Addressing association entropy by reconstructing guanidinium anchor groups for anion binding: Design, synthesis and host-guest binding studies in polar and protic solutions

Article abstract of DOI:10.1002/chem.200702036

The bicyclic hexahydropyrimidino[1,2a]pyrimidine cationic scaffold has a well-known capacity to bind a variety of oxoanions (phosphates, carboxylates, squarates, phosphinates). Based on this feature, the parent host was supplemented with sec-carboxamido substituents to generate compounds 1-3 in an effort to improve the anion-binding affinity and selectivity and to learn about the role and magnitude of entropic factors. Bicyclic guanidinium compounds were prepared by a convergent strategy via the corresponding tetraester 22 followed by catalytic amidation. Host-guest binding studies with isothermal titration calorimetry in acetonitrile probed the behavior of artificial hosts 1-3 in comparison with the tetraallyl-guanidinium compound 4 on binding p-nitrobenzoate, dihydrogenphosphate, and 2,2'-bisphenolcyclophosphate guests that showed enhanced affinities in the 10⁵-10⁶M ^-1 range. Contrary to expectation, better binding emerges from more positive association entropies rather than from stronger enthalpic interactions (hydrogen bonding). In an NMR spectroscopy titration in DMSO, o-phthalate was sufficiently basic to abstract a proton from the guanidinium function, as confirmed by an X-ray crystal structure of the product. The novel carboxamide-appended anchor groups also bind carboxylates and phosphates, but not hydrogen sulfate in methanol with affinities in excess of 10⁴ M^-1. The energetic signature of the complexation in methanol is inverted with respect to acetonitrile solvent and shows a pattern of general ion pairing with strong positive entropies overcompensating endothermic binding enthalpies. This study provides an example of the fact that bona fide decoration of a parent guanidinium anchor function with an additional binding functionality may provide the desired enhancement of the host-guest affinity, yet for a different reason than that implemented by design as guided by standard molecular modeling.

Full text of DOI:10.1002/chem.200702036

DOI: 10.1002/chem.200702036
Addressing Association Entropy by Reconstructing Guanidinium Anchor
Groups for Anion Binding: Design, Synthesis, and Host–Guest Binding
Studies in Polar and Protic Solutions
Vinod D. Jadhav, Eberhardt Herdtweck, and Franz P. Schmidtchen*[a]
Abstract: The bicyclic hexahydro-
pyrimidino[1,2a]pyrimidine cationic
guanidinium compound 4 on binding p-
nitrobenzoate, dihydrogenphosphate,
The
novel
carboxamide-appended
G
U
anchor groups also bind carboxylates
and phosphates, but not hydrogen sul-
fate in methanol with affinities in
excess of 104 mÀ1. The energetic signa-
ture of the complexation in methanol is
inverted with respect to acetonitrile
solvent and shows a pattern of general
ion pairing with strong positive entro-
pies overcompensating endothermic
binding enthalpies. This study provides
an example of the fact that bona fide
decoration of a parent guanidinium
anchor function with an additional
binding functionality may provide the
desired enhancement of the host–guest
affinity, yet for a different reason than
that implemented by design as guided
by standard molecular modeling.
scaffold has a well-known capacity to
bind a variety of oxoanions (phos-
phates, carboxylates, squarates, phos-
phinates). Based on this feature, the
parent host was supplemented with
sec-carboxamido substituents to gener-
ate compounds 1–3 in an effort to im-
prove the anion-binding affinity and se-
lectivity and to learn about the role
and magnitude of entropic factors. Bi-
cyclic guanidinium compounds were
prepared by a convergent strategy via
the corresponding tetraester 22 fol-
lowed by catalytic amidation. Host–
guest binding studies with isothermal
titration calorimetry in acetonitrile
probed the behavior of artificial hosts
1–3 in comparison with the tetraallyl-
and
2,2’-bisphenolcyclophosphate
guests that showed enhanced affinities
in the 105–106 mÀ1 range. Contrary to
expectation, better binding emerges
from more positive association entro-
pies rather than from stronger enthalp-
ic interactions (hydrogen bonding). In
an NMR spectroscopy titration in
DMSO, o-phthalate was sufficiently
basic to abstract a proton from the gua-
nidinium function, as confirmed by an
X-ray crystal structure of the product.
Keywords:
anion recognition
·
entropy · guanidinium · host–guest
systems · isothermal titration calo-
rimetry
Introduction
from spatially quite relaxed ion pairing (e.g., in the coiling
of DNA around the basic histone protein core of nucleo-
somes[7]) to a very dedicated geometric correspondence of
the two moieties, as sketched in Scheme 1 (e.g., in the recog-
nition of C-terminal carboxylate by carboxypeptidase A).[8])
The latter motif, in particular, inspired the use of this pat-
tern in artificial receptors,[9–22] although such a well-defined
structural relationship cannot be taken for granted. For
most instrumental methods, it is
Cationic guanidinium functions play prominent roles in mo-
lecular recognition of oxoanionic species in both the living
world and in abiotic applications.[1–5] In biology they are fre-
quently involved in anion binding to proteins[6] as a conse-
quence of their presence in the side chain of the ubiquitous
proteinogenic amino acid arginine. The structural definition
of the arginine–oxoanion interaction covers a broad range—
indistinguishable from a struc-
turally relaxed and dynamically
disordered, yet time-averaged
[a] Dr. V. D. Jadhav, Dr. E. Herdtweck, Prof. Dr. F. P. Schmidtchen
Department of Chemistry
Technical University of Munich
Lichtenbergstrasse 4
85747 Garching (Germany)
Fax : (+49)89-289-14698
E-mail: Schmidtchen@ch.tum.de
preference for the depicted as-
sociation mode. Distinction be-
tween the two extreme binding
Scheme 1. A sketch of the gua-
modes is of paramount impor-
nidinium oxoanion binding
motif as it occurs in many nat-
tance for the functional utility
of the anchor group and is thus ural and artificial receptors.
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
6098
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6098 – 6107

FULL PAPER
indispensable in the rational design of artificial guanidinium
receptors.
amazing fact that guest affinity in acetonitrile was indeed
enhanced in 1 relative to an analogous guanidinium host 4
that lacked the carboxamide moieties; however, for a differ-
ent reason than originally anticipated: for all guest anions
tested, the attractive enthalpy of association was diminished
and the higher affinity was exclusively due to a much more
positive association entropy component. From a consider-
ation of the main factors contributing to this surprising ob-
servation, a rationalization was advanced that the deeper
cause of this unexpected result could be attributed the in-
crease in configurational entropy of the binding partners as
a consequence of the low structural definition in the host–
guest complex. Herein we report the design, synthesis, and
host–guest binding of two more candidates (2 and 3) of the
guanidinium carboxamide variety that stand out by their
host–guest affinity, even in polar protic solutions (Kass up to
37000mÀ1 in methanol). In addition, an unprecedented
transprotonation by the o-phthalate dianion was found that
supports calls for caution voiced recently with respect to
anion binding to artificial amide receptors.[29]
Regrettably, there is no immediate and unambiguous ana-
lytical tool available to report on the geometrical fuzziness
of host–guest interactions in general. The diversity of bind-
ing modes, however, must be reflected in the entropy com-
ponent of the association energetics because the respective
configurational entropy contributes a substantial share to
the overall entropy output. Even though the entropy of as-
sociation is now readily accessible by modern calorimetry
techniques, unfolding of the configurational component is
less straightforward. At best, a qualitative estimate can be
deduced from a comparison of closely related host–guest
pairs measured under strictly identical environmental condi-
tions (solvent, solution composition, temperature, etc.) to
minimize interference from unspecific solvation effects.
Such a trend analysis may then reveal a qualitative picture
of the variation of static and dynamic disorder in the host–
guest complex with incremental structural modification.
The parent guanidinium receptor depicted in Scheme 1
offers the opportunity to evaluate a structure–binding mode
correlation in a popular supramolecular binding motif of
proven utility. This abiotic host–guest system features a
robust and easy-to-grasp recognition pattern that is unper-
turbed by remote influences endemic to the much more
complex natural pendants. The trend analysis of the energet-
ics in a series of rather rudimentary molecular recognition
motifs should provide an experimental probe to address the
basic role of entropy in molecular recognition and its poten-
tial for supramolecular design.
