Urea-functionalized M4L6 cage receptors: Anion-templated self-assembly and selective guest exchange in aqueous solutions

Article abstract of DOI:10.1021/ja300677w

We present an extensive study of a novel class of de novo designed tetrahedral M₄L₆ (M = Ni, Zn) cage receptors, wherein internal decoration of the cage cavities with urea anion-binding groups, via functionalization of the organic components L, led to selective encapsulation of tetrahedral oxoanions EO₄^n- (E = S, Se, Cr, Mo, W, n = 2; E = P, n = 3) from aqueous solutions, based on shape, size, and charge recognition. External functionalization with tBu groups led to enhanced solubility of the cages in aqueous methanol solutions, thereby allowing for their thorough characterization by multinuclear (¹H, ¹³C, ⁷⁷Se) and diffusion NMR spectroscopies. Additional experimental characterization by electrospray ionization mass spectrometry, UV-vis spectroscopy, and single-crystal X-ray diffraction, as well as theoretical calculations, led to a detailed understanding of the cage structures, self-assembly, and anion encapsulation. We found that the cage self-assembly is templated by EO₄^n- oxoanions (n ≥ 2), and upon removal of the templating anion the tetrahedral M₄L₆ cages rearrange into different coordination assemblies. The exchange selectivity among EO₄^n- oxoanions has been investigated with ⁷⁷Se NMR spectroscopy using ⁷⁷SeO₄^2- as an anionic probe, which found the following selectivity trend: PO ₄^3- ? CrO₄^2- > SO ₄^2- > SeO₄^2- > MoO ₄^2- > WO₄^2-. In addition to the complementarity and flexibility of the cage receptor, a combination of factors have been found to contribute to the observed anion selectivity, including the anions' charge, size, hydration, basicity, and hydrogen-bond acceptor abilities.

Full text of DOI:10.1021/ja300677w

Article
pubs.acs.org/JACS
Urea-Functionalized M L Cage Receptors: Anion-Templated Self-
4
6
Assembly and Selective Guest Exchange in Aqueous Solutions
,
†
†,‡
†
†
§
Radu Custelcean,* Peter V. Bonnesen, Nathan C. Duncan, Xiaohua Zhang, Lori A. Watson,
†
†
†
Gary Van Berkel, Whitney B. Parson, and Benjamin P. Hay
†
Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6119, United States
‡
Center for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6494, United States
§Department of Chemistry, Earlham College, Richmond, Indiana 47374-4095, United States
S Supporting Information
*
ABSTRACT: We present an extensive study of a novel class
of de novo designed tetrahedral M L (M = Ni, Zn) cage
4
6
receptors, wherein internal decoration of the cage cavities with
urea anion-binding groups, via functionalization of the organic
components L, led to selective encapsulation of tetrahedral
n−
oxoanions EO4 (E = S, Se, Cr, Mo, W, n = 2; E = P, n = 3)
from aqueous solutions, based on shape, size, and charge
recognition. External functionalization with tBu groups led to
enhanced solubility of the cages in aqueous methanol
solutions, thereby allowing for their thorough characterization by multinuclear ( H, 13C, 77Se) and diffusion NMR
spectroscopies. Additional experimental characterization by electrospray ionization mass spectrometry, UV−vis spectroscopy,
and single-crystal X-ray diffraction, as well as theoretical calculations, led to a detailed understanding of the cage structures, self-
1
n−
assembly, and anion encapsulation. We found that the cage self-assembly is templated by EO4 oxoanions (n ≥ 2), and upon
removal of the templating anion the tetrahedral M L cages rearrange into different coordination assemblies. The exchange
4
6
n−
77
77
2−
selectivity among EO4 oxoanions has been investigated with Se NMR spectroscopy using SeO as an anionic probe, which
4
3−
2−
2−
2−
2−
2−
found the following selectivity trend: PO4 ≫ CrO4 > SO4 > SeO4 > MoO4 > WO4 . In addition to the
complementarity and flexibility of the cage receptor, a combination of factors have been found to contribute to the observed
anion selectivity, including the anions’ charge, size, hydration, basicity, and hydrogen-bond acceptor abilities.
INTRODUCTION
Research in the area of self-assembled cages has undergone
remarkable progress in the past two decades. These molecular
gen-bonding groups, and highly efficient and selective guest
binding in aqueous environments.
■
3
1
Among the various supramolecular interactions that can be
employed for cage self-assembly, metal−ligand coordination
bonds are relatively strong and directional, thereby offering a
containers, also called capsules or flasks, present a number of
attractive features: (i) they can be easily assembled in one step
by combining relatively simple components via reversible
supramolecular interactions; (ii) they contain confined
chemical environments that can serve as platforms for host−
guest chemistry, molecular recognition, sensing, or catalysis;
and (iii) they are often aesthetically appealing. While research
on self-assembled cages was initially focused mainly on their
synthesis and characterization, more recently the attention has
shifted to the exploration of their chemical and biological
1
a
higher degree of control and predictability. The M4L6
tetrahedron, self-assembled from four metal centers and six
ditopic ligands serving as vertices and edges, respectively,
represents the prototypical coordination cage. Following the
4
discovery of the first M L cage in 1988, many other examples
4
6
5−8
of such tetrahedral assemblies have been demonstrated.
These tetrahedral cages can bear negative or positive charge,
thereby serving as potential hosts for cationic or anionic guests,
respectively. In either case, charge-diffuse monoions are
preferentially encapsulated, as multicharged anions are
associated with large dehydration energies that cannot be
fully compensated by binding inside the typically hydrophobic
1
applications. Toward this goal, parallels are often drawn
between molecular containers and protein receptors or
enzymes. Indeed, some synthetic cage molecules have been
demonstrated to mimic their natural counterparts in their
abilities to bind substrates, stabilize reactive species, or
7b,9
cavities of M L cages.
4
6
Recent work in our group has focused on the study of anion
2
10
accelerate chemical transformations. However, the synthetic
recognition and separation with self-assembled receptors.
molecular containers reported to date rarely combine the two
critical elements characteristic to natural receptors and
enzymes: internal cavities precisely functionalized with hydro-
Received: January 20, 2012
Published: April 30, 2012
©
2012 American Chemical Society
8525
dx.doi.org/10.1021/ja300677w | J. Am. Chem. Soc. 2012, 134, 8525−8534

Journal of the American Chemical Society
Article
Scheme 1. M L Cages Investigated in This Study
4
6
Along this line, we have been particularly interested in the
structures, self-assembly, and anion encapsulation chemistry of
this novel class of cage receptors. We found that, in direct
contrast to previously reported ion-binding M L cages favoring
1
1,12
development of novel metal-based anion receptors
that can
13
function selectively and efficiently in aqueous environments.
4
6
This is a challenging task, as many anions have large free
relatively hydrophobic, neutral or monocharged guests, the cage
receptors described here preferentially encapsulate strongly
hydrophilic, multicharged anions, due to hydrogen-bond
stabilization inside the urea-functionalized cavities.
energies of hydration, and thus are difficult to bind and extract
1
4
from water. Toward this end, cage receptors that can fully
encapsulate the anions, thereby shielding them from the
1
5,16
competitive aqueous environment, are especially promising.
M L cages in particular appeared to us ideal for binding
4
6
EXPERIMENTAL SECTION
■
tetrahedral anions based on the shape match between the cage
host and the anion guest. Indeed, a number of positively
charged M4L6 cages have been demonstrated to bind
Materials and Methods. All chemicals were at least reagent grade
and were used as received from the suppliers without further
purification unless otherwise noted. Solutions were prepared using
volumetric glassware, and calibrated Eppendorf pipets were used to
−
−
−
6,7,8c
tetrahedral anions, such as BF , ClO , or FeCl .
All
4
4
4
these anions, however, have relatively low charge density, and
their binding was generally observed in nonaqueous solvents.
