyama, M.; Takeuchi, T.; Mukawa, T.; Asanuma, H. Molecular Imprinting: From Fundamentals
to Applications. Weinheim: Wiley-VCH; 2003. p. 147
3. Shea KJ. Trends Polym. Sci. 1994; 2:166–173.
4. Alexander C, Andersson HS, Andersson LI, Ansell RJ, Kirsch N, Nicholls IA, O'Mahony J,
Whitcombe MJ. J. Mol. Recognit. 2006; 19:106–180. [PubMed: 16395662]
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 10
5. Kempe M, Mosbach K. J. Chromatogr. A. 1995; 694:3–13.
6. Mayes AG, Mosbach K. Trends Anal. Chem. 1997; 16:321–332.
7. Andersson LI. Bioseparation. 2002; 10:353–364. [PubMed: 12549879]
8. Ramstrom O, Mosbach K. Curr. Opin. Chem. Biol. 1999; 3:759–764. [PubMed: 10600723]
9. Wulff G. Chem. Rev. 2002; 102:1–27. [PubMed: 11782127]
10. Ansell RJ, Kriz D, Mosbach K. Curr. Opin. Biotechnol. 1996; 7:89–94. [PubMed: 8791315]
11. Haupt K, Mosbach K. Trends Biotechnol. 1998; 16:468–475. [PubMed: 9830155]
12. Haupt K, Mosbach K. Chem. Rev. 2000; 100:2495–2504. [PubMed: 11749293]
13. Kriz D, Ramstrom O, Mosbach K. Anal. Chem. 1997; 69:345A–349A.
14. Deng Q, Lun Z, Shao H, Yan C, Gao R. Anal. Bioanal. Chem. 2005; 382:51–58. [PubMed:
15900451]
15. Ye L, Haupt K. Anal. Bioanal. Chem. 2004; 378:1887–1897. [PubMed: 15064898]
16. Bouman MAE, Allender CJ, Brain KR, Heard CM. Methodological Surveys in Bioanalysis of
Drugs. 1998; 25:37–43.
17. Ramstrom O, Ye L, Mosbach K. Anal. Commun. 1998; 35:9–11.
18. Khasawneh MA, Vallano PT, Remcho VT. J. Chromatogr. A. 2001; 922:87–97. [PubMed:
11486894]
19. Vallano PT, Remcho VT. J. Chromatogr. A. 2000; 888:23–34. [PubMed: 10949469]
20. Xu X, Zhu L, Chen L. J. Chromatogr. B. 2004; 804:61–69.
21. Ye L, Yu Y, Mosbach K. Analyst. 2001; 126:760–765. [PubMed: 11445933]
22. Becker JJ, Gagne MR. Acc. Chem. Res. 2004; 37:798–804. [PubMed: 15491126]
23. Conrad, PGI.; Shea, KJ. Molecular Imprinted Materials: Science and Technology. Mingdi, Y.;
Ramstrom, O., editors. New York: Marcel Dekker; 2005. p. 123-180.
24. Shea KJ, Sasaki DY. J. Am. Chem. Soc. 1989; 111:3442–3444.
25. Forster EA, Cliffe IA, Bill DJ, Dover GM, Jones D, Y R, Fletcher A. Eur. J. Pharmacol. 1995;
281:81–88. [PubMed: 8566121]
26. Cliffe IA, Fletcher A. Current Drugs: Serotonin. 1993; 1:99–124.
27. Eisenberg D. Annu. Rev. 1984; 53:593–623.
28. Fletcher A, et al. Behav. Brain Res. 1996; 73:337–353. [PubMed: 8788530]
29. Zhuang Z, Kung M, Chumpradit S, Mu M, Kung HF. J. Med. Chem. 1994; 37:4572–4575.
[PubMed: 7799409]
30. Sellergren B, Shea KJ. J. Chromatogr. 1993; 635:31–49.
31. Shea KJ, Spivak D, Sellergren B. J. Am. Chem. Soc. 1993; 115:3368–3369.
32. The capacity factor (k') is defined as k' = [(V(t) − V(0))/V(0)], where V(t) is the retention volume
and V(0) is the dead volume or the retention volume of a nonbinding substrate.