Results and Discussion
Planning and synthesis: Before defining the guanidinium
carboxamides as synthetic targets, molecular modeling was
carried out to confirm their suitability. Geometric optimiza-
tion of the benzoate salt of anilide 3 in vacuo initially con-
verged on a structure (Figure 1a) that placed the guest
anion almost juxtaposed to the guanidinium unit in the fa-
miliar salt-bridge/hydrogen-bond pattern depicted in
Scheme 1. This motif was supplemented by two further hy-
drogen bonds donated from a carboxamide on either the
left- or right-hand side of the bicyclic molecule. The two re-
maining carboxamides were engaged in lateral amide–amide
interactions to form 6-membered hydrogen-bonded rings.
Subjecting this ensemble to a molecular dynamics run over
20 ps followed by simulated annealing produced a significant
structural change. Now the guest carboxylate group was
held by two hydrogen bonds to the guanidinium cation and
one carboxamide group in a slightly dissymmetric fashion,
while two amides again formed hydrogen-bonded rings;
however, one of them spans the bicyclic frame to form a
supramolecular macrocycle (Figure 1b). One carboxamide
Towards this goal, appending the 6,6-membered bicyclic
guanidinium scaffold with sec-carboxamide moieties in the
a,a’ positions to give guanidinium hosts 1–3 was deemed an
optimal choice because sec-carboxamides are well-estab-
lished anchor groups in anion binding[23–27] and their geminal
placement in the immediate vicinity of the binding site was
expected to enforce its sticki-
ness towards the anionic guests
by the sheer accumulation of
hydrogen-bond donor functions
and the suboptimal solvation of
the individual groups as a result
of overcrowding.
We have already reported an
example that followed this
design outline.[28] The binding
of 6,6-gua-carboxamide 1 to a
Figure 1. Energy-minimized structure of the benzoate salt of anilide 3 by using the Amber force field a) in
series of oxoanions revealed the vacuo, b) after a molecular dynamics run for 20 ps and simulated annealing, and c) rear view of structure b.
Chem. Eur. J. 2008, 14, 6098 – 6107
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6099

F. P. Schmidtchen et al.
was released from the network of hydrogen bonds and this
pattern remained untouched when the complex was embed-
ded and energy-minimized under periodic boundary condi-
tions in a water box containing more than 5800 water mole-
cules.
The modeling results support our initial idea that the car-
boxamido functions help in guest binding. Even though not
all potential hydrogen-bond donors were found to simulta-
neously interact with the bound anion in the minimum-
energy structure, they stitch together a network of hydrogen
bonds. This organizes the entire host and thus aids host–
guest binding by limiting the entropic loss expected on
merging the binding partners into an associated structure.
On this basis, guanidinium compounds 1–3 were judged to
be promising synthetic goals. As suggested by their symmet-
ric topology, a convergent strategy, such as that shown in
Scheme 2, appeared to be the most efficient for their synthe-
sis. Apart from the final standard conversion of an ester into
an amido group, our approach takes advantage of the com-
bination of two building blocks, 8 and 11, that are simple de-
rivatives of cheap commercial starting materials.
Scheme 3. Cbz=carbobenzyloxy
zation taking one deprotected amino group as the nucleo-
phile. This outcome did not change when more drastic con-
ditions in transfer hydrogenation were employed (Pd/C,
cyclohexadiene, reflux) or by the use of trimethylsilyliodide
as a demasking agent. Strong acid solvolysis (HBr/propionic
acid), however, changed the course of the reaction and
yielded monocyclic guanidinium salt 15 as the main product.
Unfortunately, this compound resisted all attempts to pro-
duce the desired bicyclic structure by amine exchange of the
guanidinium unit.[31] We envisaged an escape from this
dead-end road to be the elaboration of the cyanamido func-
tion, which was introduced at the very start to serve as an
amino protecting group as well as a synthon for the core
guanidino carbon, into a more reactive moiety susceptible
to attack by the shielded carbinylamino atom. Treatment of
cyanoamide 12 with dilute sulfuric acid cleanly gave the cor-
responding urea 16, which in turn could be deprotected by
catalytic hydrogenation to give 18 in quantitative yield. The
putatively simple monocyclization to cyclic urea 20 present-
ed a major obstacle, again testifying to the low nucleophilici-
ty of the carbinylamino group. After some optimization,
compound 20 was obtained in a moderate yield of 27%,
which suggested that we should look for a better alternative.
A viable alternative was to subject 12 to the Pinner reac-
tion to afford isourea 17 as the initial step in a sequence
that was profitably conducted without purification of the in-
Scheme 2. Retrosynthetic approach to compounds 1–3.
The synthesis of guanidinio carboxamides 1–3 followed
the route depicted in Schemes 3 and 4. The alkylative C C
À
coupling of building blocks 8[28] and 11,[30] which were ob-
tained by short literature procedures, required low reaction
temperatures (<108C) for several days to give a yield of
>70% of key intermediate 12 and to avoid substantial loss
due to the formation of elimination product 13. A 20:1 pref-
erence for the desired product was achieved under the speci-
fied conditions, whereas raising the temperature to 508C
produced a 1:1 mixture of these compounds.
Originally, deprotection of the amino functions by hydro-
genation in the next step was presumed to directly furnish
the guanidinium monocycle 15, which seemed to be an at-
tractive candidate for ring closure to the bicyclic skeleton.
Instead, with the Pd/carbon catalysts tested, the product iso-
lated in almost quantitative yield was monocyclic formami-
dine 14, which undoubtedly arose from preferential hydro-
genation of the cyanamido function followed by monocycli-
6100
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6098 – 6107

Reconstructing Guanidinium Anchor Groups
FULL PAPER
signed structures are fully supported by single-crystal X-ray
analyses of the halide salts (see the Supporting Informa-
tion), which also show the peculiar hydrogen-bonding ring
formation within the malonic amide moieties that was pre-
dicted by the modeling studies. The occurrence of this motif
in the solid state in two compounds attests to the stability of
this pattern. In conjunction with the computational results,
there is now reason to presume that this kind of host preor-
ganization will also persist in solution. All of the binding
studies used iodide as the counteranion because these salts
are readily purified and this counterion does not interfere
with oxoanion complexation.[12]
Binding studies: Initial NMR spectroscopy binding studies
of propylamido host 1 and dihydrogenphosphate in acetoni-
trile had revealed a saturation isotherm when using the
amide NH signal as a probe.[28] However, a significant non-
random swing of the experimental data points across the
fitted hyperbola describing the 1:1 stoichiometric model was
interpreted (and later confirmed by isothermal calorimetry
(ITC) measurements) as an indication of higher-order com-
plex formation. To avoid such risk and further escape any
restrictions set by insufficient solubility of the host and
guest salts, the solvent was switched to DMSO. Titration of
anilide host 3 with o-phthalate dianion (tetraethylammo-
nium salt) in this solvent, however, did not show the antici-
pated shift of the NH signals characteristic of rapid equili-
bration-averaged host–guest binding (Figure 2).
Instead, the NH signals of the amido and guanidinium
groups broadened and eventually disappeared, but both
kept their positions, indicative of a slow exchange process
on the NMR spectroscopy timescale. The aromatic resonan-
ces of either salt were unaffected at host–guest molar ratios
below 1:1, whereas the methylene signals of the 6-mem-
bered rings were gradually shifted upfield (Figure 2). Re-
grettably, this change could not be quantified because the
host signals hide behind the solvent and guest resonances
for a considerable proportion of the titration. At host–guest
molar ratios greater than 1:1, substantial redistribution of
the aromatic signals was seen (Figure 2d, e) and was decon-
voluted by means of a COSY 2D spectrum into two sets of
coupling networks. The relative proportion of these sets did
not depend on the concentration of the added guest, which
suggests that they are not derived from supramolecular asso-
ciation.
