Analyte templating: Enhancing the enantioselectivity of chiral selectors upon incorporation into organic polymer environments

Article abstract of DOI:10.1021/ac050407s

A simple strategy for preserving and enhancing the chiral recognition capacity of polymer-embedded chiral selectors is proposed, capitalizing on a temporary blockage of the receptor binding site with tightly binding analytes during the polymerization proc

Full text of DOI:10.1021/ac050407s

Anal. Chem. 2005, 77, 5009-5018
Analyte Templating: Enhancing the
Enantioselectivity of Chiral Selectors upon
Incorporation into Organic Polymer Environments
Elena Gavioli,† Norbert M. Maier,*,† Karsten Haupt,‡ Klaus Mosbach,§ and Wolfgang Lindner†
Institute of Analytical Chemistry, University of Vienna, Wa¨hringerstrasse 38, A-1090 Vienna, Austria, Department of
Bioengineering, Compie`gne University of Technology, UMR CNRS 6022, BP 20529, 60205 Compie`gne Cedex, France, and
Swiss Federal Institute of Technology, Biotechnology, ETH Ho¨nggerberg HPT E 77, CH-8093 Zu¨rich, Switzerland
materials display a considerable number of residual, noncapped
silanol functions, giving rise to nonspecific interactions and
rendering the surfaces vulnerable to attack by network-degrading
agents.
A simple strategy for preserving and enhancing the chiral
recognition capacity of polymer-embedded chiral selectors
is proposed, capitalizing on a temporary blockage of the
receptor binding site with tightly binding analytes during
the polymerization process. We demonstrate that the
copolymerization of a quinine tert-butylcarbamate selec-
tor monomer with chiral (and achiral) 3,5-dichlorobenzoyl
amino acids allows one to control to a certain extent the
binding characteristics of the resultant polymeric chiral
stationary phases. The structural and stereochemical
requirements of the templating analytes for maximizing
the chiral recognition capacity of the polymer-embedded
selectors are probed. The chromatographic chiral recogni-
tion characteristics of the analyte-templated polymeric
chiral stationary phases are analyzed with respect to
binding capacities and affinities and compared to those
obtained with a conventional silica-based surface-grafted
reference material. Changes in substrate-specific enantio-
selectivity originating from analyte templating are also
addressed.
These silica-specific limitations may be circumvented by the
use of organic polymeric supports for CSP production.3,5 Not only
do organic polymers exhibit stability over the entire pH range,
but they also show an enhanced chemical inertness to aggressive
media. Polymers are amenable to a variety of surface chemistries,
facilitating ligand immobilization and suppression of nonspecific
binding.3,5,6 A particularly attractive feature of organic polymers
is the ease by which enantiomer separation media of different
shapes can be accessed.7 Copolymerization of appropriate selector-
type monomers with cross-linking agents and comonomers may
be used to shape chiral recognition materials in the form of self-
supporting beads,8-11 membranes,12 and porous monoliths;3,13,14
considerably broadening the scope of applications compared to
conventional silica-grafted materials.
Unfortunately, the transfer of selectors originally developed
for silica-surface grafting into polymeric matrixes is far from being
straightforward. Materials produced with polymerizable versions
of selectors often exhibit inferior enantiomer separation charac-
teristics as compared to the corresponding silica-grafted binders.15-20
The vast majority of the commercially available chiral stationary
phases for direct enantiomer separation comprises chiral selectors
grafted or coated onto particulate silica-type supports.1 Silica-based
materials offer many desirable features for chromatographic
applications, such as mechanical stability, resistance to swelling,
and high efficiency. In addition, they can be precisely tailored with
respect to particle and pore characteristics, and a well-established
repertoire of (silane-based) surface chemistry is available for
selector immobilization.2 However, silica-based chiral stationary
phases suffer from several inherent drawbacks.3,4 These include
pronounced instability in acidic and basic media, restricting their
working range to pH 2-8. Generally, even densely ligand-grafted
(5) Svec, F. J. Sep. Sci. 2004, 27, 747-766.
(6) Jungbauer, A.; Pflegerl, K. In Monolithic Materials: Preparation, Properties,
and Applications; Svec, F., Tennikova, T. B., Deyl, Z., Eds.; Journal of
Chromatography Library 67; Elsevier: Boston, 2003; pp 725-743.
(7) See ref 6.
(8) Xu, M.; Brahmachary, E.; Janco, M.; Ling, F. H.; Svec, F.; Fre´chet, J. M. J.
J. Chromatogr., A 2001, 928, 25-50.
(9) Hahn, R.; Podgornik, A.; Merhar, M.; Schallaun, E.; Jungbauer, A. Anal.
Chem. 2001, 73, 5126-5132.
(10) Lewandowski, K.; Murer, P.; Svec, F.; Fre´chet, J. M. J. Anal. Chem. 1998,
70, 1629-1638.
(11) Burguete, M. I.; Fre´chet, J. M. J.; Garc´ıa-Verdugo, E.; Janco, M.; Luis, S.
V.; Svec, F.; Vincent, M. J.; Xu, M. Polym. Bull. 2002, 48, 9-15.
(12) Svec, F.; Tennikova, T. B. J. Bioact. Compat. Polym. 1991, 6, 393-405.
(13) La¨mmerhofer, M.; Peters, E. C.; Yu, C.; Svec, F.; Fre´chet, J. M. J. Anal.
Chem. 2000, 72, 4614-4622.
* Corresponding author. E-mail: norbert.maier@univie.ac.at. Phone: ++43-
1-4277-52373. Fax: ++43-1-4277-9523.
† University of Vienna.
‡ Compie`gne University of Technology.
(14) L¨ammerhofer, M.; Svec, F.; Fre´chet, J. M. J.; Lindner, W. Anal. Chem. 2000,
72, 4623-4628.
§ Swiss Federal Institute of Technology.
(1) Piras, P.; Roussel, C.; Pierrot-Sanders, J. J. Chromatogr., A 2001, 906, 443-
458.
(15) Lee, Y.-K.; Yamashita, K.; Eto, M.; Onimura, K.; Tsutsumi, H.; Oishi, T.
Polymer 2002, 43, 7539-7547.
(2) Scott, R. P. W. Silica Gel and Bonded Phases. Their Production, Properties
and Use in LC; Wiley: Chichester, 1993.
(16) Lee, Y.-K.; Nakashima, Y.; Onimura, K.; Tsutsumi, H.; Oishi, T. Macromol-
ecules 2003, 36, 4735-4742.
(3) Buchmeiser, M. R. J. Chromatogr., A 2001, 918, 233-266.
(4) Claessens, H. A.; van Straten, M. A. J. Chromatogr., A 2004, 1060, 23-41.
(17) Oishi, T.; Lee, Y.-K.; Nakagawa, A.; Onimura, K.; Tsutsumi, H. J. Polym.
Sci., Part A 2002, 40, 1726-1741.
10.1021/ac050407s CCC: $30.25 © 2005 American Chemical Society
Published on Web 06/09/2005
Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 5009

The loss in performance reflects to some extent the problems
associated with controlling important macroscopic polymer prop-
erties, such as pore size and structure, surface area, different
nonspecific binding properties, and unpredictable solvent-induced
swelling behavior.11,21,22 To a major part, however, the observed
degradation in chiral recognition capacity of polymer-embedded
selectors may arise from largely ignored molecular-level insuf-
ficiencies. As the incorporation of polymerizable selectors occurs
in a random fashion, they will become fixed within the polymeric
network in all possible orientations and environments. A certain
fraction of the selector molecules may become buried in nonpo-
rous polymer regions, rendering them completely inaccessible to
target molecules. Alternatively, disposition of the selector in
crowded polymer domains may enforce detrimental conforma-
tional changes, which will impair the selector’s ability to establish
the subtle network of noncovalent contacts essential for efficient
and selective target binding. Degeneration of chiral recognition
capabilities may also originate from unfavorable influences acting
on the selector prior to polymerization. Prepolymer formulations
generally containsapart from the selector monomersa variety of
other molecular species at considerable concentration levels, such
as comonomers, cross-linking agents, and porogenic solvents.