33. Spivak D, Gilmore MA, Shea KJ. J. Am. Chem. Soc. 1997; 119:4388–4393.
34. Baggiani, C. Molecularly Imprinted Materials. Mingdi, Y.; Ramstrom, O., editors. New York:
Marcel Dekker; 2005. p. 517-552.
35. Kim H, Guiochon G. Anal. Chem. 2005; 77:2496–2504. [PubMed: 15828786]
36. Sellergren B. J. Chromatogr. A. 2001; 906:227–252. [PubMed: 11215890]
37. Connolly ML. J. Am. Chem. Soc. 1985; 107:1118–1124.
38. Raghupathi RK, Rydelek-Fitzgerald L, Teitler M, Glennon RA. J. Med. Chem. 1991; 34:2633–
2638. [PubMed: 1652026]
39. Titeler M, Lyon RA, Davis KH, Glennon RA. Biochem. Pharmacol. 1987; 36:3265–3271.
[PubMed: 3663239]
40. Wu L, Li Y. J. Mol. Recognit. 2004; 17:567–574. [PubMed: 15386620]
41. Wu L, Sun B, Li Y, Chang W. Analyst. 2003; 128:944–949.
42. Takeuchi T, Dobashi A, Kimura K. Anal. Chem. 2000; 72:2418–2422. [PubMed: 10857615]
43. Umpleby RJ II, Bode M, Shimizu KD. Analyst. 2000; 125:1261–1265.
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 11
44. Strader CD, Sigal IS, Candelrone MR, Rands E, Hill WS, Dixon RA. J. Biol. Chem. 1988;
263:10267–10271. [PubMed: 2899076]
45. Stewart JJP. J. Comput. Chem. 1989; 10:209–220.
46. Stewart JJP. J. Comput. Chem. 1989; 10:221–264.
47. Nowak M, Kolaczkowski M, Pawlowski M, Bojarski AJ. J. Med. Chem. 2006; 49:205–214.
[PubMed: 16392805]
48. Zhuang Z, Kung M, Kung HF. J. Med. Chem. 1994; 37:1406–1407. [PubMed: 8182697]
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 12
Figure 1.
5-HT selective antagonist WAY-100635
1A
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 13
Figure 2.
Binding Isotherm of P(WAY), P(BA) and P(BL).
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 14
Figure 3.
Scatchard plot for WAY-100635 on P(WAY).
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 15
Figure 4.
Chromatogram of WAY-100635 on MIP P(WAY) and control P(BL) at a flow rate of 0.5
mL/min in 2% isopropanol in CHCl .
3
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 16
Figure 5.
a) Effect of van der Waal’s volume of the N-(pyridine-2-yl) amide substituent on the
imprinting factors from an aqueous eluent (1 mL/min. in 7/3: ACN/10mM
KH PO +K HPO (aq., pH 5.8)). b) Effect of the lipophilicity (logP) on the capacity factors
2
4
2
4
from an aqueous eluent.
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 17
Figure 6.
Relationship between 5-HT binding affinity and imprinting factors: a) 1 mL/min. in 7/3:
1A
ACN/10mM KH PO +K HPO (aq., pH 5.8). b) 0.5 mL/min. in 2% isopropanol in CHCl .
2
4
2
4
3
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 18
Figure 7.
5-HT receptor high affinity agonist and antagonists.
1A
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 19
Figure 8.
Molecular models of WAY-100635 with electrostatic potential mapping. Attraction for
positive charge is represented by blue and repulsion is represented by red.
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 20
Figure 9.
Molecular model of 3 methacrylic acid molecules bound to WAY-100635 heteroatoms: a)
N1, N3 and O1; b) N3, O1, O2
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 21
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 22
Figure 10.
Models of 5-HT1A antagonists bound to three methacrylic acids in WAY-100635 MIP.