A tentative explanation was gleaned from the serendipi-
tous crystallization of such a titration mixture from metha-
nol. Contrary to expectation, the X-ray analysis of the crys-
tals did not reveal the presence of a guest anion, but corre-
sponded to the free base form of host 3, as depicted in
Figure 3. The structure shows a similar hydrogen-bonding
pattern at one geminal dicarboxamide side, as that observed
in guanidinium salt 3 (see the Supporting Information). The
other malonic amide moiety, however, is no longer fixed in
a mutual hydrogen-bond interaction, but contains the NH
vectors pointing towards the proximal nitrogen atom of the
guanidine (NH···N ꢀ2.1 Š). The free base nature is clearly
Scheme 4.
termediates. Thus, imidate 17 was hydrogenated and cau-
tiously neutralized to give bicyclic guanidinium tetraester 22
in a one-pot reaction in a yield of 72% over four steps. The
ester-to-amide transformation was initially attempted by
using the classic method of hydrolysis to give acid 23 fol-
lowed by amidation after activation to the acid chloride or a
mixed anhydride. This stage, however, proved intractable
owing to the insolubility of the starting acid. Direct transa-
midation of ester 22 with the respective amines by using
promotion by alkylaluminium reagents proved effective.[32,33]
The acidity of the substrate and the alcohol formed required
a considerable excess of the organometallic Lewis acid, yet
target compounds 1–3 were isolated from the multistep se-
quence in a yield of 70–80% as crystalline salts. The as-
Chem. Eur. J. 2008, 14, 6098 – 6107
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6101

F. P. Schmidtchen et al.
structures adjacent to or remote
from the C=N double bond in
the guanidine. Similar switching
between genuine supramolec-
ular association or covalent
proton transfer has frequently
been found with the fluoride
anion,[29,35] which also consti-
tutes an extremely basic species
in organic solutions.[36]
Previous investigations of the
energetics and speciation of
guanidinium phosphate host–
guest binding revealed an ad-
vantage of ITC over NMR
spectroscopy for analysis in this
peculiar case.[28] Following up
on these studies, host com-
pounds 2 and 3 were probed by
p-nitrobenzoate, phosphate, and
Figure 2. 1H NMR spectra ([D6]DMSO, 250 MHz) recorded for the complex formed by the anilide 3 (2 mm)
with a) 0 mm, b) 1.2 mm, c) 2.4 mm, d) 3.6 mm, and e) 4.8 mm of o-phthalate. The designated signals arise from
~
the guest anion ( ) and methylene groups ( ) of the 6-membered rings in the anilide host 3.
*
cyclic phosphate diester 24 in acetonitrile to assess the con-
sequences and possible generalizations for the design of gua-
nidinium anchor groups. The recurring feature in these inter-
actions is the tendency to form complexes beyond 1:1 stoi-
chiometries. An illustrative example is depicted in Figure 4
for the interaction of 3 with nitrobenzoate, which shows
strong deviations from the response expected for a 1:1 bind-
ing model in the regimes of excess host over guest (i.e., at
molar ratios <1 at the beginning of the titration) as well as
at higher guest-to-host molar ratios. The analysis of such sys-
tems is greatly facilitated when experimental conditions are
found that allow simplification of the interaction model.
This is shown in Figure 4b as an example in which dilution
of the starting concentration largely suppressed the initial
formation of higher-order host–guest complexes. Under this
precondition, the deconvolution of the macroscopic heat re-
sponse into individual components is possible, as conveyed
by the fitted curve which represents the experimental data
quite well (Figure 4). The conditional association constants
collected in Tables 1 and 2 were derived on this basis.
Figure 3. ORTEP-style plot of anilide 3 showing the free base form (dC1
À
N2 =1.29 Š; dC1 N1 =1.40 Š).[37] Thermal ellipsoids are drawn at the 50%
À
probability level.
In general, the pattern observed with hosts 2 and 3 greatly
resembles the results obtained earlier with propylamido host
1 binding the same guest.[28] Relative to tetraallylguanidini-
um host 4, which displays clean 1:1 binding, all carboxamido
hosts 1–3 possess a higher affinity for these guests; however,
not as a result of more attractive binding (which was the
original foundation of the design concept), but for a more
positive output in the association entropy.[28] The only excep-
tion to this rule is the binding of anilide host 3 to dihydro-
genphosphate. Although in all cases this guest shows the
most negative interaction enthalpy among the oxoanions,
which most probably reflects the greater potential for hydro-
gen bonding because of its additional hydrogen-bond donor
capabilities, anilido host 3 stands out for its special energetic
signature. Host 3 is the only anchor group in this series that
provides a negative entropy contribution that counteracts a
dramatically enhanced attractive enthalpy of binding. On
À
reflected by a much shorter C N distance in the proximal
ring (ꢀ1.29 Š versus 1.40 Š in the other ring), indicating
double bond character, and the pronounced pyramidalized
configuration of the nitrogen atom in the distal ring. Clearly,
the basicity of the phthalate dianion in the organic solvent,
probably owing to a short and strong hydrogen bond con-
necting the carboxylate moieties in the monoanion, suffices
to deprotonate the guanidinium cation,[34] particularly if the
free base form of the host is stabilized by hydrogen bonding
to the adjacent carboxamides. Such deprotonation is impos-
sible with all of the other oxoanions used in this study (see
below) owing to their much lower basicities. The molecular
process occurring during the titration in DMSO thus ap-
pears to be a transprotonation rather than a host–guest in-
teraction. The two sets of observed aromatic proton signals
most probably refer to the respective malonic amide sub-
6102
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6098 – 6107

Reconstructing Guanidinium Anchor Groups
FULL PAPER
ergistically supported by mutual
interactions of the phenyl rings
to explain this result.
The
respectable
affinity
found for these host–guest pairs
in the polar, yet non-protic ace-
tonitrile solvent fostered the ex-
pectation that association might
also persist in protic solvents.
ITC titrations performed with
anilide host 3 in methanol did
indeed corroborate this pre-
sumption (Table 2). The ener-
getic signatures are in contrast
with the results obtained in ace-
tonitrile. In methanol, all asso-
ciations
are
endothermic
(except for one step in the
binding of p-nitrobenzoate) and
owe their affinity to an exces-
sive positive entropy contribu-
tion. Thus, they resemble the
general ion-pairing process that
features substantial desolvation
of both ionic partners without
Figure 4. ITC traces of the titration of p-nitrobenzoate into the solution of host 3 in acetonitrile at 303 K a) at
4.33 mm and b) at 1.3 mm. The solid lines represent fitting curves on the basis of a 1:1 binding model (left) or
sequential 2:1 model (right).
the balance sheet, this leaves a smaller affinity than ob-
served in any other host–guest combination; however, for
the benefit of a better structured complex.[38,39] Undoubted-
ly, the organizing role of the phosphate guest must be syn-
generating a singular host–guest complex structure. Disre-
garding hydrogen sulfate, which did not produce a sufficient
heat effect to allow analysis, the other guests showed unam-
biguous supramolecular binding behavior.
Table 1. 1:1 Host–guest binding energetics of guanidinium anchor groups (iodide salts) with oxoanions in acetonitrile at 303 K.
TBA p-nitrobenzoate
TBA H2PO4
TBA 2,2’-bisphenolcyclophosphate 24
Kass
ÀDG8
ÀDH8
TDS8
Kass
ÀDG8
ÀDH8
TDS8
Kass
ÀDG8
ÀDH8
TDS8
[mÀ1
]
[kJmolÀ1
]
[kJmolÀ1
]
[kJmolÀ1
]
[MÀ1
]
[kJmolÀ1
]
[kJmolÀ1
]
[kJmolÀ1
]
[MÀ1
]
[kJmolÀ1
]
[kJmolÀ1
]
[kJmolÀ1
]
1
2
3
4
2.0”10À5 30.8
9.3”10À4 28.8
1.2”10À6 35.5
7.1”10À4 28.2
17.6
13.2
22.1
22.3
+13.2
+15.6
+13.1
+5.9
1.3”106 35.6
1.6”106 36.1
3.8”104 26.5
2.3”104 24.6
17.5
12.9
41.9
23.4
+18.0
+23.2
À15.4
+1.1
4.8”104 27.2
1.7”104 24.4
2.7”105 31.6
1.9”104 24.7
10.3
9.9
12.1
14.7
+16.8
+14.6
+19.4
+10.0
Table 2. Energetics of oxoanion binding (tetraethylammonium/tetrabutylammonium salts) to anilide 3 (as the iodide salt) in methanol at 303 K. The stoi-
chiometry number n refers to the guest/host molar ratio as the experimentally determined fit parameter in a stepwise equilibrium system.