Being exposed to these environments, the highly functionalized
chiral selector molecules are likely to complex more or less tightly
with these nonsubstrate components. In addition, under high
concentration conditions, selectors may undergo self-aggrega-
tion23-25 to a considerable extent, forming rather stable dimers,
oligomers, or both. On polymerization, these undesired hetero-/
homomolecular complexes may also become integrated into the
macromolecular network, generating blocked and thus unproduc-
tive selector sites. The combination of these deleterious effects
may result in materials displaying extremely heterogeneous
binding characteristics, becoming macroscopically evident in low
levels of enantioselectivity, binding capacity, and lack of peak
efficiency.15-20 The availability of procedures that allow an im-
proved level of control over accessibility, conformational con-
straints, and surface representation of the polymer-embedded
selectors certainly would greatly facilitate efforts to develop
polymer-type chiral stationary phases.
in the molecular recognition event, rendering them inaccessible
to perturbing interactions with nonsubstrate and other selector
molecules. (ii) Tight analyte binding should enforce the selector
to adopt and maintain its “productive conformation(s)” throughout
the entire polymerization process, enhancing the level of analyte-
specific preorganization of the resultant polymer-embedded bind-
ing site. (iii) Incorporation of intact selector-analyte assemblies
should also help to define the specific spatial requirements for
efficient selector-analyte interaction within polymeric networks,
diminishing the fraction of selector units being nonfunctional due
to steric restrictions. (iv) As an additional benefit, embedding
selector-analyte complexes into polymers may shape secondary
binding sites around the complexes, with the potential to enhance
enantioselectivity and substrate specificity.
Analyte templating bears resemblance to classical molecular
imprinting26-29 in that both exploit template molecules to shape
(more or less) target-specific polymer-embedded binding sites.
However, there are important conceptual differences concerning
the nature of the functional monomers and the specific role of
the templates. In molecular imprinting, the template molecules
serve to spatially preassemble functional monomers lacking
inherent selectivity. Subsequent polymerization is an essential step
to stabilize the array of monomers assembled around the template
to create selective binding sites. In contrast, analyte templating
employs inherently selective functional monomers. These are
purposefully complexed with tightly binding analytes prior to
polymerization with the intention to preserve (and ideally enhance)
existing, rather than create new, binding sites upon incorporation
into the polymeric environment.
In this contribution, we test the feasibility of this new concept
by preparing a series of analyte-templated porous polymers;
employing a polymerizable version of a well-established cinchona
carbamate-type selector, in the presence of tightly binding ana-
lytes. By variation of the structure and chirality of the templating
analytes, the molecular requirements for achieving significant
levels of binding site preservation of the polymer-embedded
selectors are probed. Observed differences in the chromatographic
chiral recognition characteristics of the corresponding polymers
are analyzed with respect to binding capacities and affinities and
compared to those obtained with a conventional silica-grafted
reference material. Changes in substrate-specific enantioselectivity
originating from analyte templating are also addressed.
A simple, yet elegant approach toward this goal may consist
in a temporary blockage of the selector binding site during
polymerization, exploiting its ability to form strong and spatially
well-defined, but noncovalent and thus reversible, complexes with
favorable analytes. Employing this analyte-templating approach
as a protective measure may help to preserve selector functionality
in polymeric environments in a number of ways: (i) Specific
analyte binding should selectively “saturate” selector sites engaged
EXPERIMENTAL SECTION
Materials. Ethylene glycol dimethacrylate (EDMA), 2,2′-
azobisisobutyronitrile (AIBN), thionyl chloride, HPLC-grade metha-
nol, acetonitrile, chloroform, and dimethoxyethane were pur-
chased from Merck (Darmstadt, Germany). N-Hydroxysuccinimide,
N-ethyldiisopropylamine, tert-butylamine, methacryloyl chloride,
2-mercaptoethanol, tetrahydrofuran (THF), sodium chloride,
sodium hydrogencarbonate, magnesium sulfate, sodium hydrox-
ide, and acetic acid (analytical grade) were supplied from Fluka
(Buchs, Switzerland). 3,5-Dichlorobenzoyl chloride, triphenyl-
(18) Lee, K.-P.; Choi, S.-H.; Kim, S.-Y.; Kim, T.-H.; Ryoo, J. J.; Ohta, K.; Jin, J.-Y.;
Takeuchi, T.; Fujimoto, C. J. Chromatogr., A 2003, 987, 111-118.
(19) Mayr, B.; Schottenberger, H.; Elsner, O.; Buchmeiser, M. R. J. Chromatogr.,
A 2002, 973, 115-122.
(20) Mayr, B.; Sinner, F.; Buchmeiser, M. R. J. Chromatogr., A 2001, 907, 47-
56.
(21) Xu, M.; Peterson, D. S.; Rohr, T.; Svec, F.; Fre´chet, J. M. J. Anal. Chem.
2003, 75, 1011-1021.
(22) Buchmeiser, M. R. New J. Chem. 2004, 28, 549-557.
(23) Uccello-Barretta, G.; Bari, L. D.; Salvadori, P. Magn. Reson. Chem. 1992,
30, 1054-1063.
(26) Piletsky, S. A.; Andersson, H. S.; Nicholls, I. A. Macromolecules 1999, 32,
633-636.
(27) Wulff, G. Chem. Rev. 2002, 102, 1-27.
(28) Wulff, G.; Knorr, K. Bioseparation 2002, 10, 257-276.
(29) Hosoya, K.; Tanaka, N. In Molecular and Ionic Recognition with Imprinted
Polymers; Bartsch, R. A., Maeda, M., Eds.; American Chemical Society:
Washington, DC, 1997; pp 143-158.
(24) Duret, P.; Foucault, A.; Margraff, R. J. Liq. Chromatogr., Relat. Technol.
2000, 23, 295-312.
(25) Pirkle, W. H.; Pochapsky, T. C. J. Am. Chem. Soc. 1987, 109, 5975-5982.
5010 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

phosphine, and diethyl azodicarboxylate were from Aldrich
(Austria, Vienna). tert-Butylmethacrylamide (tBu-MAA) was pre-
pared following the procedure described in ref 43.
(R)-N-3,5-Dichlorobenzoylleucine (DCB-(R)-Leu). (R)-Leucine
(1.47 g, 11.2 mmol) and NaHCO3 (3.15 g, 56 mmol) were dissolved
in 100 mL of water with stirring. O-Succinimmidyl-3,5-dichloroben-
zoate (2.16 g, 7.5 mmol) was dissolved in 50 mL of THF and added
in single portion to the amino acid solution. The mixture was
stirred at ambient temperature for 18 h. The reaction mixture was
evaporated under reduced pressure. The solid residue was
suspended in 100 mL of water, and concentrated aqueous HCl
was added to adjust the pH to ∼2. The solid was isolated by
filtration and dried at 60 °C to give 2.22 g (97%) of a white solid:
mp after recrystallization from MeOH/H2O (4:3) 158-160 °C; IR
(KBr) 3279, 3082, 2961, 1719, 1643, 1567, 1418, 1333, 1274, 1240,
General Information. Unless otherwise stated, all reactions
were carried out under strictly anhydrous conditions under
nitrogen atmosphere. All solvents were dried and distilled by
standard procedures prior to use. 1H spectra were acquired on a
Bruker DRX 400 MHz spectrometer. Chemical shifts are given
in parts per million (δ ppm) with respect to TMS as internal
standard. FT-IR spectra were recorded on a Perkin-Elmer Spek-
trum 2000 spectrometer. Optical rotation values were acquired
on a Perkin-Elmer 341 polarimeter at 25 °C. Melting points were
determined with a Kofler apparatus, equipped with a Leica
microscope. Flash chromatography was performed on Silica 60
(0.040-0.063-mm particle size, Merck).
