19F NMR fingerprints: Identification of neutral organic compounds in a molecular container

Article abstract of DOI:10.1021/ja504110f

Improved methods for quickly identifying neutral organic compounds and differentiation of analytes with similar chemical structures are widely needed. We report a new approach to effectively fingerprint neutral organic molecules by using ¹⁹F NMR and molecular containers. The encapsulation of analytes induces characteristic up- or downfield shifts of ¹⁹F resonances that can be used as multidimensional parameters to fingerprint each analyte. The strategy can be achieved either with an array of fluorinated receptors or by incorporating multiple nonequivalent fluorine atoms in a single receptor. Spatial proximity of the analyte to the ¹⁹F is important to induce the most pronounced NMR shifts and is crucial in the differentiation of analytes with similar structures. This new scheme allows for the precise and simultaneous identification of multiple analytes in a complex mixture.

Full text of DOI:10.1021/ja504110f

Article
pubs.acs.org/JACS
Open Access on 07/22/2015
19F NMR Fingerprints: Identification of Neutral Organic Compounds
in a Molecular Container
Yanchuan Zhao, Georgios Markopoulos, and Timothy M. Swager*
Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States
S
* Supporting Information
ABSTRACT: Improved methods for quickly identifying
neutral organic compounds and differentiation of analytes with
similar chemical structures are widely needed. We report a new
approach to effectively “fingerprint” neutral organic molecules
by using 19F NMR and molecular containers. The encapsulation
of analytes induces characteristic up- or downfield shifts of 19F
resonances that can be used as multidimensional parameters to
fingerprint each analyte. The strategy can be achieved either with
an array of fluorinated receptors or by incorporating multiple
nonequivalent fluorine atoms in a single receptor. Spatial
proximity of the analyte to the 19F is important to induce the most pronounced NMR shifts and is crucial in the differentiation of
analytes with similar structures. This new scheme allows for the precise and simultaneous identification of multiple analytes in a
complex mixture.
association with fluorinated chelates or crown ethers where
characteristic shifts are generated for each metal ion.9 As the
induced 19F NMR shifts are largely dependent on the through-
bond disturbance of electron density at the fluorine atom upon
association, charged species are typically selected as target analytes.
In contrast, the detection and differentiation of neutral organic
molecules with similar structures represents a significant challenge
for most sensing methods.
To achieve our goal of unique identification of an analyte, our
platform needs to meet the following criteria: (1) The molecular
recognition event is sufficiently defined to provide a well-structured
binding complex. (2) There are a number of independently varying
19F NMR signals that shift to provide a robust multidimensional
discrimination of an analyte. (3) The shift of the 19F resonance
should be induced by spatial proximity rather than through-bond
electron density transmission so that the structure information on
the whole molecule can be accessed by spatially arranged fluorine
atoms.
Molecular containers, such as cavitands and capsules with
different levels of preorganization, have found wide-ranging
applications in molecular recognition.10 By design, the
encapsulation of an analyte induces a change of the environment
inside the container, thereby creating easily discernible 19F NMR
shifts. The multidimensional output can be achieved either
with an array of receptors bearing equivalent fluorine atoms
at different positions relative to the analyte (Scheme 1a) or by
employing a single receptor with multiple nonequivalent fluorine
atoms (Scheme 1b). As a result of the scarcity of organic fluorine
compounds in nature,6,11 it is unlikely that there will be interfering
INTRODUCTION
■
There is an increasing awareness of the need for more selective
and reliable methods to detect and rapidly identify target analytes
of interest in a variety of contexts relevant to health care, process
control, and environmental monitoring.1 Chemosensory systems
designed to assist in this process are molecular constructs that
respond to a stimulus and give a measurable change in electronic,
optical, and/or chemical/spectroscopic properties.2 Trans-
duction generally involves molecular associations or an electron
transfer process between the analyte and a receptor.3 These
interactions typically occur at a specific bonding site, and sensing
methods based on this strategy are best suited to detect classes
of structurally related analytes, but often fail in the precise
discrimination of related species. Array sensing has emerged as an
approach that increases discriminatory power by combining
signals collected by a large amount of individual sensors.4
However, without highly orthogonal discrimination between
analytes, this method often has difficulty in unambiguously
identifying analytes at unknown concentrations. Herein, we report
a sensing method based on 19F NMR and the encapsulation of an
analyte with molecular containers. The method provides a unique
spectroscopic signature (fingerprint) that allows for an output and
enables precise and simultaneous identification of multiple guest
molecules in a complex mixture.5
19F NMR has emerged as a versatile tool in biological and
pharmaceutical studies as a result of the high sensitivity and
scarcity of naturally occurring background signals.6 Libraries of
fluorinated compounds are used to identify potential ligands that
bind to target proteins.7 Fluorinated biological molecules have
utility in the determination of enzyme activity.8 In addition to
reaction monitoring, such as the hydrolysis of a fluorine-containing
substrate, various metal ions can be detected through reversible
Received: May 6, 2014
Published: July 22, 2014
© 2014 American Chemical Society
10683
dx.doi.org/10.1021/ja504110f | J. Am. Chem. Soc. 2014, 136, 10683−10690

Journal of the American Chemical Society
Article
Scheme 1. Schematic Illustration of 19F NMR Spectroscopy
Identification of Organic Molecules with Molecular
Containers
Scheme 3. Preparation of Fluorinated Calix[4]arenes 7−11
and pentafluorobenzene recently reported by Zhang and co-
workers.14 The methyl groups were subsequently removed by
treatment with BBr3 in CH2Cl2 at low temperature (Scheme 3b).