Entry
1
Guest
Model[a]
Kass [MÀ1
]
DG8 [kJmolÀ1
]
DH8 [kJmolÀ1
]
TDS8 [kJmolÀ1
]
TBA benzoate
A
K1 =180.3
K2 =3846
K1 =112
DG1 =À12.9
DG2 =À20.5
DG1 =À11.7
DG2 =À24.2
DG1 =À11.3
DG2 =À21.5
DG1 =À19.1
DG2 =À24.5
À25.0
DH1 =29.9
DH2 =49.4
DH1 =À67.5
DH2 =88.0
DH1 =6.7
DH2 =4.8
DH1 =3.0
DH2 =65.7
45.2
42.9
70.1
À55.7
112.2
18.1
26.2
22.0
90.3
70.2
2
4
4
TBA-p-nitrobenzoate
TBA H2PO4
A
A
A
K2 =1.7”104
K1 =96.7
K2 =5968
K1 =2192
K2 =1.95”104
3.3”104
TEA naphthalene dicarboxylate
5
6
7
8
9
10
11
TEA phthalate
TEA terephthalate
TEA HSO4
B
B, n=0.35
1.45”104
À22.1
90.3
112.4
insufficient heat response
TEA oxalate
B, n=0.57
B, n=0.4
B, n=0.45
B, n=0.5
3.7”104
8441
À25.0
À21.7
À24.1
À20.1
60.6
85.6
59.8
48.6
54.2
TEA fumarate
TEA isophthalate
TEA squarate
38.0
24.5
34.2
1.7”104
4077
[a] A=titration mode: guest into host solution; 2 sequential-site model; ligand-in-cell, B=titration mode: guest into host solution; one-site model.
Chem. Eur. J. 2008, 14, 6098 – 6107
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6103

F. P. Schmidtchen et al.
nected to a Knauer L 4250 UV detector, a SEDEX 55 light-scattering de-
tector, and a Kipps&Zonen two-channel recorder. The columns used in
HPLC analysis were Phenomenex, Aqua C18, 250”4.60 mm, 5 m column,
and Nucleodur-100-5 C8 ec column. The solvents were purchased in pure
analytical grade and distilled before use except for DMF, which was pur-
chased in anhydrous quality from Aldrich, and acetonitrile, which was
purchased in HPLC quality from Baker. Aqueous solutions were pre-
pared from deionized, glass distilled water. The solvents (CH2Cl2,
CH3CN) were dried by passing them through a small column of activated
alumina directly into the reaction vessel. All other chemicals were pur-
chased in reagent quality from commercial sources and used as received.
1H and 13C NMR spectra were recorded on either a Brucker AC 360
(MHz) or 250 (MHz) instrument and were referenced with respect to the
residual solvent peak. All of the melting points were measured in open
capillary tubes by using a Fisher–Jones apparatus. Mass spectra were ob-
tained on a Finnigan LQC: electrospray ionization (ESI, HPLC-MS) in-
strument. HRMS spectra were obtained on a Bruker microTOF-Q instru-
ment. Elemental analyses were carried out by the microanalytical labora-
tory of the TU Munich. Calorimetric titrations were performed on the
Isothermal Titration Calorimeter MCS-ITC from Microcal (USA). Mo-
lecular modeling was performed on Pentium PCs with HyperChem 8 mo-
lecular-modeling software from Hypercube. X-ray crystal structure analy-
ses were carried out by the Inorganic Chemistry Department of TU
Munich.
Except for phthalate, which adhered to a 1:1 stoichiomet-
ric model, the heat responses of the other anions required
more complicated binding models to reproduce the enthalpy
output. In some cases (Table 2 entries 6–11), the individual
steps could not be deconvoluted; however, the stoichiomet-
ric factor clearly indicated the involvement of two binding
events representing the association of one and two guanidi-
nium anchor groups onto the bis-anionic guest. Charge
matching is an important, yet not the exclusive cause for
higher complex formation. Also, in the case of monoanionic
guests the energetic profile cannot be fitted without the par-
ticipation of higher-order complexes. However, the absolute
affinities are somewhat smaller than those of the dianions.
Nevertheless, host–guest association between singly charged
partners in a protic solution reaches quite an impressive
level (Kass =104 mÀ1), despite the simple construction blue-
print of this anchor function. Selectivity is poor (about a
factor of 10), but may be improved by implementing restric-
tions affecting the accessibility and conformational flexibili-
ty of the binding site. A sufficient amount of binding free
energy in polar solvents, as secured in this study, is a neces-
sary prerequisite to allow tuning of the various design op-
tions.
Bis(2-iodoethyl)cyanamide (8): Bis(2-chloroethyl)amine hydrochloride
(6, 17.9 g,100 mmol) was dissolved in a mixture of water (50 mL) and
CH2Cl2 (100 mL). The mixture was cooled to 08C and 4n aqueous
NaOH (25 mL) was added. A solution of 5 (13.1 g, 125 mmol) in CH2Cl2
(to make 30 mL) and 4n aqueous NaOH (30 mL) were added alternately
portionwise with vigorous stirring while keeping the temperature below
58C. The reaction mixture was then stirred for another 30 min at RT. The
two phases were separated. The organic phase was washed with dilute
acetic acid and then with brine, dried over magnesium sulfate, and the
solvent was evaporated in vacuo to obtain a liquid residue that was dis-
tilled in a Kugelrohr apparatus (1308C/0.05 torr) to afford 7 as a colorless
viscous liquid (11.65 g, 70%). This liquid (11.65 g, 70 mmol) was dis-
solved in acetone (50 mL) to which finely powdered NaI (30 g,
200 mmol) was added, and the mixture was heated at reflux for 3 d.
After complete conversion to 8, as seen from the HPLC analysis, the
mixture was evaporated in vacuo. The resulting residue was redissolved
in toluene and filtered. The filtrate was passed through a pad of alumina
and then concentrated and dried under a high vacuum to afford 8 as a
viscous colorless liquid (18.13 g, 74%). HPLC analysis: Rv =10 mL, Nu-
cleodure-100–5 C8 ec column, UV220, flow=1 mLminÀ1, gradient from
10% CH3OH to 50% CH3OH in 10 min and then 50% CH3OH to 90%
CH3OH in next 10 min, 0.1% TFA; 1H NMR (360 MHz, CDCl3, 258C):
d=3.45 (t, J=7.0 Hz, 4H; -N-CH2), 3.29 (t, J=7.0 Hz, 4H; -CH2I);
13C NMR (90.56 MHz, CDCl3, 258C): d=115.1 (-CN), 53.75 (-N-CH2),
À0.7 (-CH2I).
Conclusion
With the goal of learning about the role of entropy in mo-
lecular recognition, the elaboration of the prominent bicy-
clic guanidinium anchor group for oxoanions was undertak-
en by supplementation of the parent moiety with sec-carbox-
amido groups. X-ray crystal structures confirmed the expect-
ation emerging from modeling studies that the variety of
conformations is limited by the build-up of an intramolecu-
lar hydrogen-bonded network. The resulting structural or-
ganization predisposes the guanidinium host for guest bind-
ing using only a fraction of the formally available hydrogen-
bond donors and allows rationalization of an unexpected
transprotonation occurring with the most basic phthalate
guest. Binding studies by ITC titrations reveal high affinity
in acetonitrile that exceeds ordinary ion pairing between
monoionic partners by a hundred thousand fold.[40] The af-
finities are somewhat diminished in methanol (Kass =103–
104 mÀ1), but more importantly show an energetic profile
that is unlikely to emerge from geometrically dedicated
complex formation. In all of the cases studied, entropy con-
tributes a substantial share to the free energy of molecular
recognition and even marks the dominant role in the protic
solvent methanol. This experimental fact calls the validity of
exclusively enthalpy-based design approaches into question.
Benzyloxycarbonylamino malonic acid diethyl ester (11): Diethyl 2-ami-
nomalonate hydrochloride (10, 21.2 g, 100 mmol) and bis(trimethylsilyl)
acetamide (BTSA; 24.8 mL, 100 mmol) were taken up in anhydrous di-
ethyl ether (200 mL) under an inert atmosphere. The reaction mixture
was cooled to 08C in an ice bath and 9 (14.4 mL, 100 mmol) was rapidly
added. After addition was complete, another batch of BTSA (24.8 mL,
100 mmol) was added. The resulting cloudy reaction mixture was then al-
lowed to warm slowly to RT and stirred for 30 min. On completion of the
reaction, as monitored by HPLC analysis, the reaction mixture was
washed with 0.1n aqueous HCl (100 mL), dried over magnesium sulfate,
and then evaporated in a rotary evaporator to give a gummy residue that
was recrystallized from hexane/ether (1:1) to afford 11 as a white solid
(26.5 g, 85%). HPLC analysis: Rv =13 mL, Phenomenex, Aqua C18, 250”
4.6 mm, 5 m column, UV220
, , gradient from 50%
flow=1 mLminÀ1
CH3OH to 90% CH3OH in 10 min and then 90% CH3OH for a further
5 min, 0.1% TFA; m.p. 30–328C (hexane/ether), (lit.[41] m.p. 32–338C);
1H NMR (360 MHz, CDCl3): d=7.36 (s, 5H; aromatic-H), 5.85 (d, J=
6.8 Hz, 1H; -NH), 5.14 (s, 2H; PhCH2O-), 5.03 (d, J=7.4 Hz, 1H; -CH-
N), 4.29 (q, J=6.5 Hz, 4H; -OCH2CH3), 1.32 (t, J=7.0 Hz, 6H;
-OCH2CH3); 13C NMR (90.56 MHz; CDCl3): d=166.3 (-CO, ester), 155.4
Experimental Section
All of the experiments in organic solvents were performed under a nitro-
gen atmosphere and monitored by HPLC. The HPLC analyses were per-
formed on a Merck-Hitachi instrument L 6200A or 655 A-11 pump con-
6104
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6098 – 6107

Reconstructing Guanidinium Anchor Groups
FULL PAPER
(-NCO-), 135.9, 128.5, 128.2, 128.1 (aromatic carbons), 67.3 (PhCH2O-),
62.6 (-OCH2CH3), 57.7 (-CH-), 13.9 (-OCH2CH3); MS: m/z (%): 310.2
(100) [M+H]+.