Instruments. The mechanic ball mill (MM 2000) used for
polymer grinding was from Retsch (Haan, Germany). A Merck-
Hitachi L-6000 preparative pump was used for column packing of
the polymers. Chromatographic data were acquired on a Merck-
Hitachi L-7000 HPLC system consisting of an L-7150 semiprepara-
tive pump, a L-7250 autosampler, and a L-7455 diode array
detector. A D-7000 software was employed for chromatogram
interpretation and data processing. All mobile phases were
composed from HPLC-grade solvents.
Synthesis of Templating Analytes and Monomers. O-Suc-
cinimidyl-3,5-dichlorobenzoate. Dry N-hydroxysuccinimide (15.6 g,
134 mmol) was dissolved in 200 mL of dry THF in the presence
of N-ethyldiisopropylamine (23 mL, 134 mmol). To the precooled
solution (ice/H2O bath) 3,5-dichlorobenzoyl chloride (23.5 g, 112
mmol) in 50 mL of dry THF was added dropwise with stirring.
After 2 h the reaction mixture was filtered to remove precipitated
amine hydrochloride. The filtrate was evaporated under reduced
pressure at 40 °C. The solid residue was extracted with 2 M HCl
and ethyl acetate (100 mL/350 mL). The organic phase was
washed with brine (2 × 100 mL) and dried (MgSO4). Evaporation
of the solvent under reduced pressure gave 31.2 g (97%) of white
crystals: mp 131-134 °C; IR (KBr) 3082, 2946, 1750, 1571, 1427,
1405, 1354, 1304, 1227, 1197, 1118 cm-1; 1H NMR (CDCl3) δ 8.01
(2H, s), 7.67 (1H, s), 2.91 (4H, s).
1
1148 cm-1; H NMR (CDCl3) δ 7.65 (2H, s), 7.50 (1H, s), 7.55
(1H, d), 4.83 (1H, m), 1.76 (3H, m), 0.98 (6H, d); optical rotation
[R]589 ) +11.2°, [R]546 ) + 12.5°, [R]436 ) + 19.2° (c ) 1.0,
MeOH).
(S)-N-3,5-Dichlorobenzoylleucine (DCB-(S)-Leu). This com-
pound was prepared from (S)-leucine analogous to the procedure
described above for DCB-(R)-Leu: yield 2.12 g (93%) of a white
solid; optical rotation [R]589 ) -11.1°, [R]546 ) -12.6°, [R]436 ) -
19.3° (c ) 1.0, MeOH). All other physical properties were identical
with those of the (R)-enantiomer.
(R,S)-N-3,5-Dichlorobenzoylphenylalanine (DCB-(R,S)-Phe). This
compound was prepared from (R,S)-phenylalanine analogous to
the procedure described above for DCB-(R)-Leu: yield 2.45 g
(96%) of a white solid recrystallized from MeOH/H2O (4:3).
3,5-Dichlorobenzoylglycine (DCB-Gly). This compound was
prepared from glycine analogous to the procedure described above
for DCB-(R)-Leu: yield 1.79 g (96%) of a white solid; mp after
recrystallization from MeOH/H2O 1/1, 188-190°C; IR (KBr):
1
3293, 3079, 1716, 1638, 15552, 1423, 1335, 1239 cm-1; H NMR
(CDCl3) δ 7.71 (2H, s), 7.52 (1H, s), 3.98 (2H, s).
11-[(2-Hydroxyethyl)thia]-9-tert-butylcarbamoyl dihydroquinine
(1). Quinine tert-butylcarbamate (20.00 g, 47.3 mmol) was dis-
solved in 150 mL of dry CHCl3. 2-Mercaptoethanol (6.60 mL, 94.6
mmol) and AIBN (0.25 g) were added. The mixture was refluxed
with stirring for 20 h. The reaction mixture was extracted with 3
M aqueous NaOH (3 × 200 mL). The organic phase was dried
(MgSO4) and the solvent evaporated under reduced pressure to
give a yellowish oil, which crystallized on addition of ethyl
acetate: yield 22.8 g (96%) of white crystals; mp 179-182 °C; IR
(KBr) 3339, 2915, 1731, 1619, 1591, 1509, 1457, 1391, 1365, 1267,
(30) Maier, N. M.; Nicoletti, L.; L¨ammerhofer, M.; Lindner, W. Chirality 1999,
11, 522-528.
(31) La¨mmerhofer, M.; Lindner, W. J. Chromatogr., A 1996, 74, 33-48.
(32) Krawinkler, K. H.; Maier, N. M.; Sajovic, E.; Lindner, W. J. Chromatogr., A
2004, 1053, 119-131.
(33) La¨mmerhofer, M.; Maier, N. M.; Lindner, W. Am. Lab. 1998, 30, 71.
(34) Lah, J.; Maier, N. M.; Lindner, W.; Vesnaver, G. J. Phys. Chem., B 2001,
105, 1670-1678.
(35) Maier, N. M.; Schefzick, S.; Lombardo, G. M.; Feliz, M.; Rissanen, K.;
Lindner, W.; Lipkowitz, K. J. Am. Chem. Soc. 2002, 124, 8611-8629.
(36) Czerwenka, C.; Zhang, M. M.; Ka¨hlig, H.; Maier, N. M.; Lipkovitz, K. B.;
Lindner, W. J. Org. Chem. 2003, 68, 8315-8327.
(37) Wirz, R.; Bu¨rgi, T.; Lindner, W.; Baiker, A. Anal. Chem. 2004, 76, 5319-
5330.
(38) Lesnik, J.; La¨mmerhofer, M.; Lindner, W. Anal. Chim. Acta 1999, 401,
3-10.
(39) Franco, P.; La¨mmerhofer, M.; Klaus, P. M.; Lindner, W. J. Chromatogr., A
2000, 869, 111-127.
(40) Franco, P.; Klaus, P. M.; Minguillo´n, C.; Lindner, W. Chirality 2001, 13,
177-186.
(41) Sellergren, B. J. Chromatogr., A 2001, 906, 227-252.
(42) Haupt, K.; Mosbach, K. Chem. Rev. 2000, 100, 2495-2504.
(43) Jodlbauer, J.; Maier, N. M.; Lindner, W. J. Chromatogr., A 2002, 945, 45-
63.
1
1227 cm-1; H NMR (CDCl3) δ 8.73 (1H, d), 7.99 (1H, d), 7.48
(1H, s), 7.35 (2H, m), 6.42 (1H, d), 4.73 (1H, s), 3.95 (3H, s), 3.70
(2H, t), 3.28 (1H, d), 3.06 (2H, m), 2.72 (2H, t), 2.62 (1H, m), 2.48
(2H, t), 2.33 (1H, d), 1.28 (9H, s); optical rotation [R]589 ) -24.6°,
[R]546 ) - 29.5°, [R]436 ) - 67.6° (c ) 1.0, MeOH).
11-[(2-Aminoethyl)thia]-9-tert-butylcarbamoyl-dihydroquinine (2).