Furthermore, their encapsulation inside the cages’ hydrophobic
cavities is primarily based on optimal filling of the cavity
volume, rather than on specific interactions. We reasoned that
selective and efficient binding of the more hydrophilic,
1
deliver specific volumes of reagent solutions. All H (499.717 MHz),
13
31
C, and P NMR experiments were performed on a Varian VNMRS
5
00 NMR spectrometer at 23 °C using a 5-mm AutoX DB probe, with
spinner turned off, unless otherwise noted. Spectra recorded in CDCl3
were referenced to internal tetramethylsilane (0 ppm) for proton
spectra, and to 77.23 ppm for carbon spectra. Spectra obtained in
1
7
2−
2−
CD OD−D O mixtures were referenced to 3.31 and 49.15 ppm unless
multicharged tetrahedral oxoanions (e.g., SO4 , SeO4
,
3
2
CrO42 , PO4 ) from water requires a new generation of M4L6
cage receptors strategically functionalized with complementary
binding groups. Besides advancing the basic understanding of
self-assembled cages, effective implementation of this approach
may also lead to anion separation applications, as many
tetrahedral oxoanions are relevant to environment and
−
3−
otherwise noted. Low-resolution ESI-MS was performed on 5 μM cage
solutions in 1:1 water−methanol (v/v) using a Sciex API 165
instrument operated at 5 kV, with a sample injection rate of 20 μL/
min. High-resolution electrospray ionization ion mobility time-of-flight
mass spectrometry (ESI-IM-TOFMS) was performed on 2.5 μM
solutions in 1:1 water−methanol (v/v) using a SYNAPT HDMS
(Waters Corp., Manchester, UK) equipped with a NanoMate 100
1
7
energy.
system (Advion BioSciences, Inc. Ithaca, NY). The sample flow rate
was 500 nL/min. A voltage of 1.4 kV and a gas pressure of 0.3 psi were
employed. Spectra were acquired in positive polarity mode, externally
calibrated, and processed using the MassLynx software version 4.1
Guided by de novo computer-aided design, we recently
developed the sulfate-encapsulating tetrahedral cage
6+
18
[
Ni (L1) (SO )] (1a, Scheme 1). Functionalization of the
4 6 4
(
Waters Corp.). UV−vis spectra were collected with a Cary 50 UV−
L1 edges with ureas resulted in an internal cavity decorated
with 12 NH hydrogen-bond donors complementary to the
tetrahedral sulfate anion. This led to exceptionally strong sulfate
binding in water, on a par with sulfate-binding protein.
However, a number of fundamental questions remained
unanswered: (1) Does the cage display, as designed, shape
selectivity for tetrahedral oxoanions? (2) Does the encapsulated
anion act as a template to cage self-assembly? (3) What
fundamental factors govern anion selectivity is this class of cage
receptors? These questions will be addressed in the present
paper based on an extensive study of the prototype cage 1a and
vis spectrophotometer.
77
77
77
Se NMR Experiments. Na SeO and Zn SeO (99.2 atom %
4
4
77
77
Se) were prepared as described in the Supporting Information. All
Se experiments (except diffusion experiments described below) were
acquired at a frequency of 95.4 MHz at 23.0 °C on the Varian VNMRS
77
5
00. All Se chemical shifts were referenced externally to neat Me Se
2
77
(
0.0 ppm) contained in a sealed tube. For the [Zn (tBuL1) ( SeO )]-
4 6 4
7
7
( SeO ) cage complex in 2:1 v/v CD OD−D O, the selenate anion
4
3
3
2
inside the cage is generally observed as a singlet at 1044.2 ppm,
whereas the three selenates outside the cage are generally observed as
a somewhat broader singlet at 1046.5 ppm. The pw90 was measured
to be 13.0 μs at a transmitter power level of 56 for selenate both inside
(8−n)+
the next-generation cages [Zn (tBuL1) (EO )]
(E = S, Se,
4
6
4
and outside the cage. Inversion−recovery (T ) experiments on the
1
Cr, Mo, W, n = 2; E = P, n = 3) (2). Experimental investigation
77
77
1
13
77
[Zn (tBuL1) ( SeO )]( SeO ) cage complex in 2:1 v/v CD OD−
4 6 4 4 3 3
by multinuclear ( H, C, Se) and diffusion NMR spectros-
copies, electrospray ionization mass spectrometry (ESI-MS),
UV−vis spectroscopy, and single-crystal X-ray diffraction
77
D O showed the T for SeO outside and inside the cage to be 6.78
2
1
4
77
±
0.66 and 15.38 ± 1.66 s, respectively. Accordingly, for Se
acquisition parameters, a recycle delay of 120 s and a 10.83 μs (75°)
pulse at a transmitter power level of 56 were used in all exchange
experiments. Using these parameters, typically a 3.4(1):1 ratio for
selenate outside to selenate inside the cage was observed, following the
(
XRD), as well as theoretical analysis by density functional
theory (DFT) calculations and molecular dynamics (MD)
simulations, led in the end to a detailed understanding of the
8
526
dx.doi.org/10.1021/ja300677w | J. Am. Chem. Soc. 2012, 134, 8525−8534

Journal of the American Chemical Society
Article
collection of at least 500 transients, and the average of at least five
separate integrations.
average of at least five integrations) was compared to the baseline
ratio.
Diffusion NMR Spectroscopy. Diffusion NMR measurements
were carried out at 298 K on a Bruker Avance 400 spectrometer using
a gradient amplifier with a maximum gradient of 54.1 G/cm. Reference
RESULTS AND DISCUSSION
Anion-Templated Self-Assembly of [Ni (L1) (EO )]-
NO3)8−n Cages (E = S, P; n = 2, 3). A solution containing
■
1
77
4
6
4
H and Se (76.31 MHz) NMR spectra of the samples were taken
(
prior to diffusion experiments. The pulse length used for observation
of the selenium signal was 14.0 μs for a 90° pulse with a recycle delay
of 45s. The Stebpgp1s (STimulated Echo BiPolar Gradient Pulse)
program from Bruker Biospin was used for the DOSY NMR using
gradients varied linearly from 5% up to 95% in 16 steps, with 64 scans
L1, Ni(NO ) , and Na SO in a 6:4:1 molar ratio in H O/
3
2
2
4
2
MeOH led upon slow evaporation to crystallization of
Ni (L1) (SO )](NO ) (1a(NO ) ). Single-crystal XRD
[
4
6
4
3
6
3 6
analysis revealed the formation of a cage with quasitetrahedral
symmetry (C true symmetry), with slightly different Ni−Ni
per step for 1H, and 300 scans per step for 77Se. For the 1
H
3
measurements, the diffusion time (Δ) was set at 200 ms, and the
gradient length (δ) was set at 3.6 ms. For the 77Se measurements, Δ
and δ values were in the range of 0.4−2.5 s and 3.6−9.6 ms,
respectively. The diffusion coefficients were determined using the
Simfit algorithm from Bruker Biospin.
distances of 11.69 and 11.70 Å, and Ni−Ni−Ni angles of 59.96,
60.00, and 60.02°. This minor distortion from the ideal
tetrahedral symmetry is likely a consequence of packing forces
in the crystal. However, the observed perturbation in this case is
significantly smaller compared to the previously reported cage
NMR Analysis of Cage Self-Assembly. All stock solutions of
18
1
a(SO ) . While the crystal packing in the previous structure
4 3
tBuL1 at 30−33 mM in CD OD, Zn(NO ) ·4H O at 41−44 mM in
3
3
2
2
was dominated by relatively strong interactions between the
external sulfate anions and the cages, the extra-cage nitrate
anions in the present structure are highly disordered, resulting
in a more loosely packed crystal. Once again, as predicted by
molecular design, the encapsulated sulfate accepts 12 hydrogen
bonds from the six urea groups functionalizing the cage cavity,
D O, and the sodium salt of the anion of interest at 27−65 mM in
2
D O were prepared and used within a 2-week time frame. The
2
experiments were designed such that 0.500 mL of tBuL1 in CD OD is
3
first delivered to the NMR tube, followed by addition of the
Zn(NO ) ·4H O in D O solution (0.240 to 0.250 mL, the exact
3
2
2
2
volume depending on the concentration of the stock solution), such
19
that the CD OD−D O is close to a v/v ratio of 2:1. The sodium salts
with each urea binding an edge of the sulfate. The observed
N−H···O hydrogen bond distances and angles are very similar
with the corresponding values in 1a(SO ) , ranging between
3
2
were at a concentration in D O such that 40−105 μL quantities would
2
deliver 1 equiv of anion. An example of a typical NMR experiment
4
3
designed to probe the cage self-assembly is as follows: to 16.1 μmol (6
2
.04 and 2.08 Å, and 165.2 and 174.2°, respectively.
equiv) of tBuL1 in 0.500 mL of CD OD in a 5-mm tube was added
3
Having established the sulfate encapsulation by the Ni (L1)
4
6
1
0.7 μmol (4 equiv) of Zn(NO ) ·4H O in 0.241 mL of D O, and the
3 2 2 2
cage, we next turned our attention to investigating the ability of
the cage to encapsulate other tetrahedral oxoanions. The
phosphate anion, with size and hydrogen-bonding preferences
similar to sulfate, was selected as the next target. Slow
proton spectrum was recorded to verify the formation of the
intermediate coordination complex between zinc nitrate and tBuL1.