CH3OH, over the next 10 min, 0.1% TFA; m.p. 1368C (ethyl acetate/
CH2Cl2); 1H NMR (360 MHz, CD3CN): d=9.64 (s, 2H; guanidinium-H),
4.26 (q, J=7.05 Hz, 8H; -OCH2CH3), 3.38 (t, J=6.1 Hz, 4H; -N-CH2),
2.41 (t, J=6.1 Hz, 4H; -CCH2-), 1.28 (t, J=7.02 Hz, 12H; -OCH2CH3);
13C NMR (90.56 MHz, CD3CN): d=167.7 (-CO, esters), 151.1 (guanidi-
nium carbon), 64.1 (-OCH2CH3), 63.1 (quaternary carbons), 44.4 (-N-
CH2), 26.1 (-CCH2-), 14.2 (-OCH2CH3); MS: m/z (%): 428 (100)
N,N-Bis(3-benzyloxycarbonylamino-3,3-diethoxycarbonyl)propyl-1-cyan-
ACHTREUNGamide (12): Carbamate 11 (26.48 g, 85.7 mmol) in dry DMF (60 mL) was
added to a suspension of NaH (3.59 g, 90 mmol) (60% dispersion in min-
eral oil) in dry DMF (20 mL) cooled to 58C. The rate of addition was
such that the temperature of the reaction mixture was maintained below
128C (increase in temperature results in self-condensation of the ester).
When the evolution of H2 ceased and the reaction mixture became clear,
a solution of 8 (10 g, 28.5 mmol) in DMF (30 mL) was added in two por-
tions, half of the solution was added by syringe pump over a period of
6 h and the remaining half was added after stirring for 24 h (it is impor-
tant to maintain the temperature of the reaction mixture below 208C
throughout the reaction time to avoid formation of the unwanted elimi-
nation product 13). After complete addition of 8, stirring was continued
for 3 d, then the reaction mixture was poured into cold water containing
1 mL of acetic acid. The aqueous layer was extracted with CH2Cl2. The
organic phase was washed several times with water, dried over magnesi-
um sulfate, and concentrated in vacuo to give gummy substance that was
further dried in a high vacuum to remove residual DMF. Recrystalliza-
tion from ether/hexane (2:1) afforded 12 as a white crystalline compound
(14.2 g, 72%). HPLC analysis: Rv =16.4 mL, Phenomenex, Aqua C18,
250”4.60 mm, 5 m column, UV220, flow=1 mLminÀ1, gradient from 50%
CH3OH to 90% CH3OH in 10 min and then 90% CH3OH for a further
5 min, 0.1% TFA; m.p. 72–738C (ether/hexane); 1H NMR (360 MHz,
CDCl3, 258C): d=7.34 (s, 10H; aromatic), 6.18 (s, 2H; -NH), 5.09 (s,
4H; PhCH2O-), 4.24 (m, 8H; -OCH2CH3), 2.93 (t, J=6.5 Hz, 4H; -N-
CH2-), 2.63 (t, J=7.2 Hz, 4H; -CCH2-), 1.24 (t, J=7.0 Hz, 12H;
-OCH2CH3); 13C NMR (90.56 MHz, CDCl3, 258C): d=167.2 (-CO,
esters), 154.5 (-NCO-), 135.9, 128.5, 128.3, 128.1 (aromatic carbons),
116.2 (-CN), 67.2 (PhCH2O-), 64.7 (quaternary carbons), 63.0
AHCTREUNG
[M+H]+; anal. calcd. (%) for C19H29N3O8·HBr: C 44.89, H 5.95, N 8.27,
Br 15.72; found: C 45.03, H 5.77, N 8.23, Br 15.87.
Compound 17: HPLC analysis: Rv =15 mL, Phenomenex, Aqua C18,
250”4.60 mm, 5 m column, UV220, flow=1 mLminÀ1, gradient from 50%
CH3OH to 90% CH3OH in 10 min and then 90% CH3OH for a further
5 min, 0.1% TFA; MS: m/z (%): 759.5 (100) [M+H]+.
Compound 19: HPLC analysis: Rv =17 mL, Phenomenex, Aqua C18,
250”4.60 mm, 5 m column, UV detection at 220 nm, flow=1 mLminÀ1
,
gradient from 10% CH3OH to 50% CH3OH in 10 min and then to 90%
CH3OH over the next 10 min, 0.1% TFA; MS: m/z (%): 491.2 (100)
[M+H]+.
2,2,8,8-Tetra(propylcarbamoyl)-3,4,6,7,8,9-hexahydro-2H-pyrimido
ACHTREUNG[1,2-
a]pyrimidine hydrobromide (1): solution of trimethylaluminium
A
A
(3.5 mL, 7 mmol, 2m solution in toluene) was added dropwise to a solu-
tion of n-propylamine (0.493 mL, 6 mmol) in dry CH2Cl2 (8 mL) under a
nitrogen atmosphere. After stirring at RT for 40 min, a solution of 22
(508 mg, 1 mmol) in CH2Cl2 (7 mL) was added dropwise, and the reaction
mixture was heated at reflux for 12 h. After cooling, the reaction mixture
was quenched by addition of an aqueous HBr solution (47%) and ex-
tracted with CH2Cl2. The CH2Cl2 layer was washed with water, and the
aqueous layer was reextracted with CH2Cl2. The combined organic layers
were dried over magnesium sulfate and concentrated in vacuo to give a
white solid that was recrystallized from acetonitrile/ether to afford 1 as
white micro crystals (475 mg, 85%). HPLC analysis: Rv =16 mL, Nucleo-
dur-100–5 C8 ec column, UV220, flow=1 mLminÀ1, gradient from 10%
CH3OH to 50% CH3OH in 10 min and then 50% CH3OH to 90%
CH3OH over the next 10 min, 0.1% TFA; m.p. 2228C (ether/acetoni-
trile); 1H NMR (360 MHz, CD3OD, 258C): d=8.41 (brs, guanidinium-
H), 8.27 (t, J=5.67 Hz, amide protons), 3.38 (t, J=5.6 Hz, 4H; -N-CH2),
3.21 (t, J=7.03 Hz, 8H; -NCH2CH2CH3), 2.47 (t, J=6.1 Hz, 4H;
-CCH2-), 1.55 (h, J=7.2 Hz, 8H; -NCH2CH2CH3), 0.88 (t, J=7.49 Hz,
12H; -NCH2CH2CH3); 13C NMR (90.56 MHz, CD3OD, 258C): d=169.2
(-CO, amides), 150.8 (guanidinium carbon), 64.9 (-CCH2-), 45.5 (-N-
CH2), 43.0 (-NCH2CH2CH3), 28.6 (-CCH2-), 23.5 (-NCH2CH2CH3), 11.6
(-NCH2CH2CH3); MS: m/z (%): 480.5 (100) [M+H]+; anal. calcd. (%)
for C23H41N7O4·HI (as an iodide salt): C 45.47, H 6.97, N 16.14, I 20.89;
found: C 45.47, H 6.53, N 15.83, I 19.94; HRMS (microTOF-Q) calcd for
C23H42N7O4+: 480.3293; found: 480.3289.