The corresponding tungsten−imido complexes 1−5 were
obtained using a previously reported “one-pot” procedure from
calixarenes 7−11 by reaction with WOCl4 and iminophosphor-
ane (Ph3PNR) reagent.15
signals, and our scheme provides an efficient method to fingerprint
a chosen analyte.12
RESULTS AND DISCUSSION
■
We chose calix[4]arene tungsten−imido complexes as a scaffold
from which to produce partially fluorinated molecular containers
on the basis of their synthetic accessibility and the fact that the
Lewis acidic nature of the metal center gives predictable binding
structures with Lewis basic analytes.13 To evaluate the feasibility
of the strategy based on encapsulation and the chemical shift
induced by spatial proximity, we examined calixarene tungsten−
imido complexes appended with spatially varying trifluoromethyl
(CF3) and trifluoromethoxy (OCF3) groups at the upper rim
(Scheme 2, complexes 1−4). In addition to an array of complexes
that can be employed together to output a fingerprint, receptors
5 and 5a with multiple nonequivalent fluorine atoms are also
prepared (Scheme 2).
NMR Fingerprinting with an Array of Receptors. To
evaluate the fidelity of this strategy in the precise identification of
structurally similar molecules, we selected a series of nitriles with
an interest in differentiating pesticides and pharmaceuticals.16
The Lewis basic nitrile can be encapsulated in the molecular
containers 1−5 and 5a through the formation of a coordination
bond with the tungsten atom. Sensing experiments are
performed by adding analytes to chloroform solutions of 1 at
ambient temperature. The formation of a static complex with 1 is
critical to create a clear shift rather than a dynamic structure that
will produce shifts that are more akin to a solvent effect.15 In this
way, the fluorine atoms provide discrete signals at precise shifts
that are uniquely assignable to the encapsulated analytes.
Notably, the −OCF3 group in tungsten complex 1 appears as a
singlet at −56.63 ppm (Figure 1a), which is very close to the shift
found with parent calix[4]arene 7 (−56.51 ppm), indicating the
remote through-bond effects are not efficient to induce a 19F
NMR shift. In contrast, the binding of nitriles to 1 produces
0.2−0.9 ppm downfield shifts in 19F NMR as a result of the
disturbance of the environment through replacement of solvent
molecules by the analyte. Consistent with this model, acetonitrile
induces a much smaller shift than less electron-donating
3-bromopropionitrile. All of our results are consistent with the
differences in 19F NMR of free and bound complex 1 being
caused by spatial proximity rather than through-bond electron
transmission (Figure 1d,g). The precision in the identification of
molecules is illustrated by comparison of the differences induced
by the binding of acetonitrile, propionitrile, and nonanenitrile
with 1. In this experiment, nonanenitrile induces a more
pronounced downfield shift than propionitrile and acetonitrile
(Figure 1d−f). The power of this method was further evaluated
by the analysis of a mixture with a number of potential guest
molecules. In this experiment, a mixture of nine different nitriles
and 1 gave the same spectrum as obtained by superimposing the
spectra recorded with each analyte independently (Figure 1b,c).
It is notable that the precise identification of the multiple neutral
organic analytes in a mixture represents a powerful advance in
chemical sensing.
Scheme 2. Fluorinated Calix[4]arene−Tungsten Complexes
Employed in This Study
Synthesis. The −CF3 and −OCF3-substituted calix[4]arenes
7−10 were prepared through a Suzuki−Miyaura coupling of
diiodocalix[4]arene (6) and various organoboronic acids
followed by a demethylation with Me3SiI (Scheme 3a). The
target bis(pentafluorophenyl)-substituted calix[4]arene (11)
was prepared through a silver-mediated direct coupling of 6
We next explored the sensing properties of 2-CF3-substituted
complex 2. Interestingly, although the encapsulation of alkyl
nitriles (Figure 2d−h) and benzyl nitriles (Figure 2i,j) produces
downfield shifts which are also observed in the experiments with 1,
10684
dx.doi.org/10.1021/ja504110f | J. Am. Chem. Soc. 2014, 136, 10683−10690

Journal of the American Chemical Society
Article
Figure 2. 19F NMR spectra (64 scans) of complex 2 alone and mixtures
of complex 2 (1.0 mM in CDCl3) and different analytes (2.0 mM):
(a) complex 2 alone, (b) eight nitriles added to a solution of 2 in CDCl3,
(c) superimposition of the spectra of complex 2 with each of the eight
nitriles from (b) collected independently, (d)−(o) complex 2 bound to
various nitriles.
