Synthesis, spectroscopic, and analyte-responsive behavior of a polymerizable naphthalimide-based carboxylate probe and molecularly imprinted polymers prepared thereof

Article abstract of DOI:10.1021/jo3019522

A naphthalimide-based fluorescent indicator monomer 1 for the integration into chromo- and fluorogenic molecularly imprinted polymers (MIPs) was synthesized and characterized. The monomer was equipped with a urea binding site to respond to carboxylate-containing guests with absorption and fluorescence changes, namely a bathochromic shift in absorption and fluorescence quenching. Detailed spectroscopic analyses of the title compound and various models revealed the signaling mechanism. Titration studies employing benzoate and Z-l-phenylalanine (Z-l-Phe) suggest that indicator monomers such as the title compound undergo a mixture of deprotonation and complex formation in the presence of benzoate but yield hydrogen-bonded complexes, which are desirable for the molecular imprinting process, with weakly basic guests like Z-l-Phe. Compound 1 could be successfully employed in the synthesis of monolithic and thin-film MIPs against Z-l-Phe, Z-l-glutamic acid, and penicillin G. Chromatographic assessment of the selectivity features of the monoliths revealed enantioselective discrimination and clear imprinting effects. Immobilized on glass coverslips, the thin-film MIPs of 1 displayed a clear signaling behavior with a pronounced enantioselective fluorescence quenching dependence and a promising discrimination against cross-analytes.

Full text of DOI:10.1021/jo3019522

Article
pubs.acs.org/joc
Synthesis, Spectroscopic, and Analyte-Responsive Behavior of a
Polymerizable Naphthalimide-Based Carboxylate Probe and
Molecularly Imprinted Polymers Prepared Thereof
Ricarda Wagner,† Wei Wan,‡ Mustafa Biyikal,‡ Elena Benito-Pena,§ María Cruz Moreno-Bondi,§
̃
Issam Lazraq,†,# Knut Rurack,*,‡ and Borje Sellergren*,†,⊥
̈
†Institute of Environmental Research (INFU), Faculty of Chemistry, Technical University of Dortmund, Otto-Hahn Strasse 6,
D-44221 Dortmund, Germany
‡Division 1.9 Sensor Materials, BAM Federal Institute for Materials Research and Testing, Richard-Willstatter Strasse 11, D-12489
̈
Berlin, Germany
§Chemical Optosensors & Applied Photochemistry Group (GSOLFA), Department of Analytical Chemistry, Faculty of Chemistry,
Universidad Complutense de Madrid, E-28040 Madrid, Spain
⊥Department of Biomedical Sciences, Malmo University, SE-205 06 Malmo, Sweden
̈
̈
S
* Supporting Information
ABSTRACT: A naphthalimide-based fluorescent indicator
monomer 1 for the integration into chromo- and fluorogenic
molecularly imprinted polymers (MIPs) was synthesized and
characterized. The monomer was equipped with a urea binding
site to respond to carboxylate-containing guests with
absorption and fluorescence changes, namely a bathochromic
shift in absorption and fluorescence quenching. Detailed
spectroscopic analyses of the title compound and various
models revealed the signaling mechanism. Titration studies
employing benzoate and Z-L-phenylalanine (Z-L-Phe) suggest
that indicator monomers such as the title compound undergo a
mixture of deprotonation and complex formation in the
presence of benzoate but yield hydrogen-bonded complexes, which are desirable for the molecular imprinting process, with
weakly basic guests like Z-L-Phe. Compound 1 could be successfully employed in the synthesis of monolithic and thin-film MIPs
against Z-L-Phe, Z-L-glutamic acid, and penicillin G. Chromatographic assessment of the selectivity features of the monoliths
revealed enantioselective discrimination and clear imprinting effects. Immobilized on glass coverslips, the thin-film MIPs of 1
displayed a clear signaling behavior with a pronounced enantioselective fluorescence quenching dependence and a promising
discrimination against cross-analytes.
directly into the MIP. These molecular indicators should
undergo a change in either absorption and/or preferably
fluorescence upon rebinding of the analyte in the MIP.
Although several studies on using conventional MIPs for the
detection of fluorescently tagged analytes,5,6 which again is at
least a two-step process, on employing MIPs with simple
fluorophores embedded for the detection of quenchers7,8 or on
competitive assays involving fluorescently labeled imprints9,10
have been published in the past, few examples of the integration
of a polymerizable chromo- and/or fluorogenic probe have
been reported until now. In particular, examples that show
directional recognition at a designated binding site11,12 or
responses in the analytically advantageous visible spectral range
are scarce.12−14
INTRODUCTION
■
Analyte-responsive polymers1,2 in general and molecularly
imprinted polymers (MIPs)3,4 in particular are very promising
materials in many areas of analytical chemistry. Particularly for
purpose-fit sensing systems such as portable diagnostic devices,
wearable sensors, or the chemically active components in
microfluidic assemblies, tunable molecular recognition abilities
and the format diversity of MIPs are very interesting and allow
for tailored integration. However, for many such applications, it
would be advantageous for MIPs to take a step from being a
pure separation matrix to serving as an actual detection matrix,
potentially reducing the detection protocol to a single step,
continuous and direct. Such advancements require the
integration of a signaling element into the imprinted polymer.
Because many sensor formats as alluded to above operate with
light as the interrogating medium, a promising approach is the
integration of chromo- and/or fluorogenic indicator molecules
Received: September 8, 2012
Published: January 28, 2013
© 2013 American Chemical Society
1377
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
In our present work, we chose the 4-amino-substituted
naphthalimide (NI) chromophore as the scaffold because such
dyes usually emit strong fluorescence in the analytically
advantageous region >500 nm and can be excited in the near
visible at ca. 420−450 nm,15,16 a range for which low-cost
excitation sources are readily available. In addition, the NI unit
is a popular building block in fluorescent probes for various
analytes,17,18 can be integrated into polymers,19,20 and has also
been successfully employed in anion-responsive probes.21,22 On
the basis that many of the key targets as templates/analytes in
our MIP research carry at least one carboxylic acid group,23,24
the latter features are very interesting and motivated us to
pursue the synthesis of the polymerizable NI-based anion probe
1 (Chart 1). Here, we report on the synthesis, spectroscopic,
reproduce the syntheses of 3 published in refs 25 and 26
failed, we modified this step of the reaction sequence. The
successful synthesis of 1−3 was proven by NMR and MS
spectroscopic methods (Figure 1A). As a polymerizable probe
monomer, 1 is composed of three functional units. The N-
ethyl-4-amino-1,8-naphthalimide fragment presents the actual
chromophoric entity. The directly electronically fused urea
group is responsible for the binding of the analyte. The direct
conjugation between both units should guarantee an efficient
transduction of the binding event into a spectroscopic signal.
The styrene moiety, again fused directly to the urea unit, has
two roles. On one hand, it contains the reactive linker for the
copolymerization into the MIP matrix. On the other hand, aryl
substitution at the urea commonly increases the receptor’s
affinity toward anionic guests because the attachment of an aryl
moiety leads to a higher acidity of the respective NH group.27
To assess whether the probe under consideration is suitable
for the molecular imprinting process and whether it supposedly
performs in the desired analytical fashion, the following
requirements should be met. For the first task it is important
that the complex between fluorescent monomer and template is
stable under conventional imprinting conditions. This is
commonly the case when the complex stability constant lies
on the order of K = 1000 M−1 or higher in polar organic
solvents such as acetonitrile or DMSO. The second require-
ment concerns the spectroscopic behavior of the probe
monomer as such and in the presence of the analyte in the
respective solvents mentioned previously.
Chart 1. Chemical Structure of Title Monomer 1
and analyte-responsive behavior of 1 and its integration into
monolithic and thin-film MIPs for a first assessment of the
separation and sensing performance of such intrinsically
fluorogenic molecularly imprinted polymers.
Spectroscopic Properties of 1. Representative absorption
and fluorescence spectra of 1 are shown in Figure 1B. It is
apparent that the typical features of 4-aminonapththalimide
chromophores are retained in 1, i.e., a broad and nonstructured
absorption band with a maximum at ca. 400 nm and a strongly
Stokes shifted, broad emission band at ca. 500 nm (for detailed
data, see Table 1), in contrast to the structured and blue-shifted
bands of 1,8-naphthalimides lacking an amino group in the 4-
position such as 4 and 5 with λabs ca. 345 and λem ca. 375
nm28,29 (for chemical structures, see Chart 2). Moreover,
detailed spectroscopic studies of 1 as a function of solvent
polarity revealed a positive solvatochromism, i.e., absorption
and emission bands gradually shift to longer wavelength with
increasing polarity of the solvent (Table 1). In accordance with
established NI photophysics,15 these properties hint at an
intramolecular charge transfer (ICT) process being active in 1.