After a minimum of 5 min, 1 equiv (2.69 μmol) of Na SeO contained
2
4
in 48.3 μL of D O was added, and after mixing for 20 s, the proton
2
spectrum was recorded within 3−4 min. Cages of the form
evaporation of a H
O/MeOH solution containing L1, Ni-
2
[
Zn (tBuL1) (EO )](EO ) (E = S, Se) were also directly prepared
4 6 4 4 3
(NO ) , and NaH PO in a 6:4:1 molar ratio resulted in
3 2
2
4
by adding the stoichiometric amount of ZnEO dissolved in D O to
crystallization of [Ni (L1) (PO )](NO ) (1b(NO ) ), as
4
2
4 6 4 3 5 3 5
0
.500-mL solutions of tBuL1 in CD OD. For the more basic
3
determined by single-crystal XRD (Figure 1). This structure
phosphate anion, the added D O solutions of Zn(NO ) ·4H O and
2
3
2
2
Na HPO were buffered at pH 9.5 with borax (0.100 M Na B O
2
4
2
4
7
9
9.5%).
77Se NMR Analysis of Anion Exchange Using 77
Se-Enriched
Selenate. For each exchange experiment, a solution of the
77
77
[
Zn (tBuL1) ( SeO )]( SeO ) cage complex was first prepared by
4 6 4 4 3
77
adding 118.0 μL of 88.9 mM stock solution of Zn SeO in D O to
4
2
0
.500 mL of a 31.45 mM solution of tBuL1 in CD OD in 5-mm NMR
3
tubes. After mixing and allowing to age for at least 30 min, the proton
spectrum was recorded to ensure that the cage complex had been
cleanly generated. Next, a specific volume of 0.100 M Na B O in D O
2
4
7
2
was added (between 92 and 113 μL, the precise volume depending on
the concentration of the exchanging anion solution), and the resulting
solution was thoroughly mixed. The proton spectrum was then
recorded, followed by an overnight 77Se acquisition (minimum of 500
transients, using the acquisition parameters described above) to
determine the baseline ratio of selenate outside to selenate inside the
cage (average of at least five integrations.) To perform the anion
exchange, a precise volume between 19 and 40 μL of freshly (≤3 h)
prepared solutions of anhydrous Na SO , Na CrO ·4H O,
Figure 1. Crystal structure of 1b(NO ) . (a) Tetrahedral Ni (L1)
6
cage encapsulating a PO4 anion. (b) Phosphate binding by 6 urea
groups forming 12 NH···O hydrogen bonds.
3
5
4
3
−
2
4
2
4
2
is essentially isomorphous with 1a(NO ) , containing a
3
6
Na MoO ·2H O, Na WO ·2H O, or Na HPO ·2H O in 0.100 M
2
4
2
2
4
2
2
4
2
quasitetrahedral Ni (L1) cage with observed Ni−Ni distances
4
6
Na B O in D O was added to the tube, and the resulting solution was
2
4
7
2
of 11.63 and 11.69 Å, and Ni−Ni−Ni angles of 59.83, 60.00,
and 60.34°. The encapsulated phosphate is fully deprotonated,
despite the fact that under the acidic conditions used for
crystallization (pH ≈ 4), the phosphate anion exists
preponderantly in its diprotonated form. This suggests that
the hydrogen-bonding microenvironment inside the cage has a
thoroughly mixed. In the cases of chromate and hydrogen phosphate,
equiv of anion was added; for the other anions, 2 equiv of anion was
added. The CD OD/D O volume ratio of the final solutions was
1
3
2
nominally 2:1 in all experiments. Proton spectra were recorded at least
twice over a period of 20 min to observe any changes, and then the
7
7
Se acquisition was recorded overnight, with a minimum of 600
3
−
transients being collected. The peaks for the selenate outside and
selenate inside the cage were then integrated, and the ratio (as the
strongly stabilizing effect on PO
4
, essentially shifting its
2
0
basicity by a few orders of magnitude. As with the sulfate cage
8
527
dx.doi.org/10.1021/ja300677w | J. Am. Chem. Soc. 2012, 134, 8525−8534

Journal of the American Chemical Society
Article
analogue, and in agreement with the molecular design, the
phosphate orientation is inverted with respect to the tetrahedral
cage, with the P−O bonds pointing to the centers of the four
Ni (L1)3 in crystalline form and prove unequivocally its
2
formation by XRD analysis.
Considering that the latest data suggests the Ni (L1) cage
8
6
+
4
triangular faces rather than to the Ni(bpy) vertices. The
does not persist in solution after the removal of the
encapsulated sulfate, a reevaluation of the sulfate binding
constant is in order here. Based on competition experiments
3
phosphate accepts 12 urea NH hydrogen bonds with relatively
linear geometry (N−H···O angles range between 165.8 and
2
+
1
74.2°), and N−H···O contact distances of 1.95−2.06 Å, which
with Ba , the apparent sulfate association constant (K ) had
app
6
−1 18
are slightly shorter than those observed for the sulfate analogue.
This structure thus represents a rare example of a fully
deprotonated phosphate coordinatively saturated by 12 hydro-
been estimated around (6 ± 1) × 10 M . Since it now
8+
appears that the Ni (L1)6 cage rearranges into other
4
4+
3
coordination assemblies, such as Ni
(L1)
2
(Scheme 2), the
2
1
gen bonds.
While the tetrahedral cages form in the presence of the
Scheme 2. Putative Reaction Equilibria Involving the
SO42 or PO4 anions, spectroscopic data indicates that the
−
3−
6+
a
[Ni (L1) (SO )] Cage in Aqueous Solutions
4
6
4
empty Ni (L1) cage does not persist upon removal of the
4
6
anionic guest, but it rearranges into other coordination
4n+
1h,22
assemblies with the [Ni (L1)3]n
stoichiometry.
The
2
visible absorption spectrum of a solution containing 6 equiv of
L1 and 4 equiv of Ni(NO ) , but no sulfate, is virtually
3
2
indistinguishable from that of the sulfate-containing cage
a(NO3)6 in H O/MeOH (1:1), with absorption maxima
1
2
observed at 522 and 795 nm (Supporting Information). These
values are very similar to the absorption maxima of 519 and 789
2
+
2+
nm reported for Ni(bpy)3 , thereby confirming that all Ni
cations in solution are chelated by three bpy groups from L1.
However, these data cannot distinguish among the various
4n+
[
Ni (L1) ]
stoichiometries possible. More informative was
2
3 n
ESI-MS analysis of H O/MeOH (1:1) mixtures containing L1
2
and Ni(NO ) in a 3:2 ratio, as required for the formation of
3
2
4n+
the [Ni (L1) ]
coordination assemblies. Specifically, high-
2
3 n
resolution ESI-IM-TOFMS identified a relatively intense peak
4+
at m/z = 326.6 corresponding to the Ni (L1)3 ion, as
2
indicated by the good match between the observed and
calculated isotopic distributions (Figure 2).
a
All structures were obtained by taking representative snapshots from
the MD simulations (Supporting Information). C, H, N, O, Ni, and S
atoms are shown in cyan, white, blue, red, silver, and yellow,
respectively.
apparent constant can be defined as Kapp = K K(SO42−), where
r
K is the equilibrium constant for the rearrangement of 2 equiv
r
4+
8+
2−
of Ni (L1)3 into Ni (L1)6 , and K(SO4 ) is the actual
sulfate binding constant by the putative Ni (L1)
2
4
8
6
+
cage.
4
Unfortunately, the value of K remains unknown at this time,
r
2−
which precludes us from obtaining an exact value for K(SO ).
However, considering that no trace of the Ni (L1) ion could
4
8
6
+
4
be detected by mass spectrometry, we can assume the
concentration of the empty cage in solution is negligible,
2
−
which implies that K ≪ 1. It thus follows that K(SO ) ≫ (6
r
4
6
−1
±
1) × 10 M . This value thus represents a lower limit
estimate of the association constant for sulfate binding by
Figure 2. High-resolution ESI-MS: overlay of the calculated (red) and
4+
observed (black) peaks for the Ni (L1)3 ion.
