FULL PAPER

Design of a Neutral Macrocyclic Ionophore: Synthesis and Binding Properties

for Nitrate and Bromide Anions

Rainer Herges,*[a] Anton Dikmans,[a] Umasish Jana,[a] Felix Köhler,[a] Peter G. Jones,[b]
Ina Dix,[b] Tom Fricke,[c] and Burkhard König*[c]

In memory of Prof. W. Grahn[‡]

Keywords: Anions / Receptors / Density functional calculations / Environmental chemistry / Macrocycles

A macrocyclic neutral ionophore 8 (X = O) capable of binding bound with an even higher affinity. Chloride is obviously too
weakly coordinating anions such as nitrate and bromide in small, and iodine too large, to form 1:1 complexes. The bind-
DMSO solution has been designed by a stepwise, deductive ing motif of 8 (X = O) was compared with related molecules
approach. The optimum geometrical arrangement of the hy- of similar structure, such as 8 (X = S) and 19. As predicted
drogen bond donor sites in the target ionophore was deter- from calculations, the small structural variations give rise to
mined by DFT calculations. From these data, a suitable mac- a complete loss of nitrate and bromide ion binding ability in
rocyclic molecular framework was constructed. The 24-mem- DMSO. This sensitivity to geometrical changes and the affin-
bered macrocyclic ionophore was synthesized by standard ity of 8 (X = O) to nitrate and bromide ions, which are poor
macrocyclization methods. NMR titrations revealed molecu- hydrogen bond acceptors, confirm the predicted
lar complexes with defined 1:1 stoichiometries in DMSO for complementarity of ionophore binding site and anion geo-
8 (X = O) with nitrate, hydrogensulfate, acetate, cyanide, iod- metry. According to DFT and MD calculations the higher af-
ide, and bromide ions, while dihydrogenphosphate, sulfate, finity of 8 (X = O) to bromide than to nitrate is mainly due to
and chloride ions yielded aggregates of higher stoichiometry. the greater flexibility of the bromide complex and thus to its
The nitrate binding constants of 8 (X = O) are substantial for higher entropy.
a neutral ionophore with defined binding sites in pure DMSO ( Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany,
solution. Bromide ions, which have a similar ion radius, are 2002)

Introduction
the shape, charge distribution, and hydrogen bonding pat-
The selective detection of anions is of importance in en-
tern of nonspherical anions. High selectivities can only be
vironmental monitoring, medicinal diagnostics, and the
achieved in neutral receptor molecules, because the merely
analysis of biological samples.[1] While spherical cations can
distance-dependent Coulomb forces are much stronger than
be distinguished by their diameter, the shape-selective re-
the spatially more specific hydrogen bonding.
cognition of anions is more demanding:[2] anions are larger,
with smaller charge-to-radius ratios. Attractive electrostatic Commercial anion-selective electrodes separate anions
forces in recognition events are therefore much weaker than mostly by their hydration energy, and the observed selectiv-
with cations. Other implications are protonation, and thus ities strictly follow the Hofmeister series.[3] This results in
neutralization, at low pH and the need for large and struc- strong interference by anions of similar hydration energy,
turally well-defined receptor molecules complementary to such as nitrate and chloride. While several neutral artificial
receptors[4⫺7] and sensors have been reported for selective
[a]
Institut für Organische Chemie der Universität Kiel binding of halides,[8⫺12] phosphates[13,14] or carb-
Otto-Hahn-Platz 4, 24098 Kiel, Germany oxylates[15,16] in competitive solvents, recognition of the
Fax: (internat.) ⫹ 49-(0)431/880-1558
E-mail: rherges@oc.uni-kiel.de weakly basic anions, such as the nitrate ion, remains a chal-
[b]
Institut für Anorganische und Analytische Chemie der lenge,[17] in particular with neutral, uncharged receptor mo-
Technischen Universität Braunschweig, lecules. Anslyn et al. have reported a cage-like receptor 1,
38106 Braunschweig, Germany
[c]
Institut für Organische Chemie, Universität Regensburg which binds acetate and nitrate ions in CH3CN/CH2Cl2 and
93040 Regensburg, Germany MeOH/CH2Cl2 solution (Figure 1).[18,19] Lippert et al.
Fax: (internat.) ⫹ 49-(0)941/943-1717
E-mail: Burkhard.Koenig@chemie.uni-regensburg.de showed that a macrocyclic platinum palladium complex
[‡]
Institut für Organische Chemie, TU Braunschweig binds nitrate and PF6 ions simultaneously,[20] while a plat-

3004  WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002 1434⫺193X/02/0917⫺3004 $ 20.00⫹.50/0 Eur. J. Org. Chem. 2002, 3004⫺3014
Design of a Neutral Macrocyclic Ionophore FULL PAPER
inum complex bearing nicotinamide ligands was presented We introduced thiourea units as hydrogen bond donors,
by Loeb et al.[21] In a recent paper Hamilton et al. reported since this motif is often found in crystal structures of nitrate
the macrocyclic receptor 2,[22] which binds iodine, chloride, salts (usually with one urea or thiourea unit per nitrate) and
nitrate, and tosylate ions in a chloroform/DMSO solvent because thiourea provides the correct hydrogen bond angles
mixture in a stepwise 2:1 and 1:1 equilibrium. X-ray struc- (a pair of parallel N⫺H bonds each, see b in Figure 2). This
tures of cage compounds with inside-bound nitrate have approach is crude, since the graphite pattern provides a very
been published by Bowman-James et al.[23] and Barbour et coarse grid and we could hardly expect to obtain the op-
al.[24] Jurczak et al. reported a macrocyclic polylactam-type timum geometry for binding at this first attempt. However,
receptor for anions, which binds acetate in DMSO solu- there are three positions X (see b in Figure 2) that can be
tion.[25] Strong binding of weakly basic anions, with select- used for ‘‘fine tuning’’ of the geometry. Depending on the
ivity for spherical anions such as chloride and bromide, was group (X ⫽ S, O, NH, CH2), the size of the cavity should
found for the urea-substituted porphyrins 3.[26] However, change slightly. To find the optimum arrangement of the
the binding cavity of this ionophore contains solvent molec- three urea units relative to the anion, we performed density
ules as well as the anionic guests, so that no direct assess- functional theory calculations on the 3:1 complex of thio-
ment of the geometry of the binding motif is possible. urea with nitrate (Figure 3). The optimized structure is
Acyclic ionophores[13,27,28] for anion binding show affinity
to basic anions, such as phosphate or acetate, in DMSO.
Interaction with weakly basic anions, such as nitrate or hal-
ides, is generally not observed.

Figure 1. Structures of recently reported neutral macrocyclic an-

ion receptors
Figure 2. ‘‘Chicken wire’’ approach to the design of a nitrate recep-
We report here our results from a rational design ap- tor
proach to a neutral macrocyclic ionophore with binding af-
finity towards the weakly coordinating anions nitrate and
bromide in polar solvents.