For comparison, 6 (Chart 2) shows absorption and
fluorescence maxima of 413 and 513 nm in acetonitrile and
420 and 500 nm in THF.15 Alkyl substitution shifts these values
slightly to 415 and 528 nm (MeCN) and 408 and 508 nm
(THF) for 7 (Chart 2) due to the higher donor strength of the
dimethylamino compared with the amino group.15 The
attached urea-styryl moiety thus has only a minor influence
on the ICT process in the NI chromophore. (Note that the
different nature of the alkyl substituents at the imide nitrogen,
n-butyl in ref 15 and ethyl in our case, has virtually no influence
on the photophysics.29)
RESULTS AND DISCUSSION
■
The synthesis of the urea-based monomer 1 follows the path
illustrated in Scheme 1. Starting from commercially available 4-
bromo-1,8-naphthalic anhydride, 1 is obtained in three steps in
an overall acceptable yield. Because initial attempts to
Scheme 1. Schematic Representation of the Synthesis of 1
With regard to fluorescence quantum yields and lifetimes, the
substitution pattern at the 4-amino nitrogen atom, however, is a
more decisive factor. Going back to 4 and 5 with electronically
rather weakly interacting substituents at the 4-position, these
NI dyes display largely quenched fluorescence because of
efficient coupling of energetically close-lying nπ* states.28,29 In
contrast, when a primary or secondary amino group is in the 4-
position like in 6 and 8, high fluorescence quantum yields
1378
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
Figure 1. (A) NMR spectra of 1 (bottom) and 3 (top) in DMSO. (B) Absorption and fluorescence spectra of 1 in acetonitrile at 298 K.
Saha and Samanta attributed this behavior to the formation of a
highly emissive ICT state in 6 and 8 which is quenched in 7
because of accelerated inversion at the nitrogen center in the 4-
position and/or the influence of a close-lying dark state.15 If we
now consider 3, the fluorescence is also quenched although the
molecule should be more reminiscent of 6. Since 3 is planar
and entirely π-conjugated, photoinduced electron-transfer-type
quenching can be ruled out as an active process. Moreover,
because the spectroscopically important transitions are
distinctly lower in energy as for 4 and 5, nπ* state interaction
is also unlikely. However, if we also take into account the
behavior of other NI derivatives with highly electron-rich
nitrogen-containing substituents at the 4-position, i.e., virtually
no fluorescence in the case of 930 and a moderate fluorescence
of Φf = 0.21 in MeCN in the case of 10 (for chemical
structures, see Chart 2), which are all π-conjugated and not
particularly sterically hindered, the quenching process should
be related to intrinsic electronic properties of the molecules.
Indeed, quantum chemical calculations performed on 1, 3, and
6−10 revealed that the order of emissivity seen in these dyes (9
< 3 ∼ 1 ∼ 7 < 10 < 8 < 6) basically corresponds to the singlet−
triplet energy gap for the lowest and second lowest excited
states and hence the probability of accelerated nonradiative
deactivation via intersystem crossing.31 (Note: Recently, Φf =
0.65 and 0.51 in acetonitrile, which are almost 2 orders of
magnitude higher than Φf of 3 determined by us here, have
been reported for 11 and 12 (Chart 2), i.e., for closely related
analogues of 3 varying only in the spectroscopically silent alkyl
substituent at the imide nitrogen atom.26 Because we were not
able to synthesize 3 under the conditions employed by those
authors and all our other experimental and theoretical
considerations do not support such high fluorescence for this
type of molecule, we tentatively assume that the earlier
reported values are erroneous.)
Table 1. Spectroscopic Properties of 1 and 3 in Selected
Solvents
compd
solvent
λabs /nm
λem /nm
Φf
τf /ns
1
1
1
1
1
3
DMSO
EtOH
MeCN
THF
420
412
404
402
392
425
542
535
517
512
500
516
0.018
0.013
0.015
0.020
0.040
0.008
0.11
0.07
0.12
0.13
0.46
0.01
dioxane
MeCN
Chart 2. Chemical Structures of Model Compounds
Discussed in the Text (Me = Methyl, Et = Ethyl, nPr = n-
Propyl, nBu = n-Butyl)
Spectrophotometric Response of 1 toward Carbox-
ylates. Having established the underlying photophysical
processes in 1, we investigated the response of the functional
monomer toward oxyanionic guest molecules by different types
of experiments and employing tetrabutylammonium benzoate
(TBAB) and Z-L-phenylalanine (Z-L-Phe) as model templates.
To assess the ground-state interaction of 1 and benzoate (Bz−),
spectrophotometric titrations were performed in THF, MeCN,
and DMSO. Figure 2 shows the color changes in DMSO visible
to the naked eye, with the color developing from bright yellow
through dark yellow and green to brownish-orange upon
addition of increasing amounts of TBAB up to 10 equiv.
Upon closer inspection of the guest-induced spectroscopic
changes by spectrophotometric measurements as shown in
emerge (e.g., ≥0.6 in ethanol15,29), yet with a tertiary amine like
in 7, Φf is again considerably lower (e.g., 0.01 in ethanol15).
1379
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
deprotonate 1 yet the lack of the solvent to interact with the
urea group might lead to the formation of two different species.
One possibility is an equilibrium between the intramolecularly
stabilized and the nonstabilized species in Scheme 2, route C.
(Note that according to quantum chemical calculations the two
anionic structures on the product side of route C are the most
stable structures of the various anion isomers/tautomers that
are theoretically possible because of the three NH groups and
respective tautomeric equilibria in the hydrazinecarboxamide
system.31) The other possibility is that at higher excess, Bz−
coordinates to the urea group in the designated fashion
(Scheme 2, route A).
Further insight into the possible mechanisms at play can be
obtained from an analysis of the association constants K
extracted from the different titrations. These data, derived by
nonlinear regression fitting, are collected in Table 2. As would
be expected, the association constant strongly depends on the
solvent used. The affinity of the host monomer toward the
oxyanionic guest is most pronounced in the least competitive
solvent, namely THF. However, it is also surprising that
analysis of the titration data at different spectral positions yields
different results in the case of MeCN and especially THF. This
discrepancy, which is reflected by the fact that the isosbestic
points in MeCN and particularly THF are not as sharp as in
DMSO, suggests that more than two species are involved in the
processes happening during the titration. Accordingly, we
subjected the titration spectra to a nonlinear fitting process to
extract the equilibrium constants for deprotonation and/or
hydrogen bonding complex formation. In the case of DMSO, a
1:1 de- or reprotonation model with the global equilibrium MH
+ B− ⇌ M− + BH (with MH equaling 1, M− = 1−, B− = Bz−;
the TBA counterion was not taken into account) yielded an
acceptable fit and a log K = 3.93 0.01.34 This value is in good
Figure 2. Visible color changes induced by Bz− addition to a solution
of 1 in DMSO.
Figure 3, it is apparent that the addition of the anionic guest
leads to a pronounced decrease of the typical NI absorption
band at 420 nm in DMSO with a concomitant development of
a new broad band with a global maximum at ca. 570 nm (Figure
3A). Similar features though less pronounced are found in
MeCN (Figure 3B). On the other hand, the behavior in THF is
different. Again, a strongly red-shifted broad band appears as a
shoulder at ca. 530−570 nm. However, in this solvent the main
ICT absorption band of 1 is shifted from 402 to 434 nm
(Figure 3C). Dramatic bathochromic shifts of >100 nm as seen
here point to a significant electronic modulation of an ICT
chromophore which is usually not obtained by simple
coordination of an anionic species through two hydrogen
bonds, the anticipated binding mode between ureas and
carboxylates, especially since the urea moiety is not directly
fused to the NI core (Scheme 2, route A).32 The behavior in
DMSO thus suggests that the secondary 4-amino group in 1 is
rather acidic and is gradually deprotonated with no apparent
hints of the formation of a hydrogen bonded complex with
benzoate, presumably also because the solvent can successfully
compete (Scheme 2, route B). Similar deprotonation-induced
red-shifts have also been found for other NI derivatives
substituted with extended functional groups at a secondary 4-
amino nitrogen.33 In contrast, in less strongly coordinating
solvents such as MeCN or THF, TBAB is still basic enough to
Figure 3. Absorption titration spectra of 1 and TBAB in DMSO (A), acetonitrile (B), and THF (C) (starting point spectra in blue, end point spectra
in red) and representative titration curve (fit as dashed line) obtained from a plot of A435 vs cTBAB in THF (D); c1 = 7 μM, cTBAB = 1.5−600 μM
(different step sizes).