8+
2
Ni (L1)6
4
.
A n i o n - T e m p l a t e d S e l f - A s s e m b l y o f
Furthermore, ion mobility analysis, a technique that allows
ions with the same m/z to be separated based on their size
difference, found no additional ionic species with the mass-to-
charge ratio of 326.6 (Supporting Information). Thus, these
[Zn (tBuL1) (EO )](NO ) Cages (E = S, Se, Cr, Mo, W,
4
6
4
3 8−n
n = 2; E = P, n = 3). While the prototype L1 ligand and the
6+
corresponding [Ni (L1) (SO )] cage provided the proof of
4
6
4
principle for de novo design of self-assembled cage receptors for
encapsulation of tetrahedral oxoanions, the limitations of this
system quickly became apparent. In particular, the presence of
the paramagnetic Ni2+ centers rendered the use of NMR
spectroscopy impractical, thereby precluding a detailed study of
the cage structure and anion selectivity in solution. Our
results provide supporting evidence for the tetrahedral Ni (L1)6
4
cage rearrangement into Ni (L1) upon removal of the
2
3
templating anion. A similar guest-controlled interconversion
between an M L tetrahedral cage and an M L helicate had
4
6
2 3
2
3
previously been observed by Raymond et al. Unfortunately,
despite numerous attempts, we were not able to isolate the
2
+
2+
attempts to self-assemble L1 with diamagnetic Zn or Fe
8
528
dx.doi.org/10.1021/ja300677w | J. Am. Chem. Soc. 2012, 134, 8525−8534

Journal of the American Chemical Society
Article
metal cations resulted in the formation of intractable
which will be represented here generically as
2
4
precipitates. Due to these limitations, we sought to prepare
a derivative of L1 that would self-assemble with the diamagnetic
[Zn (tBuL1) ] (NO ) .
2
3 n
3 4n
To probe the formation of the tetrahedral
2+
n−
Zn into soluble tetrahedral cages amenable to NMR analysis.
We hypothesized that a derivative of L1 containing tert-butyl
groups attached to the bpy moieties would display improved
solubility. We therefore proceeded with the synthesis of this
second-generation ligand (Supporting Information). The
resulting ligand (tBuL1) is indeed substantially more soluble
than L1 in polar organic solvents such as methanol, enabling us
to monitor the cage self-assembly by NMR.
[Zn (tBuL1) (X )](NO )
cages (2(NO ) ), we added 1
4
6
3 8−n 3 6
equiv of various NaX or Na X salts in D O to NMR tubes
2
2
containing the [Zn (tBuL1) ] (NO ) complex in 2:1 v/v
2
3 n
3 4n
CD OD−D O. In the cases where X was a dinegatively charged
3
2
2
−
tetrahedral oxoanion (EO4 ; M = S, Se, Cr, Mo, W), in the
time needed to add the salt solution to the NMR tube, shake it,
and acquire the proton NMR spectrum (typically 4−5 min after
addition), a new set of sharp resonances appeared, indicating
the formation of a high-symmetry complex (Figure 3c). The
proton and carbon NMR spectra in this series of complexes are
very similar, with only very slight differences in the chemical
shifts among them (Supporting Information). However, there
are noteworthy changes in the chemical shifts when compared
to those of the initial [Zn (tBuL1) ] (NO ) complex.
When dissolved in 2:1 v/v CD OD−D O, the free tBuL1
3
2
ligand features a broad NMR singlet at δ 4.45 attributable to
the methylene protons. Upon addition of six equivalents of zinc
1
nitrate in D O to 4 equiv of tBuL1 in CD OD, the H NMR
2
3
revealed the formation of a stable but nonsymmetrical complex
within a few minutes (Figure 3b). The broad singlet at δ 4.45
2
3 n
3 4n
Particularly in the aromatic region, there are now six sharp
resonances corresponding to the six aromatic protons. The
−
CH − group displays a typical AB pattern consisting of two
2
symmetrical doublets centered at δ 4.71 and 3.97, indicative of
the diastereotopic relationship of the two methylene protons.
These spectra are consistent with the formation of tetrahedral
2
−
[Zn (tBuL1) (EO )](NO ) cages templated by the EO
4 6 4 3 6 4
anions. The 2− charge of the tetrahedral anions appears to
be critical for the cage formation, as no cage was templated by
the ClO4− or ReO4− tetrahedral monoanions, with their NMR
spectra remaining virtually unchanged compared to that of the
[
Zn (tBuL1) ] (NO ) complex even after several days. Other
2 3 n 3 4n
−
−
−
−
−
monoanions of different shapes, like F , Cl , Br , I , NO
(
3
−
−
−
−
−
excess), BF , PF , CH CO , CH SO , or CF SO , also
4 6 3 2 3 3 3 3
did not form the cage. Also, no cage formation was observed in
the presence of dinegative anions with shapes other than
2
−
tetrahedral, such as the trigonal planar CO
and the
3
pyramidal SeO32 or SO3
−
2− 25
.
Therefore, it is evident that
there is a high degree of shape and charge selectivity in the cage
formation, with only tetrahedral oxoanions with charges greater
than 1− being able to template the cage self-assembly.
The phosphate anion was also found to template the
formation of the tetrahedral cage under buffered conditions
(
borax, pH ≈ 9). Thus, the [Zn (tBuL1) ] (NO ) complex
2
3 n
3 4n
was generated as before, but using about two times more
concentrated Zn nitrate solution in D O. To that solution was
2
added a nearly equal volume of 100 mM Na B O in D O. The
2
4
7
2
[
Zn (tBuL1) ] (NO ) complex appeared to remain un-
2 3 n 3 4n
changed in the resulting alkaline solution containing 15.7
1
mM of the Na B O buffer, as judged by H NMR. In this case,
2
4
7
upon the addition of 1 equiv of Na HPO (together with 100
2
4
mM Na B O in D O), resonances consistent with the
2
4
7
2
[
Zn (tBuL1) (PO )](NO ) cage were observed within 5
4 6 4 3 5
1
min (Supporting Information). It was also noted, however,
that resonances due to the free tBuL1 ligand also appeared
within the same time, with an observed ratio of free tBuL1 to
cage-bound tBuL1 of approximately 11(±1):89(±1). In the
same time, a small amount of flocculant precipitate formed,
which was likely zinc phosphate based on its very low solubility
(Ksp = 9.0 × 10− ). We thus surmise that the free tBuL1
observed by NMR could be due to the loss of Zn by
competitive precipitation of zinc phosphate. It appears that
the system reached the equilibrium, as there was no apparent
increase in the amount of precipitate or free ligand relative to
the cage-bound ligand over the course of 2 days.
Figure 3. Aromatic regions of the H NMR spectra of (a) free tBuL1,
(
b) [Zn (tBuL1) ] (NO ) , and (c) [Zn (tBuL1) (SO )](NO ) in
2 3 n 3 4n 4 6 4 3 6
2
:1 CD OD−D O.