Results and Discussion

Design of the Receptor Structure
The starting point of our design approach was to con-
struct a neutral ligand to bind nitrate. In the NO3⫺ ion,
there are six hydrogen bond acceptor sites arranged in D3h
symmetry. It was thus straightforward to place the ion in
the C3 axis of a hexagonal grid and in a first approach to Figure 3. B3LYP/6-31G*-optimized structure of NO3⫺ complexed
cut the molecular frame from the chicken wire pattern.[28b] with three thiourea units (D3 symmetry)

Eur. J. Org. Chem. 2002, 3004⫺3014 3005

FULL PAPER R. Herges, B. König et al.

slightly twisted out of plane (HONO dihedral angle 13.6°) Table 2. DFT/PCM-calculated enthalpies of complexation
and has D3 symmetry. The ‘‘cavity size’’ (average distance
of hydrogen atoms involved in hydrogen bonding from the ∆Hcompl. 8 (X ⫽ O) ⫹ X⫺ 씮 8 (X ⫽ O)·Xⴚ
center of mass of the molecule) is 2.799 Å. X⫺ Vacuum DMSO Water
For optimum preorientation of the ligand, we therefore
NO3⫺ ⫺75.4 ⫺28.8 ⫺22.9
had to find a spacer that would connect the thiourea units Br⫺ ⫺80.4 ⫺25.2 ⫺16.0
in or close to this geometry. ∆∆Hcompl. ⫺5.0 3.6 6.8
Density functional theory (DFT) calculations (at the
B3LYP/6-31G* level of theory) on the macrocycles with
X ⫽ S and O and with C3 symmetry (the highest point work of the Kohonen type[37] with 25 ⫻ 25 neurons. The
group in which all macrocycles are minima) predict a cavity structure represented by each neuron was energy-minim-
size of 2.727 Å for X ⫽ O and 2.922 Å for X ⫽ S. Interest- ized. The potential energy of each of these 625 conforma-
ingly, the macrocycles with X ⫽ NH and X ⫽ CH2 or CH, tions is coded by color (low energies blue, high energies red;
such as Göbels macrocycle 19[29] (the size of which is be- Figure 4). The energies are relative with respect to the low-
tween that of the O- and that of S-macrocycle) can be ruled est-energy structure and the potential energy range covered
out as potential receptors, because the hydrogen atoms in from blue to red is 10 kcal mol⫺1.
the spacer group X would sterically interfere with a bound
nitrate and with the thiourea N⫺H groups. This leaves the
macrocycles with X ⫽ O and S as possible candidates for
synthesis. According to the DFT calculations, the cavity
diameter of the oxygen derivative is closer to the optimum
value and thus provides better preorientation.
The six hydrogen bond donor sites in 8 (X ⫽ O) and 8
(X ⫽ S) are located in a suitable position to bind NO3⫺,
although ions with spherical symmetry and of similar size
to nitrate should also fit in the cavity. In the series of hal-
ogen ions, the ion radius of bromide is very close to that of
nitrate. Thus Br⫺ is a good candidate to bind, whereas
chloride is too small and iodide too large to form a 1:1
complex (Table 1).

Table 1. Ion radii of halogens and nitrate[30]

Ion Ion radii [Å]

F⫺ 1.24
Cl⫺ 1.80
Br⫺ 1.98 Figure 4. Potential energy hypersurfaces of the conformations of 8
NO3⫺ eq 2.06 (X ⫽ O)·NO3⫺, 8 (X ⫽ O)·Br⫺, and 8 (X ⫽ O); both complexes
(NO3⫺ ax) (1.26) and the ligand were simulated by molecular dynamics (1000 ps,
I⫺ 2.25 600 K) and the resulting conformational space was projected in two
dimensions through use of Kohonen artificial neural networks (25
⫻ 25 neurons); the colors indicate the relative potential energies
(Epot,rel in kcal mol⫺1) of the conformations represented by the
According to DFT calculations[31] (B3LYP/6-31G*) in- corresponding neuron; relative energies above 10 kcal mol⫺1 are
cluding solvent effects by use of the polarized continuum in white
model (PCM),[32,33] the enthalpy of complexation [8 (X ⫽
O) ⫹ X⫺ 씮 8 (X ⫽ O)·Xⴚ] within the C3 point group The Kohonen map of the nitrate complex 8 (X ⫽
strongly depends on the dielectric constant of the solvent O)·NO3⫺, in comparison with that of the bromide complex
(DMSO: ε ⫽ 46.7; water: ε ⫽ 78.39). In the absence of 8 (X ⫽ O)·Br⫺, exhibits discrete and much larger areas for
solvent interactions, the formation of the bromide complex conformations with low potential energy (blue). The nitrate
is more exothermic by 5.0 kcal mol⫺1, while in water nitrate complex thus occupies conformations of low potential en-
is bound with a more negative enthalpy of complexation ergy more frequently than the bromide complex. Inspection
(∆∆Hcompl. ⫽ 6.8 kcal mol⫺1) (Table 2). of the conformations in both Kohonen maps reveals that
To estimate entropy effects we performed MD simula- the nitrate complex structures are relatively planar, whereas
tions, based on the MM2 force field, on 8 (X ⫽ O)·Br⫺ and the bromide complexes do not exhibit this kind of restric-
8 (X ⫽ O)·NO3⫺.[34,35] The molecular dynamics calcula- tion. The fact that the bromide complex at the same tem-
tions were performed for 1000 ps at 600 K (bath relaxa- perature passes a larger number of conformations of higher
tion).[36] The conformational spaces of both complexes and potential energy can be interpreted in terms of a larger con-
the free ligands defined by the trajectories were projected formation space and a higher entropy than for the nitrate
onto 2D hypersurfaces by the use of an artificial neural net- complex. The free ligand is even more flexible.

3006 Eur. J. Org. Chem. 2002, 3004⫺3014

Design of a Neutral Macrocyclic Ionophore FULL PAPER
Reoptimization of the most stable conformations within be isolated in 90% yield. However, heating of the solution
the low potential energy areas from the Kohonen map of 8 at reflux in chloroform overnight, without intermediate
(X ⫽ O)·Br⫺ at the B3LYP/6-31G* level of DFT theory workup, resulted in the formation of the desired protected
confirmed three conformations of C3, Cs, and C1 symmetry diamine 5 in 70% yield. Deprotection to give the diamine 6
as the lowest minima on the energy hypersurface, the C1 (dihydrochloride) was achieved with 1.0  HCl in acetic
and Cs structures being about 0.5 kcal mol⫺1 more stable acid/dichloromethane, in 85% yield.
than the C3 structure. Bis(thioisocyanate) 7, as the second reaction component
To prove our theoretical concept, we decided to synthes- for macrocyclization, was obtained from 2 by treatment
ize 8 (X ⫽ O) and 8 (X ⫽ S) and to investigate their anion- with CS2 according to a modified procedure reported by
binding properties. Luk’yanenko et al.[39] Macrocyclization was performed by
dissolving the dihydrochloride salt of 6 with triethylamine
Synthesis or NaOH and addition of the diisothiocyanate 7 in aceton-
Our strategy for the formation of 8 (X ⫽ O) and 8 (X ⫽ itrile at room temperature. The target compound 8 (X ⫽
S) is similar to that reported by Lehn et al.,[38] which fo- O) was obtained in 50⫺55% yield (Scheme 1).
cused on the addition of amines and isothiocyanates, and The structure was confirmed by all spectroscopic data.
that reported by Göbel et al.,[29] in which a thiourea unit Proton NMR spectra suggest unrestricted intramolecular
bearing two terminal amines reacts with a diisothiocyanate motion of the macrocycle in solution. Crystals suitable for
in the final macrocyclization step. Treatment of a twofold X-ray analysis were obtained from acetone. The structure
excess of diamine 2 with Boc anhydride, with use of a con- in the solid state shows three intramolecular hydrogen
tinuous extraction method for product isolation, gave a bonds of N⫺H to the ether oxygen atoms (Figure 5).
mixture of 97% monoprotected 3 and 3% diprotected di- As a minor reaction product compound 9 was isolated.
amine 4. Treatment of 3 with thiophosgene at room temper- The diprotected amine 4, formed as a by-product in 3%
ature yielded the corresponding isothiocyanate, which could yield in the first Boc protection step, survives treatment

Scheme 1. Synthesis of macrocycles 8 (X ⫽ O) and 9

Eur. J. Org. Chem. 2002, 3004⫺3014 3007

FULL PAPER R. Herges, B. König et al.

also confirmed by X-ray analysis. Other spectroscopic prop-

erties are in agreement with known data.[40]
Bis(2-aminoethyl) thioether dihydrobromide (14) was
synthesized by slight modification of the published
procedures[41⫺45] (Scheme 2). The components 17 and 18
for macrocyclization were prepared similarly to the oxa
analogues 6 and 7. Treatment of 15 (unlike the oxo derivat-
ive 3) with thiophosgene at room temperature directly af-
forded the diprotected diamine 16. Macrocyclization of the
thia compound required higher temperatures (70 °C) than
needed for the oxa derivative (20 °C).