1380
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
Scheme 2. Possible Hydrogen Bonding and Deprotonation Pathways Involving 1, Benzoate, and Specific Solvents
Table 2. Apparent Association Constants K [M−1] for the Interaction between 1 and TBAB in DMSO, MeCN, and THF As
Derived from the Spectrophotometric Titration Data in the Range of 1−600 μM TBAB (c1 = 7 μM) Fitted to a 1:1 Interaction
Model at Different Wavelengths; the Latter in (nm) Are Given in Parentheses (r Was Always >0.99)
DMSO
MeCN
THF
1.1 0.1 × 105 (400)
1.5 0.1 × 106 (403)
5.6 0.2 × 105 (435)
3.0 0.1 × 106 (565)
1.1 0.1 × 104 (430)
1.0 0.1 × 104 (560)
1.6 0.3 × 105 (555)
Figure 4. Fitting results for the titrations in THF and DMSO shown in Figure 3. Species spectra (A) and relative concentrations during the course of
the titration (B) of 1 (black), 1− (red), and 1⊂Bz− (blue) are shown; solvents are indicated.
agreement with the data in Table 2 and the sharp isosbestic
points in Figure 3; i.e., they reveal a stoichiometric equilibrium.
The corresponding species spectra are shown in Figure 4A, and
the development of the relative concentration of species 1
(black) and 1− (red) is shown in Figure 4B. In the case of THF,
a 1:1 model does not lead to an acceptable fit. However, when
employing a second global equilibrium MH + B− ⇌ MHB−
(with MHB− equaling 1⊂Bz−)34 the fit converges, yielding a
log K = 4.79 0.02 for the formation of the hydrogen bonded
complex. When analyzing the species spectra of 1 (black, Figure
4A), 1− (red), and 1⊂Bz− (blue) it is apparent that the nature
of the blue species has to lie in between those of 1 and 1− with
regard to electron donor properties, i.e., an increased electron
richness at the hydrazinecarboxamide moiety should lead to the
bathochromic shift to 435 nm.35 Thus, a complex between 1
and a carboxylate anion is assumed to be responsible for this
species. Turning to acetonitrile, a detailed analysis also supports
the observations made in Figure 3; i.e., three species 1, 1−, and
1⊂Bz− (with a maximum at 411 nm) are involved. As would be
expected, complex formation is weaker in MeCN than in THF
(log K = 4.26 0.02).
To get better access to this third species, the desirable one
from the point of view of the recognition reaction, it is
necessary to suppress deprotonation while maintaining hydro-
1381
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
Figure 5. (A) Spectrophotometric titration spectra of 1 (c1 = 7 μM) upon addition of increasing amounts of a 1:1 mixture of Z-L-Phe and PMP (1−
3300 equiv) in MeCN; starting and end point fluorescence spectra are also included. (B) Species spectra of the fitting results (1:1 model) for the
corresponding spectrophotometric titration. Spectroscopic changes as a function of addition sequence: 1 followed by an excess Z-L-Phe and the same
excess PMP (C) and 1 followed by an excess PMP and the same excess Z-L-Phe (D); excitation in panel (A) at isosbestic point (404 nm).
gen-bond complex formation. We thus turned to a weaker base,
Z-L-phenylalanine (Z-L-Phe). Since a carboxylic acid group does
not form a complex with a urea binding site, Z-L-Phe has to be
turned into the carboxylate in situ. For this purpose, we
employed 1,2,2,6,6-pentamethylpiperidine (PMP) in stoichio-
metric amounts with respect to Z-L-Phe and titrated 1 with this
1:1 mixture of Z-L-Phe and PMP in MeCN (Figure 5A).
Apparently, under these conditions and with this analyte, no
deprotonation occurs as judged from the virtual absence of the
broad band at 570 nm. In addition, the titration shows clear
isosbestic points and the development of a new, red-shifted
absorption band, which is tentatively ascribed to the hydrogen-
bonded complex 1⊂Z-L-Phe−. Fitting of the titration spectra to
a 1:1 binding model reveals two species with maxima of 400
and 412 nm (Figure 5B), the first belonging to the component
that is consumed during the titration and the second to the
species that is formed (see insert, Figure 5B). We then carried
out another set of experiments to confirm these findings. When
first an excess of Z-L-Phe is added to 1 in MeCN, the
absorption spectrum is virtually unchanged (cf. black and red
spectra in Figure 5C). However, when an equal excess of PMP
is added, the moderate bathochromic shift is observed (cf. red
and blue spectra in Figure 5C). In contrast, when first PMP is
added to 1, a strongly red-shifted band reminiscent of
deprotonation is seen (cf. black and red spectra in Figure
5D). Further addition of an excess of Z-L-Phe then reverses the
shift and yields the moderately red-shifted band (cf. red and
blue spectra in Figure 5D), presumably because of reprotona-
tion of 1 by Z-L-Phe accompanied by complex formation. The
complex 1⊂Z-L-Phe− thus absorbs at 412 nm in MeCN.
These findings strongly support our assignments made in the
TBAB case, i.e., that the band at 435 nm (blue species in Figure
4) belongs to 1⊂Bz−. The discrepancy between the absolute
maxima of the two bands found for 1⊂Bz− and 1⊂Z-L-Phe−, ca.
20 nm, is presumably due to the different electron-donating
properties of benzoate and the anion of Z-L-Phe. In addition,
the counterion also plays an important role with PMPH+ being
able to donate a further directional hydrogen bond but TBA+
being unable to do so, most likely facilitating deprotonation.
Plotting and analyzing titration curves of 1 and Z-L-Phe/PMP
in MeCN yields log K = 3.08 0.01. Although this value is
rather moderate, the combination of 1/Z-L-Phe/PMP should
be suitable for molecular imprinting in polar organic solvents
such as acetonitrile.
Fluorescence and NMR Response of 1 toward
Carboxylates. Because of their large Stokes shift and
considerable brightness, naphthalimide-based probes are well
suited for fluorescence assays. If we recall the above-mentioned
considerations on the active photophysical processes in such NI
dyes and the results shown in Table 1 that the fluorescence
quantum yield is reduced with increasing polarity of the solvent,
i.e., with an acceleration of the ICT process, the addition of
excess electron density to 1, either through deprotonation or
anion binding, is expected to lead to a reduction of the
fluorescence of 1 as the exploitable signaling mechanism.
Indeed, when the samples shown in Figure 3 are excited into
the long-wavelength band at ca. 570 nm, no fluorescence can be
recorded. 1− is thus virtually nonemissive. In a similar way, the
fluorescence spectra shown in Figure 5A suggest that also 1⊂Z-
L-Phe− is weaker emissive than 1. The absence of pronounced
spectral shifts in emission between 1 and 1⊂Z-L-Phe− further
suggests that upon excitation the NI-based CT process takes
place in the complex, yet that it is less efficient, leading to the
reduced fluorescence. The fluorogenic indication mode there-
fore follows an ON−OFF behavior.
Having established the interaction between 1 and anionic
1
templates by optical spectroscopic means, H NMR experi-
ments were conducted in DMSO-d6 to get further insight into
the mechanistic features in the case of benzoate. For the NMR
titration, an increasing amount of TBAB was thus added at
ratios of 0, 0.25, 0.5, 0.75, 1, 2, 6, 8, and 10 equiv to a constant
amount of host monomer. As can be seen from Figure 6, the
1382
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
Figure 6. NMR spectra of 1 upon addition of 0, 0.25, and 6 equiv of
TBAB in DMSO-d6.
Figure 7. Colors of polymers synthesized from 1 without template
(NIP) and in the presence of Z-L-Phe, Z-L-Glu, and PenG prior to and
after Soxhlet extraction.
hydrazide proton as well as the urea protons show no
complexation induced shift upon addition of increasing
amounts of the guest but severe peak broadening accompanied
by a decrease in intensity. Apparently, the NMR data support
our assumption derived from the UV/vis titrations that the
dominant mode of interaction between the fluorescent
monomer and benzoate is deprotonation.
Molecularly Imprinted Polymers (MIPs) Prepared
from 1. The most common application of MIPs nowadays is
still (chromatographic) separation.36 To be suitable for such
purposes, bulk polymers are usually prepared and are
subsequently grounded into small (micrometer-sized) particles
which are packed into columns. On the other hand, for sensory
applications especially with optical interrogation, MIPs are best
used in thin-film or (de novo-synthesized) nanoparticle
formats.37,38 To assess the performance of 1 in both formats,
we synthesized intrinsically fluorescent monolithic and thin-film
MIPs containing 1 and investigated their analyte recognition
and discrimination features.
changes shown in Figure 2, the much darker colors of MPhe and
MGlu point to a significant degree of deprotonation; i.e.,
although used in stoichiometric amounts, Z-L-Glu and Z-L-Phe
apparently generated significant amounts of 1− sites in the
monoliths. The orange color after Soxhlet extraction then
indicates that all the templates were successfully removed from
the polymers and 1 is retained in its neutral form.