Hydrogen bonding interactions are high-lighted. a) NAN-190; b) Spiperone; c) Buspirone.
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 23
Figure 11.
47
Calculated interactions of NAN-190 bound to the 5-HT binding site.
1a
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 24
Scheme 1.
Schematic diagram of the molecular imprinting process.
(a) Formation of prepolymerization complex between template T and functional monomers
M; (b) Copolymerization with an excess of cross-linking monomer; (c) Extraction of
template from imprinted binding sites using porogen; (d) Rebinding of the template to
imprinted binding sites.
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 25
Scheme 2.
Synthesis of WAY-100635 and analogues.
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 26
Scheme 3.
Synthesis of Molecular Imprinted Polymer for WAY-100635
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 27
Table 1
Physical characteristics of Molecular Imprinted and Control Polymers.
Characteristics
P(WAY) P(BA) P(BL)
Surface Area (m2/g)
5.1
1.1
~0
5.9
~0
Pore Volume (ml/g) ~0
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 28
Table 2
Chromatographic analysis on P(WAY) and P(BL) at a flow rate of 0.5 mL/min in 2% isopropanol in CHCl .
3
Compound
Substituent
P(WAY) (k´) P(BL) Imprinting
(k´)
Factor (Iorg)
2
N.A.
0.2
3.4
2.6
2.7
3.1
0.2
0.3
0.6
0.4
0.4
0.5
0.4
0.4
0.5
0.6
1.0
11.3
4.3
6.7
7.7
7.6
4.2
4.0
5.6
6.3
WAY-100635 Cyclohexyl
5
6
Adamantyl
4-iodophenyl
Phenyl
7
8
4-propylphenyl 3.8
4-pentylphenyl 1.7
4-heptylphenyl 1.6
9
10
11
12
4-fluorophenyl
2-furoyl
2.8
3.8
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 29
Table 3
Chromatographic analysis on P(WAY) and P(BL) supports at a flow rate of 1 mL/min. in 7/3: ACN/10mM
KH2PO4+K2HPO4 (aq., pH 5.8).
Compound
Substituent
P(WAY) (k´) P(BL) (k´) Imprinting
Factor (Iaq
)
2
N.A.
0.4
5.5
6.1
4.4
5.0
0.3
1.9
2.8
1.7
1.9
2.8
4.0
5.3
1.3
2.9
2.2
2.6
2.6
2.1
2.0
1.9
WAY-100635 Cyclohexyl
5
6
Adamantyl
4-iodophenyl
Phenyl
7
8
4-propylphenyl 5.9
4-pentylphenyl 7.8
4-heptylphenyl 9.9
9
10
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.

O’Connor et al.
Page 30
Table 4
Imprinting factors and 5-HT receptor binding affinities for WAY-100635 and analogues.
1A
Compound
Substituent
Imprinting
Imprinting
a
Ki (nM)
b
c
Factor (Iorg
)
Factor (Iaq)
2
N.A.
1.0
11.3
4.3
6.7
7.7
1.3
2.9
2.2
2.6
2.6
2.1
2.0
1.9
N.A.
N.A.
N.A.
0.84
N.A.
2.6
WAY-100635 Cyclohexyl
5
6
Adamantyl
4-iodophenyl
Phenyl
7
2.5
8
4-propylphenyl 7.6
4-pentylphenyl 4.2
4-heptylphenyl 4.0
10.3
75.0
81.8
3.3
9
10
11
12
4-fluorophenyl
2-furoyl
5.6
6.3
N.A.
a
b
125
Inhibition constants on the binding of [
29
I]-(R)-(+)-trans-8-OH-PIPAT to Rat Hippocampal Homogenates.
2% isopropanol in CHCl at a flow rate of 0.5 mL/min.
3
c
7/3: ACN/10mM KH PO +K HPO (aq., pH 5.8) at a flow rate of 1 mL/min.
2
4
2
4
J Am Chem Soc. Author manuscript; available in PMC 2011 October 14.