11-[(2-Hydroxyethyl)thia]-9-tert-butylcarbamoyldihydroquinine (10.00
g, 19.9 mmol) was dissolved in 200 mL of dry THF. Triphen-
ylphosphine (6.27 g, 23.9 mmol) and hydrazoic acid (2 M in
toluene, 11.90 mL, 23.9 mmol) were added. The mixture was
cooled (ice/H2O), and diethylazodicarboxylate (3.77 mL, 23.9
mmol) in 30 mL of dry THF was added dropwise with stirring.
After 3 h another portion of triphenylphosphine (5.23 g, 23.92
mmol) was added. The reaction mixture was stirred at 50 °C for
2 h. Then 3 mL of H2O were added, and stirring was continued at
the same temperature for 20 h. The solvent was removed under
Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 5011

Table 1. Composition of the Templated (P1-P4) and Control (CP1-CP4) Polymers
tBuCQN-MAA
(mmol)
tBu-MAA
(mmol)
EDMA
(mmol)
AIBN
MeOH
(mL)
material
template/mmol
(mmol)
P1
1.00
3.00
3.00
3.00
3.00
4.00
4.00
4.00
4.00
DCB-(S)-Leu/1.00
DCB-(R)-Leu/1.00
DCB-Gly/1.00
none
DCB-(S)-Leu/1.00
DCB-(R)-Leu/1.00
DCB-Gly/1.00
none
20.00
20.00
20.00
20.00
20.00
20.00
20.00
20.00
0.24
0.24
0.24
0.24
0.24
0.24
0.24
0.24
5.60
5.60
5.60
5.60
5.60
5.60
5.60
5.60
P2
1.00
P3
P4
CP1
CP2
CP3
CP4
1.00
1.00
none
none
none
none
reduced pressure to give a yellowish oil. This was taken into 200
mL of CHCl3 and extracted with 2 M HCl (250 mL). The organic
phase was discarded. The aqueous phase was washed with 50
mL of CHCl3, which again was discarded. Ice (50 mL) was added
to the aqueous phase, and solid sodium hydroxide was added in
small portions to adjust to pH 14. The strongly basic mixture was
extracted with CHCl3 (1 × 200 mL, 1 × 100 mL). The combined
organic phases were dried (MgSO4), and the solvent was removed
under reduced pressure to yield 7.85 g (79%) of a yellowish foam.
An analytically pure sample was obtained by flash chromatography
(aminopropyl-modified silica/ethyl acetate): mp 81-85 °C; IR
(KBr) 3351, 2934, 1713, 1621, 1592, 1509, 1474, 1393, 1364, 1267
cm-1; 1H NMR (CDCl3) δ 8.74 (1H, s), 8.00 (1H, d), 7.47 (1H, s),
7.35 (2H, m), 7.26 (1H, s), 6.43 (1H, d), 4.78 (1H, s), 3.95 (3H, s),
3.25 (1H, m), 3.06 (2H, m), 2.84 (2H, t), 2.59 (3H, t), 2.46 (2H, t),
several layers of a plastic insulation tape. The tubes were then
incubated for thermal polymerization in a water bath at 65 ( 0.1
°C for 24 h. The resultant polymer rods were manually broken
into small pieces using a mortar and subsequently subjected to
mechanical grinding employing a ball mill. For size classification,
the ground polymers were suspended in acetone and wet-sieved
through a 25-µm steel sieve by gentle agitation with a brush. To
achieve the removal of fines, the fractions of <25 µm were
subjected to five suspension-sedimentation cycles in acetone/
acetic acid (100:5, v/v). For this purpose, the polymers were
transferred into high-walled 250-mL beakers and suspended in
200 mL of sedimentation medium. After 30 min, the turbid
supernatants containing primarily fines were removed from the
settled materials and discarded. The resultant materials were dried
at ambient temperature for 72 h and then 24 h at 60 °C in high
vacuum to give polymers P1-P4 and CP1-CP4 with a mean
particle size of 15-25 µm, with yields ranging between 1.5 and
2.1 g. To establish the receptor incorporation levels of the
cinchona-type polymers P1-P4, exhaustively washed samples
(MeOH/AcOH, 95:5) were subjected to elemental analysis. The
receptor incorporation levels calculated on the basis on the sulfur
content were 116, 109, 101, and 123 µmol/g for P1, P2, P3, and
P4, respectively.
Preparation of the Silica-Grafted tBuCQN-Type CSP. The
silica-grafted chiral stationary phase displaying a receptor surface
loading similar to the receptor incorporation level of polymers
P1-P4 was prepared by controlled immobilization of tBuCQN onto
mercaptopropyl-modified silica gel.30 Briefly, a 3.00-g amount of
mercaptopropyl-modified silica (ProntoSIL 120 (5 µm) comprising
0.90 mmol of SH groups/g) was suspended in 40 mL of methanol.
tBuCQN (150 mg, 0.35 mmol) and 5 mg AIBN were added. The
mixture was refluxed with gentle mechanical stirring for 10 h.
The modified silica gel was isolated by filtration through a glass
filter funnel, washed with hot methanol (5 × 50 mL), and dried
in high vacuum at 60 °C. Elemental analysis of the modified silica
gel gave 8.50% C, 0.46% N, and 2.80% S, corresponding to a
receptor loading level of 110 µmol/g (calculated based on N).
Column Packing. For chromatographic evaluation, polymers
P1-P4, CP1-CP4, and the silica-grafted CSP were slurry packed
into 125 × 4 mm i.d. stainless steel HPLC columns. The polymers
(∼1.0 g) were suspended in 10 mL of methanol/acetic acid
(95:5) and allowed to equilibrate for 2 h. Then the slurry was
sonicated for 10 min and packed at a constant pressure of 150
bar using methanol as pressuring solvent. All polymer columns
were washed with methanol comprising 5% acetic acid (1 mL/
min for 15 h) to remove any trace of templates and nonincorpo-
rated compounds. The silica-grafted CSP was packed under
2.33 (1H, d), 1.47 (1H, m), 1.28 (9H, s); optical rotation [R]589
-22.0°, [R]546 ) -27.6°, [R]436 ) -64.4° (c ) 1.0, MeOH).
)
11-[(2-(Methacrylamido)ethyl)thia]-9-tert-butylcarbamoyldi-
hydroquinine (tBuCQN-MAA). The resulting amine (7.80 g, 15.6
mmol) was dissolved in 50 mL of CHCl3 and cooled in ice/H2O
bath. Methacryloyl chloride (3.02 mL, 31.15 mmol) was dissolved
in 25 mL of CHCl3 and added dropwise to the amine solution with
stirring. After 20 h the solvent was removed under reduced
pressure at ambient temperature. The resultant white foam was
dissolved in 150 mL of CHCl3 and washed with 1 M aqueous
NaOH (3 × 100 mL). The organic phase was dried (MgSO4) and
evaporated under reduced pressure. The product was purified by
flash chromatography (150 g of silica, CHCl3/MeOH, 10:1) to yield
4.87 g (55%) of a white solid. An analytically pure sample was
obtained by recrystallization from ethyl acetate: mp 168-171 °C;
IR (KBr) 3310, 2932, 2361, 1713, 1660, 1621, 1511, 1456, 1393,
1365, 1267, 1228 cm-1; 1H NMR (CDCl3) δ 8.74 (1H, d), 8.00 (1H,
d), 7.49 (1H, s), 7.35 (2H, m), 6.43 (1H, d), 6.18 (1H, s), 5.69 (1H,
s), 5.34 (1H, s), 4.77 (1H, s), 3.96 (3H, s), 3.49 (2H, m), 3.29 (1H,
m), 3.07 (2H, m), 2.63 (3H, m), 2.50 (2H, t), 2.35 (1H, d), 1.95
(3H, s), 1.28 (9H, s); optical rotation [R]589 ) -22.1°, [R]546
-27.6°, [R]436 ) -62.0° (c ) 1.0, MeOH).