ACHTREUNG
Benzyl
1,1-di(ethoxycarbonyl)-3-(N-cyano-N-vinylamino)propylcarba-
mate (13): HPLC analysis Rv =14 mL, Phenomenex, Aqua C18, 250”
4.60 mm, 5 m column, UV220, , gradient from 50%
flow=1 mLminÀ1
CH3OH to 90% CH3OH in 10 min and then 90% CH3OH over the next
5 min, 0.1% TFA; 1H NMR (360 MHz; CDCl3): d=7.34 (s, 5H; aromat-
ic-H), 6.24 (s, 1H; -NH), 5.88 (dd, 1H; CH2=CH-), 5.10 (s, 2H;
PhCH2O-), 4.63 (dd, 1H; CH2=CH-), 4.44 (dd, 1H; CH2=CH-), 4.21 (m,
4H; -OCH2CH3), 3.32 (t, J=6.8 Hz, 4H; -N-CH2), 2.72 (t, J=7.2 Hz,
4H; -CCH2-); 1.23 (t, J=7.02 Hz, 12H; -OCH2CH3); 13C NMR
(90.56 MHz; CDCl3): d=168.6 (-CO, esters), 156.1 (-NCO-), 137.4 (CH2=
CH-), 135.4, 130.0, 129.7, 129.6 (aromatic carbons), 114.1 (-CN), 96.7
(CH2=CH-), 68.8 (PhCH2O-), 66.2 (quaternary carbons), 64.7
ACHTREUNG(-OCH2CH3), 47.8 (-N-CH2), 32.7 (-CCH2-), 15.3 (-OCH2CH3); MS: m/z
2,2,8,8-Tetra(2-methoxyethylcarbamoyl)-3,4,6,7,8,9-hexahydro-2H-
pyrimidoCAHTRE[UNG 1,2-a]pyrimidine hydrochloride (2): A solution of trimethylalu-
(%): 404.3 (25) [M+H]+.
minium (3.5 mL, 7 mmol, 2m solution in toluene) was added dropwise to
a solution of 2-methoxyethylamine (0.522 mL, 6 mmol) in dry CH2Cl2
(8 mL) under a nitrogen atmosphere. After stirring at RT for 40 min, a
solution of 22 (508 mg, 1 mmol) in CH2Cl2 (7 mL) was added slowly, and
the mixture was heated at reflux for 12 h. After cooling, the reaction was
cautiously quenched by adding 6n aqueous HCl. The CH2Cl2 layer was
washed with water, and the aqueous layer was reextracted with CH2Cl2.
The combined organic layers were washed with water, brine, and dried
with magnesium sulfate. Evaporation of the organic layer left a crude
solid that was recrystallized from acetonitrile/ether to yield 2 as a white
crystalline solid (435 mg, 75%). HPLC analysis: Rv =10.8 mL, Nucleo-
dur-100–5 C8 ec column, UV220, flow=1 mLminÀ1, gradient from 10%
CH3OH to 50% CH3OH in 10 min and then 50% CH3OH to 90%
CH3OH over the next 10 min, 0.1% TFA; m.p.: 178–1798C (acetonitrile/
ether); 1H NMR (360 MHz, CD3OD, 258C): d=3.35–3.44 (m, 28H), 3.29
(s, 12H; -CH2OCH3), 2.41 (t, J=5.01 Hz, 4H; -CCH2-); 13C NMR
(90.56 MHz, CD3OD, 258C): d=169.3 (-CO, amide), 151.1 (guanidinium
carbon), 71.5 (-CH2CH2OCH3), 64.8 (-CCH2-), 58.9 (-CH2CH2OCH3),
45.4 (-NCH2-), 40.9 (-CH2CH2OCH3), 29.2 (-CCH2-); MS: m/z (%): 544.5
(100) [M++H]; anal. calcd. (%) for C23H41N7O8·HI (as an iodide salt): C
2,2,8,8-Tetraethoxycarbonyl-3,4,6,7,8,9-hexahydro-2H-pyrimidoACHTRE[UGN 1,2-a]pyr-
imidine hydrobromide (22): A saturated solution of HCl in ethanol
(10 mL) was added to a solution of 12 (10 g, 14 mmol) in absolute etha-
nol (250 mL). The reaction mixture was stirred overnight at 458C under
a nitrogen atmosphere to give 17 (confirmed by ESIMS). After cooling
to RT, 15% Pd/C (500 mg) was added and the reaction mixture was
stirred for 2–3 h under an H2 atmosphere. The mixture was then filtered
through a pad of Celite and concentrated to give 19 as a solid residue
(confirmed by ESIMS).
Compound 17 was redissolved in absolute ethanol (250 mL) and triethyl-
amine (2.5 mL, 18.2 mmol) was added (Caution: Do not use excess base
because it leads to the formation of monocyclic guanidinium compound
15), and the reaction mixture was stirred at 508C for 2 h. After cooling to
RT, the solvent was evaporated in vacuo, and the residue was taken up in
CH2Cl2, washed with aqueous ammonium bromide solution (3”). The or-
ganic layer was dried and concentrated to give slightly brownish solid 22,
which was recrystallized from ethyl acetate/CH2Cl2 (9:1) to afford 22
(5.1 g, 74% in four successive steps). HPLC analysis: Rv =20 mL, Phe-
nomenex, Aqua C18, 250”4.60 mm, 5 m column, UV220, flow=1 mLminÀ1
gradient from 10% CH3OH to 50% CH3OH in 10 min and then to 90%
,
Chem. Eur. J. 2008, 14, 6098 – 6107
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6105

F. P. Schmidtchen et al.
41.14, H 6.30, N 14.60, I 18.90; found: C 41.26, H 6.23, N 14.62, I 18.51;
HRMS (microTOF-Q) calcd for C23H42N7O8+: 544.3089; found: 544.3095.
stirred in an oil bath at 808C. After stirring for 30 min, the reaction mix-
ture was cooled to RT, neutralized with 6n aqueous NaOH, concentrated
to half its volume, and freeze-dried to give a colorless solid. The solid
was taken up in acetonitrile (10 mL), and the insoluble fraction was re-
moved by filtration. The filtrate was then evaporated under vacuo to give
urea 16 as a gummy substance (2.3 g, 98%). HPLC analysis: Rv =14 mL,
2,2,8,8-Tetra(phenylcarbamoyl)-3,4,6,7,8,9-hexahydro-2H-pyrimido
ACHTRE[UNG 1,2-
a]pyrimidine hydroiodide (3): solution of trimethylaluminium
A
A
(3.5 mL,7 mmol, 2m solution in toluene) was added dropwise to a solu-
tion of aniline (0.546 mL, 6 mmol) in dry CH2Cl2 (8 mL) under an inert
atmosphere. After stirring at RT for 40 min, a solution of 22 (508 mg,
1 mmol) in CH2Cl2 (7 mL) was added dropwise. Then the resulting mix-
ture was heated at reflux for 5 h. After cooling, the reaction was cau-
tiously quenched by addition of aqueous HBr solution (47%). The
CH2Cl2 layer was washed with water, and the aqueous layer was reex-
tracted with CH2Cl2. The combined organic layers were washed with
aqueous sodium iodide solution (3”10 mL), dried over magnesium sul-
fate, and concentrated in vacuo to give a white solid. Recrystallization
from acetonitrile/ether afforded 3 as a white crystalline solid (572 mg,
77%). HPLC analysis: Rv =20.2 mL, Nucleodur-100–5 C8 ec column,
UV220, flow=1 mLminÀ1, gradient from 10% CH3OH to 50% CH3OH in
10 min and then 50% CH3OH to 90% CH3OH over the next 10 min,
0.1% TFA; m.p. 178–1808C (acetonitrile/ether); 1H NMR (360 MHz;
CD3CN, 258C): d=9.19 (s, 4H; amide protons), 8.43 (brs, 2H; guanidini-
um-H), 7.66 (d, J=8.1 Hz, 8H; aromatic), 7.32 (t, J=7.9 Hz, 8H; aro-
matic), 7.15 (t, J=7.37 Hz, 4H; aromatic), 3.42 (t, J=5.90 Hz, 4H; N-
CH2-), 2.67 (t, J=6.1 Hz, 4H; -CH2-); 13C NMR (90.56 MHz; CD3CN,
258C): d=166.5 (-CO, amide), 149.8 (guanidinium carbon), 138.2, 129.7,
126.1, 122.0 (aromatic carbons), 65.4 (-CCH2-), 45.5 (-N-CH2), 27.9
Phenomenex, Aqua C18, 250”4.60 mm, 5 m column, UV220
, flow=
1 mLminÀ1, gradient from 50% CH3OH to 90% CH3OH in 10 min and
then 90% CH3OH extended for a further 5 min, 0.1% TFA; 1H NMR
(360 MHz, CDCl3): d=7.25–7.34 (m, 10H, aromatic), 6.29 (s, 2H, -NH),
5.74 (brs, 2H, -CONH2), 5.07 (s, 4H, PhCH2O-), 4.15-4.23 (m, 8H,
-OCH2CH3), 3.13 (t, J=7.2 Hz, 4H, -NCH2), 2.46 (t, J=7.9 Hz, 4H,
-CCH2-), 1.19 (t, J=7.1 Hz, 12H, -OCH2CH3); 13C NMR (62.9 MHz,
CDCl3): d=167.54 (-CO, ester), 158.39 (-CO-, urethane), 154.73
AHCTREUNG
(-CONH2À), 135.99, 128.42 128.15, 128.99 (aromatic carbons), 66.99
(PhCH2-), 64.9 (quaternary carbons), 62.83 (-OCH2CH3), 42.55 (-NCH2-),
31.84 (-CCH2-), 13.76 (-OCH2CH3); MS: m/z (%): 731.3 (100) [M+H]+.