3
2
corresponding to the −CH − of the free ligand tBuL1 split into
2
33
two unsymmetrical broad doublets centered at δ 4.29 and 4.07,
as the two methylene protons become diastereotopic. All of the
previously sharp aromatic resonances of the free ligand shifted
and broadened substantially. Thus, it is apparent that some low-
symmetry coordination complex or mixture of complexes is
formed between the six tBuL1 ligands and the four Zn cations,
26
8
529
dx.doi.org/10.1021/ja300677w | J. Am. Chem. Soc. 2012, 134, 8525−8534

Journal of the American Chemical Society
Article
Table 1. ESI-MS Peaks Observed (Calculated) for the [Zn (tBuL1) (EO )](NO ) (EO ) Cages from 5 μM Solutions in 1:1
4
6
4
3 n
4 m
(
v/v) H O/MeOH
2
m/z
E
n = 0, m = 0
n = 0, m = 1
n = 1, m = 0
n = 2, m = 0
n = 3, m = 0
S
568.2 (568.2)
575.9 (576.1)
571.6 (571.6)
578.7 (578.9)
−
876.1 (876.0)
899.9 (899.8)
886.1 (886.3)
908.2 (908.1)
−
694.3 (694.3)
703.7 (703.7)
698.1 (698.3)
707.0 (707.1)
−
883.3 (883.4)
1198.4 (1198.5)
Se
Cr
Mo
W
P
895.0 (895.1)
1213.5 (1213.8)
−
−
−
−
−
−
−
−
681.6 (681.5)
−
867.3 (867.6)
Complementary evidence for the formation of the
Table 2. Pertinent Geometric Parameters for the Crystal
Structures of 2a−c
n+
[
Zn (tBuL1) (EO )] tetrahedral cages was provided by ESI-
4
6
4
MS. Thus, analysis of solutions containing 6:4:1 molar ratios of
2
a (E = S)
2b (E = Cr)
2c (E = Mo)
tBuL1, Zn(NO ) , and Na EO (or Na HPO ) in 1:1 (v/v)
3
2
2
4
2
4
d(Zn−Zn) (Å)
11.79−12.34
av. 11.95
11.76−12.28
av. 11.97
11.74−12.30
av. 11.92
H O/MeOH displayed peaks corresponding to various
2
(
8−n−2m)+
[
1
Zn (tBuL1) (EO )](NO ) (EO )
cations (Table
4
6
4
3
n
4 m
d(NH···O) (Å)
1.97−2.27
av. 2.09
1.98−2.28
av. 2.07
1.98−2.39
av. 2.10
). For example, in the case of sulfate, peaks were observed
6
+
at m/z values corresponding to [Zn (tBuL1) (SO )] ,
4
6
4
4
+
5+
∠(N−H···O) (deg)
∠(E−O···H) (deg)
151.1−175.3
av. 162.3
145.7−175.0
av. 161.4
143.6−174.1
av. 159.2
[
[
(
(
Zn (tBuL1) (SO )](SO ) , [Zn (tBuL1) (SO )](NO ) ,
4 6 4 4 4 6 4 3
4
+
Zn (tBuL1) (SO )](NO ) , and [Zn (tBuL1) (SO )]-
4
6
4
3
2
4
6
4
3+
n−
96.1−149.3
av. 119.7
92.2−145.0
av. 115.9
81.6−145.1
av. 112.5
NO3)3 (Supporting Information). The other EO4 anions
n = 2, 3) showed similar results, with the exception of WO42−
,
which displayed no peaks corresponding to the cage. We
assume that in this case the cage either does not persist at the
very low concentration of the solution under study (5 μM), or
of the oxoanion, and the sixth urea binding an O vertex (Figure
b). The hydrogen-bonding parameters are very similar in the
27
it decomposes under the electrospray ionization conditions.
4
No cage formation was evident from dinegative anions of
three structures, despite the significant difference in the anion
sizes. For example, the shortest average NH···O contact
distance of 2.07 Å is observed for chromate in 2b, but the
corresponding contact distances for sulfate and molybdate in 2a
and 2c are only 0.02 and 0.03 Å longer, respectively. Similarly,
the average hydrogen-bond donor angles (∠N−H···O) vary
only by 3.1° among the three structures. However, there are
2−
2−
shapes different than tetrahedral, such as SO3 or CO3
.
Single crystals suitable for XRD were obtained for the
[
Zn (tBuL1) (EO )](NO ) tetrahedral complexes (E = S
4
6
4
3 6
(2a), Cr (2b), and Mo (2c)), thereby allowing for a detailed
structural characterization of the cage geometry and anion
encapsulation. The three cages have isomorphous structures,
with the monoclinic P2 /n space group. The structure of 2a is
1
larger differences among the average hydrogen-bond acceptor
illustrated in Figure 4, and relevant geometric parameters for
2−
angles (E−O···H), which vary from 119.7° for SO to 112.5°
4
2a−c are listed in Table 2.
2−
for MoO4 . Overall, these X-ray data indicate that the
[
Zn (tBuL1) (EO )](NO ) tetrahedral cages are structurally
4 6 4 3 6
quite flexible, distorting their frames to accommodate the
different sized EO42 anions and optimize their hydrogen
bonding by the urea groups.
−
Diffusion NMR spectroscopy was carried out to obtain
further information about the identity of the various
coordination assemblies formed in solution. The translational
self-diffusion coefficient (D) for the [Zn (tBuL1) (SO )]-
4
6
4
(
NO ) cage (2a) in 2:1 v/v CD OD−D O, measured by the
3
6
3
2
1 −10 2
stimulated echo diffusion H NMR, was 1.01(7) × 10 m /s.
For comparison, at 1.89(16) × 10− m /s, the corresponding
D value measured for tBuL1 was almost twice as large.
Furthermore, for 2a, a 2D DOSY (diffusion ordered spectros-
copy) plot showed that all the cage peaks have the same
diffusion coefficient (Supporting Information), indicating the
presence of a single supramolecular aggregate in solution. Based
on these data, a hydrodynamic radius (r) of 13.6 ± 1.4 Å was
estimated for 2a using the Stokes−Einstein equation (eq 1),
where k is the Boltzmann constant, T the temperature, and η
10
2
Figure 4. Crystal structure of 2a. (a) Tetrahedral Zn (tBuL1) cage
4
6
encapsulating a SO42 anion. (b) Sulfate binding by six urea groups,
−
forming 12 NH···O hydrogen bonds.
The three cages have distorted tetrahedral shapes and almost
identical sizes, with the average Zn−Zn edge length varying
only slightly among the cages, from 11.92 to 11.95 Å. Each of
28
the solvent viscosity. This experimental value for r is in good
agreement with the 14.2(2) Å calculated value for the
tetrahedral cage [Zn (tBuL1) (SO )], obtained from MD
2
−
the three EO4 anions is encapsulated in the center of the cage
via 12 hydrogen bonds from six urea groups. However, the
anion coordination is less symmetrical than in the Ni cage
analogues, with five urea groups each binding an O−E−O edge
4
6
4
29
simulations.
D = k T/6πηr
(1)
B
8
530
dx.doi.org/10.1021/ja300677w | J. Am. Chem. Soc. 2012, 134, 8525−8534

Journal of the American Chemical Society
Article
Diffusion NMR spectroscopy is often used as a powerful tool
for probing encapsulation phenomena in molecular cages, by
demonstrating that the encapsulated guest has a similar
30
diffusion coefficient as the cage host. This could be achieved
7
7
2−
in our case by employing SeO4 as an NMR-active anionic
probe. Se is a spin 1/2 nucleus with generally narrow
linewidths over a wide chemical shift range. However, its
receptivity at natural abundance of 7.6% is moderately low
about 3.1 times greater than for C), and its spin−lattice
relaxation times tend to be long. Initial Se NMR testing was
77
1
3
(
77
performed on the [Zn (tBuL1) (SeO )](SeO ) cage using
4
6
4
4 3
natural abundance zinc selenate in 2:1 v/v CD OD−D O.
3
2
Under these conditions it was possible to detect separate peaks
for selenate outside (ca. 1046 ppm) and inside (ca. 1044 ppm)
the cage in a roughly 3:1 ratio, although the signal-to-noise ratio
7
7
was quite low. We therefore prepared Zn SeO from elemental
4
7
7
Se that was enriched to 99.2 atom % in Se (Supporting
Information), and subsequently converted it into the
Zn (tBuL1) ( SeO )]( SeO ) cage (2-Se).
7
7
77
[
4 6 4 4 3
1
H DOSY of 2-Se in 2:1 v/v CD OD−D O yielded a
3
2
−10
2
diffusion coefficient D of 0.98(8) × 10 m /s for this cage,
which, as expected, is close to the value found for the analogous
7
7
31
sulfate-encapsulating cage 2a. Se DOSY of 2-Se in the same
solvent mixture yielded a very similar diffusion coefficient of
.94(5) × 10− m /s for one of the two distinct SeO4
10
2
77
2−
0
anions observed, thereby providing strong evidence for its
encapsulation inside the cage. On the other hand, a significantly
−
10
2
larger diffusion coefficient of 1.8(2) × 10
m /s was
7
7
2−
measured for the SeO4 anions outside the cage. For
comparison, the diffusion coefficient of “free” selenate,
7
7
measured from a solution of Zn SeO in the same solvent
mixture, is 2.7 × 10
4
−10
2
m /s, indicating some degree of ion
pairing between the 6+ cage cation and the outside selenate
anions.
(
8−n)+
Anion Exchange by the [Zn (tBuL1) (EO )]
Cages
4
6
4
7 7
SeO4 2 − by CrO
2 −
Figure 5. Exchange of
[
in the
4
(
[
E = S, Se, Cr, Mo, W, n = 2; E = P, n = 3). Although the
77
77
77
Zn (tBuL1) ( SeO )]( SeO ) cage, monitored by Se NMR: (a)
(8−n)+
4 6 4 4 3
Zn (tBuL1) (EO )]
cages do not persist in the absence
of the encapsulated anions, thereby preventing the measure-
ment of the absolute values for the anion binding constants,
competitive exchange of the EO4 anions offers an alternative
way for measuring the relative anion encapsulation selectivity.