Evaluation of Binding Properties

The anion-binding properties of macrocycle 8 (X ⫽ O)
were investigated by NMR titration in deuterated DMSO.
Association constants were derived from the observed com-
plex-induced shifts (see Exp. Sect. for details) and are sum-
marized in Tables 3⫺6 (Figure 6).[46] The binding of 8 (X ⫽
O) with nitrate, acetate, cyanide, bromide, iodide, and hy-
drogensulfate ions yield data (see Table 3) consistent with
1:1 stoichiometries, as confirmed by Job’s plot analysis and
Scatchard plots. In the case of HSO4⫺ it is most likely that
protonation of 8 (X ⫽ O) by the acidic anion gives rise to
a charged receptor 8 (X ⫽ O) H⫹. From the titration of 8
(X ⫽ O) with NaNO3 an association constant K11 ⫽ 23.2 ⫾
Figure 5. Structure of 8 (X ⫽ O) in the solid state 0.6 L mol⫺1 was derived. The increase of binding strength
relative to that of Bu4NNO3 may be attributable to the co-
ordination of the sodium counter-ion to the ether oxygen
with thiophosgene and workup. In the subsequent acid atoms, which should result in higher anion affinity of the
treatment it is deprotected together with 5 and reacts in the ionophore. The binding constants of 8 (X ⫽ O) for nitrate
final macrocyclization step to give 9. The structure of 9 was ions in pure DMSO are among the highest values reported

Scheme 2. Synthesis of 8 (X ⫽ S)

3008 Eur. J. Org. Chem. 2002, 3004⫺3014

Design of a Neutral Macrocyclic Ionophore FULL PAPER
Table 3. Association constants K11 of 8 (X ⫽ O) with tetrabutylammonium salts in deuterated DMSO

Anion[a] K11 pKS [47]

R[b] ∆δmax (obsd.) ∆δmax (calcd.) ∆G°
[L·mol⫺1] [kJ·mol⫺1]

NO3⫺ 17.1 ⫾ 0.4 ⫺1.32 0.9999 0.23 (16 equiv.) 0.32 ⫺7.0
OAc⫺ 1260 ⫾ 260 ⫹4.75 0.9987 0.76 (5.0 equiv.) 0.76 ⫺17.7
CN⫺ 1300 ⫾ 570 ⫹9.31 0.9956 0.31 (3.0 equiv.) 0.31 ⫺17.8
Br⫺ [c] 400 ⫾ 40 ca. ⫺9 0.9991 0.26 (10 equiv.) 0.26 ⫺14.9
I⫺ ⬍2 ca. ⫺10 0.9984 0.03 (15 equiv.) ⫺ ⫺
HSO4⫺ 58 ⫾ 4 ⫹1.92 0.9978 0.24 (30 equiv.) 0.24 ⫺10.1

All anions were used as their Bu4N⫹ salts. [b] Correlation coefficient of regression.
[a] [c]
Binding constant and stoichiometry of binding
were independently confirmed by calorimetric titration.

Table 4. Association constants of more complex aggregates of 8 (X ⫽ O) and tetrabutylammonium salts in DMSO as determined by
NMR titration

Anion[a] Model[b] K pKa [47]

∆δmax (obsd.) ∆δmax (calcd.)

SO42⫺ 2:3 ⫺[c] ca. ⫺3 1.8 (15 equiv.) ⫺

H2PO4⫺ 2:1 5.3·104 L2·mol⫺2 ⫹2.12 1.3 (7.5 equiv.) 1.4
Cl⫺ 1:2 200 L2·mol⫺2 ca. ⫺6 0.41 (40 equiv.) 0.44
[a]
All anions were used as their Bu4N⫹ salts. [b] Stoichiometry of binding used to emulate the experimental data by mathematical model.
[c]
The very strong association does not allow meaningful data to be derived.

Table 5. Association constants K11 of tetrabutylammonium nitrate binding of 8 (X ⫽ O), 8 (X ⫽ S), and 19 in DMSO

Compound K11 [L·mol⫺1] R[a] ∆δmax (obsd.) ∆δmax (calcd.) ∆G° [kJ·mol⫺1]

8 (X ⫽ O) 17.1 ⫾ 0.4 0.9999 0.23 (16 equiv.) 0.32 ⫺7.0

8 (X ⫽ S) ⬍1 0.9982 0.10 (40 equiv.) ⫺ ⫺
19 ⬍1 0.9972 0.08 (40 equiv.) ⫺ ⫺
[a]
Correlation coefficient of regression.

Table 6. Association constants K11 of tetrabutylammonium bromide binding of 8 (X ⫽ O), 8 (X ⫽ S), and 19 in DMSO as determined
by NMR titration

Compound K11 [L·mol⫺1] R[a] ∆δmax (obsd.) ∆δmax (calcd.) ∆G° [kJ·mol⫺1]

8 (X ⫽ O) 400 ⫾ 40 0.9991 0.26 (10 equiv.) 0.26 ⫺14.9

8 (X ⫽ S) 2.6 ⫾ 0.1 0.9994 0.25 (40 equiv.) 0.37 ⫺2.4
19 2.8 ⫾ 0.2 0.9993 0.18 (40 equiv.) 0.33 ⫺2.6
[a]
Correlation coefficient of regression.

for neutral ionophores with defined binding motifs.[18,19,22] that macrocycle 8 (X ⫽ O), although well suited to form
The formation of molecular aggregates of 8 (X ⫽ O) with hydrogen bonds to the weakly basic nitrate anion even in
stoichiometries other than 1:1 was observed for H2PO4⫺, competition with the polar solvent DMSO, does not exhibit
Cl⫺, and SO42⫺ anions in deuterated DMSO. While the ex- exceptional nitrate binding selectivity. On the contrary, it
perimental data (see Table 4) are in good agreement with a shows a remarkable binding selectivity for bromide ions,
2:1 binding model for H2PO4⫺ and a 1:2 model for chlor- which are even poorer hydrogen bond acceptors.
ide,[46] a tight association of 8 (X ⫽ O) with sulfate did not How important, though, is the exact geometry of the
allow meaningful values to be derived.[48] binding site of 8 (X ⫽ O) for defined association with ni-
The ability of anions to act as hydrogen bond acceptors trate and bromide ions in DMSO? To answer this question,
is correlated with the pKa value of their corresponding acid. the nitrate-binding abilities of macrocycles 8 (X ⫽ S) and
Although there is no direct relationship,[49] the plot of K11 19[29] (Scheme 3)⫺ with related, but slightly different struc-
association constants of 8 (X ⫽ O) vs. pKa should show an tures ⫺ were determined in DMSO by NMR titration. For
approximately linear interdependence. Figure 7 shows that comparison, compound 9, which has only two thiourea
this is indeed the case, with one exception: the binding of binding sites, was also tested. Compounds 8 (X ⫽ S) and 19
bromide ions is much stronger than expected. This suggests showed weak nitrate ion binding in DMSO with association

Eur. J. Org. Chem. 2002, 3004⫺3014 3009

FULL PAPER R. Herges, B. König et al.