)
Preparation and Size Classification of Polymers P1-P4/
CP1-CP4. All of the cinchona receptor-type polymers P1-P4
and corresponding control polymers CP1-CP4 were prepared
according to the following general procedure:
The respective amounts of monomers, initiator, and porogenic
solvent indicated in Table 1 were transferred into in glass screw-
cap centrifugation tubes (25 mL, 120 × 15 mm i.d.). The solutions
were degassed by passing a vigorous stream of dry nitrogen
through the viscous mixtures for 5 min. The tubes were tightly
closed with Teflon-lined screw caps and additionally sealed with
5012 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

Figure 1. Synthesis of the polymerizable quinine tert-butylcarbamate-type selector.
analogous conditions, but employing a constant packing pressure
of 600 bar.
analyte must be sufficiently soluble in the given polymerization
environment, avoiding poor integration into the polymer due to
premature physical precipitation.
Single-Component Binding Isotherms. Single-component
binding isotherms for the individual DCB-Leu enantiomers on
polymers P1 and P2 and the silica-grafted CSP were established
by batch extraction experiments. Briefly, 50-mg amounts of
polymeric and silica-type adsorbents, respectively, were incubated
for 24 h at 25°C with 4 mL of DCB-(S)-Leu/DCB-(R)-Leu methanol
stock solutions, covering concentrations ([DCB-Leu]0) ranging
from 0.25 to 150 mM. HPLC analysis of the supernatants of the
equilibrated samples gave the concentrations of the free enanti-
omer, [DCB-Leu]free.. The amounts of DCB-Leu enantiomer bound
by the polymers/CSP ([DCB-Leu]bound) were calculated from the
difference [DCB-Leu]0 - [DCB-Leu]free. Each data point reported
in this study represents the mean value of a duplicate determi-
nation, generally with a deviation of <5%. The corresponding data
were plotted in the form [DCB-Leu]bound versus [DCB-Leu]free and
fitted to a single-site Langmuir model. For HPLC quantification
of the DCB-Leu enantiomers, a 6′-hydroxy-tBuCQN chiral station-
ary phase (5 µm, 150 × 4 mm i.d.) was employed. The mobile
phase consisted of a mixture of methanol and 0.1 M aqueous
ammonium acetate (80:20 v/v), pHa was 6.00. The flow rate was
set at 1 mL/min, and peaks were detected at 254 nm.
These requirements are ideally fulfilled by cinchona carbamate-
type selectors.31,32 Silica-grafted versions of these selectors are
capable of binding a broad assortment of acidic chiral analytes33
with considerable affinity (Kd in the low mM range34) and
exceptionally high enantioselectivity (R up to 3035), even in
competing polar protic media. Cinchona carbamates also offer the
benefits of being fully compatible with and chemically inert under
free radical polymerization conditions.13,14,30 Moreover, a substan-
tial body of mechanistic knowledge on analyte binding with this
class of selectors is available,35-38 facilitating the design of
polymerizable versions with minimal perturbation of molecular
recognition capacity.
Guided by the advanced molecular-level understanding of its
chiral recognition process, we chose the exceptionally successful
quinine tert-butylcarbamate selector motif for exploratory analyte-
templating studies.30,33 For this selector, X-ray structure and NMR
spectroscopic data evidence the existence of a spatially well-
defined analyte binding site, being located within a cleft-shaped
void generated by the carbamate and quinoline groups.30,35,36
However, even minor structural changes at, or in proximity to
this pocket were found to alter and, in most cases, to impair
enantioselective analyte binding properties.39,40 Considering the
sensitive nature of the binding site architecture, we devised a
noninvasive strategy to create a suitable polymerizable version
of this selector, entailing the introduction of a polymerizable
functional group at the rather distant C11 position. Attaching a
polymerizable functionality at this remote position was anticipated
to confer little, if any, interference to the C6/O9 binding domain
of the selector.
RESULTS AND DISCUSSION
Selection of Model System. Synthesis of Polymerizable
Selector. The aim of analyte templating is the preservation of
selector functionality upon incorporation into a porous polymeric
environment by temporary blocking of its binding site via an
appropriate analyte. An essential requirement for maximizing
these protective interactions between the polymerizable selector
and the templating analyte is a high level of noncovalent associa-
tion, which ideally should be maintained during polymerization.
In addition, both polymerizable selector and templating analyte
must be chemically inert under polymerization conditions and
should be devoid of polymerization-inhibiting properties. Finally,
the complexes formed by the polymerizable selector and the
The polymerizable version of quinine tert-butylcarbamate was
synthesized in three-steps as outlined in Figure 1. For this
purpose, quinine tert-butylcarbamate was reacted with mercapto-
ethanol under free radical addition conditions, transforming the
Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 5013

quinuclidine-linked vinyl function into the corresponding hydroxy
thioether derivative (96% yield). This intermediate 1 was then
converted via a one-pot Mitsunobu-substitution/Staudinger-reduc-
tion protocol into the corresponding amino compound 2, employ-
ing hydrazoic acid as nucleophilic nitrogen source (79%). Finally,
acylation of amine 2 with methacryloyl chloride provided the
polymerizable methacrylamide-type selector tBuCQN-MAA in 55%
overall yield.
macroporous polymers.44,45 Using methanol as the porogenic
solvent was also anticipated to favor chiral recognition by mimick-
ing the polar protic interaction environment in which cinchona
carbamate-type selectors perform best.
1
Standard H NMR titration experiments performed with the
polymerizable selector tBuCQN-MAA and the DCB-amino acids
in deuterated methanol revealed clean 1:1 complex stoichiometry
for all combination and dissociation constants in the low-micro-
molar range (DCB-(S)-Leu, Kd ) 1.42 mM; DCB-(R)-Leu, Kd )
7.45 mM; DCB-Gly, Kd ) 2.21 mM). These dissociation constants
were used to estimate the degrees of selector-analyte complex-
ation in the polymerization mixture. Assuming selector and analyte
concentrations both being in the 0.1 M range (and neglecting the
influence of all other components of the polymerization mixture),
roughly 75-90% of the polymerizable selector should exist as
analyte-templated species.
Attempts to perform photochemical polymerization in the
presence of a free radical initiator (AIBN) were frustrated by low
reaction rates and formation of deeply colored mixtures; indicating
some degree of degradation. Under thermal conditions (AIBN,
60 °C), however, polymerization proceeded smoothly without any
indication of decomposition. Consequently, efforts to achieve
polymerization photochemically were abandoned and all polymeric
materials discussed in the forthcoming sections were produced
under thermal reaction conditions.
Employing the protocol outlined above, a series of analyte-
templated porous polymers and nontemplated control materials
was prepared, the compositions of which are summarized in Table
1. Polymers P1-P3 were produced by copolymerizing the selector
monomer in the presence of the selected templates. Thus, P1 was
templated with DCB-(S)-Leu, the analyte that showed the tightest
binding for the employed selector in polar protic environments.