In the previous section we had shown that Se NMR is a viable
method for probing anion encapsulation with this class of self-
assembled cages. In this section, we demonstrate that the same
technique may be effectively employed for quantitative
measurement of anion exchange selectivity.
The exchange experiments were performed by first preparing
solutions of the [Zn (tBuL1) ( SeO )]( SeO ) cage complex
that were buffered by addition of small volumes (92−113 μL)
of 0.100 M Na B O in D O, to bring the pH of the solution to
around 9. Next, a Se NMR spectrum was collected to obtain
the [ SeO ] /[ SeO ] baseline before adding 1 or 2 equiv
of Na SO , Na CrO , Na MoO , or Na WO in 0.100 M
Na B O in D O. After addition of the exchanging anion
before and (b) after the addition of 1 equiv of CrO42−.
4
6
4
n−
77
77
Based on the initial and the equilibrium [ SeO ] /[ SeO ]
4
out
4 in
ratios measured by NMR, K was then calculated according to
7
7
ex
eq 3 (see Supporting Information for details of the
calculations):
2
−
6+
[77SeO ]eq
2−
[Zn (tBuL1) (EO )]
4
6
4
4
eq
Kex
=
7
7
6+
2−
[EO4 ]eq
eq
[
Zn (tBuL1) ( SeO )]
4
6
4
(3)
7
7
77
4
6
4
4 3
77
77
The initial and equilibrium [ SeO ] /[ SeO ] ratios
4
out
4 in
measured by NMR, and the K values corresponding to
ex
2
4
7
2
2−
selenate exchange by the EO4 anions (E = S, Cr, Mo, W) are
3
2
77
shown in Table 3. According to these data, the relative cage
affinity for the EO anions is CrO4 > SO4 > SeO4
MoO4 > WO4 . Using a similar approach, we also measured
the PO4 /SeO4 anion exchange with the cage, and obtained
an apparent K = 15.82. This exchange equilibrium constant is
denoted as apparent, as under these conditions there are other
coexisting protonation equilibria involving phosphate. Though
the exact value of K for phosphate cannot be determined from
77
77
4
out
4 in
2−
2−
2−
2−
>
4
2−
2−
2
4
2
4
2
4
2
4
3−
2−
2
4
7
2
solution, the proton NMR spectrum was recorded, followed by
an overnight Se acquisition to determine the equilibrium
77
ex
7
7
77
[
SeO4]out/[ SeO ] (Figure 5). The anion exchange
4 in
reactions under consideration are shown in eq 2.
ex
the available data, we can estimate that log Kex > 4.5
7
7
6+
2−
3
−
[
Zn (tBuL1) ( SeO )] + EO4
4
Kex
6
4
(Supporting Information). Thus, the binding of PO
is
4
much stronger compared to the EO42 anions, as expected
−
2
−
6+
77SeO
4
2−
HooI [Zn (tBuL1) (EO )]
+
(2)
4
6
4
based on the enhanced electrostatic attraction of the 8+ cage
8
531
dx.doi.org/10.1021/ja300677w | J. Am. Chem. Soc. 2012, 134, 8525−8534

Journal of the American Chemical Society
Article
Table 3. 77SeO42− Exchange by EO (E = S, Cr, Mo, W) in
2−
other anions. On the other hand, both the average hydrogen-
4
7
7
77
the [Zn (tBuL1) ( SeO )]( SeO ) Cage, Measured by
bond donor (N−H···O) and acceptor (E−O···H) angles are
4
6
4
4 3
77
2− 35
Se NMR
most favorable for SO4
.
These geometric differences,
however, are too small to fully account for the observed
[77
SeO ]out/
6+
4
7
7
anion exchange selectivity in the [Zn (tBuL1) (EO )] cage
4
6
4
[
SeO4]in
series, and other factors must play significant roles.
no. equiv added per
a
b
c
d
The calculated gas-phase free energy of reactions for the
anion
equiv cage
initial
final
Kex
2−
binding of the EO4 anions by six urea groups decreases when
SO 2
−
2
1
2
2
3.522
3.569
3.485
3.561
13.65
15.42
4.115
3.670
7.21
40.10
0.27
4
going from the smallest sulfate to the largest tungstate anion.
This is expected based on the corresponding decrease in
anions’ charge density with the increase in their size. The
observed anion exchange selectivity generally follows the same
trend, which is consistent with an anti-Hofmeister bias. Anti-
Hofmeister selectivity in anion binding from water corresponds
to a more stabilizing environment for the anions in the binding
cavity of the receptor relative to the external aqueous
CrO 2
−
4
MoO42
−
WO42
−
0.04
a
In all cases, after addition of the anion solution, the total D O volume
was 0.250 mL and the total CD OD volume was 0.500 mL. From the
average of at least five integrations for spectra obtained after a
minimum of 500 transients; maximum uncertainty in all integrals is
5%. From the average of at least 5 integrations for spectrum
obtained after a minimum of 600 transients; all solutions were clear
with no precipitates; maximum uncertainty in all integrals is ±5%. Kex
2
b
3
c
±
14
environment. Such a behavior is expected for the urea-
6+
d
functionalized [Zn (tBuL1) (EO )] cages, whose binding
4
6
4
calculated with eq 3; uncertainty is ±14%.
cavities offer a highly complementary, strongly stabilizing
environment to tetrahedral EO42 anions. The only anion in
the series that does not follow the anti-Hofmeister selectivity
trend is CrO4 . In this regard, it should be noted that
chromate is by far the strongest base in the series, based on the
−
for the higher-charged phosphate, as well as the greater basicity
of this anion.
2−
To rationalize the observed selectivity trend among the
−
known pK values of HEO . It is most instructive to compare
a
4
2−
EO4 anions, several factors need to be taken into account.
First, regarding the cage host, the important factors are its
rigidity and the complementarity of the binding cavity for the
various encapsulated anions. Second, concerning the anionic
guests, there are some intrinsic properties associated with the
anions, such as their size (expressed here as the E−O bond
2−
2−
CrO4 with SeO4 , which are essentially identical in size, and
have similar free energies of hydration. However, CrO
preferred over SeO4 by a factor of 40, which could be
attributed to the much stronger basicity of the former. A similar
comparison can be made between MoO4 and WO4 , which
2−
is
4
2−
2−
2−
2−
have virtually identical size, but the more basic MoO
is
4
−
33
length), their basicity (pK of HEO ), and their free energy
2−
a
4
preferred over WO4 by a factor of 7. On the other hand,
3
4
of hydration (Table 4). Also included in Table 4 are the gas-
2−
2−
between SO4 and SeO4 , which have similar basicity, the
2−
phase thermodynamic parameters for EO4 binding by six urea
groups, calculated using DFT at the B3LYP level of theory;
these parameters define the intrinsic hydrogen-bond-accepting
2−
smaller SO4 is bound more strongly.
2−
CONCLUSIONS
abilities of the EO4 anions toward the urea functional groups.
All these intrinsic factors define the baseline selectivity, on top
of which operate any existing host−guest recognition
phenomena.
■
A novel class of M L tetrahedral cages acting as selective anion
4
6
receptors in competitive aqueous solutions has been thoroughly
investigated experimentally by multinuclear NMR spectrosco-
py, DOSY, ESI-MS, UV−vis, and X-ray crystallography, as well
as theoretically by DFT calculations and MD simulations.
Internal functionalization of the cages’ cavities with six urea
anion-binding groups, guided by de novo computer-aided
design, resulted in selective encapsulation of tetrahedral
As indicated by the X-ray structural analysis of 2a−c, the
tetrahedral cage has a high degree of complementary for
tetrahedral oxoanions, manifested by the formation of 12
NH···O hydrogen bonds. There is, however, relatively little
difference in binding geometry among the three EO42− anions
analyzed (Table 3), despite their significantly different sizes.