Figure 6. Observed induced chemical shifts of 8 (X ⫽ O) upon

titration with NaNO3 (䉱) and Bu4NNO3 (•); observed induced
chemical shifts of 8 (X ⫽ S) (䉬) and 19 (䊏) upon titration with
Bu4NNO3; see Exp. Sect. for details of measurements

Scheme 3. Structures of macrocycles evaluated for nitrate binding

firmed its affinity for nitrate ions even in DMSO, a solvent

that strongly competes for intermolecular interactions. The
derived association constants for nitrate binding in DMSO
are among the highest values so far reported for neutral
defined ionophores. Small structural variations, such as in
8 (X ⫽ S), which is slightly larger than our first target, or
19, which exhibits steric interactions with bound NO3⫺, re-
sult in almost complete loss of binding activity.
The Br⫺ ion, which has an ion radius similar to that of
nitrate, exhibits an even larger affinity for 8 (X ⫽ O),
whereas iodide is obviously too large and chloride (which
Figure 7. Association constants K11 of tetrabutylammonium salts is bound in a 2:1 complex) is too small to fit in the cavity.
with 8 (X ⫽ O) vs. pKa of the corresponding acids of the anions The high affinity for bromide relative to nitrate cannot be
explained on enthalpic grounds. According to molecular
dynamics simulations, the bromide complex undergoes
constants K11 ⬍ 1 L mol⫺1, but a defined 1:1 stoichiometry rapid conformational changes and thus has a higher en-
(see Figure 6 for titration curves, Table 3 for data). Com- tropy than the nitrate complex.
pound 9 bound nitrate ions very weakly without defined
stoichiometry.
The same strong sensitivity of binding ability to struc- Experimental Section
tural changes is observed in bromide binding. The associ-
ation constants drop by a factor of more than 100 on com- General Methods: Melting points were taken with a hot-plate mi-
croscope apparatus and are uncorrected. NMR spectra were re-
parison of 8 (X ⫽ O) with 8 (X ⫽ S) and 19. The stoichi-
corded with a Bruker AC 400 spectrometer at 400 MHz (1H) and
ometry of binding clearly remains 1:1 in all three cases. 100 MHz (13C) in CDCl3 solutions unless otherwise stated. The
multiplicity of the 13C signals was determined with the DEPT tech-
nique and quoted as: (⫹) for CH3 or CH, (⫺) for CH2 and (Cquat)
Conclusion for quaternary carbon atoms. Chemical shifts are reported on the
δ scale relative to tetramethylsilane. Mass spectra were determined
By means of a novel systematic approach we have de- with a MAT 8430 mass spectrometer at an ionizing voltage of
signed a neutral receptor molecule for binding of weakly 70 eV.
basic anions such as nitrate and bromide ions. The most Crystal Structure Determination of 8 (X ⴝ O): Crystal data: Mono-
promising structure 8 (X ⫽ O) and two other macrocycles clinic, space group C2/c, a ⫽ 20.975(4), b ⫽ 9.5033(16), c ⫽
were synthesized. Binding studies of 8 (X ⫽ O) indeed con- 10.934(2) Å, β ⫽ 94.207(4)°, V ⫽ 2173.6 Å3, Z ⫽ 4, T ⫽ ⫺130 °C.