The high binding affinity and the excellent molecular-level
complementarity of this selector-analyte combination was antici-
pated to give rise to a high concentration of spatially well-defined
analyte-template complexes in the polymerization mixture. As a
consequence, this “match case” polymer should comprise a
significant population of templated selector sites, manifesting
themselves in enhanced molecular recognition capacity for the
templating analyte. In contrast, P2 was prepared in the presence
of the less strongly bound DCB-(R)-Leu enantiomer. Here, the
rather low affinity to the employed selector should lead to a
relatively low number of loose analyte-selector associates. This
“mismatch case” polymer was predicted to harbor a low number
of templated selector sites and, thus, to exhibit poor molecular
recognition capabilities. P3, templated with DCB-Gly, was prepared
to probe the impact of chirality and analyte size on analyte
templating. Showing an intermediate affinity to the selector, but
lacking the amino acid side chain, DCB-Gly may still successfully
trigger the formation of selector binding sites; however, they are
expected to be spatially restricted. These “nonchirally templated”
selector binding sites in P3 may display reduced chiral recognition
capacity for spatially more demanding analytes. P4 was produced
in absence of any analyte to serve as a reference to assess
presence and nature of analyte templating effects in P1-P3. In
addition to these polymers, the corresponding control materials
Template analytes were selected based upon their selector
binding characteristics (affinity and enantioselectivity) and phys-
icochemical criteria (compatibility with polymerization conditions,
chemical inertness, solubility of selector-analyte complexes).
Initially, N-3,5-dinitrobenzoyl amino acids, which exhibit outstand-
ing levels of enantioselectivity and binding affinity with cinchona
carbamate-type selectors,30,35,36 were evaluated as potential tem-
plates. In the course of pilot studies, however, this class of
compounds proved to be unsuitable for analyte-templating ap-
plications due to formation of sparingly soluble selector-analyte
salts and strong polymerization inhibition effects. Suspecting that
these prohibitive analyte properties may arise from the aromatic
nitro groups, the corresponding 3,5-dichlorobenzoyl (DCB) amino
acid derivatives were prepared and evaluated. Indeed, the DCB-
amino acids were fully compatible with free radical polymerization
conditions and showed satisfactory solubility in the presence of
the selector, as well as in a broad range of solvents. Subsequent
chromatographic screening of a comprehensive set of 3,5-dichlo-
robenzoyl amino acids on a silica-grafted quinine tert-butylcar-
bamate chiral stationary phase identified DCB-leucine (DCB-Leu)
as the analyte exhibiting the highest level of enantioselectivity in
combination with strong retention. Due to these favorable interac-
tion characteristics, the DCB-Leu enantiomers were selected to
serve as chiral templates for the production of analyte-templated
polymers. In addition, DCB-glycine was selected as a structure-
related, but nonchiral template molecule to assess the impact of
chirality on the outcome of an analyte-templating experiment.
Preparation of Analyte-Templated and Control Polymers.
To obtain self-supporting, macroporous analyte-templated bulk
polymers suitable for chromatographic screening, experimental
protocols traditionally employed in noncovalent molecular imprint-
ing technology were adapted28,41,42 (see Figure 2).
Polymer formulations were prepared comprising, apart from
the polymerizable selector-analyte complex, a cross-linking agent
(20 mol equiv), a lipophilic comonomer (3 mol equiv), and a
porogenic solvent. The cross-linking agent (EDMA) was employed
in large excess to ensure the formation of shape-persistent bulk
polymers and a sufficient level of network rigidity to preserve
favorable selector arrangements and microenvironments gener-
ated by analyte templating. tBu-MMA was included as a potential
interaction-enhancing agent.43 We speculated that in the relatively
polar polymerization environment this lipophilic compound may
preferentially associate onto exposed hydrophobic domains of
selector-analyte complexes. This solvophobic effect was expected
to enhance the electrostatic stabilization of selector-analyte
complexes and support the creation of shape-selective microen-
vironments. Methanol was chosen as porogenic solvent, primarily
due to its excellent solvation properties for all monomeric
components and its known tendency to promote the formation of
(44) Sellergren, B.; Shea, K. J. J. Chromatogr. 1993, 635, 31-49.
(45) Kempe, M. Chiral Recognition. Doctoral Thesis, Lund University, 1994.
5014 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

Figure 2. Schematic representation of the polymerization protocol employed for the preparation of the analyte-templated polymers.
(CP1-CP4) were generated in the absence of the polymerizable
cinchona derivative but with the templating analyte(s) for probing
analyte binding resulting from polymer components other than
selector.
Table 2. Characterization of the Templated Polymers
P1-P4
SO loadinga BET surf. areab pore volc pore diamd
material
(mmol/g)
(m2/g)
(mL/g)
(Å)
The resultant porous bulk polymers were grained and sub-
jected to size classification by sieving to isolate the 10-25-µm
particle size fraction. After template removal by exhaustive
washing, the selector loading levels of the polymers were
determined based upon the sulfur content by elemental analysis.
A rather uniform selector loading of 110 ( 10 µmol/g of polymer
was detected for P1-P4, providing evidence that the presence of
the templating analytes does not interfere chemically with the
polymerization process. The rate of selector incorporation was in
the range of 53-63% of the theoretically possible amounts.
Determination of the porosity characteristics of the analyte-
templated polymers was carried out by BET nitrogen adsorption/
desorption and mercury intrusion porosimetry, respectively, and
the results are reported in Table 2. Significant differences in total
BET surface areas were observed, with P1 exhibiting a surface
area three times that of P2 and approximately two times that of
P1
P2
P3
P4
116
112
113
123
93.29
28.68
50.58
56.06
0.86
0.95
0.96
0.93
2500
2800
3000
2400
a Calculated based on the sulfur content. b Surface area determined
from N2 adsorption/desorption isotherm (using BET equation). c Pore
volume measured by mercury intrusion porosimetry. d Pore size at
maximum of distribution curve.
P3 and P4. However, by mercury intrusion porosimetry, all
polymers showed comparable values for pore volume and pore
size. From these data, it can be concluded that the observed
differences in total surface area are merely a reflection of the
varying levels of microporosity and that the presence of the
templating analyte does not change the inherently macroporous
nature of the polymers.
Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 5015

Table 3. Enantiomer Separation Data of DCB-Leu and
DCB-Phe on Reference Silica CSP, Templated
Polymers P1-P4, and Control Polymers CP1-CP4
DCB-Leu
DCB-Phe
a
b
a
b
k(R)
k(S)
Rc
k(R)
k(S)
Rc
CSP
P1
1.90
1.39
1.44
1.02
0.95
nr
16.79
13.01
9.08
7.18
6.8
8.83
9.33
6.31
7.04
7.13
5.18
4.26
4.22
3.33
3.03
nr
30.41
21.67
16.24
14.15
13.3
nr
5.87
5.08
3.89
4.25
4.39
P2
P3
P4
CP1
CP2
CP3
CP4
nr
Figure 3. Single-component binding isotherms for (S)- and (R)-
DCB-Leu enantiomers for polymers P1, P2, and the silica-grafted
reference CSP.
nr
nr
nr
nr
nr
nr
nr
nr
nr
nr
nr
nr
a k(R) ) (t(R) - t0)/t0. k(S) ) (t(S) - t0)/t0. c R ) k(S)/k(R). Chro-
matographic conditions, 125 × 4 mm i.d. columns; mobile phase,
MeOH + 1% AcOH; flow, 1 mL/min; T ) 25 °C; UV detection at 254
nm. nr, not retained.
b
improvements in enantioselectivity originate primarily from co-
generated molecular imprinted cavities rather than sophisticated
analyte-templated selector sites. This concern, however, can be
safely ruled out on basis of the lack of chiral recognition of the
control polymers CP1-4. As is evident from the data in Table 3,
none of these selector-free polymers showed any retention under
the given polar protic mobile-phase condition. Molecular imprinted
cavities, in addition to analyte-templated binding domains, may
still exist in these polymers. Their contributions to the observed
changes in the global chiral recognition characteristics, however,
appear to be negligible considering the strongly competing mobile-
phase environments employed in this study.