Figure 1. 19F NMR spectra (64 scans) of complex 1 alone and mixtures
of complex 1 (1.0 mM in CDCl3) and different analytes (2.0 mM):
(a) complex 1 alone, (b) nine nitriles added to a solution of 1 in CDCl3,
(c) superimposition of the spectra of complex 1 with each of the nine
nitriles from (b) collected independently, (d)−(o) complex 1 bound to
various nitriles.
aromatic nitriles (Figure 2k−o) induced upfield shifts upon
binding, thus providing a facile way to determine the identity of the
analyte. Unlike the trend observed in the experiments with 1, the
bonding of 3-bromopropionitrile with 2 gives a smaller downfield
shift than that of acetonitrile and nonanenitrile (Figure 2d−g). This
result indicates receptors/sensors with orthogonal discriminatory
power can be easily produced by incorporating fluorine atoms at
different positions. Similarly, complex 2 also shows the ability to
identify a series of nitriles in a complex mixture (Figure 2b).
The differences observed for individual analytes are shown in
Figures 1 and 2, and the characteristic up- and downfield shifts
induced by each analyte are given in a two-dimensional plot, with
the 19F resonances of 1 and 2 as the axes (Figure 3). Simple
inspection of these data reveals the ability of the combined sensor
system to resolve all the alkyl and benzyl nitriles. In contrast, the
discrimination of benzonitriles with the para-substituents
investigated is still not satisfactory probably because the remote
substituent only results in a minimal magnetic influence on the
fluorine atoms in receptors 1 and 2. Consistent with this assumption,
3-iodobenzonitrile with the substituent closer to the fluorine atom
displays behavior different from that of para-substituented nitriles
(Figure 3). It should be mentioned that a difference of 0.03 ppm
leads to a baseline separation of singlet peaks in our 19F NMR
spectra, which correlates to a magnitude of 30 on the axes used in
Figure 3.
Figure 3. 2D scatter of analytes based on the shifts of 19F resonances
upon bonding: x axis, OCF3 fluorine (1) (−Δδ × 1000); y axis, CF3
fluorine (2) (−Δδ × 1000).
10685
dx.doi.org/10.1021/ja504110f | J. Am. Chem. Soc. 2014, 136, 10683−10690

Journal of the American Chemical Society
Article
To achieve better resolution of benzonitriles, we examined
complexes 3 and 4 with −OCF3 and −CF3 groups in the
meta-position, respectively (Scheme 2). By design, the fluorine
atoms in these complexes are closer to the para-substituent of the
nitrile guests, which allows discrimination of this remote structural
difference that was not achieved by 1 and 2. As shown in Figures 4
Figure 5. 19F NMR spectra (64 scans) of complex 4 alone and mixtures
of complex 4 (1.0 mM in CDCl3) and different analytes (2.0 mM): (a)
complex 4 alone, (b) four aromatic nitriles added to a solution of 4 in
CDCl3, (c) superimposition of the spectra of complex 4 with each of the
four nitriles from (b) collected independently, (d)−(o) complex 4
bound to various nitriles.
Figure 4. 19F NMR spectra (64 scans) of complex 3 alone and mixtures
of complex 3 (1.0 mM in CDCl3) and different analytes (2.0 mM): (a)
complex 3 alone, (b) four aromatic nitriles and propionitrile added to a
solution of 3 in CDCl3, (c) superimposition of the spectra of complex 3
with each of the five nitriles from (b) collected independently, (d)−(o)
complex 3 bound to various nitriles.
variance, this combination provides better resolution than that
shown in Figure 3 wherein 1 and 2 were employed. Moreover,
the resolution can be further enhanced by using signals collected
by a third receptor. The use of 1, 2, and 4 enables an interpretable
3D differentiation of all the analytes. As shown in Figure 7, all
aromatic nitriles appear below the xy plane, benzyl nitriles give
pronounced x values, and alkyl nitriles give smaller x values.
Simple inspection of these figures reveals utility for the facile
classification of analytes.
NMR Fingerprinting with a Single Receptor. The
preceding studies enable the development of a receptor with
multiple nonequivalent fluorine atoms that can fingerprint
organic nitriles. In this regard, in addition to pentafluorophenyl
groups, we have also incorporated a fluorine atom on the
arylimido group, which has been shown to differentiate the
electron-donating ability of the bound analytes by 19F NMR
shifts.15 By design, the pentafluorophenyl group of 5a spatially
arranges fluorine groups in a polarizable π-system to create an
environment capable of differentiating structurally similar
analytes (Figure 1). The NMR experiments were carried out
similarly to those of complexes 1−4. As shown in Figure 8, the
imido-fluorine of 5a appears as a triplet at around −100 (t) ppm,
and 5, the differences in 19F NMR of free and bound complexes are
within the range of <0.3 ppm, which is smaller than those observed
with 1 and 2, suggesting spatial proximity is crucial to induce shifts.