3010 Eur. J. Org. Chem. 2002, 3004⫺3014

Design of a Neutral Macrocyclic Ionophore FULL PAPER
Data collection: A ca. 0.4 ⫻ 0.3 ⫻ 0.1 mm crystal was used to 1,3-Bis{2-[2-(tert-butoxycarbonylamino)ethoxy]ethyl}thiourea (5):
record 7081 intensities with a Bruker SMART 1000 CCD diffracto- An orange solution of thiophosgene (0.15 mL, 2.04 mmol, 1.0
meter (Mo-Kα radiation, 2θmax ⫽ 56.6°). Structure refinement: The equiv.) in 10 mL of chloroform was added dropwise over 0.5 h to
structure was refined anisotropically on F2 (G. M. Sheldrick, a clear solution of tert-butyl [2-(2-aminoethoxy)ethyl]carbamate (3,
SHELXL-97, Univ. of Göttingen). The molecule possesses crystal- 1.25 g, 6.12 mmol, 3 equiv.) and 4-(dimethylamino)pyridine (0.75 g,
lographic twofold symmetry. The atoms C7, C8, O9, C10, C11, and 6.12 mmol) in 100 mL of chloroform. After the mixture had been
S2 are disordered over two positions, but a suitable disorder model stirred for 1 h at room temperature, the yellow mixture was heated
could be refined by use of similarity restraints. Hydrogen atoms under reflux overnight and the chloroform was removed in vacuo,
were included by use of a riding model. Refinement proceeded to to yield a white solid as crude product. Flash chromatography (50%
wR2 ⫽ 0.120, R1 ⫽ 0.047 for 192 parameters, 247 restraints, and dichloromethane/50% ethyl acetate) then afforded 0.64 g of the de-
2691 unique reflections; S ⫽ 1.03, max. ∆ρ ⫽ 0.54 e Å⫺3. sired product, 1,3-bis{2-[2-(tert-butoxycarbonylamino)ethoxy]-
ethyl}thiourea (5), as a white solid (m.p. 99⫺101 °C), in 70% yield.
1
H NMR: δ ⫽ 1.39 (s, 18 H), 3.20⫺3.30 (m, 4 H), 3.49 (t, J ⫽
1
H NMR Titrations in [D6]DMSO: All titrations were performed 5.1 Hz, 4 H), 3.57 (t, J ⫽ 4.6 Hz, 4 H), 3.65 (br. s, 4 H) ppm. 13C
at room temperature with ionophore concentrations of 30 mmol NMR: δ ⫽ 28.2 (⫹), 40.0 (⫺), 44.3 (⫺), 69.6 (⫺), 70.2 (⫺), 79.2
L⫺1 and were repeated three times to confirm results.[50] Anions (⫺), 155.9 (C⫽O), 182.8 (C⫽S) ppm. MS(EI): m/z (%) ⫽ 450 (100),
were added, unless otherwise stated, as their tetrabutylammonium 307, 289, 164, 146, 57, 44.
salts from stock solutions (0.9 mol L⫺1). NMR spectra were re-
corded with a 400 MHz spectrometer. Association constants were 1,3-Bis[2-(2-aminoethoxy)ethyl]thiourea Dihydrochloride (6): A so-
derived from the induced chemical shifts of several protons with lution of 1,3-bis{2-[2-(tert-butoxycarbonylamino)ethoxy]ethyl}-
the program HYP NMR 2000.[46] A minimum of 10 data points thiourea (5, 0.29 g, 0.64 mmol) in 10 mL of dichloromethane was
with p values (probability of binding) between 0.2 and 0.8 were stirred for 15 min, and acetic acid (2.0 , 20 mL) and HCl (1.0 ,
used for data analysis in each case. To exclude self association of 20.0 mL) were added. The solution was stirred vigorously for 3 h.
the investigated ionophores, NMR spectra of [D6]DMSO solutions All solvents were removed under high vacuum. The crude product
were recorded over a wide rage of concentrations. The observed was purified by recrystallization from methanol/ethyl acetate (1:1)
shifts of proton resonance are very small and can be neglected. For to yield a white solid (0.175 g, 85%), m.p. 186⫺188 °C. IR (KBr):
Job’s plot analysis solutions of ionophore and salt (each 30 mmol ν̃ ⫽ 3429, 3277, 3010, 2966, 2963, 1560, 1119, 1101 cm⫺1. 1H NMR
L⫺1) were mixed in different ratios. ([D4]methanol): δ ⫽ 3.18 (t, J ⫽ 5.1 Hz, 4 H), 3.67 (t, J ⫽ 5.3 Hz,
4 H), 3.74 (t, J ⫽ 5.1 Hz, 4 H), 3.77 (br. s, 4 H) ppm. 13C NMR
([D4]methanol): δ ⫽ 40.7 (⫺), 44.9 (⫺), 67.6 (⫺), 70.7 (⫺), 191.5
tert-Butyl [2-(2-Aminoethoxy)ethyl]carbamate (3): The solid dihy- (C⫽S).
drochloride salt of diamine 2 (2.0 g, 11.3 mmol, 1 equiv.) was added 1-Isothiocyanato-2-[2-(isothiocyanato)ethoxy]ethane (7): 1-Isothi-
to a suspension of NaOH (0.81 g, 20.3 mmol, 1.8 equiv.) in meth- ocyanato-2-[2-(isothiocyanato)ethoxy]ethane (7) was prepared ac-
anol (120 mL). The mixture was heated to 60⫺65 °C with a heat cording to the procedure described in the literature, with a few
gun for 2⫺3 min and then stirred for 0.5 h to allow the solid to minor changes.[39] Triethylamine (0.74 g, 1 mL, 2 equiv.) was added
dissolve. A solution of di-tert-butyl dicarbonate (1.24 g, 5.7 mmol, by syringe at room temperature to a stirred solution of bis(2-amin-
0.5 equiv.), dissolved in THF (40 mL), was then added dropwise oethyl) ether dihydrochloride (1, 0.65 g, 3.67 mmol) and NaOH
over 15 min to the methanol mixture. After the mixture had been (0.279 g, 6.97 mmol, 1.9 equiv.) in methanol (30 mL). After 0.5 h
stirred at room temperature for 24 h, all solvents were removed in of stirring, carbon disulfide (1.12 g, 14.7 mmol, 4 equiv.) was slowly
vacuo, which yielded a white, oily solid. This was dissolved in dis- added dropwise to the solution, which was stirred for another 2 h.
tilled water and placed in a continuous extractor apparatus already At this point, the reaction mixture was cooled to 0 °C, and ethyl
containing dichloromethane. After heating under reflux for 24 h, chloroformate (1.6 g, 1.4 mL, 14.7 mmol, 4 equiv.) was added drop-
the dichloromethane was removed in vacuo to yield 1.13 g (97%) wise. This was followed by slow warming of the solution to room
of pure tert-butyl [2-(2-aminoethoxy)ethyl]carbamate (3), as a yel- temperature and further stirring for 2 h. The resulting mixture was
lowish oil. 1H NMR: δ ⫽ 1.43 (s, 9 H), 2.77 (t, J ⫽ 5.4 Hz, 2 H), then extracted with dichloromethane (3 ⫻ 100 mL), the organic
3.26 (t, J ⫽ 5.4 Hz, 2 H), 3.54 (t, J ⫽ 5.4 Hz, 2 H), 3.57 (t, J ⫽ phase was dried with MgSO4, and the solvent was removed in va-
5.4 Hz, 2 H) ppm. 13C NMR: δ ⫽ 28.5 (⫹), 40.5 (⫺), 40.7 (⫺), cuo. The residue was heated at 100⫺130 °C under vacuum
70.1 (⫺), 72.7 (⫺), 81.5 (Cquat), 159.1 (Cquat) ppm. (40 mbar) until the evolution of ethanol and carbon oxide sulfide
ceased. The temperature was then raised to 160 °C at 1 mbar pres-
sure, and the desired product distilled off as a clear, colorless liquid
[2-(tert-Butoxycarbonylamino)ethoxy]-2-isothiocyanatoethane (3b):
(0.62 g) with a yield in the range of 80⫺90% (ref.[39] 75%), b.p. 150
An orange solution of thiophosgene (2.04 mmol, 1.0 equiv.) in
°C at 9 mbar (ref.[39] 150 °C/1 Torr). 1H NMR (200 MHz, CDCl3):
10 mL of chloroform was added dropwise over 0.5 h to a clear solu-
δ ⫽ 3.69 (bs) ppm. 13C NMR (50.3 MHz, CDCl3): δ ⫽ 45.3 (⫺),
tion of tert-butyl [2-(2-aminoethoxy)ethyl]carbamate (3, 1.25 g,
69.3 (⫺) ppm.
6.12 mmol, 3 equiv.) and 4-(dimethylamino)pyridine in 100 mL of
chloroform. After 6 h of stirring at room temp., the chloroform was 1,9,17-Trioxa-4,6,12,14,20,22-hexaazacyclotetracosane-5,13,21-tri-
removed in vacuo to give a crude yellow oil. Flash chromatography thione [8 (X ⴝ O)]. Method A: 1,3-Bis[2-(2-aminoethoxy)ethyl]thi-
(90% dichloromethane/10% methanol) afforded the desired product ourea dihydrochloride (6, 0.2 g, 0.62 mmol, 1 equiv.) was added to
4, as a clear, colorless oil in 90% yield. IR (neat): ν̃ ⫽ 3360 (N⫺H), a stirred solution of triethylamine (1.9 g, 2.6 mL, 18.6 mmol, 30
2980, 2930, 2880, 2195 (N⫽C⫽S), 2110 (N⫽C⫽S), 1705 (C⫽O), equiv.) in acetonitrile (250 mL). The resulting suspension was
1520, 1185, 1130 cm⫺1. 1H NMR (200 MHz, CDCl3): δ ⫽ 1.45 (s, sonicated for 30 min until it completely dissolved, and 1 equiv. of
9 H), 3.32⫺3.35 (m, 2 H), 3.57⫺3.60 (m, 2 H), 3.66 (br. s, 4 H), 1-isothiocyanato-2-[2-(isothiocyanato)ethoxy]ethane (7) was then
5.05 (vbs, 1 H) ppm. 13C NMR (50.3 MHz, CDCl3): δ ⫽ 28.3, 40.2, added by syringe. The mixture was stirred for 16 h and the solvent
45.3, 68.9, 70.2, 79.2, 134.0 (S⫽C⫽N), 155.9 ppm. and triethylamine were then removed in vacuo to afford a white

Eur. J. Org. Chem. 2002, 3004⫺3014 3011

FULL PAPER R. Herges, B. König et al.