Binding Capacity and Binding Affinity of Templated
Polymers. The chromatographic data provide unambiguous
support that the chiral recognition capacity of polymer-embedded
selectors can be controlled by analyte templating. However, they
do not provide detailed information on the underlying mecha-
nisms. Chromatographically observed changes in retention (bind-
ing) and relative retention (selectivity) may originate from different
sources. Variations in retention may be brought about by a change
in binding affinity (thermodynamic binding), an alteration of
number of binding sites (capacity), or a combination of both
effects.
To dissect the changes in chiral recognition for the templated
enantiomers with respect to binding affinity and binding site
density, the adsorption behavior of the polymers showing most
pronounced analyte-templating effects, P1 and P2, and the silica-
grafted reference CSP, were studied in detail. For this purpose,
batch adsorption experiments with the individual DCB-Leu enan-
tiomers from methanol under controlled temperature conditions
were performed. Methanol was chosen as binding environment
to mimic the conditions under which the polymers were created,
with the intention of reestablishing the original solvation and
interaction scenario. Incubation of defined amounts of polymeric
and silica-grafted materials, respectively, with methanolic solutions
of DCB-(R)-Leu and DCB-(S)-Leu (0.25-150 mM) provided
equilibrium binding data that could be fitted with satisfactory
quality to a simple single-site binding model (see Figure 3). This
finding adds further support to the assumption that analyte binding
occurs at discrete analyte-templated selector sites, without sig-
nificant contributions of coexisting imprinted sites confering
heterogeneous binding characteristics.28,41,47,48 The values for the
Chromatographic Evaluation of Chiral Recognition Char-
acteristics of Templated Polymers. To assess the impact of the
analyte-templating procedure on the chiral recognition properties
of the polymer-embedded selector, the polymers were packed into
HPLC columns and chromatographically evaluated with racemic
DCB-Leu with methanol/acetic acid as mobile phase. The chro-
matographic results are summarized in Table 3. It is evident that
P1-P4 show considerable differences with regard to their enan-
tiomer separation characteristics, with which, the trends are
largely consistent with our predictions.
Among the selector-containing polymers, P1 displays the
highest enantioselectivity (R ) 9.33) and the strongest retention
factors. This finding is consistent with the existence of favorably
templated selector binding sites, enhancing both the affinity and
the selectivity of the polymer-embedded selector. Also consistent
with our expectations are the poor chiral recognition properties
found with P2 (R ) 6.31), which actually exhibited the lowest
enantioselectivity and retention factors among the investigated
polymers. The fact that its affinity and selectivity are inferior to
those of the nontemplated P4 (R ) 7.13) suggests that the
presence of the mismatch analyte induces anticooperative effects
to chiral recognition. Interestingly, the nonchirally templated P3
and the nontemplated P4 show very similar enantiomer separation
performances. The binding of DCB-Gly to the selector does not
give rise to any significant changes of its chiral recognition
capabilities upon incorporation into a polymeric environment. A
possible explanation for this observation may be that the relatively
small DCB-Gly template creates polymer-embedded selector sites,
which are too restricted in size to accommodate spatially more
demanding analytes. A particularly striking argument of the
existence of favorable analyte-templating effects in P1 can be
derived by comparing its enantiomer separation performance with
that of a silica-grafted reference CSP. This CSP, grafted with the
identical selector at a comparable loading level, outperforms P2-
P4 but is inferior to P1 with respect to enantioselectivity
(R ) 8.83 vs 9.33). Thus, polymer-embedded binding sites shaped
in the presence of a templating analyte may successfully compete
with their structurally identical, conformationally unrestricted, and
readily accessible surface-grafted congeners.
Considering the conceptual similarities between analyte tem-
plating and molecular imprinting, one may argue that the achieved
(46) Takeuchi, T.; Haginaka, J. J. Chromatogr., B 1999, 728, 1-20.
5016 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

Table 4. Binding of DCB-(R)- and DCB-(S)-Leu to Templated Polymers P1, P2, and Silica-Grafted Reference CSP from
Methanol;
P1
P2
CSP
DCB-(R)-Leu
DCB-(S)-Leu
DCB-(R)-Leu
DCB-(S)-Leu
DCB-(R)-Leu
DCB-(S)-Leu
Kd (mM)a
Bmax (µmol/g)b
R2c
5.11 ( 0.35
62.9 ( 2.0
0.995
1.32 ( 0.07
80.2 ( 1.2
0.996
3.38 ( 0.26
53.7 ( 1.6
0.992
1.20 ( 0.13
73.6 ( 2.2
0.98
7.41 ( 0.62
105.7 ( 4.1
0.993
1.20 ( 0.09
99 ( 1.9
0.99
a Dissociation constant. b Maximal binding capacity. c R2, regression factor indicating goodness of fit. Data extracted from best fit to [analyte]bound
) Bmax[analyte]free/(Kd + [analyte]free).
maximal binding capacity (Bmax) and corresponding dissociation
constants (Kd) extracted from the best-fit isotherms are listed in
Table 4.
individual enantiomers, as compared to the unhindered selector
presentation in the silica-grafted reference material. This effect
becomes particularly evident when comparing the binding capacity
of P1 with those of P2 and the silica-grafted CSP for the templating
enantiomer. Thus, P1 exhibits for the templating DCB-(S)-Leu a
significantly larger binding capacity as compared to P2 but still
less sites than the silica-grafted reference material.
Analyte templating also leads to significant changes in binding
affinities, with direct consequences on the chiral recognition
performance. The binding parameters in Table 4 again suggest
that these shifts in binding affinity are triggered in favor of the
templating analyte. Thus, templating with DCB-(S)-Leu for P1
causes a drop in binding affinity for DCB-(R)-Leu as compared to
P2. The opposite is true for DCB-(R)-Leu-templated P2, where
binding affinity becomes enhanced for the templating analyte
relative to P1. The overall effect of analyte templating, however,
emerges from the sum of both effects. For instance, the favorable
combination of a large number of binding sites and a high affinity
for the templating analyte makes P1 a superior binder for DCB-
(S)-Leu than P2, which offers comparable affinity, but a lower
number of binding sites for this analyte.
Substrate Specificity of Templated Polymers. Copolymer-
ization of an analyte (more or less tightly) associated with a
polymerizable selector can be expected to have some structural
influence on the surrounding polymeric microenvironment. Many
applications of noncovalent molecular imprinting reported in the
literature show that even rather weak complex-stabilizing interac-
tion forces can successfully induce the formation of size and shape
complementary, polymer-embedded binding sites. It is justified
to expect that analyte templating may also form spatially well-
defined cavities around the selector-analyte complex, with the
ability to memorize specific structural features, particularly size
and shape, of the templating analyte. Thus, templated polymers
should show an enhanced level of substrate specificity, preferen-
tially recognizing the original template and expressing a sort of
size exclusion effect for spatially more demanding analytes.