Chromatographic Evaluation of Monoliths. After
appropriate pretreatment as detailed in the Experimental
Section, the crushed polymers (particle size of 25−50 μm)
were chromatographically investigated by HPLC analysis
employing mobile phases with increasing water content and
in the presence and absence of a base modifier. To determine
the success of the imprinting step and the cross-selectivity of
the polymers, solutions of the respective template and
structurally related cross-analytes were injected resulting in
the retention results shown in Table 3.
Monolithic Polymers via Bulk Polymerization. Bulk
polymers were prepared following the noncovalent stoichio-
metric approach39 using a 1:1 ratio between monomer 1 and
the respective template. The polymers were formed by free-
radical solution polymerization via thermal instead of UV
initiation to avoid photobleaching of the dye monomer in the
presence of an excess of the cross-linker EDMA (ethylene
glycol dimethacrylate). Taking into account the insight gained
from the molecular solution studies of 1, three weakly basic
molecules were chosen to serve as templates in the bulk
synthesis, namely Z-L-glutamic acid (Z-L-Glu), Z-L-Phe, and
penicillin G (PenG). Whereas the amino acids were
deprotonated in situ employing PMP, PenG was used as its
procain salt. Further details are given in the Experimental
Section.
Results for MGlu Columns. Showing a retention time of 5.38
min for the L-enantiomer of Z-Glu and 4.37 min for the D-
enantiomer, MGlu is clearly able to discriminate between its
template and the D-enantiomer as well as Z-L-Phe in MeCN/
H2O (50:50); for retention factors k, see Table 3. In contrast,
MNIP shows no retention of both enantiomers. Also in pure
MeCN, the retention of the template is stronger for the MIP
than for the NIP with a chromatographic imprinting factor IFC
= kZLGlu(MGlu)/kZLGlu(MNIP) of 3.32. However, the MIP is not
able to discriminate between its template and the cross-analytes
Z-D-Glu and Z-L-Phe in pure MeCN, although a clearly
enhanced retentivity of Z-Glu on the MIP compared to the NIP
was observed. The separation factor for Z-L-Glu, aGlu
=
k
ZLGlu(MGlu)/kZDGlu(MGlu)), equals 1.14. Upon base modifica-
tion of MeCN with 1% triethylamine (TEA), the retention
times are shortened and a worse imprinting factor IFC = 1.11 is
obtained. Based on the results discussed above for molecular 1,
we assume that the proneness of the secondary amine function
of 1 to deprotonation is enhanced under basic conditions,
rendering the designated hydrogen bond-mediated discrim-
ination process less efficient. This contrasts with the “expected”
enhanced retentivity observed in presence of basic modifiers for
Z-Glu-imprinted polymers prepared using a less acidic urea
monomer.40 In the mobile phase with a high water content of
90% no retention was observed.
Figure 7 shows that monolithic MIPs could successfully be
prepared using 1 as the urea-expressing recognition moiety.
Moreover, a comparison of the color of the polymers prepared
in the absence (monolithic NIP, termed MNIP in the following)
and in the presence of a template (monolithic MIPs, termed
M
Phe, MGlu and MPen) reveals distinct color differences
depending on the template employed. Whereas both MNIP
and MPen show a bright orange color prior to and after Soxhlet
extraction, the polymer imprinted with Z-L-Phe is brownish and
the Z-L-Glu polymer is green. The orange color of MNIP
signifies the reference point of 1 incorporated into an organic
polymer matrix. Hence, the only slightly deeper color of MPen
suggests that this template is primarily incorporated as a
hydrogen-bonded complex with 1. Reconsidering the color
Results for MPhe Columns. Different than MGlu, MPhe shows
no enantioselective discrimination in a water-containing mobile
phase with aPhe = 1.06 and 0.99 in MeCN/H2O (10:90) and
(50:50) (Table 3). The imprinting factor is small in MeCN/
1383
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
Table 3. Chromatographic Performance of MGlu, MPhe, and MPen in Selected Solvents upon Injection of the Analytes
Investigated Here
a
b
c
d
mPh
pol
kZLGlu
kZDGlu
aGlu
kZLPhe
kZDPhe
aPheu
kPenG
kBA
MeCN/H2O 10:90
MeCN/H2O 50:50
MeCN
MGlu
MPhe
MPen
MGlu
MPhe
MPen
MGlu
MPhe
MPen
MGlu
MPhe
MPen
0.00
0.00
0.00
0.00
1.44
1.36
1.06
100.00
100.00
0.00
0.00
5.23
0.42
4.06
2.33
0.59
1.29
1.14
1.02
1.30
1.00
1.01
0.55
1.92
0.99
14.15
1.00
2.66
0.00
2.25
7.78
MeCN (1% TEA)
0.60
0.00
0.27
1.92
0.00
0.00
a
b
c
Mobile phase. Polymer. Retention factor k = (tanalyte−tvoid marker)/tvoid marker with t = retention time [min] and acetone injected as void marker.
d
Separation factor, see text.
H2O (10:90), yet again substantial in MeCN/H2O (50:50)
with IFC = 14.3. Again, the retention time for the template Z-L-
Phe is significantly higher than for the D-enantiomer in MeCN
(aPhe = 14.1), and a high IFC = 9.85 is also obtained for Z-L-
Phe. Like for MGlu, addition of base also leads to a diminution
of the imprinting factor (IFC = 2.02) here, and the
enantioselectivity is lost. It is assumed that the significantly
higher retention of the template in the Z-L-Phe case is due to
enhanced side chain stiffness or aromatic interactions between
the template and the monomer.
Results for MPen Columns. For solubility reasons, PenG was
injected as either its procain or potassium salt. The template is
fully retained using mobile phases with water content, whereas
benzoic acid used as a control is not retained. As in the case of
the Z-Phe-imprinted polymer a strong imprinting effect was
observed in MeCN−water mixtures, whereas in MeCN and the
base-modified MeCN both analytes are eluting fast.
Table 4. Position of Maximum Emission Wavelength of
M
NIP, MGlu, MPhe, and MPen in Selected Solvents
λem /nm
solvent
MNIP
MGlu
MPhe
MPen
dioxane
THF
502
508
510
514
498
500
508
510
500
502
510
514
500
506
508
516
MeCN
EtOH
quenching should be more efficient and as F∞ becomes smaller
(F0 − F∞)/F0 becomes larger; IFQ > 1 are to be expected. For
example, MGlu (vs MNIP) shows an IFQ = 1.4.
MIP
F0 − F∞
F0
(
(
)
IFQ
=
NIP
F0 − F∞
F0
)
(1)
Collectively, the above results demonstrate successful
imprinting for all templates with particularly strong imprinting
effects seen using MeCN as the mobile phase and Z-Phe and
PenG as templates.
Here, F0 = fluorescence intensity in the absence of analyte and
F∞ = fluorescence intensity in the presence of analyte at
saturation (concentration c→∞) of (superscript) MIP or NIP.
No quenching was observed upon addition of only PMP.
Thus, the quenching is not a mere deprotonation phenomenon.
Nevertheless, none of the grounded monoliths showed
discrimination against the cross-analytes employed, for
example, Z-D-Glu, Z-L-Phe and PenG for MGlu. Investigation
of the swelling behavior in five different solvents, MeCN,
MeCN/H2O (10:90), MeCN/H2O (50:50), MeCN/TEA
(99:1), and water, shows that the swelling in water is not
significantly different from that in the other solvent systems to
explain the different quenching behavior. The same behavior
was also observed for MPhe and MPen. At present, the peculiar
spectroscopic responses of the MX MIPs are not entirely clear.
However, it has to be noted that the dye concentration is rather
high in the bulks so that dye−dye interactions such as
reabsorption and self-quenching phenomena, which can differ
in different solvents, cannot be ruled out.
Spectroscopic Studies of Monoliths. As can be seen in
Figure 7, the monoliths as such are very deeply colored and do
not qualify for spectroscopic studies. They are not transparent,
precluding conventional absorption and fluorescence measure-
ments in transmission modes. Moreover, if one would use
reflectance or front-face setups, only the spectroscopic features
of the first nano- or micrometers could be assessed and not the
features of the entire bulk. The monoliths thus had to be used
in grounded form by suspending 10 mg of polymer in a 96-well
plate prior to the analytical measurements in a fluorescence
plate reader. First, the free accessibility of 1 in the polymer
network was confirmed by recording the fluorescence spectra in
solvents of different polarity, revealing also a positive
solvatochromism as found for 1 before (Table 4).