solid. The solid was dried under high vacuum for several hours. 1H Detosylation of 13:[45] A mixture of 13 (9 g, 21 mmol), phenol
and 13C NMR spectra of this white solid revealed that 2 equiv. of (18 g), and HBr in acetic acid solution (30%, 500 mL) was gently
Et3N were probably bound by the desired product macrocycle. 1H heated under reflux for 50 h. The solvent was completely removed
NMR (200 MHz, [D6]DMSO): δ ⫽ 1.25 (t, J ⫽ 6.7 Hz, 6 H), 3.14 under vacuum. The residue was washed successively with ethyl
(q, J ⫽ 6.7 Hz, 4 H), 3.50 (t, J ⫽ 5.2 Hz, 12 H), 3.65 (t, J ⫽ 5.2 Hz, acetate and decanted, giving 14 as a brown solid. Yield 4.8 g (81%);
12 H), 7.93 (s, 6 H) ppm. 13C NMR (50.3 MHz, CDCl3): δ ⫽ 9.5, m.p. 122⫺125 °C. 1H NMR (200 MHz, D2O): δ ⫽ 2.92 (t, J ⫽
44.2, 46.0, 69.5, 182.5 ppm. Flash chromatography (20% methanol/ 6.5 Hz, 4 H), 3.27 (t, J ⫽ 6.5 Hz, 4 H) 4.81 (s, 6 H).
80% ethyl acetate) afforded 0.22 g (55%) of the desired product as
a white solid. Method B: 1,3-Bis[2-(2-aminoethoxy)ethyl]thiourea [5-(tert-Butoxycarbonylamino)-3-thiapentyl]amine (15): A mixture
dihydrochloride (6, 0.3 g, 0.93 mmol, 1 equiv.) was added to a of 14 (5 g, 17.9 mmol) and NaOH (1.42 g, 35.7 mmol) in methanol
stirred solution of NaOH (74 mg, 1.86 mmol, 2 equiv.) in acetonitr- (100 mL) was stirred for 0.5 h to obtain the neutral diamine. A
ile (250 mL). The resulting suspension was sonicated for 30 min, solution of di-tert-butyl dicarbonate (1.55 g, 7.14 mmol) in CHCl3
and 5.0 mL of water was added to dissolve all reagents. 1-Isothiocy- (4 mL) was then added dropwise. The reaction mixture was stirred
anato-2-[2-(isothiocyanato)ethoxy]ethane (7, 160 mg, 0.93 mmol) for 24 h, methanol was removed under vacuum, and water (50 mL)
was added to this solution by syringe. The mixture was stirred vig- was added. The residue was extracted with CH2Cl2 (2 ⫻ 100 mL),
orously for 16 h and the solvent was removed in vacuo to yield a and the organic layer was dried with anhydrous Na2SO4 and con-
white solid. Flash chromatography (methanol/ethyl acetate, 1:4) centrated under vacuum. The solid residue was subjected to column
gave 210 mg (50%) of the desired product as a white solid; m.p. chromatography on silica gel (ethyl acetate/methanol, 8:2) to afford
231⫺232 °C. 1H NMR (400 MHz, [D6]DMSO): δ ⫽ 3.54 (br. s, 12 the monoprotected amine as a colorless oil. Yield 1.02 g (65%). IR
H), 3.58 (vbs, 12 H), 7.56 (br. s, NH, 6 H) ppm. 13C NMR (neat): ν̃ ⫽ 3355, 1696 cm⫺1. 1H NMR (400 MHz, D2O): δ ⫽ 1.40
(100.6 MHz, [D6]DMSO): δ ⫽ 44.0 (⫺), 68.9 (⫺), 182.5 (C⫽S) (s, 9 H), 2.63(t, J ⫽ 6.6 Hz, 4 H), 2.78 (t, J ⫽ 6.6 Hz), 3.23(t, J ⫽
ppm. MS(FAB): m/z (%) ⫽ 439.2 (100) [M ⫹ H]⫹. 6.6 Hz), 4.73 (b, 3 H) ppm. 13C NMR (100 MHz, D2O): δ ⫽ 30.3
(⫹), 33.5 (⫺), 36.2 (⫺), 45.3 (⫺), 83.6 (Cquat), 160.75 (Cquat) ppm.
2-(p-Tolylsulfonamido)ethyl p-Toluenesulfonate (11):[41] A mechanic- MS(FAB): m/z (%) ⫽ 221 (100) [M ⫹ H]⫹.
ally stirred suspension of p-toluenesulfonyl chloride (80.6 g, 0.42
1,3-Bis[5-(tert-butoxycarbonylamino)-3-thiapentyl]thiourea (16): A
mol) in 50 mL of pyridine was cooled to ⫺15 °C with an ice/salt
CH2Cl2 solution of thiophosgene (2.5 m, 0.7 mL 1.6 mmol), dis-
bath. A pre-cooled (0 °C) solution of 2-aminoethanol (12.2 g, 0.2
solved in 4 mL chloroform, was added dropwise at room temper-
mol) in 20 mL of pyridine was then added dropwise over a period
ature to a stirred solution of mono-Boc-protected amine 15
of 0.5 h, and the mixture was stirred at this temperature for 2 h
(750 mg, 3.2 mmol) in dry CHCl3 (40 mL) and triethylamine (1.4 g,
and at 0 °C for 5 h, and kept at 0 °C overnight. Ice/water was added
12.7 mmol). The mixture was stirred overnight and then washed
to the reaction mixture and the residue was filtered off. The solid
with 4  acetic acid (3 ⫻ 15 mL), saturated NaHCO3 solution, and
was dissolved in CHCl3 (500 mL), washed with water (3 ⫻
brine. After drying with anhydrous Na2SO4, the filtrate was con-
100 mL), and dried with anhydrous Na2SO4. The solvent was re-
centrated to dryness in vacuo. The residue was purified by column
moved under vacuum and the product was crystallized from CCl4
chromatography (ethyl acetate/hexane, 3:7; Rf ⫽ 0.35 in ethyl acet-
to afford 11 as a white solid (60 g, 81%); m.p. 82 °C. 1H NMR
ate/hexane, 1:1) to yield 16 (580 mg, 76%) as a soft, light yellow
(200 MHz, CDCl3): δ ⫽ 2.42 (s, 3 H), 2.45 (s, 3 H), 3.17⫺3.26 (m,
solid. IR (neat): ν̃ ⫽ 3330, 1686 cm⫺1. 1H NMR: δ ⫽ 1.38 (s, 18
2 H), 4.04 (t, J ⫽ 5.0 Hz, 2 H), 4.84 (t, 1 H), 7.27⫺7.37 (m, 4 H),
H), 2.59 (t, J ⫽ 6.5 Hz, 4 H), 2.71 (t, J ⫽ 6.4 Hz, 4 H), 3.25 (q,
7.67⫺7.76 (m, 4 H) ppm.
J ⫽ 6.4 Hz, 4 H), 3.65⫺3.75 (m, 4 H), 4.97⫺5.23 (brs, 2 H), 7.12
(brs, 2 H) ppm. 13C NMR: δ ⫽ 28.4 (⫹), 31.5 (⫺), 32.5 (⫺), 40.0
N-(p-Tolylsulfonyl)aziridine (12):[42⫺44] A solution of 20% aqueous
(⫺), 43.8 (⫺), 79.7 (Cquat), 156.3 (Cquat), 181.66 (C⫽S) ppm. MS
KOH was rapidly added to 11 (25 g, 67 mmol) in 700 mL of freshly
(FAB): m/z (%) ⫽ 483 (100) [M ⫹ H]⫹.
distilled benzene. The two-phase mixture was vigorously stirred.
After a pink color and a white solid had appeared, stirring was 1,3-Bis(5-amino-3-thiapentyl)thiourea Dihydrochloride (17): HCl so-
continued for 0.5 h. The benzene layer was washed with water and lution (4 , 10 mL) was added to a stirred solution of 16 (482 mg,
dried with anhydrous Na2SO4. The solvent was removed under va- 1 mmol) in 15 mL of methanol. The reaction was monitored by
cuum in the presence of catalytic amounts of 4-tert-butylcatechol TLC and was complete within 8 h. The solvent was removed under
to avoid polymerization. The brownish solid was crystallized from vacuum to afford 17 as a colorless semi-solid. Yield 340 mg (95%).
a mixture of dichloromethane and hexane to provide a white solid, IR (neat): ν̃ ⫽ 3251, 2966, 2924 cm⫺1. 1H NMR (400 MHz,
yield 11 g (93%); m.p. 51 °C. 1H NMR (200 MHz, CDCl3): δ ⫽ [D4]MeOH): δ ⫽ 2.82 (t, J ⫽ 6.8 Hz, 4 H), 2.91 (t, J ⫽ 6.8 Hz, 4
2.36 (s, 4 H), 2.44 (s, 3 H), 7.32⫺7.36 (m, 2 H), 7.82⫺7.84 (m, 2 H). H), 3.22 (t, J ⫽ 6.8 Hz, 4 H), 3.73 (b, 4 H), 4.89 (brs, 8 H) ppm.
13
C NMR (100 MHz, [D4]MeOH): δ ⫽ 29.3 (⫺), 30.9 (⫺), 39.6
1,5-Bis(p-tolylsulfonamido)-3-thiapentane (13):[41,45] N-(p-Tosyl)azi- (⫺), 44.3 (⫺).
ridine (12, 11 g, 0.067 mol) was added to a solution of Na2S·9H2O
(9 g, 33 mmol) in 30% aqueous ethanol. The mixture was heated 3-Thiapentane-1,5-diyl Diisothiocyanate (18): The synthesis of 18
under reflux for 5 h, the ethanol was removed under reduced pres- was performed according to general literature procedures[39] for the
sure, and the mixture was acidified with aqueous 5% HCl. The synthesis of diisothiocyanates from diamines. A mixture of 14 (1 g,
product was extracted with ethyl acetate (300 mL) and the organic 3.54 mmol) and NaOH (300 mg, 7.18 mmol) in methanol (25 mL)
layer was washed successively with water, NaHCO3 solution, and was stirred for 1 h. Triethylamine (725 mg, 7.19 mmol) and carbon
brine and dried with anhydrous Na2SO4. The solvent was evapor- disulfide (0.85 mL, 14.2 mmol) were added dropwise to this solu-
ated under vacuum to afford a light yellow semi-solid, which was tion at room temperature. After 2 h of stirring, the reaction mixture
sufficiently pure for use in the next step without further purifica- was cooled to 0 °C, and ethyl chloroformate (1.53 g, 14.2 mmol)
tion. Yield 9.5 g (65%). 1H NMR (200 MHz, CDCl3): δ ⫽ 2.40 (s, was added dropwise from a dropping funnel. Stirring was con-
6 H), 2.51 (t, J ⫽ 6.0 Hz, 4 H), 3.03 (q, J ⫽ 6.0 Hz, 4 H), 5.32 (t, tinued for another 2 h at 20 °C. The methanol was removed under
J ⫽ 6.0 Hz, 2 H), 7.25⫺7.32 (m, 4 H), 7.68⫺7.76 (m, 4 H). vacuum and the residue was dissolved in water and CHCl3. The