This can be explained by the high flexibility of the cage, which−
distorts its structure to accommodate the different sized EO42
anions and optimize their binding by the urea groups. There are
some subtle geometric differences, though, in the observed
hydrogen-bonding parameters. Thus, the average NH···O
n−
EO4 oxoanions (E = S, Se, Cr, Mo, W, n = 2; E = P, n =
−
3) against anions of different shapes and charges, including F ,
−
−
−
−
−
−
−
−
−
Cl , Br , I , NO , BF , ClO , ReO , PF , CH CO ,
3
4
4
4
6
3
2
−
−
2−
2−
2−
CH SO , CF SO , CO , SO3 , and SeO3 . In a notable
3
3
3
3
3
contrast to previously reported ion-binding M L cages that
4
6
favored encapsulation of relatively hydrophobic monocharged
ions, the cage receptors described here showed preferential
2
−
contact distance is slightly shorter for CrO4 than for the
Table 4. Properties of the EO42− Anions
EO 2
−
d(E−O),
a
Å
ΔG°h, kJ/mol
b
ΔGrxn(urea ), kJ/mol
c
ΔHrxn(urea ), kJ/mol
c
pK (HEO4−),d
4
6
6
a
SO 2
−
1.47(4)
1.64(4)
1.63(4)
1.75(2)
1.75(3)
−1080
−950
−900
−
−530.5
−519.0
−501.9
−486.0
−485.9
−770.8
−762.3
−745.7
−729.0
−732.2
1.99
6.51
1.70
4.24
4
CrO42
−
SeO42
−
MoO42
−
WO 2
−
−
3.50
4
a
b
34
2−
Average values obtained from Cambridge Structural Database, Version 5.32 (Nov 2010). Free energies of hydration; no values for MoO and
4
WO42 are available. Calculated thermodynamic parameters for the equilibrium reaction: EO + 6(urea) ⇆ EO (urea)6 . Values obtained from
−
c
2−
2−
d
4
4
ref 33.
8
532
dx.doi.org/10.1021/ja300677w | J. Am. Chem. Soc. 2012, 134, 8525−8534

Journal of the American Chemical Society
Article
encapsulation of strongly hydrophilic, multicharged anions, as a
consequence of the presence of stabilizing hydrogen-bond
donor groups decorating the internal cavities of the cages.
REFERENCES
■
(
1) Recent reviews on self-assembled containers: (a) Chakrabarty, R.;
Mukherjee, P. S.; Stang, P. J. Chem. Rev. 2011, 111, 6810. (b) Laughrey,
Z.; Gibb, B. C. Chem. Soc. Rev. 2011, 40, 363. (c) Avram, L.; Cohen,
Y.; Rebek, J. Chem. Commun. 2011, 47, 5368. (d) Breiner, B.; Clegg, J.
K.; Nitschke, J. R. Chem. Sci. 2011, 2, 51. (e) Pluth, M. D.; Bergman,
R. G.; Raymond, K. N. Acc. Chem. Res. 2009, 42, 1650. (f) Michito, Y.;
Klosterman, J. K.; Fujita, M. Angew. Chem., Int. Ed. 2009, 48, 3418.
g) Rebek, J. Acc. Chem. Res. 2009, 42, 1660. (h) Ward, M. D. Chem.
Commun. 2009, 4487.
2) (a) Hastings, C. J.; Backlund, M. P.; Bergman, R. G.; Raymond,
K. N. Angew. Chem., Int. Ed. 2011, 50, 10570. (b) Hastings, C. J.;
Backlund, M. P.; Bergman, R. G.; Raymond, K. N. J. Am. Chem. Soc.
010, 132, 6938. (c) Bao, X.; Rieth, S.; Stojanovic, S.; Hadad, C. M.;
Badjic, J. D. Angew. Chem., Int. Ed. 2010, 49, 4816. (d) Crisostomo, F.
R. P.; Lledo, A.; Shenoy, S. R.; Iwasawa, T.; Rebek, J. J. Am. Chem. Soc.
2009, 131, 7402. (e) Iwasawa, T.; Hooley, R. J.; Rebek, J. Science 2007,
17, 493. (f) Pluth, M. D.; Bergman, R. G.; Raymond, K. N. Science
2007, 316, 85. (g) Yoshizawa, M.; Tamura, M.; Fujita, M. Science 2006,
n−
Tetrahedral EO4 oxoanions (n = 2, 3) have been found to act
as templates for the cage self-assembly. In the absence of such
templates, no M4L6 cages could be observed, and other
coordination assemblies formed in solution.
Despite their lack of preorganization, the M4L6 cage
receptors studied here display remarkable affinities and
(
n−
selectivities for tetrahedral EO4 oxoanions in aqueous
(
environments. Though the exact anion binding constants
could not be measured due to the instability of the “empty”
6
−1
cages, we estimated a lower limit of (6 ± 1) × 10 M for the
2
association constant corresponding to sulfate binding by the
8+
6+
Ni (L1)6 cage receptor. In the case of [Zn (tBuL1) (EO )]
4
4
6
4
cages, the relative anion encapsulation selectivity could be
77
3
assessed from anion exchange experiments monitored by Se
7
7
2−
NMR spectroscopy, using SeO4 as an NMR-active anionic
3−
312, 251.
probe. The following anion selectivity trend was found: PO
4
≫
CrO42 > SO4 > SeO4 > MoO4 > WO4 . The
−
2−
2−
2−
2−
(3) (a) Ferrand, Y.; Crump, M. P.; Davis, A. P. Science 2007, 318,
19. (b) Rehm, T. H.; Schmuck, C. Chem. Soc. Rev. 2010, 39, 3597.
6
observed trend mostly parallels the decrease in anions’ charge
densities, which corresponds to an anti-Hofmeister selectivity.
Such a behavior is consistent with the presence of a more
stabilizing environment for the anions inside the urea-
functionalized cages relative to the external aqueous solution.
Another factor contributing to the observed selectivity is the
structural flexibility of the cages, which can distort their frames
to accommodate different-sized anions and optimize their
binding by the urea groups. These results suggest that with
improved preorganization and structural rigidity, even stronger
anion binding and more prominent selectivities may be
achievable with this class of cage receptors, potentially leading
to applications related to separation of environmentally relevant
(
4) (a) Saalfrank, R. W.; Stark, A.; Peters, K.; von Schnering, H. G.
Angew. Chem., Int. Ed. 1988, 27, 851. (b) Saalfrank, R. W.; Maid, H.;
Scheurer, A. Angew. Chem., Int. Ed. 2008, 47, 8794.
(5) (a) Caulder, D. L.; Powers, R. E.; Parac, T. N.; Raymond, K. N.
Angew. Chem., Int. Ed. 1998, 37, 1840. (b) Caulder, D. L.; Raymond, K.
N. Acc. Chem. Res. 1999, 32, 975. (c) Davis, A. V.; Fiedler, D.; Ziegler,
M.; Terpin, A.; Raymond, K. N. J. Am. Chem. Soc. 2007, 129, 15354.
(
6) (a) Fleming, J. S.; Mann, K. L. V.; Carraz, C.-A.; Psillakis, E.;
Jeffery, J. C.; McCleverty, J. A.; Ward, M. D. Angew. Chem., Int. Ed.
998, 37, 1279. (b) Paul, R. L.; Bell, Z. R.; Jeffery, J. C.; McCleverty, J.
1
A.; Ward, M. D. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4883. (c) Paul,
R. L.; Argent, S. P.; Jeffery, J. C.; Harding, L. P.; Lynam, J. M.; Ward,
M. D. Dalton Trans. 2004, 3453. (d) Tidmarsh, I. S.; Taylor, B. F.;
Hardie, M. J.; Russo, L.; Clegg, W.; Ward, M. D. New J. Chem. 2009,
33, 366. (e) Hall, B. R.; Manck, L. E.; Tidmarsh, I. S.; Stephenson, A.;
Taylor, B. F.; Blaikie, E. J.; Vander Griend, D. A.; Ward, M. D. Dalton
Trans. 2011, 40, 12132.
36
anions (e.g., sulfate separation from nuclear waste).
ASSOCIATED CONTENT
■
(
7) (a) Glasson, C. R. K.; Meehan, G. V.; Clegg, J. K.; Lindoy, L. F.;
*
S
Supporting Information
Turner, P.; Duriska, M. B.; Willis, R. Chem. Commun. 2008, 1190.
(b) Glasson, C. R. K.; Clegg, J. K.; McMurtrie, J. C.; Meehan, G. V.;
Lindoy, L. F.; Motti, C. A.; Moubaraki, B.; Murray, K. S.; Cashion, J.