Minimum 19F NMR shifts are observed for acetonitrile as a result
of its smaller size (Figures 4d and 5d). Interestingly, despite the
smaller shifts produced, complexes 3 and 4 display improved
resolution of benzonitriles relative to complexes 1 and 2 as shown in
Figures 4k−o and 5k−o. Our collective results indicate it is possible
to rationally design sensors with the desired selectivity by optimizing
the position of the fluorine atoms. Simultaneous discrimination of
diverse benzonitriles in a mixture is further demonstrated by the
well-dispersed peaks shown in Figures 4b and 5b.
Multiple sensors with orthogonal discriminatory properties
allow for higher analyte resolution through a combined analysis
of signals from multiple receptors. Figure 6 is a plot using the
19F NMR differences observed with 1 and 4. As a result of
the orthogonal selectivity imparted by the spatial distribution
10686
dx.doi.org/10.1021/ja504110f | J. Am. Chem. Soc. 2014, 136, 10683−10690

Journal of the American Chemical Society
Article
and provide a differentiation from the alkyl nitriles investigated
(Figure 8k−o). In contrast, benzyl nitrile did not induce a shift
of m-19F signals, thereby indicating the importance of the
precise position of the aromatic group in the molecular container
(Figure 8i). It is also notable that 4-iodobenzonitrile induces less
pronounced upfield shifts of m- and p-19F NMR signals as
compared to benzonitrile (part n vs part k of Figure 8), indicating
a downfield shifting effect with halide substitution. This trend is
also observed for 4-iodobenzyl cyanide and 3-bromopropionitrile,
which induce downfield shifts of m- and p-19F NMR signals.
The downfield shifts relative to the shifts for the uncomplexed
receptor are not surprising because only very small upfield shifts
are produced by their nonhalogenated analogues (part j vs part i
and part g vs part d of Figure 8). Electron-rich aromatic nitriles
produce a more pronounced upfield shift of m- and p-19F NMR
signals as compared to electron-deficient aromatic nitriles, and
this trend is also displayed by the shifts in the imido-19F NMR
signals, which are solely dependent upon the electron-donating
ability of the nitriles (Figure 1k−n). Owing to the polarizable
π-system, 5a is more sensitive to the electronic properties of
aromatic nitrile than 1−4. As a result of the multiplets of the
19F NMR, the overlap of signals produced by each analyte is more
likely in the analysis of a complex mixture (Figure 8b).
Figure 6. 2D scatter of analytes based on the shifts of 19F resonances
upon bonding: x axis, 2-OCHF3 fluorine (1) (−Δδ × 1000); y axis, 3,5-
CF3 fluorine (4) (−Δδ × 1000).
The selective detection/identification of insecticides is
important considering the widespread usage and toxicity of
these chemicals. Cyanophos [O-(4-cyanophenyl) O,O-dimethyl
phosphorothioate] is an organophosphorus-based insecticide
that is effective against various plant pests.17 It is a powerful
cholinesterase inhibitor and represents a threat to human health.
Traditional chemosensing methods typically rely on bonding or
reactions with the Lewis acidic phosphorus group, which is not
readily distinguished from structurally related compounds.18 In
contrast, our method generates a fingerprint that precisely
distinguishes this compound from all other analytes (Figure 8p).
Notably, the characteristic upshift of m-fluorine enables a fast
assignment of cyanophos as an aromatic nitrile. This method was
able to provide unambiguous detection of the cyanophos signals
(S/N > 15) at an analyte concentration of 100 μM using a
400 MHz spectrometer and an acquisition time of 24 min
(800 scans) (for details, see Figure S1 in the Supporting Information).
A three-dimensional plot is shown in Figure 9, with the o-, p-,
and m-19F NMR signals as the axes and the relative shift of
the imido-19F NMR signal represented by the radius of a sphere.
The highly dispersed data points demonstrated the ability of
5a to resolve all the analytes. As expected, nitriles with similar
structures display 19F NMR signals that are close to one another.
For example, acetonitrile and propionitrile (Figure 8d,e) induce
similar but differentiated responses. It should be mentioned that
the radii of the spheres in Figure 9 correlate with the shift of
imido-fluorine and can further differentiate analytes that produce
similar spectral differences in the other 19F NMR signals, such as
ethyl (R)-4-cyano-3-hydroxybutyrate (Figure 8h) and C8H17CN
(Figure 8f).
Figure 7. 3D scatter of analytes based on the shifts of 19F resonances
upon bonding: x axis, 3,5-CF3 fluorine (4) (−Δδ × 1000); y axis,
2-OCF3 fluorine (1) (−Δδ × 1000); z axis, 2-CF3 fluorine (2) (−Δδ ×
1000).
and the peaks at −143 (dd), −156 (t), and −162 (m) ppm are
identified as o-, p-, and m-fluorine, respectively (Figure 8a).