N,N-Bis[(3-amino-3,3-diethoxycarbonyl)propyl]urea (18): 15% Pd/C
(250 mg) was added to a solution of 16 (3.27 g, 4.47 mmol) in absolute
ethanol (35 mL). The obtained suspension was stirred under an atmos-
phere of H2 for 3 h. Then the reaction mixture was filtered through a pad
of Celite, and the evaporation of the filtrate under vacuo gave 18 as a
gummy substance (2.04 g, 99%). HPLC analysis Rv =16 mL, Phenomen-
ex, Aqua C18, 250”4.60 mm, 5 m column, UV220, flow=1 mLminÀ1, gradi-
ent from 10% CH3OH to 50% CH3OH in 10 min and then to 90%
ACHTREUNG
(-CCH2-); MS: m/z (%): 616.4 (100) [M+H]+; HRMS (micrOTOF-Q)
1
calcd for C35H34N7O4+: 616.2667; found: 616.2676.
CH3OH over the next 10 min, 0.1% TFA; H NMR (250 MHz, CD3CN):
d=5.35 (brs, 2H, -CONH2), 4.14 (q, J=7.0 Hz, 8H, -OCH2CH3), 3.20 (t,
J=7.3 Hz, 4H, -NCH2-), 2.3 (brs, 4H, -NH2), 2.08 (t, J=7.3 Hz, 4H,
-CCH2-), 1.20 (t, J=6.9 Hz, 12H, -OCH2CH3); 13C NMR (90.56 MHz,
CD3CN): d=172.24 (-CO, ester), 159.99 (-CONH2), 65.23 (quaternary
carbons), 62.78 (-OCH2CH3), 42.50 (-NCH2-), 33.96 (-CCH2-), 14.29
3-(3-Amino-3,3-diethoxycarbonyl)propyl-6,6-diethoxycarbonyl-3,4,5,6-tet-
rahydropyrimidine (14): 15% Pd/C (150 mg) was added to a solution of
12 (1 g, 1.4 mmol) in absolute ethanol (25 mL). The suspension was
stirred in an atmosphere of H2 for 2 h. The reaction mixture was then fil-
tered through a pad of Celite. The filtrate was evaporated in a vacuum to
give a gummy substance that was recrystallized from ether/hexane to
afford 14 as a colorless crystalline solid (600 mg, 99%). HPLC analysis:
AHCTREUNG
(-OCH2CH3); MS: m/z (%): 463.2 (100) [M+H]+.
1-[(3-Amino-3,3-diethoxycarbonyl)propyl]-2-oxo-4,4-diethoxycarbonyl-
1,4,5,6-tetrahydropyrimidine (20): A solution of 18 (80 mg, 0.17 mmol) in
nitropropane (2 mL) was heated in an oil bath at 1358C for 2 h. After
cooling the reaction mixture to RT, the solvent was removed under re-
duced pressure, and the residue was purified by column chromatography
over silica gel (50% ethyl acetate/hexane) to give 20 as a gummy sub-
stance (20 mg, 27%). HPLC analysis: Rv =16 mL, Phenomenex, Aqua
C18, 250”4.60 mm, 5 m column, UV220, flow=1 mLminÀ1, gradient from
10% CH3OH to 50% CH3OH in 10 min and then to 90% CH3OH over
the next 10 min, 0.1% TFA; 1H NMR (250 MHz, CDCl3): d=6.4 (s, 1H,
-NH-), 5.1 (s, 2H, -NH2), 4.2–4.3 (m, 8H, -OCH2CH3), 3.7–3.9 (m, 4H,
-NCH2-), 2.1–2.6 (m, 4H, -CCH2-), 1.2–1.3 (m, 12H, -OCH2CH3);
13C NMR (62.9 MHz, CDCl3): d=169.86, (-CO, ester), 169.69 (ring -CO),
65.29, (quaternary carbons), 62.54 (-OCH2CH3), 62.50 (-OCH2CH3),
45.14, 39.30 (-NCH2-), 31.8, 31.16 (-CCH2-), 13.9 (-OCH2CH3); MS: m/z
(%): 446.2 (100) [M+H]+
Rv =17 mL, Phenomenex, Aqua C18, 250”4.60 mm, 5 m column, UV220
,
flow=1 mLminÀ1
,
gradient from 10% CH3OH to 50% CH3OH in
10 min and then to 90% CH3OH over the next 10 min, 0.1% TFA;
1H NMR (250 MHz, CDCl3): d=7.06 (s, 1H, -CH=N-), 4.04–4.15 (m,
8H, -OCH2CH3), 3.18 (t, J=7.3 Hz, 2H, -NCH2), 3.08 (t, J=5.7 Hz, 2H,
-NCH2), 2.09 (t, J=6.1 Hz, 2H, -CCH2-), 1.99 (t, J=7.1 Hz, 2H,
-CCH2-), 1.13 (t, J=7.0 Hz, 12H, -OCH2CH3); 13C NMR (62.9 MHz,
CDCl3): d=170.50, 169.93 (-CO, ester), 150.53 (-CH=N-), 64.60, 63.96
(quaternary carbons), 61.86, 61.41 (-OCH2CH3), 47.92, 39.66 (-NCH2-),
33.70, 25.39 (-CCH2-), 13.73, 13.66 (-OCH2CH3); MS: m/z (%): 430.4
(100) [M+H]+.
1-[(3-Amino-3,3-diethoxycarbonyl)propyl]-2-amino-4,4-diethoxycarbonyl-
1,4,5,6-tetrahydropyrimidine bishydrobromide (15): A solution of 33%
HBr in propionic acid (1 mL) was added to a solution of 12 (40 mg,
0.056 mmol) dissolved in dry CH2Cl2 (10 mL), and the mixture was
stirred at RT for 1 h. The excess HBr was removed by a jet of nitrogen,
and the residue was redissolved in acetonitrile (5 mL). The solvent was
again removed by a stream of nitrogen to remove residual HBr. The resi-
due was dried under a high vacuum to give 15 as a yellow powder
(32 mg, 95%). HPLC analysis: Rv =16 mL, Phenomenex, Aqua C18, 250”
2,2,8,8-Tetracarboxy-3,4,6,7,8,9-hexahydro-2H-pyrimidoACHTER[UNG 1,2-a]pyrimidine
hydrobromide (23): A 4n aqueous solution of NaOH (16 mL) was added
to a solution of 22 (1 g, 2 mmol) in ethanol (16 mL) and cooled to 08C.