Spectroscopic studies of the polymers in the presence of their
templates revealed that a change in the emission intensity is
only induced in water. To quantify the performance in
fluorescence mode, we calculated a (fluorescence) quenching
imprinting factor IFQ by dividing the reduced fluorescence
signal (F0 − F∞)/F0 obtained for a MIP at the saturation
concentration (plateau at the end of a titration) of a particular
analyte by the corresponding reduced fluorescence signal of the
NIP (eq 1). If the analyte is bound better in the MIP,
Thin-Film Polymers on Glass Coverslips. For optical-
sensing applications, thin transparent polymer films covalently
anchored to glass substrates are an ideal format to handle such
responsive layers in actual devices. On the basis of the results
obtained with the monolithic polymers (cf. Figure 7), it is
essential to have a straightforward and defined optical response,
i.e., to use sufficiently low dye concentrations and to avoid as
1384
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
much any undesired deprotonation reactions during the
polymerization of the sensory MIP. Although the analyte
might be retained in a polymer containing certain degrees of
deprotonated dye monomer, like the behavior of the deeply
colored MGlu and MPhe discussed above suggests, all 1− sites
would not be available for the desired indication reaction
(hydrogen bond complex formation between 1 and carbox-
ylate) in a sensory MIP. The quest was thus to find a solvent
for the MIP preparation that maintains a high degree of
hydrogen-bonded complex between 1 and template, that is
suitable for the synthetic steps as such, and that yields
transparent polymer films. Because the equilibria between
complex formation and deprotonation depend not only on
solvent/porogen but also on species concentration and the
nature of additional reaction partners (comonomer, cross-
linker), it was necessary to spectroscopically study the
prepolymerization mixtures and assess the degree of deproto-
nation of 1 to be actually expected in the reaction mixture.
After various attempts, toluene crystallized as the best
suitable solvent for our purposes. The absorption maximum of
1 dissolved in toluene is centered at 384 nm (Figure 8, red
assess the reusability of the films and the robustness of the
preparation procedure the films were repeatedly extracted and
reused. Furthermore, the obtained results were compared to
the ones of a second batch of films yielding deviations lower
than 5% in all cases.
Spectroscopic investigation of the films in different solvents
revealed that the position of the emission maximum is distinctly
bathochromically shifted with respect to the values obtained for
1 molecularly dissolved in solution (compare Tables 5 and 1).
M/E
NIP
Table 5. Position of Maximum Emission Wavelength of F
,
B/E
FMPh/eE, FBN/IEP, and F in Selected Solvents
Phe
λem /nm
M/E
NIP
M/E
Phe
B/E
NIP
B/E
solvent
F
F
F
F
Phe
dioxane
THF
466
470
480
480
467
471
480
480
465
472
481
481
465
471
480
481
MeCN
EtOH
Apparently, when the dye monomer is integrated into the
polymer full relaxation of the excited state of 1 is precluded by a
confinement in the polymer matrix or by a reduced local
polarity of the microenvironment. Nevertheless, the films
exhibit positive solvatochromism like the free dye monomer in
solution; the sites should thus be well-accessible. Sensing tests
with imprinted Z-L-Phe as analyte in acetonitrile in a custom-
build quartz cuvette (measured with a fluorometer in front-face
mode) revealed a positive response, that is, the fluorescence of
M/E
the F
MIPs was significantly stronger quenched than that of
Phe
the corresponding NIPs, FMNI/PE; terminology: F
= Films
M/E
Phe
prepared with MMA/EDMA (superscript) as comonomer and
cross-linker and Z-L-Phe (subscript) as template. A clear
imprinting effect was observed (Figure 9). However, because
Figure 8. Absorption spectra of prepolymerization mixtures of 1⊂Z-L-
Phe in toluene/MMA/EDMA (blue) and HEMA/EDMA (black);
reference points are 1⊂Z-L-Phe in toluene (green) and 1 in toluene
(red).
curve), still hypsochromically shifted compared with the slightly
more polar 1,4-dioxane (Table 1). The successful complex
formation with Z-L-Phe is then evident from the shift of the
absorption band to 433 nm (Figure 8, green curve).
Furthermore, upon addition of the other reaction partners
MMA and EDMA at their respective concentrations, a shift
from 433 to 423 nm is induced (Figure 8, blue curve),
indicating a slight weakening of the hydrogen bonded complex
by the presence of these compounds of higher polarity.
Nevertheless, the complex should be sufficiently stable for film
formation and, what is even more important, the virtual absence
of the band at ca. 570 nm suggests that deprotonation of 1 is
completely avoided especially in the final reaction mixture. A
HEMA/EDMA mixture on the other hand leads to significant
dissociation of the complex (Figure 8, black curve, maximum at
402 nm) and was discarded here. Thus, thin-film polymers were
prepared on glass coverslips in toluene using TBA Z-L-Phe as
template, MMA as comonomer and EDMA as cross-linker.
Subsequent evaluation was carried out in MeCN. In the present
approach, MMA and EDMA were used at ratios that
correspond to a cross-linking level of 25%, which is high
enough to endow the films with sufficient stability for reuse and
which yields promising sensing results, vide infra. In order to
M/E
Figure 9. Titration curves of F in the presence of Z-L-Phe (squares)
Phe
M/E
and Z-D-Phe (circles) as well as F
in the presence of Z-L-Phe
NIP
(crosses) in acetonitrile; encircled data points at higher concentrations
belong to spectra that show already contributions of deprotonation;
i.e., the quenching recorded for these points is disproportionally large.
the spectra at concentrations above 0.5 mM indicated
deprotonation side reactions, IFQ = 3.4 was determined at c
= 0.5 mM from Figure 9. Since a clear onset of fluorescence
quenching was recorded only for c = 0.1 mM, the dynamic
working range is rather limited (Figure 9). The next step was to
investigate the enantioselectivity of the MMA/EDMA films.
For this purpose, the fluorescence signals observed for Z-L-Phe-
upon addition of Z-L-Phe and Z-D-Phe were
analyzed according to eq 2, arriving at the discrimination factor
M/E
Phe
imprinted F
1385
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
derived from fluorescence quenching q (which corresponds
phenomenologically to a of the HPLC section).
To assess the influence of base addition and the possibility to
further increase the discrimination ability, the studies were
repeated in the base-modified solvent mixture. Again,
fluorescence quenching was observed, arriving at q = 1.2, 1.6,
and 1.3 for Z-L-Phe against Z-D-Phe, Z-L-Glu, and Z-L-Tyr,
which are not as high as in the neat polar solvent, reflecting the
behavior observed in the monoliths above. Most likely
deprotonation plays also a role in this case. Detailed analyses
however were precluded by the low optical density of the thin
films which made reliable quantitative absorption measure-
ments impossible.
BMA/EDMA MIP Films against Penicillin G Potassium
Salt. In a final attempt to assess the first performance criteria of
intrinsically fluorescent MIP thin films, penicillin G potassium
salt was imprinted according to the same recipe as used before
in the films for amino acids. The spectroscopic features were
again evaluated in MeCN, base-modified MeCN and mixtures
with increasing water content. However, different from the
monoliths, fluorescence quenching in the presence of PenG
could only be observed in MeCN, arriving at an IFQ of 2.5
(Figure 11).
LMIP
F0 − F∞
F0
(
(
)
L
q =
LMIP
F0 − F∞
F0
)
(2)
D
Here, F0 = fluorescence intensity of (superscript) L-enantiomer-
imprinted MIP in the absence of analyte and F∞ = fluorescence
intensity in the presence of analyte (L- or D-enantiomer,
subscripts L and D) at saturation (concentration c → ∞). For
M/E
Phe
F
in MeCN, this discrimination factor is obtained to qPhe =
1.7 (again at c = 0.5 mM) for Z-L-Phe vs Z-D-Phe, suggesting
favorable enantioselectivity (Figure 9).
Exchange of MMA for BMA. Taking into account that Z-L-
Phe contains an aromatic moiety, the system was fine-tuned by
substituting the aliphatic comonomer MMA with an aromatic
one, benzyl methacrylate (BMA). We assumed that the affinity
of the aromatic template toward the polymer should be
enhanced by additional ππ interactions. As displayed above for
the MMA/EDMA films the emission bands of the BMA/
EDMA films are also shifted to shorter wavelengths with
respect to molecularly dissolved 1 (Table 5).
The response of the films toward their template and the
cross-selectivity to structurally closely related compounds was
investigated in pure MeCN, mixtures with increasing water
content, and a base-modified system. In MeCN, the MIP films
showed remarkable quenching upon addition of the template,
and a clear imprinting effect of IFQ = 18.6 was obtained (Figure
10). Similar results were found in MeCN/TEA (99:1). To
B/E
Phe
B/E
NIP
Figure 11. Titration curves of F (squares) and F (circles) in the
presence of PenG in acetonitrile; the lines are only guides to the eye.