These distinctive chemical shifts provide a multidimensional
spectroscopic signature without complexity from overlapping 19F
NMR signals. Binding of nitriles to 5a produces upfield shifts in
the 19F NMR of the pentafluorophenyl group as a result of the
shielding effects of the encapsulated molecules. Alkyl nitriles with
varying chain length from acetonitrile to nonanenitrile display
increasing upfield shifts for the imido-fluorine which correlate
with the electron-donating ability of these molecules to the
tungsten center (Figure 8d−g), and the same trend is observed
for substituted aromatic nitriles (Figure 8k−n). Pronounced
upfield shifts of m-19F signals are observed with aromatic nitriles
Association Constants and Detection Limits. The
association constants were measured in chloroform. The
concentrations of free and bound complexes are determined by
the integration of the 19F NMR signal, and the concentration of
free nitrile is calculated accordingly. As shown in Table 1, the
magnitude of the bonding constant varies significantly toward
different nitriles. For 1 and 2, the constants decrease in the
sequence acetonitrile, benzonitrile, and benzyl nitrile. Significant
bonding enhancement of benzonitrile is observed with 4, 5, and
5a, indicating the favorable π−π interactions between the phenyl
10687
dx.doi.org/10.1021/ja504110f | J. Am. Chem. Soc. 2014, 136, 10683−10690

Journal of the American Chemical Society
Article
Figure 8. 19F NMR spectrum (typically 128 scans) of a mixture of complex 5a (2 mM in CDCl3) and different analytes (5.0 mM). (a) Five nitriles were
added to a solution of 5a in CDCl3. (b) Superimposition of the spectrum collected independently. 19F NMR spectra (typically 128 scans) of complex 5a
alone and mixtures of complex 5a (2.0 mM in CDCl3) and different analytes (5.0 mM): (a) complex 5a alone, (b) five nitriles added to a solution of 5a in
CDCl3, (c) superimposition of the spectra of complex 5a with each of the five nitriles from (b) collected independently, (d)−(p) complex 5a bound to
various nitriles.
Table 1. Association Constants (K/M−1) of Various Nitriles
a
with a Tungsten−Imido Complex
1
2
3
4
5
5a
K (CH3CN)
K (PhCN)
945
345
177
815
372
97
b
b
618
852
318
786
1360
600
279
897
K (PhCH2CN)
118
219
a
Determined by 19F NMR in CDCl3. Three measurements at different
concentrations are taken, and the average is given in the table, error
b
<15%. Not determined because the signals overlap in both 1H NMR
and 19F NMR.
on the arylimido group is beneficial to the binding as a result of
the increased Lewis acidity of the tungsten center. It should be
mentioned that the dissociation of the internal bound ligand of a
metalated cavitand tends to be slow because of the isolation of
the ligand from the bulk solution. This explains why a relatively
low binding constant is accompanied by a slow exchange
system.19 Notably, with the association constants, the simulta-
neous and quantitative measurements of multiple analytes can
be achieved on the basis of signal integrations (for details, see
Figure S8 in the Supporting Information). According to eq 1, the
Figure 9. 3D scatter of analytes based on the shifts of 19F resonances
upon bonding: x axis, o-19F (−Δδ × 1000); y axis, p-19F (−Δδ × 1000); z
axis, m-19F (−Δδ × 1000). The sphere radius is correlated to imido-19F
(−Δδ × 1000) with a factor of 0.04.
K
CalixW(NR) + analyte ⇌ CalixW(NR):(analyte)
[CalixW(NR):(analyte)]
[CalixW(NR)][analyte]
K =
ring and electron-deficient 3,5-bis(trifluoromethyl)phenyl or
pentafluorophenyl group. Changing the methyl group to fluorine
(1)
10688
dx.doi.org/10.1021/ja504110f | J. Am. Chem. Soc. 2014, 136, 10683−10690

Journal of the American Chemical Society
Article
ratio of bound to free analyte is equal to K[CalixW(NR)]. For
the detection of analyte in the presence of excess receptor,
[CalixW(NR)] is the total complex concentration employed in
the analysis. This means, for example, about 45% of the analyte is
in the complexed form when a trace amount of benzonitrile is
detected in the presence of 2 M tungsten complex 1. Owing to
the six equivalent fluorine atoms and singlet peak, the detection
limit of benzonitrile in the presence of 2 M 1 is determined to be
down to 10 μM using a 400 MHz spectrometer and an
acquisition time of 24 min (800 scans) in contrast to the 100 μM
detection limit of cyanophos obtained with 5a (for details, see
Figure S7 in the Supporting Information).
To gain more insight into the transduction of the current
method, the X-ray single-crystal structures of 1, 2, and 5a were
obtained. Interestingly, 2:CH3CN is found to be perfectly
isostructural to 1:CH3CN, and the only difference is the OCF3
group is replaced by a CF3 group (Figure 11). This result
The robust sensing power is further demonstrated by the
analysis of a complex mixture of various nitriles in the presence of
an excess amount of hexane, ethyl acetate, and acetone with 1. As
shown in Figure 10, noncoordinating analytes, such as hexane,
Figure 10. 19F NMR spectrum (64 scans) of a mixture of complex 1 (ca.