3012 Eur. J. Org. Chem. 2002, 3004⫺3014

Design of a Neutral Macrocyclic Ionophore FULL PAPER
[15]
organic layer was washed successively with water and brine, dried A. Metzger, V. M. Lynch, E. V. Anslyn, Angew. Chem. Int. Ed.
with anhydrous Na2SO4, and concentrated under vacuum. The res- Engl. 1997, 36, 862⫺865.
[16]
idue was distilled at 120⫺130 °C/2⫺3 mbar until the evolution of P. D. Beer, M. G. B. Drew, Chem. Commun. 1997, 107⫺108.
[17]
ethanol and carbon oxide sulfide ceased. The deep brown residue Nitronate anion recognition with neutral receptor: B. R. Lin-
was then immediately subjected to column chromatography on ton, M. S. Goodman, A. D. Hamilton, Chem. Eur. J. 2000,
6, 2449⫺2455.
silica gel with hexane and ethyl acetate (7:3) as the eluent to afford [18]
A. P. Bisson, V. M. Lynch, M.-K. C. Monahan, E. V. Anslyn,
18 (550 mg, 76%) as an orange oil; Rf ⫽ 0.75 (ethyl acetate/hexane, Angew. Chem. 1997, 109, 2435⫺2437; Angew. Chem. Int. Ed.
3:7). IR (neat): ν̃ ⫽ 2102⫺2188 (⫺N⫽C⫽S) cm⫺1. 1H NMR: δ ⫽ Engl. 1997, 36, 2340⫺2342.
2.92 (t, J ⫽ 6.8 Hz, 4 H), 3.77 (t, J ⫽ 6.8 Hz, 4 H) ppm. 13C NMR: [19]
K. Niikura, A. P. Bisson, E. V. Anslyn, J. Chem. Soc., Perkin
δ ⫽ 32.6 (⫺), 45.4 (⫺), 133.2 (Cquat) ppm. MS (EI): m/z (%) ⫽ 204 Trans. 2 1999, 1111⫺1114.
(100). HRMS: C6H8N2S3 (203.9848) [M⫹] ⫽ 203.9849 ⫾ 0.3 ppm. [20]
R.-D. Schnebeck, E. Freisinger, B. Lippert, Angew. Chem. Int.
Ed. 1999, 38, 168⫺171.
1,9,17-Trithia-4,6,12,14,20,22-hexaazacyclotetracosane-5,13,21-tri- [21]
C. R. Bondy, P. A. Gale, S. J. Loeb, Chem. Commun. 2001,
thiane [8 (X ⴝ S)]: A mixture of 17 (355 mg, 1 mmol) and NaOH 729⫺730.
(90 mg, 2.2 mmol) in dioxane was stirred with 4 mL of water for [22]
K. Choi, A. D. Hamilton, J. Am. Chem. Soc. 2001, 123,
1 h. Compound 18 (204 mg, 1 mmol), dissolved in 4 mL of dioxane, 2456⫺2457.
was then added to this solution. After stirring for 12 h, the mixture [23]
S. Mason, T. Clifford, L. Seib, K. Kuczera, K. Bowman-James,
was heated to 70 °C for 4 h, and the solvent was then removed J. Am. Chem. Soc. 1998, 120, 8899⫺8900.
under vacuum. The residue was dissolved in acetone and subjected [24]
L. J. Barbour, G. W. Orr, J. L. Atwood, Nature 1998, 393,
to column chromatography (ethyl acetate/hexane, 7:3; Rf ⫽ 0.54, 671⫺673.
ethyl acetate) to yield 8 (X ⫽ S) (150 mg, 30%) as a white solid, [25]
A. Szumna, J. Jurczak, Eur. J. Org. Chem. 2001, 4031⫺4039.
m.p. 175⫺176 °C. IR (neat): ν̃ ⫽ 3200, 1558 cm⫺1. 1H NMR [26]
R. C. Jagessar, M. Shang, W. R. Scheidt, D. H. Burns, J. Am.
(400 MHz, [D6]DMSO): δ ⫽ 2.71 (t, J ⫽ 6.8 Hz, 12 H), 3.52⫺3.62 Chem. Soc. 1998, 120, 11684⫺11692.
[27]
(m, 12 H) ppm. 13C NMR (100 MHz, [D6]DMSO): δ ⫽ 30.1 (⫺), S. Nishizawa, P. Bühlmann, M. Iwao, Y. Umezawa, Tetrahedron
43.4 (⫺), 181.0 (Cquat) ppm. MS(FAB): m/z (%) ⫽ 487 (100) [M ⫹ Lett. 1995, 36, 6483⫺6486.
[28] [28a]
H]⫹. C15H30N6S6 (486.80): calcd. C 37.03, H 6.22, N 17.28, S 39.46; P. Bühlmann, S. Nishizawa, K. P. Xiao, Y. Umezawa, Tetra-
hedron Lett. 1997, 53, 1647⫺1654.[28b] T. W. Bell, N. M. Hext,
found C 37.04, H 6.14, N 16.75, S 38.72.
A. B. Khasanov, Pure Appl. Chem. 1998,, 70, 2371⫺2377.
[29]
R. Gross, G. Dürner, M. W. Göbel, Liebigs Ann. Chem. 1994,
49⫺58.
[30]
Acknowledgments Y. Marcus, Chem. Rev. 1988, 88, 1475⫺1498, and ref. cited
there.
This work was supported by the Deutsche Bundesstiftung Umwelt, [31]
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M.
which is gratefully acknowledged. T. F. thanks the Grand Duchy A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgom-
of Luxembourg and the Free State of Bavaria for graduate fellow- ery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Mil-
ships. We thank Prof. Michael Göbel for providing us a sample of lam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J.
compound 19. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Po-
melli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P.
Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck,
[1]
K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A.
T. S. Snowden, E. V. Anslyn, Curr. Opin. Chem. Biol. 1999, G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
3, 740⫺746. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith,
[2]