CONCLUSION
■
A novel polymerizable naphthalimide monomer was designed
and synthesized, incorporating a urea binding site and showing
the favorable spectral features of naphthalimide chromophores
such as broad and strongly Stokes’ shifted absorption and
emission bands in the visible spectrum. Detailed investigations
of the title compounds and several other model NI dyes
allowed us to unravel the quenching mechanisms in N-nitrogen
substituted 4-amino-NIs. Being rather acidic in nature, titrations
of 1 with TBA benzoate in various solvents and a
comprehensive analysis of the titration spectra provided deeper
insight into the dualism between deprotonation and complex
formation. Whereas the fluorescent monomer is not able to
bind rather strong bases like benzoate only in the designated
fashion, the formation of a directionally oriented hydrogen-
bonded complex should yield much better imprinting results
than the nondirectional purely electrostatic association of ionic
species, the experiments employing Z-L-Phe yielded very
promising results in terms of the imprinting of such weakly
basic carboxylates. Accordingly, MIP synthesis was attempted
and title compound 1 could be successfully integrated in
polymers of monolithic and thin-film format. Upon chromato-
graphic evaluation, the bulk polymers showed remarkable
recognition behavior; enantioselective discrimination was
B/E
Figure 10. Titration curves of F in the presence of Z-L-Phe (solid
Phe
squares), Z-D-Phe (circles), Z-L-Glu (crosses), and Z-L-Tyr (triangles)
B/E
NIP
as well as F
in the presence of Z-D-Phe (open squares) in
acetonitrile; lines are only guides to the eye for FBPh/Ee /Z-L-Phe, F /Z-
B/E
Phe
B/E
NIP
D-Phe and F /Z-D-Phe; AA = amino acid.
assess the performance against the other enantiomer and
closely related analytes, the response of F and F toward
B/E
Phe
B/E
NIP
TBA Z-D-Phe, TBA Z-L-Glu, being aliphatic in nature yet
carrying two carboxylate groups, and TBA Z-L-Tyr as an
aromatic competitor was investigated. The enantioselective
discrimination with qPhe = 2.8 was found to be better than in
the case of FMPh/eE. Even more pronounced was the discrimination
obtained against the structural analogues, i.e., qPhe/Glu of 4.7 and
Phe/Tyr of 3.3 were found against Z-L-Glu and Z-L-Tyr. As could
be expected, the difference is more pronounced in the case of
the aliphatic amino acid (Figure 10).
q
1386
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
was cooled to 0 °C, and 100 mL 0.5 M NaOH was added dropwise.
The mixture was stirred 48 h at room temperature. The solid was
removed by filtration and washed with 1 M NaOH solution. The
filtrate was acidified with 10% HCl solution to precipitate the product.
The product was dried and recrystallized from ethanol/water 3:7.
Yield: 89%. 1H NMR (300 MHz, DMSO-d6) δ: 5.37 (d, 1H, J = 9 Hz),
5.94 (d, 1H, J = 12 Hz), 6.80 (dd, 1H, J = 12 Hz, 9 Hz), 7.54 (d, 2H, J
= 6 Hz), 7.89 (d, 2H, J = 6 Hz), 8.39 (s, 1H). 13C NMR (DMSO-d6)
δ: 117.9, 127.1, 130.5, 130.8, 136.7, 142.1, 167.9. MS, m/z: found mass
148.55 g mol−1, calcd mass 148.16 g mol−1. For the subsequent
transformation of 4-vinylbenzoic acid into 1-isocyanato-4-vinyl-
benzene, ethyl chloroformate as the chlorination agent and NaN3
were used in accordance with a protocol published in ref 42.Yield:
obtained. However, spectroscopic studies of the crushed
monoliths (as microparticles) revealed that these polymers
are not suitable for sensing applications, leaving the
spectroscopic features of the materials rather unchanged in
presence of oxyanionic analytes.
The step to thin-film polymers covalently attached to glass
supports composed of monomer 1, MMA or BMA,
respectively, and EDMA then allowed to obtain sensory MIPs
with a rather low cross-linking level of 25% yet with sufficient
stability of the films for reuse in analytical experiments.
Investigation of the spectroscopic and analyte-responsive
behavior of the films revealed that in all cases the imprinting
step was successful with the polymers being able to chirally
discriminate between their respective template and structurally
closely related cross-analytes especially when employing the
aromatic comonomer. Finally summarizing the observations
made during the present studies of the various MIP formats,
i.e., good enantioselective separation but weak sensing
responses for monolithic MIPs which contain a considerable
degree of deprotonated dye monomer (see for instance the
deep colors of some of the monoliths in Figure 7) and a good
1
62%. H NMR (CDCl3) δ: 5.24 (dd, 1H), 5.71 (dd, 1H), 6.65 (dd,
2H), 7.02 (d, 2H), 7.34 (d, 2H); MS, m/z: found mass 145.10 g
mol−1, calcd mass 145.01 g mol−1.
N-Ethyl-4-(N-[4-vinylphenyl]hydrazinecarboxamidyl)-1,8-
naphthalimide (1). The reaction was carried out under inert gas
atmosphere. Compound 3 (150 mg, 0.59 mmol) was dissolved in 12
mL of THF/pyridine (50:50, v/v). 1-Isocyanato-4-vinylbenzene (85
mg, 0.59 mmol) in 12 mL of THF/pyridine (50:50, v/v) was added
dropwise. The solution was stirred at 28 °C for two days under
nitrogen atmosphere, precipitated from n-hexane, filtrated, and dried,
yielding a yellow solid. Yield = 136.88 mg (0.34 mmol, 58%). Mp:
sensing behavior yet lower discrimination ability (e.g., qPhe = 2.8
1
245−247 °C. H NMR (400 MHz, DMSO-d6) δ: 1.15 (t, J = 7.0 Hz,
B/E
Phe
for F in MeCN vs aPhe = 14.15 for MPhe in MeCN) for the
3H), 4.03 (q, J = 7.0 Hz, 2H), 5.08 (d, J = 11.7 Hz, 1H), 5.64 (d, J =
17.6 Hz, 1H), 6.60 (dd, J = 17.7, 10.8 Hz, 1H), 7.01 (d, J = 8.5 Hz,
1H), 7.39 (q, J = 29.6, 8.5 Hz, 4H), 7.73 (m, 1H), 8.33 (d, J = 8.4 Hz,
1H), 8.46 (d, J = 6.7 Hz, 1H), 8.64 (s, 1H), 8.68 (d, J = 8.6 Hz, 1H),
9.04 (s, 1H), 9.55 (s, 1H). 13C NMR (500 MHz, DMSO-d6) δ: 13.3,
34.2, 103.9, 107.3, 112.0, 118.0, 118.4, 121.7, 124.0, 126.6, 128.2,
129.2, 130.4, 130.8, 134.1, 136.2, 139.3, 152.2, 153.1, 162.6, 163.5.
HRMS: calcd for C23H21O3N4 [M + H]+ 401.1608, found 401.1608.
Tetrabutylammonium Benzoate (TBAB). Benzoic acid (272 mg,
2.2 mmol) was dissolved in 10 mL of dry methanol. A 1 M solution of
tetrabutylammonium hydroxide (1 equiv, 2.23 mL) in methanol was
added dropwise. The mixture was stirred for 2 h at room temperature.
After solvent removal, the product was dried over P2O5.
Bulk Polymers. The bulk polymers were prepared according to the
ratios displayed in Table S1 (Supporting Information) and the
following procedure (with the example of MGlu). Z-L-Glu (0.25 mmol)
and PMP (0.5 mmol) were dissolved in anhydrous DMF (5.6 mL).
Subsequently, 1 (0.5 mmol) and EDMA (20 mmol) were added. After
addition of the azo-initiator ABDV (1% w/w total monomers), the
prepolymerization mixture was transferred to a polymerization tube,
purged with dry nitrogen, and polymerized at 45 °C for 24 h. The
obtained polymers were crushed and extracted by Soxhlet for 24 h
with MeOH/HCl (0.5 mM) (80:20 v/v) and 24 h with MeOH. Anal.