0.8 mM in CH2Cl2), various nitriles (each ca. 1.6 mM), hexane (5 μL),
ethyl acetate (5 μL), and acetone (5 μL).
ethyl acetate, and acetone, did not give signals, while various
nitriles can be unambiguously identified simultaneously even in
nondeuterated solvent.
The detection of pollution in water is crucial to environmental
monitoring. Although many sensing methods are capable of
detecting a specific target in domestic water, the analysis of a
more complex matrix, such as river water, is still challenging. To
mimic a sample in the environment, water taken from the Charles
River between Boston and Cambridge, MA, was contaminated
with cyanophos at various concentrations. To use a minimum
amount of organic solvent, river water (5 mL) was extracted with
a solution of receptor 1 in dichloromethane (2 M, 0.6 mL), and
the resulting dichloromethane phase was analyzed by 19F NMR.
The detection limit of cyanophos is determined to be 5 μM by
using this method (for details, see Figure S10 in the Supporting
Information). Enrichment by extraction is often employed when
detecting nanomolar range neutral organic molecules in water.
As the process is not selective, a complex mixture with a number
of components at much higher concentrations than the target
analyte is often obtained. To test our method in the analysis of a
mixture obtained from enrichment, river water (500 mL) was
extracted with dichloromethane (100 mL × 3) and concentrated.
The extract was then redissolved in a solution of receptor 1 in
dichloromethane (2 M, 0.5 mL) and analyzed by 19F NMR.
Detection of cyanophos at 20 nM in river water was achieved by
this method (for details, see Figure S11 in the Supporting
Information). It is worth noting that a number of unidentified
species at much higher concentrations than that of cyanophos
Figure 11. X-ray structures of 1, 2, and 5a (1:1 cocrystal with CH3CN or
PhCN): black, carbon; green, fluorine; blue, nitrogen; red, oxygen;
purple, tungsten. Note: The methyl groups of the acetonitriles in
1:CH3CN and 2:CH3CN are disordered about the crystallographic 2-
fold axis.
suggests that it is valid to estimate the structures of related
complexes. Although the nonlinear geometry of acetonitrile in
2:CH3CN is unusual, it is not unprecedented and has been
observed in a variety of metal complexes.20 Another observation
is that fluorinated groups face inward for the cavity in 2:CH3CN
whereas the opposite is true for 2:PhCN. Probably as a result of
the larger size of benzonitrile, the cavity of calixarene expands to
fit the analyte. The discrete behaviors found in 2:CH3CN and
2:PhCN in the crystal structure also shed light on the chemical
shift induced with 2 wherein alkyl nitrile produces a downfield
shift whereas aromatic nitrile induces an upfield shift (Figure 2).
The distance of tungsten to the nitrogen of the nitrile in 2:PhCN
is significantly longer than that of 2:CH3CN (2.310 Å vs 2.287 Å),
suggesting a weaker bonding of PhCN. This observation is
consistent with the trend of association constants found in Table 1.
It should be mentioned that the NMR signals are collected in
solution; therefore, the shifts are largely dependent on the average
distance between the fluorine atom and the analyte in all of the
conformational isomers.
1
were observed in H NMR, which makes the identification of
cyanophos unsuccessful in 1H NMR (for details, see Figure S12
in the Supporting Information). The preceding studies are
intended to illustrate that the method is sufficiently robust for
demanding applications. A more efficient extraction process
could be achieved by immobilization of 1 or analogues in a
concentrator/filter assembly.
CONCLUSIONS
■
In summary, we have demonstrated a new sensing scheme based
on 19F NMR and the encapsulation of analytes with molecular
containers. The method collects extensive interactions between
the analyte and receptor/container to provide measurable signals
10689
dx.doi.org/10.1021/ja504110f | J. Am. Chem. Soc. 2014, 136, 10683−10690

Journal of the American Chemical Society
Article
8649. (c) Miranda, O. R.; Creran, B.; Rotello, V. M. Curr. Opin. Chem.
Biol. 2010, 14, 728. (d) Wang, F.; Swager, T. M. J. Am. Chem. Soc. 2011,
133, 11181.
(5) For selected examples of NMR sensing to identify organic
molecules, see: (a) Perrone, B.; Springhetti, S.; Ramadori, F.; Rastrelli,
F.; Mancin, F. J. Am. Chem. Soc. 2013, 135, 11768. (b) Teichert, J. F.;
Mazunin, D.; Bode, J. W. J. Am. Chem. Soc. 2013, 135, 11314.
(6) For a review discussing applications of 19F NMR, see: Yu, J.-X.;
Hallac, R. R.; Chiguru, S.; Mason, R. P. Prog. Nucl. Magn. Reson. Spectrosc.
2013, 70, 25.
with sufficient dimensionality (information) to uniquely identify
or “fingerprint” analytes that have only small structural differences.