To probe the presence of enhanced shape selectivity in
templated polymers, we compared the chromatographic enanti-
omer separation behaviors of P1-P4 and the silica-grafted refer-
ence CSP for DCB-Leu and DCB-Phe. The chromatographic
results are given in Table 2. Changes in chiral recognition trends
were observed, being most significant for P1 and the silica-grafted
CSP. With DCB-Leu as analyte, polymer P1 exhibited higher
enantioselectivity than the silica-grafted CSP. In contrast, with the
structurally very closely related, but sterically more demanding
DCB-Phe as analyte, this trend is inverted. Evidently, the DCB-
These parameters provide valuable insights into the complex
interplay of binding site density and affinity contributions that
control the macroscopic chiral recognition performance of analyte-
templated polymers. The silica-grafted reference material was
found to offer the highest level (∼100 µmol/g) of binding sites
to the analytes, being practically identical for the individual DCB-
Leu enantiomers within the experimental error. These numbers
are also in excellent agreement with the theoretical loading level,
indicating ready accessibility of the surface-immobilized selector
units for both analytes. This complete availability of surface-
immobilized selector units is a reflection of the well-defined
porosity of the supporting silica material, allowing fast and efficient
mass transport. This situation was changed in the analyte-
templated organic polymers. Both polymeric materials displayed
binding capacities significantly lower than the numbers expected
from incorporation of the tBuCQN-MAA functional monomer and
also distinct numbers of accessible binding sites for the individual
enantiomers. With P1, the number of sites for DCB-(S)-Leu and
DCB-(R)-Leu were 80 (corresponding to 69% of the incorporated
selector units) and 63 µmol/g (54%), respectively. An even less
favorable situation was encountered with P2, where sites acces-
sible to DCB-(S)-Leu and DCB-(R)-Leu were further diminished
to 74 (66%) and 54 µmol/g (48%), respectively.
Specific differences were also observed for the investigated
materials with respect to binding affinities. While the dissociation
constants for the more tightly bound DCB-(S)-Leu were roughly
identical for silica-grafted and the polymeric materials (1.25 mM),
the values for the less strongly bound DCB-(R)-Leu were subject
to variation. Thus, the silica-grafted materials showed the lowest
affinity for DCB-(R)-Leu (Kd )7.4 mM), followed by P1 with
somewhat increased binding (Kd ) 5.1 mM) and P2 with
substantially higher affinity (Kd ) 3.4 mM).
Incorporation of polymerizable selectors into polymeric envi-
ronments compromises analyte accessibility with unfavorable
consequences with respect to capacity as well as selectivity.
Analyte templating appears to be able to improve this situation
by modulating the immediate selector environment in a favorable
fashion. The binding parameters extracted from the isotherms
suggest that this modulation results from a differential reduction
in accessibility of polymer-embedded selector sites for the
(47) Stanley, B. J.; Szabelsky, P.; Chen, Y. B.; Sellergren, B.; Guiochon, G.
Langmuir 2003, 19, 772-778.
(48) Rampey, A. M.; Umpleby, R. J.; Rushton, G. T.; Iseman, J. C.; Shah, R. N.;
Shimizu, K. D. Anal. Chem. 2004, 76, 1123-1133.
Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 5017

embedded selectors in P1, featuring an efficient, but restricted-
access type binding motif for DCB-amino acids. Generated in the
presence of the tight DCB-(S)-Leu-selector complex, the compact
dimensions of this binding cavity may allow access to the template
and smaller analytes, but accommodation of larger analytes with
difficulty only. For these reasons, the spatially well-defined
environment may impair efficient analyte-selector interactions
with the more demanding size of DCB-Phe.
CONCLUSIONS
Proof of principle for the possibility to preserve and enhance
the chiral recognition capacity of chiral selectors on incorporation
into porous functionalized bulk polymers by analyte templating
has been provided. Copolymerization of a poylmerizable metacryl-
amide derivative of the well-established quinine tert-butylcarbam-
ate-type selector in the presence of a tightly binding N-(S)-
dichlorobenzoylleucine as template gave macroporous bulk poly-
mers, which showed enhanced enantioselectivity relative to a
silica-grafted reference CSP under chromatographic conditions.
High affinity of the template for the monomer selector was found
to be a crucial requirement to successfully shape superior polymer-
embedded selector sites. Analysis of the corresponding adsorption
behavior suggests that the observed improvements in chiral
recognition arise from an increase in population of template-
specific binding sites, in combination with a decrease in binding
affinity for the nontemplating enantiomer. Polymer-embedded
templated selector sites were also shown to exhibit a moderate
enhancement in substrate specificity relative to the silica-grafted
congeners.
Although the observed effects are still in a modest range, the
results are of fundamental interest considering the broad reper-
toire of possible improvements. Thus, replacing thermal by low-
temperature polymerization protocols, deployment of higher
affinity templates, and using dedicated copolymer formulations
and interaction-enhancing porogenic solvents are just a few aspects
that may allow for optimization of the performance characteristics
of templated porous polymers. It should be emphasized that
analyte templating is not restricted to the production of chiral
recognition materials, but is of broad applicability for many other
fields challenged with the task to preserve the delicate functional-
ity of high-affinity binders on polymer incorporation.
Figure 4. Relative changes of differential free energies ∆(∆∆G)
for DCB-Leu and DCB-Phe on P1-P4 (∆(∆∆G) ) -RT ln(RCSP) -
RT ln(RP)). Chromatographic conditions: see Table 3.
(S)-Phe-templated selectors in P1 are confined in a spatially well-
defined environment and express some level of size discriminatory
effects for analytes large than the templating molecule.
Size-discriminating effects can be visualized more clearly by
expressing them as the relative changes of differential free energy
of enantioselective binding (∆(∆∆G)) of the polymers relative to
that of the silica-grafted CSP (see Figure 4). These ∆(∆∆G) values
give an energy measure for gains (negative values) and losses
(positive values) in chiral recognition achieved on transferring
selector units from the silica surface into a polymeric environment,
with or without analyte templating.
The ∆(∆∆G) values for DCB-Leu underline the crucial role of
the templating analyte for shaping polymer-embedded selectors
with favorable chiral performance characteristics. Only match
analyte-templated polymer P1 expresses enhanced chiral recogni-
tion ∆(∆∆G) ) -0.12 kJ/mol) relative to the silica-grafted CSP.
The use of a mismatch template (P2) severely deteriorates the
chiral recognition (∆(∆∆G) ) +0.85 kJ/mol); most probably by
enforcing and stabilizing unfavorable selector orientations and
conformations. A template with poor selector complementarity (P3,
∆(∆∆G) ) +0.58 kJ/mol)) may fail to produce a significant level
of templating effect, exhibiting a behavior similar to nontemplated
polymers (P4, ∆(∆∆G) ) +0.55 kJ/mol).
ACKNOWLEDGMENT
This project was funded by grants of the Austrian Research
Fund (FWF), Projects P-14179-CHE and P-16300-B1, and of the
European Community under the Industrial & Materials Technolo-
gies Program (Brite Euram III), Project BE 96-3159. We are
grateful to Roland Mu¨ller and Junyan Wu for technical assistance.
We are indebted to Dr. Dieter Lubda (Merck KGaA, Darmstadt,
Germany) for providing porosity data, to Prof. Michael La¨mmer-
hofer for a critical review of the manuscript, and to Dr. Kevin
Schug for linguistic advice. This work is dedicated to Prof. Volker
Schurig on the occasion of his 65th birthday.
A comparison of the DCB-Leu/DCP-Phe ∆(∆∆G) values
provides some insight into the substrate-specificity enhancing
effects of analyte templating. The ∆(∆∆G) values for P1-P4 are
all larger for DCB-Phe than those observed for DCB-Leu, indicat-
ing that the polymeric environment compromises the chiral
recognition capacity of the selector for this particular analyte. The
relative magnitudes of these analyte-specific losses in chiral
recognition, however, vary for the individual polymers. Polymers
2-4 show small and rather uniform losses in the range of +0.18
kJ/mol, most probably a reflection of nonspecific binding effects.
For polymer P1, however, a more pronounced decrease of +0.47
kJ/mol is observed. A possible explanation for this marked drop
in enantioselectivity may be the enhanced quality of polymer-
Received for review March 8, 2005. Accepted May 11,
2005.
AC050407S
5018 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005