Calcd for MGlu: C, 61.00; H, 7.02; N, 0.67. Found: C, 60.24; H, 7.51;
N, 0.70. Anal. Calcd for MPhe: C, 61.00; H, 7.02; N, 0.67. Found: C,
60.57; H, 7.54; N, 0.75. Anal. Calcd for MPen: C, 61.00; H, 7.02; N,
0.67. Found: C, 60.43; H, 7.74; N, 0.76. Anal. Calcd for MNIP: C,
61.00; H, 7.02; N, 0.67. Found: C, 60.30; H, 7.43; N, 0.75. All
polymers show good agreement between measured and calculated
values. Thus, it can be concluded that the polymer synthesis was
successful and the monomers were integrated into the polymers
quantitatively.
thin-film MIPs with intact (protonated) functional urea
monomers it seems obvious that a lot of the cavities formed
in the bulks retain the template without the directional
hydrogen-bonding participation of the designated urea host
monomer. These deprotonated ureas will mainly interact
through nondirectional electrostatic forces, i.e., the ureas
contribute to template binding during the polymerization
process yet the cavity formed might preclude the spectroscopi-
cally desired H bond complex formation during the rebinding
step. These rebound molecules will remain spectroscopically
silent and the desired fluorescence quenching response is
triggered only from that fraction of cavities which can and will
accommodate the template in the desired fashion through H-
bonding. On the other hand, the cavities in the thin-film MIPs
are not only able to discriminate between different analytes, but
also show enantioselective discrimination. Further research in
our laboratories primarily addresses the detailed unraveling of
the specific vs unspecific and designated vs bulk recognition
and the improvement of the 3D rigidity of the thin-film MIPs.
EXPERIMENTAL SECTION
■
N-Ethyl-4-hydrazino-1,8-naphthalimide (3). Compound 2
(120 mg, 0.39 mmol) was dissolved in 30 mL of 1,4-dioxane.
Hydrazine monohydrate (2.5 mL) was added dropwise. The solution
was stirred at 60 °C for 18 h. After being cooled to room temperature,
the reaction mixture was poured into water. The precipitated solid was
collected by filtration, washed with water, and dried, yielding a yellow
solid. Yield = 90 mg (0.35 mmol, 90%). 1H NMR (300 MHz, DMSO-
d6) δ: 1.14 (t, 3H), 4.01 (q, 2H), 4.63 (s, 2H), 7.20 (d, 1H), 7.58 (dd,
1H), 8.24 (d, 1H), 8.37 (d, 1H), 8.56 (dd, 1H), 9.07 (s, 1H). MS, m/
z: found 256.10 g mol−1 [M + H]+, calcd mass 255.10 g mol−1.
1-Isocyanato-4-vinylbenzene. Following a modified procedure
of ref 41 in a first step, p-bromomethylbenzoic acid (11.50 g, 54
mmol) and triphenylphosphine (14.10 g, 54 mmol) were dissolved in
250 mL of acetone and refluxed for 3 h. The warm mixture was
precipitated from 1 L of diethyl ether. The precipitated salt was filtered
off and dried under vacuum. Yield: 97%. 1H NMR (300 MHz, DMSO-
d6) δ: 5.26 (d, 2H, J = 12 Hz), 7.09 (q, 2H), 7.74 (m, 14H), 7.92 (m,
3H), 13.11 (s, 1H). MS, m/z: found 397.14 g mol−1 [M − Br]−, calcd
Thin-Film Polymers on Glass Coverslips. Pretreatment.
Three hundred coverslips were precleaned for 12 h with concd
KOH solution and subsequently carefully rinsed with water. To
activate the surface, the coverslips were left in piranha solution
overnight, rinsed with water, acetone, and toluene, and dried. The
linker 3-methacryloxypropyltrimethoxysilane was immobilized by
stirring the coverslips in a solution of the linker (20 mL) and
triethylamine (7 mL) in toluene (150 mL). Subsequently, the
coverslips were washed with toluene and methyl tert-butyl ether
(MTBE) and dried.
m a s s 3 97 . 1 4
g
m o l − 1
.
N e x t , ( 4 - c a r b o x y p h e n y l ) -
Polymer-Film formation. Compound 1 (0.06 mg, 0.15 μmol),
TBA Z-L-Phe (equimolar to 1), MMA or BMA, respectively (12.86 μL,
0.11 mmol), and ethylene glycol dimethacrylate (EDMA, 8.58 μL, 0.04
methyltriphenylphosphonium bromide (24.6 g, 50 mmol) was
dissolved in 100 mL of formaldehyde in water (37%). The solution
1387
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
mmol) were dissolved in 25 μL of toluene. The prepolymerization
mixtures were flushed with nitrogen, and the azo-initiator ABDV was
added. Subsequently, 2 μL of the solutions was dropped onto a glass
slide and covered with a 3-methacryloxypropyltrimethoxysilane-
modified coverslip. Polymerization was initiated thermally by heating
with a heat-gun three times for 10 s and subsequently in the oven at 70
°C for 1 h. The obtained polymer films were extracted using a Soxhlet
apparatus with MeOH. The NIPs were prepared accordingly, without
addition of the template. Representative SEM images can be found in
the Supporting Information.
h, the height of the polymer was measured again. The swelling factor
was calculated by dividing the height in swollen state by the height in
dry state.
ASSOCIATED CONTENT
■
S
* Supporting Information
13C NMR spectrum of 1, tables with details of polymer
synthesis, SEM images of thin film MIPs. This material is
available free of charge via the Internet at http://pubs.acs.org.
Spectroscopic Studies. Reagents. Compund 10 was a gift from
Dipl.-Chem. S. Ast and Prof. H.-J. Holdt (University of Potsdam).
Instruments and Methods. Absorption spectra and UV/vis
spectrophotometric titrations were recorded on a UV/vis spectropho-
tometer, and fluorescence spectra and titrations were measured with a
spectrofluorometer (molecular dye, thin-film polymers) or a micro-
AUTHOR INFORMATION
■
Corresponding Author
*Tel: +49 30 81041154 (K.R.), +49 231 7554082 (B.S.). Fax:
+49 30 81041157 (K.R.), +49 231 7554084 (B.S.). E-mail:
(K.R.) knut.rurack@bam.de, (B.S.) b.sellergren@infu.tu-
dortmund.de.
plate reader (crushed monoliths, in bottom reading mode) at 298
1
K. Unless otherwise noted, only dilute solutions with an absorbance of
less than 0.1 at the absorption maximum were used for the molecular
studies. The fluorescence quantum yields (Φf) were determined
relative to coumarin 153 in ethanol (Φf = 0.544),43 respectively. The
uncertainties of measurement were determined to 5% (for Φf > 0.2),
10% (for 0.2 > Φf > 0.02), and 20% (for 0.02 > Φf). Fluorescence
lifetimes (τf) were determined by a unique customized laser impulse
fluorometer with picosecond time resolution.44,45 Further details on
the fluorescence decay time measurements can also be found in ref 45.
The uncertainty of measurement amounted to 3 ps. The
fluorescence lifetime profiles were analyzed with a PC using the
software package Global Unlimited V2.2 (Laboratory for Fluorescence
Dynamics, University of Illinois). The goodness of the fit of the single
Present Address
#Dow Europe GmbH, Bachtobelstrasse 3, CH-8810 Horgen,
Switzerland.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the DFG (RU 1622/1-1, SE 777/
15-1), the Innovationsfonds of BAM/BMWi, and MICINN (ref
CTQ2009-14565-C03). We thank Dr. Marco Emgenbroich for
2
decays as judged by reduced χ-squared (χR ), and the autocorrelation
2
synthetic advice, Dipl.-Chem. Sandra Ast and Prof. Hans-Jurgen
̈
function C(j) of the residuals was always below χR < 1.2.
Holdt (University of Potsdam) for provision of 10, Dr. Dietmar
Pfeifer (Div. 1.3, BAM) for NMR, and Peter Gans (Protonic
Software) for support and further refinement of HypSpec.
Data Analysis. Fitting of the titration data to reaction equilibria
was done with HypSpec 1.1.39 from Protonic Software (Leeds).
HPLC Evaluation of Bulk Polymers. The monolithic polymers
were crushed to a particle size between 25 and 50 μm. To remove
smaller particles, this size fraction was repeatedly sedimented (80/20
methanol/Millipore water) and then slurry-packed into stainless-steel
HPLC columns (2.5 × 0.4 cm) using the same solvent mixture as
pushing solvent. Prior to usage the columns were first washed with
methanol to ensure the complete removal of template. Subsequent
analysis of the polymers was performed with a HPLC system with a
diode array-UV detector. Analyte detection was performed at 205 and
260 nm (Z-Glu, Z-Phe) and 220 nm PenG. To evaluate the polymers
ability to discriminate between its template, the other enantiomer and
cross-analytes, solutions of Z-L-Glu, Z-D-Glu, and Z-L-Phe were
injected. The Glu solutions were prepared in the corresponding
mobile phases, concentration of 10 mM, injection volume 10 μL.
Because of the poor solubility of Z-L-Phe in mobile phases with high
water content, the concentration was lowered to 1 mM, injection
volume 100 μL.
Spectroscopic Evaluation of MIPs. Bulks. Twelve × 10 mg of
MNIP and 24 × MGlu were incubated for 4 h in a 96-well plate with
concentrations between 0 and 1 mM solutions of Z-L-Glu/PMP and
Z-D-Glu/PMP in water. Subsequently, emission scans were performed
(excitation wavelength: 377 nm; emission wavelength: 390 −800 nm).