The strategy can be achieved either with an array of receptors or by
incorporating multiple nonequivalent fluorine atoms in a single
receptor. This new scheme allows for an informative and
interpretable output and enables a precise and simultaneous
identification of multiple potential guest molecules in a complex
mixture. The structures we report herein are only representative
examples and can be extended to many other structural scaffolds,
including those targeting complex and/or larger biomolecular
species that cannot be readily identified by conventional analytical
methods (e.g., mass spectrometry). Critical to this latter prospect is
the development of receptors/probes that incorporate 19F groups
that are sensitive to their environment and produce relatively static
complexes. We envision these more complex recognition elements
will produce powerful detection schemes relevant to environmental
and biomedical sensing.
(7) (a) Tengel, T.; Fex, T.; Emtenas, H.; Almqvist, F.; Sethson, I.;
Kihlberg, J. Org. Biomol. Chem. 2004, 2, 725. (b) Dalvit, C.; Vulpetti, A.
Magn. Reson. Chem. 2012, 50, 592.
(8) (a) Tanaka, K.; Kitamura, N.; Naka, K.; Chujo, Y. Chem. Commun.
2008, 6176. (b) Tanaka, K.; Kitamura, N.; Chujo, Y. Bioconjugate Chem.
2011, 22, 1484. (d) Stockman, B. J. J. Am. Chem. Soc. 2008, 130, 5870.
(e) Albert, M.; Repetschnigg, W.; Ortner, J.; Gomes, J.; Paul, B. J.;
Illaszewicz, C.; Weber, H.; Steiner, W.; Dax, K. Carbohydr. Res. 2000,
327, 395. (f) Mendz, G. L.; Lim, T. N.; Hazell, S. L. Arch. Biochem.
Biophys. 1993, 305, 252. (g) Yu, J.; Liu, L.; Kodibagkar, V. D.; Cui, W.;
Mason, R. P. Bioorg. Med. Chem. 2006, 14, 326. (h) Yu, J.; Mason, R. P. J.
Med. Chem. 2006, 49, 1991. (i) Yu, J.-X.; Kodibagkar, V. D.; Liu, L.;
Mason, R. P. NMR Biomed. 2008, 21, 704.
(9) (a) Smith, G. A.; Hesketh, R. T.; Metcalfe, J. C.; Feeney, J.; Morris,
P. G. Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 7178. (b) Schanne, F. A. X.;
Dowd, T. L.; Gupta, R. K.; Rosen, J. F. Proc. Natl. Acad. Sci. U.S.A. 1989,
86, 5133. (c) Levy, L. A.; Murphy, E.; Raju, B.; London, R. E.
Biochemistry 1988, 27, 4041. (d) Smith, G. A.; Kirschenlohr, H. L.;
Metcalfe, J. C.; Clarke, S. D. J. Chem. Soc., Perkin Trans. 2 1993, 1205.
(e) Jiang, Z.-X.; Feng, Y.; Yu, Y. B. Chem. Commun. 2011, 47, 7233.
(10) (a) Rudkevich, D. M.; Rebek, J. J. Eur. J. Org. Chem. 1999, 1999,
ASSOCIATED CONTENT
* Supporting Information
■
S
Crystallographic data in CIF format, synthetic procedures and
characterization of the compounds, NMR spectra, association
constants, and detection limit studies. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
tswager@mit.edu
■
1991. (b) Asfari, Z.; Bohmer, V.; Harrowfield, J. M.; Vicens, J.
̈
Notes
Calixarenes 2001; Kluwer Academic Publishers: Dordrecht, The
Netherlands, 2001. (c) Rudkevich, D. M. Chem.Eur. J. 2000, 6,
2679. (d) Cram, D. J. Science 1983, 219, 1177.
The authors declare the following competing financial
interest(s): A patent has been filed on the use of this method.
(11) (a) Furuya, T.; Kamlet, A. S.; Ritter, T. Nature 2011, 473, 470.
(b) Harper, D. B.; O’Hagan, D. Nat. Prod. Rep. 1994, 11, 123.
(12) A topology-based fluorine fingerprint descriptor has been used to
ACKNOWLEDGMENTS
■
This work was supported by a National Institutes of Health
(National Institute of General Medical Sciences, NIGMS) grant
(GM095843). Y.Z. acknowledges the Shanghai Institute of
Organic Chemistry (SIOC), Zhejiang Medicine, and Pharmaron
for a joint postdoctoral fellowship. G.M. thanks the German
Academic Exchange Service (DAAD, postdoctoral fellowship).
predict the 19F NMR chemical shift: Vulpetti, A.; Landrum, G.; Rudisser,
̈
S.; Erbel, P.; Dalvit, C. J. Fluorine Chem. 2010, 131, 570.
(13) (a) Gramage-Doria, R.; Armspach, D.; Matt, D. Coord. Chem. Rev.
2013, 257, 776. (b) Kotzen, N.; Vigalok, A. Supramol. Chem. 2008, 20,
129. (c) Radius, U.; Attner, J. Eur. J. Inorg. Chem. 1999, 2221.
(d) Guillemot, G.; Solari, E.; Floriani, C.; Rizzoli, C. Organometallics
2001, 20, 607.