For a recent review on anion recognition, see: P. D. Beer, P. A. M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez,
Gale, Angew. Chem. 2001, 113, 502⫺532; Angew. Chem. Int. M. Challacombe, P. M. W. Gill, B. G. Johnson, W. Chen, M.
Ed. 2001, 40, 486⫺516. W. Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle, J. A.
[3]
F. Hofmeister, Arch. Exp. Pathol. Pharmakol. 1888, 24, 247. Pople, Gaussian 98, Revision A.x, Gaussian, Inc., Pittsburgh
[4]
For a recent review on neutral anion receptors, see: M. M. G. PA, 1998.
Antonisse, D. N. Reinhoudt, Chem. Commun. 1998, 443⫺448. [32]
S. Miertus, E. Scrocco, J. Tomasi, Chem. Phys. 1981, 55, 117.
[5]
Sulfate recognition: M. J. Berrocal, A. Cruz, I. H. A. Badr, L. [33]
M. Cossi, V. Barone, R. Cammi, J. Tomasi, Chem. Phys. Lett.
G. Bachas, Anal. Chem. 2000, 72, 5295⫺5299.
[6] 1996, 255, 327.
S. S. Y. Tobe, M. Mizuno, K. Naemura, Chem. Lett. 1998, [34]
N. L. Allinger, J. Am. Chem. Soc. 1977, 89, 8127.
835⫺836. [35]
[7]
J.-I. Hong, W.-S. Yeo, Tetrahedron Lett. 1998, 39, 8137⫺8140. DFT calculations on the corresponding D3-symmetric com-
[8]
H. Miyaji, W. Sato, J. L. Sessler, Angew. Chem. Int. Ed. 2000, plexes with thiourea were used to parameterize the H···Br and
39, 1777⫺1780. H···N hydrogen bond interactions by use of partial charges and
[9]
A. P. Davis, J. J. Perry, R. P. Williams, J. Am. Chem. Soc. 1997, additional potentials
[36]
119, 1793⫺1794. H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren, A.
[10]
K. Kavallieratos, C. M. Bertao, R. H. Crabtree, J. Org. Chem. Di Nola, J. R. Haak, J. Chem. Phys. 1984, 81, 3684⫺3690.
[37]
1999, 64, 1675⫺1683. T. Kohonen, ‘‘Self-Organization and Associative Memory’’, in
[11]
K. Heuzé, C. Mézière, M. Fourmigué, P. Batail, C. Coulon, Springer Series in Information Sciences, Springer, Heidelberg,
E. Canadell, P. Auban-Senzier, D. Jérome, Chem. Mater. 2000, 1984, vol. 8.
[38]
12, 1898⫺1904. B. Dietrich, T. M. Fyles, J.-M. Lehn, L. G. Pease, D. L. Fyles,
[12]
P. D. Beer, P. K. Hopkins, J. D. McKinney, Chem. Commun. J. Chem. Soc., Chem. Commun. 1978, 934⫺936.
[39]
1999, 1253⫺1254. N. G. Luk’yanenko, V. V. Limich, S. V. Shcherbakov, T. I. Kir-
[13] ichenko, J. Gen. Chem. USSR (Engl. Transl.) 1985, 55,
B. H. M. Snellink-Ruel, M. M. G. Antonisse, J. F. J. Engbersen,
P. Timmerman, D. N. Reinhoudt, Eur. J. Org. Chem. 2000, 1864⫺1866; Zh. Obshch. Khim. 1985, 55, 2100⫺2103.
[40]
165⫺170. A. V. Bogatskii, N. G. Luk’yanenko, T. I. Kirichenko, J. Org.
[14]
S. Sasaki, M. Mizuno, K. Naemura, Y. Tobe, J. Org. Chem. Chem. USSR (Engl. Transl.) 1980, 16, 1124⫺1129; Zh. Org.
2000, 275⫺283. Khim. 1980, 16, 1301⫺1307.

Eur. J. Org. Chem. 2002, 3004⫺3014 3013

FULL PAPER R. Herges, B. König et al.

[41] [48]
S. Chandrasekhar, A. McAuley, J. Chem. Soc., Dalton Trans. Reinhoudt et al. have recently reported the formation of an
1992, 20, 2967⫺2970. aggregate, with a stoichiometry of 2:3, of a tetraurea com-
[42]
B. Dietrich, M. W. Hosseini, J.-M. Lehn, R. B. Sessiions, Helv. pound with H2PO4⫺ anions. No binding constant could be de-
Chim. Acta 1985, 68, 289⫺299. rived: B. H. M. Snellink-Ruël, M. M. G. Antonisse, J. F. J.
[43]
A. E. Martin, T. M. Ford, J. E. Bulkowski, J. Org. Chem. 1982, Engbersen, P. Timmerman, D. N. Reinhoudt, Eur. J. Org.
47, 412⫺415. Chem. 2000, 165⫺170.
[44] [49]
M. Pietraszkiewicz, J. Jurczak, Tetrahedron 1984, 40, The pKa values used in the plot are those valid for water. Al-
2967⫺2970. though the same order may be expected in DMSO, significant
[45]
S. Ohtsuka, M. Kodera, K. Motoda, M. Ohba, H. Okawa, J. alterations are possible; pKa values describe an equilibrium be-
Chem. Soc., Dalton Trans. 1995, 16, 2599⫺2604. tween the anion and protons, which has similarities to the
[46]
The program HYP NMR was used to derive association con- formations of a hydrogen bond, but is obviously not the same.
[50]
stants: C. Frassineti, S. Ghelli, P. Gans, A. Sabatini, M. S. Mo- Some of the titrations were repeated at different ionophore
ruzzi, A. Vacca, Anal. Biochem. 1995, 231, 374⫺382. concentration to confirm the observations.
[47]
G. Jander, K. F. Jahr, Maßanalyse, 15th ed., de Gruyter, Berlin, Received April 5, 2002
1989, p. 81. [O02192]

3014 Eur. J. Org. Chem. 2002, 3004⫺3014