For the cross-linking studies, emission scans of the respective polymers
were performed in DMSO in a 96 well plate in presence of TBA
benzoate. The theoretical amount of monomer present was calculated
from the preparation recipe, Table S2 (Supporting Information). Thin
Films. The coverslips with the extracted films were mounted in a
custom-made quartz cuvette equipped with a Teflon holder. The
fluorescence emission of the films was monitored during titrations with
increasing concentration of selected analytes between 0 and 1 mM
(excitation wavelength: 380 nm, emission wavelength range: 395−550
nm). The obtained results were background corrected.
REFERENCES
■
(1) Chaterji, S.; Kwon, I. K.; Park, K. Prog. Polym. Sci. 2007, 32, 1083.
(2) Reppy, M. A.; Pindzola, B. A. Chem. Commun. 2007, 4317.
(3) Molecularly Imprinted Polymers: Man-Made Mimics of Antibodies
and Their Application in Analytical Chemistry; Sellergren, B., Ed.;
Elsevier: Oxford, 2000.
(4) Moreno-Bondi, M. C.; Navarro-Villoslada, F.; Benito-Pena, E.;
Urraca, J. L. Curr. Anal. Chem. 2008, 4, 316.
(5) Subrahmanyam, S.; Piletsky, S. A.; Piletska, E. V.; Chen, B. N.;
Day, R.; Turner, A. P. F. Adv. Mater. 2000, 12, 722.
(6) Chow, C. F.; Lam, M. H. W.; Leung, M. K. P. Anal. Chim. Acta
2002, 466, 17.
(7) Liao, Y.; Wang, W.; Wang, B. H. Bioorg. Chem. 1999, 27, 463.
(8) Liu, R. Y.; Guan, G. J.; Wang, S. H.; Zhang, Z. P. Analyst 2011,
136, 184.
(9) Benito Pena, E.; Moreno-Bondi, M. C.; Aparicio, S.; Orellana, G.;
̃
Cedefur, J.; Kempe, M. Anal. Chem. 2006, 78, 2019.
(10) Urraca, J. L.; Moreno-Bondi, M. C.; Orellana, G.; Sellergren, B.;
Hall, A. J. Anal. Chem. 2007, 79, 4915.
(11) Nguyen, T. H.; Ansell, R. J. Org. Biomol. Chem. 2009, 7, 1211.
(12) Kubo, H.; Yoshioka, N.; Takeuchi, T. Org. Lett. 2005, 7, 359.
(13) Ng, S. M.; Narayanaswamy, R. Anal. Bioanal. Chem. 2006, 386,
1235.
(14) Southard, G. E.; Van Houten, K. A.; Ott, E. W.; Murray, G. M.
Anal. Chim. Acta 2007, 581, 202.
(15) Saha, S.; Samanta, A. J. Phys. Chem. A 2002, 106, 4763.
(16) Qi, Q.; Ha, Y.-Q.; Sun, Y.-M. Int. J. Quantum Chem. 2011, 111,
2234.
(17) Niu, C. G.; Zeng, G. M.; Chen, L. X.; Shen, G. L.; Yu, R. Q.
Swelling Behavior of the Bulk Polymers. A certain amount of
the respective grounded polymer was filled in a NMR tube. The height
of the polymer in the tube was measured. Subsequently, 1 mL of the
respective solvent was added and the polymers well dispersed. After 24
Analyst 2004, 129, 20.
(18) Wang, J. B.; Qian, X. H. Chem. Commun. 2006, 109.
(19) Grabchev, I.; Qian, X.; Xiao, Y.; Zhang, R. New J. Chem. 2002,
26, 920.
1388
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389

The Journal of Organic Chemistry
Article
(20) Jiang, J. B.; Leng, B.; Xiao, X.; Zhao, P.; Tian, H. Polymer 2009,
50, 5681.
(21) Gunnlaugsson, T.; Kruger, P. E.; Lee, T. C.; Parkesh, R.; Pfeffer,
F. M.; Hussey, G. M. Tetrahedron Lett. 2003, 44, 6575.
(22) Veale, E. B.; Tocci, G. M.; Pfeffer, F. M.; Kruger, P. E.;
Gunnlaugsson, T. Org. Biomol. Chem. 2009, 7, 3447.
(23) Hall, A. J.; Quaglia, M.; Manesiotis, P.; De Lorenzi, E.;
Sellergren, B. Anal. Chem. 2006, 78, 8362.
(24) Urraca, J. L.; Moreno-Bondi, M. C.; Hall, A. J.; Sellergren, B.
Anal. Chem. 2007, 79, 695.
(25) Gunnlaugsson, T.; Kruger, P. E.; Jensen, P.; Tierney, J.; Ali, H.
D. P.; Hussey, G. M. J. Org. Chem. 2005, 70, 10875.
(26) Gan, J.; Tian, H.; Wang, Z.; Chen, K.; Hill, J.; Lane, P. A.; Rahn,
M. D.; Fox, A. M.; Bradley, D. D. C. J. Organomet. Chem. 2002, 645,
168.
(27) Li, A.-F.; Wang, J.-H.; Wang, F.; Jiang, Y.-B. Chem. Soc. Rev.
2010, 39, 3729.
(28) Wintgens, V.; Valat, P.; Kossanyi, J.; Biczok, L.; Demeter, A.;
Berces, T. J. Chem. Soc., Faraday Trans. 1994, 90, 411.
(29) Alexiou, M. S.; Tychopoulos, V.; Ghorbanian, S.; Tyman, J. H.
P.; Brown, R. G.; Brittain, P. I. J. Chem. Soc., Perkin Trans. 2 1990, 837.
(30) Sawa, M.; Hsu, T.-L.; Itoh, T.; Sugiyama, M.; Hanson, S. R.;
Vogt, P. K.; Wong, C.-H. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 12371.
(31) Manuscript in preparation.
́
(32) Boiocchi, M.; Del Boca, L.; Esteban Gomez, D.; Fabbrizzi, L.;
Licchelli, M.; Monzani, E. J. Am. Chem. Soc. 2004, 126, 16507.
(33) Leng, B.; Tian, H. Aust. J. Chem. 2010, 63, 169.
(34) The program employed only allows the fitting of microscopic
association equilibria of A + B ⇌ AB-type, i.e., M− + H+ ⇌ MH, B− +
H+ ⇌ BH, MH + B− ⇌ MHB, etc. The spectrum of 1 in DMSO and
THF (black species in Figure 4A) and the spectrum of 1− in THF (red
species; obtained by deprotonation of 1 with the spectroscopically
silent TBA hydroxide) were included as known species in the fits.
Without including 1− or M as known component in the THF case,
convergence of the fit to chemically meaningful result was much more
difficult.
(35) The shape of the absorption spectra of 1− varies somewhat
between DMSO and THF which is tentatively ascribed to the various
possible isomers/tautomers of 1− (see also Scheme 2).
(36) Sellergren, B.; Martin Esteban, A., In: Handbook of Sample
Preparation; Pawlitszyn, J., Lord, H. L., Eds.; John Wiley & Sons Inc.:
NJ, 2010, pp 445−475.
(37) Mayes, A. G.; Whitcombe, M. J. Adv. Drug Delivery Rev. 2005,
57, 1742.
(38) Tokonami, S.; Shiigi, H.; Nagaoka, T. Anal. Chim. Acta 2009,
641, 7.
(39) Andersson, L.; Sellergren, B.; Mosbach, K. Tetrahedron Lett.
1984, 25, 5211.
(40) Hall, A. J.; Manesiotis, P.; Emgenbroich, M.; Quaglia, M.; De
Lorenzi, E.; Sellergren, B. J. Org. Chem. 2005, 70, 1732.
(41) Porcal, W.; Hernan
Merlino, A.; Ferreira, A.; Di Maio, R.; Castro, A.; Gonzal
Cerecetto, H. Bioorg. Med. Chem. 2007, 15, 2768.
́
dez, P.; Aguirre, G.; Boiani, L.; Boiani, M.;
́
ez, M.;
(42) Luebke, M.; Whitcombe, M. J.; Vulfson, E. N. J. Am. Chem. Soc.
1998, 120, 13342.
(43) Rurack, K.; Spieles, M. Anal. Chem. 2011, 83, 1232.
(44) Resch, U.; Rurack, K. Proc. SPIE-Int. Soc. Opt. Eng. 1997, 3105,
96.
(45) Shen, Z.; Rohr, H.; Rurack, K.; Uno, H.; Spieles, M.; Schulz, B.;
̈
Reck, G.; Ono, N. Chem.Eur. J. 2004, 10, 4853.
1389
dx.doi.org/10.1021/jo3019522 | J. Org. Chem. 2013, 78, 1377−1389