We thank Dr. Peter Muller for collecting and solving the X-ray
̈
(14) Chen, F.; Min, Q.-Q.; Zhang, X. J. Org. Chem. 2012, 77, 2992.
(15) Zhao, Y.; Swager, T. M. J. Am. Chem. Soc. 2013, 135, 18770.
(16) (a) Fleming, F. F. Nat. Prod. Rep. 1999, 16, 597. (b) Fleming, F. F.;
Yao, L.; Ravikumar, P. C.; Funk, L.; Shook, B. C. J. Med. Chem. 2010, 53,
7902.
(17) (a) Tomlin, C. D. S. The Pesticide Manual: A World Compendium;
The British Crop Protection Council: Farnham, England, 1997; pp
282−283. (b) Romeh, A. A. J. Environ. Health Sci. Eng. 2014, 12, 38.
(18) (a) Obare, S. O.; De, C.; Guo, W.; Haywood, T. L.; Samuels, T. A.;
Adams, C. P.; Masika, N. O.; Murray, D. H.; Anderson, G. A.; Campbell,
structure data.
REFERENCES
■
(1) (a) Ho, C. K.; Robinson, A.; Miller, D. R.; Davis, M. J. Sensors 2005,
5, 4. (b) Krantz-Rulcker, C.; Stenberg, M.; Winquist, F.; Lundstrom, I.
Anal. Chim. Acta 2001, 426, 217. (c) Du, J.; Hu, M.; Fan, J.; Peng, X.
Chem. Soc. Rev. 2012, 41, 4511. (d) Pejcic, B.; Eadington, P.; Ross, A.
Environ. Sci. Technol. 2007, 41, 6333. (e) Jun, Y.-W.; Lee, J.-H.; Cheon, J.
Angew. Chem., Int. Ed. 2008, 47, 5122. (f) Kobayashi, H.; Ogawa, M.;
Alford, R.; Choyke, P. L.; Urano, Y. Chem. Rev. 2010, 110, 2620.
(g) Domaille, D. W.; Que, E. L.; Chang, C. J. Nat. Chem. Biol. 2008, 4,
168. (h) Lavis, L. D.; Raines, R. T. ACS Chem. Biol. 2008, 3, 142.
(2) (a) Czarnik, A. W. Fluorescent Chemosensor for Ion and Molecule
Recognition; ACS Symposium Series 538; American Chemical Society:
Washington, DC, 1993. (b) de Silva, A. P.; Gunaratne, H. Q. N.;
Gunnlaugsson, T.; Huxley, A. J. M.; McCoy, C. P.; Rademacher, J. T.;
Rice, T. E. Chem. Rev. 1997, 97, 1515. (c) Thomas, S. W., III; Joly, G. D.;
Swager, T. M. Chem. Rev. 2007, 107, 1339. (d) Leray, I.; Valeur, B. Eur. J.
Inorg. Chem. 2009, 2009, 3525. (e) Lange, U.; Mirsky, V. M. Anal. Chim.
Acta 2011, 687, 105.
K.; Fletcher, K. Sensors 2010, 10, 7018. (b) Aragay, G.; Pino, F.; Merkoci
̧ ,
A. Chem. Rev. 2012, 112, 5317.
(19) Le Poul, N.; Douziech, B.; Zeitouny, J.; Thiabaud, G.; Colas, H.;
Conan, F.; Cosquer, N.; Jabin, I.; Lagrost, C.; Hapiot, P.; Reinaud, O.; Le
Mest, Y. J. Am. Chem. Soc. 2009, 131, 17800.
(20) (a) Feng, S. G.; Gamble, A. S.; Philipp, C. C.; White, P. S.;
Templeton, J. L. Organometallics 1991, 10, 3504. (b) Hutchinson, D. J.;
Cameron, S. A.; Hanton, L. R.; Moratti, S. C. Inorg. Chem. 2012, 51,
5070. (c) Li, C.-P.; Chen, J.; Du, M. CrystEngComm 2010, 12, 4392.
(d) Fox, S.; Stibrany, R. T.; Potenza, J. A.; Knapp, S.; Schugar, H. J. Inorg.
Chem. 2000, 39, 4950. (e) Fernandez, E. J.; Laguna, A.; Lopez-de-
Luzuriaga, J. M.; Monge, M.; Montiel, M.; Olmos, M. E.; Rodriguez-
Castillo, M. Dalton Trans. 2009, 7509.
(3) Binghe, W.; Eric, V. A. Chemsosensor: Principles, Strategies, and
Applications; John Wiley & Sons: Hoboken, NJ, 2011.
(4) (a) Diehl, K. L.; Anslyn, E. V. Chem. Soc. Rev. 2013, 42, 8596.
(b) Askim, J. R.; Mahmoudi, M.; Suslick, K. S. Chem. Soc. Rev. 2013, 42,
10690
dx.doi.org/10.1021/ja504110f | J. Am. Chem. Soc. 2014, 136, 10683−10690