D. Chem. Sci. 2011, 2, 540.
1
Detailed experimental and synthetic procedures; NMR ( H,
1
3
77
C, Se), DOSY, UV−vis, and ESI-MS spectra; energies and
coordinates of DFT-optimized structures; details of the MD
simulations; and X-ray crystallographic data including CIF files.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(
8) (a) Mal, P.; Schultz, D.; Beyeh, K.; Rissanen, K.; Nitschke, J. R.
Angew. Chem., Int. Ed. 2008, 47, 8297. (b) Meng, W.; Clegg, J. K.;
Thoburn, J. D.; Nitschke, J. R. J. Am. Chem. Soc. 2011, 133, 13652.
(c) Hristova, Y. R.; Smulders, M. M. J.; Clegg, J. K.; Breiner, B.;
Nitschke, J. R. Chem. Sci. 2011, 2, 638.
(
9) Parac, T. N.; Caulder, D. L.; Raymond, K. N. J. Am. Chem. Soc.
AUTHOR INFORMATION
■
1998, 120, 8003.
(10) Custelcean, R. Top. Curr. Chem. 2012, 322, 193.
Corresponding Author
custelceanr@ornl.gov
(11) Recent reviews on anion receptors: (a) Wenzel, M.; Hiscock, J.
R.; Gale, P. A. Chem. Soc. Rev. 2012, 41, 480. (b) Gale, P. A. Chem.
Commun. 2011, 47, 82. (c) Gale, P. A. Chem. Soc. Rev. 2010, 39, 3746.
(12) Recent reviews on metal-based self-assembled anion receptors:
a) Mercer, D. J.; Loeb, S. J. Chem. Soc. Rev. 2010, 39, 3612.
b) Amendola, V.; Fabbrizzi, L. Chem. Commun. 2009, 513. (c) Steed,
Notes
The authors declare no competing financial interest.
(
(
ACKNOWLEDGMENTS
■
J. W. Chem. Soc. Rev. 2009, 38, 506.
(13) Kubik, S. Chem. Soc. Rev. 2010, 39, 3648.
This research was sponsored by the Division of Chemical
(14) (a) Custelcean, R. Curr. Opin. Solid State Mater. Sci. 2009, 13,
Sciences, Geosciences, and Biosciences, Office of Basic Energy
Sciences, U.S. Department of Energy. The synthesis of 77Se-
68. (b) Custelcean, R.; Moyer, B. A. Eur. J. Inorg. Chem. 2007, 1321.
15) Recent reviews on cage receptors for anions: (a) Kang, S. O.;
Llinares, J. M.; Day, V. W.; Bowman-James, K. Chem. Soc. Rev. 2010,
9, 3980−4003. (b) Ballester, P. Chem. Soc. Rev. 2010, 39, 3810.
16) (a) Rajbanshi, A.; Moyer, B. A.; Custelcean, R. Cryst. Growth.
(
labeled materials, and NMR experiments involving cage
formation and anion exchange were conducted at the Center
for Nanophase Materials Sciences, which is sponsored at Oak
Ridge National Laboratory by the Scientific User Facilities
Division, Office of Basic Energy Sciences, U.S. Department of
Energy.
3
(
Des. 2011, 11, 2702. (b) Custelcean, R.; Bock, A.; Moyer, B. A. J. Am.
Chem. Soc. 2010, 132, 7177. (c) Custelcean, R.; Remy, P. Cryst.
Growth. Des. 2009, 9, 1985.
8
533
dx.doi.org/10.1021/ja300677w | J. Am. Chem. Soc. 2012, 134, 8525−8534

Journal of the American Chemical Society
Article
(
17) Katayev, E. A.; Ustynyuk, Y. A.; Sessler, J. L. Coord. Chem. Rev.
006, 250, 3004.
18) Custelcean, R.; Bosano, J.; Bonnesen, P. V.; Kertesz, V.; Hay, B.
P. Angew. Chem., Int. Ed. 2009, 48, 4025.
19) Custelcean, R.; Moyer, B. A.; Hay, B. P. Chem. Commun. 2005,
971.
20) Busschaert, N.; Gale, P. A.; Haynes, C. J. E.; Light, M. E.;
2
(
(
5
(
Moore, S. J.; Tong, C. C.; Davis, J. T.; Harrell, W. A., Jr. Chem.
Commun. 2010, 46, 6252.
(
21) (a) Caltagirone, C.; Hiscock, J. R.; Hursthouse, M. B.; Light, M.
E.; Gale, P. A. Chem.Eur. J. 2008, 14, 10236. (b) Jia, C.; Wu, B.; Li,
S.; Yang, Z.; Zhao, Q.; Liang, J.; Li, Q. S.; Yang, X. J. Chem. Commun.
2
010, 46, 5376. (c) Jia, C.; Wu, B.; Li, S.; Huang, X.; Zhao, Q.; Li, Q.
S.; Yang, X. J. Angew. Chem., Int. Ed. 2011, 50, 486. (d) Dey, S. K.; Das,
G. Dalton. Trans. 2011, 40, 12048.
(
22) Stephenson, A.; Argent, S. P.; Riis-Johannessen, T.; Tidmarsh, I.
S.; Ward, M. D. J. Am. Chem. Soc. 2011, 133, 858.
23) Scherer, M.; Caulder, D. L.; Johnson, D. W.; Raymond, K. N.
Angew. Chem., Int. Ed. 1999, 38, 1588.
24) Reaction of L1 with ZnSO4 in MeOH/H O led at room
(
(
2
temperature to the immediate formation of an insoluble precipitate,
which was converted very slowly (over a few months) into the
crystalline [Zn (L1) (SO )](SO ) cage, as confirmed by a partially
4
6
4
4 3
refined single crystal X-ray structure.
25) Sulfite appears to get oxidized to sulfate over time, which then
templates the cage.
26) A similar observation of metal phosphate formation leading to
(
(
decomposition of a self-assembled tetrahedral cage was made recently:
El Aroussi, B.; Guenee, L.; Pal, P.; Hamacek, J. Inorg. Chem. 2011, 50,
8588.
(
27) Within the dinegative oxoanion series, the intensity of the ESI-
2
−
MS peaks corresponding to the cage decreased in the order SO
>
4
SeO42 > CrO4 > MoO4
28) A viscosity η value of (1.54 ± 0.09) × 10 Pa/s was estimated
for the 2:1 v/v CD OD−D O solvent mixture at 298 K: Wolf, D;
−
2−
2−
.
−3
(
3
2
Kudish, A. I. J. Phys. Chem. 1980, 84, 921.
29) Such a comparison needs to be done with caution, as the
(
Stokes−Einstein equation applies strictly to spherical molecular
species. For 2a, r was estimated here as the radius of the sphere
defined by the random tumbling of the tetrahedral cage in solution. As
such, r is the distance between the center of the tetrahedron and the
farthest H atom belonging to the tBu groups.
(
30) Cohen, Y.; Avram, L.; Frish, L. Angew. Chem., Int. Ed. 2005, 44,
20.
31) Hallwass, F.; Silva, R. O.; Goncalves, S. M. C.; Menezes, P. H.;
Simas, A. M. Ann. Magn. Reson. 2004, 3, 45.
32) The buffering was necessary to ensure the EO42− anions,
5
(
(
particularly the more basic chromate, molybdate, and tungstate, were
completely deprotonated.
(
33) Smith, R. M.; Martell, A. E., Critical Stability Constants; Plenum:
New York, 1975.
(
34) Marcus, Y. J. Chem. Soc., Faraday Trans. 1991, 87, 2995.
(
35) The average hydrogen-bond acceptor angle for tetrahedral
oxoanions was found to be about 122 ± 12° from a Cambridge
Structural Database survey. Hay, B. P.; Dixon, D. A.; Bryan, J. C.;
Moyer, B. A. J. Am. Chem. Soc. 2002, 124, 182.
(
36) Moyer, B. A.; Delmau, L. H.; Fowler, C. J.; Ruas, A.; Bostick, D.
A.; Sessler, J. L.; Katayev, E.; Pantos, G. D.; Llinares, J. M.; Hossain, M.
A.; Kang, S. O.; Bowman-James, K. Supramolecular Chemistry of
Environmentally Relevant Anions. In Advanced Inorganic Chemistry;
van Eldik, R., Bowman-James, K., Eds.; Elsevier: Oxford, 2006; Vol. 59,
p 175.
8
534
dx.doi.org/10.1021/ja300677w | J. Am. Chem. Soc. 2012, 134, 8525−8534