The resulting mixture was stirred at RT for 24 h. The reaction mixture
was acidified to pH 1 with 47% HBr solution. Evaporation of the solvent
in vacuo left a colorless solid residue that was then taken up in isopropa-
nol (15 mL) and stirred vigorously for 15 min. The insoluble salt residue
was removed by filtration, and the filtrate was evaporated in vacuo to
give a colorless residue that was recrystallized from acetonitrile/H2O to
afford 23 as a colorless crystalline solid (700 mg, 88%). HPLC analysis:
4.60 mm, 5 m column, UV220, , gradient from 10%
flow=1 mLminÀ1
CH3OH to 50% CH3OH in 10 min and then to 90% CH3OH over the
next 10 min, 0.1% TFA; 1H NMR (250 MHz, CD3CN): d=7.96 (s, 1H,
guanidinium-H); 7.49 (s, 2H, guanidinium-H); 4.21–4.38 (m, 8H,
-OCH2CH3); 3.63–3.70 (m, 2H, -NCH2); 3.42–3.51 (m, 2H, -NCH2);
2.40–2.51 (m, 4H, -CCH2-); 1.22–1.31 (m, 12H, -OCH2CH3); 13C NMR
(62.9 MHz, CD3CN): d=167.64, 165.20 (-CO, ester); 153.82 (guanidinium
carbon); 65.62 (-OCH2CH3), 65.52, (quaternary carbon); 64.85
Rv =2 mL, Phenomenex, Aqua C18, 250”4.60 mm, 5 m column, UV220
,
flow=1 mLminÀ1
,
gradient from 10% CH3OH to 50% CH3OH in
ACHTREUNG(-OCH2CH3); 63.17 (quaternary carbon); 46.80, 44.14 (-NCH2-); 29.58,
10 min and then to 90% CH3OH over the next 10 min, 0.1% TFA;
1H NMR (250 MHz, D2O): d=3.5 (t, J=5.5 Hz, 4H, -NCH2-), 2.51 (t, J=
5.4 Hz, 4H, -CCH2-); 13C NMR (62.9 MHz, D2O): d=176.01 (CO, acid),
152.10 (guanidinium carbon), 67.23 (quaternary carbons), 47.46
26.24 (-CCH2-); 14.18, 14.09 (-OCH2CH3); MS: m/z (%): 445.3 (100)
[M+H]+.
N,N-Bis[(3-benzyloxycarbonylamino-3,3-diethoxycarbonyl)propyl]urea
(16): A 6n aqueous solution of H2SO4 (9 mL) was added to a solution of
12 (2.29 g, 3.21 mmol) in CH3CN (20 mL), and the resulting mixture was
AHCTREUNG
6106
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6098 – 6107

Reconstructing Guanidinium Anchor Groups
FULL PAPER
[30] L. A. Jacob, B. L. Chen, D. Stec, Synthesis 1993, 611–614.
Acknowledgements
[31] T. Suhs, B. Koenig, Mini-Rev. Org. Chem. 2006, 3, 315–331.
[32] J. M. Hoerter, K. M. Otte, S. H. Gellman, S. S. Stahl, J. Am. Chem.
Soc. 2006, 128, 5177–5183.
This work was supported by grants from the Deutsche Forschungsge-
meinschaft and by the Hans-Fischer-Gesellschaft, Munich (Germany).
[33] J. I. Levin, E. Turos, S. M. Weinreb, Synth. Commun. 1982, 12, 989–
993.
[34] The successive pKa values of phthalic acid in water are 2.92 and
5.39, respectively; (S. Espinosa, E. Bosch, M. Roses, Anal. Chem.
2002, 74, 3809–3818), whereas the parent bicyclic guanidinium has a
pKa of 11.2 and thus is about a million times more basic (A. F.
McKay, M. E. Kreling, Can. J. Chem. 1962, 40, 1160–1163). In
DMSO, the pKa values of phthalic acid expectedly become more
basic: pKa =6.4 and 17.4 (P. J. Choi, K. A. Petterson, J. D. Roberts,
J. Phys. Org. Chem. 2002, 15, 278–286); corresponding pKa values
for the guanidinium hosts in DMSO are not available. Although cat-
ionic acids regularly experience a less dramatic shift of their acidity
towards higher pKa values on moving from water into DMSO, the
total inversion of the sequence is rather amazing in view of the acid-
ity differences in water.
[35] S. O. Kang, D. van der Velde, D. Powell, K. Bowman-James, J. Am.
Chem. Soc. 2004, 126, 12272–12273.
[36] R. Schwesinger, R. Link, P. Wenzl, S. Kossek, Chem. Eur. J. 2006,
12, 438–445.
[37] a) Crystal structure analysis of 3 (free base): C35H33N7O4; Mr =
615.68; orthorhombic; space group P212121 (no. 19); a=10.2551(5),
[1] a) P. Blondeau, M. Segura, R. Perez-Fernandez, J. de Mendoza,
Chem. Soc. Rev. 2007, 36, 198–210; b) P. Blondeau, J. Benet-Buch-
holz, J. de Mendoza, New J. Chem. 2007, 31, 736–740.
[2] K. A. Schug, W. Lindner, Chem. Rev. 2005, 105, 67–113.
[3] B. P. Orner, A. D. Hamilton, J. Inclusion Phenom. Macrocyclic
Chem. 2001, 41, 141–147.
[4] B. T. Nguyen, E. V. Anslyn, Coord. Chem. Rev. 2006, 250, 3118–
3127.
[5] C. L. Hannon, E. V. Anslyn in Bioorganic Chemistry Frontiers Vol. 3
(Ed.: H. Dugas), Springer, New York, 1993, pp. 193–255.
[6] M. M. Islam, H. Hayashi, H. Mizuguchi, H. Kagamiyama, Biochem-
istry 2000, 39, 15418–15428.
[7] M. Garcia-Perez, M. Pinto, J. A. Subirana, Biopolymers 2003, 69,
432–439.
[8] S. Nakagawa, H. Umeyama, J. Am. Chem. Soc. 1978, 100, 7716–
7725.
[9] A. Gleich, F. P. Schmidtchen, Chem. Ber. 1990, 123, 907–915.
[10] W. Peschke, P. Schiessl, F. P. Schmidtchen, P. Bissinger, A. Schier, J.
Org. Chem. 1995, 60, 1039–1043.
[11] V. Kral, F. P. Schmidtchen, K. Lang, M. Berger, Org. Lett. 2002, 4,
51–54.
b=13.7374(8), c=21.5253(11) Š; V=3032.5(3) Š3; Z=4; 1calcd
=
1.349 gcmÀ3; F000 =1296; m=0.091 mmÀ1; Rint =0.1189; 3177 (1899:
Io >2s(Io)) independent reflections; 419 parameters; R1=0.0466
(Io >2s(Io)); wR2=0.1237 (all data); GOF=1.063. The low scatter-
ing power of the very tiny crystal forced us to cut the data set at q=
[12] a) M. Haj-Zaroubi, N. W. Mitzel, F. P. Schmidtchen, Angew. Chem.
2002, 114, 111–114; Angew. Chem. Int. Ed. 2002, 41, 104–107; b) M.
Haj-Zaroubi, F. P. Schmidtchen, ChemPhysChem 2005, 6, 1181–
1186.
[13] C. R. Bondy, S. J. Loeb, Coord. Chem. Rev. 2003, 240, 77–99.
[14] M. Rekharsky, Y. Inoue, S. Tobey, A. Metzger, E. Anslyn, J. Am.
Chem. Soc. 2002, 124, 14959–14967.
20.838. b) Crystal structure analysis of
3 (bromide, acetonitrile
adduct): C37H37BrN8O4; Mr =737.65; monoclinic; space group C2/c
(no. 15); a=20.0476(1), b=18.4443(1), c=20.4080(1) Š; b=
98.2102(2)8; V=7468.80(7) Š3; Z=8; 1calcd =1.312 gcmÀ3
; F000 =
3056; m=1.150 mmÀ1; Rint =0.043; 6841 (6145: Io >2s(Io)) independ-
ent reflections; 599 parameters; R1=0.0269 (Io >2s(Io)); wR2=
0.0704 (all data); GOF=1.057. In addition to the solvent molecules
well located in the difference Fourier maps, two unresolved solvent
molecules remained and had to be removed by the PLATON calc
squeeze procedure. c) Crystal structure analysis of 23 (trihydrate):
[15] M. D. Best, S. L. Tobey, E. V. Anslyn, Coord. Chem. Rev. 2003, 240,
3–15.
[16] C. Schmuck, V. Bickert, J. Org. Chem. 2007, 72, 6832–6839.
[17] C. Schmuck, L. Geiger, J. Am. Chem. Soc. 2004, 126, 8898–8899.
[18] X. Salvatella, M. Martinell, M. Gairi, M. G. Mateu, M. Feliz, A. D.
Hamilton, J. de Mendoza, E. Giralt, Angew. Chem. 2004, 116, 198–
200; Angew. Chem. Int. Ed. 2004, 43, 196–198.
¯
C11H19N3O11; Mr =369.29; triclinic; space group P1 (no. 2); a=
6.5084(1), b=8.6005(1), c=14.3244(3) Š; a=83.9319(5), b=
[19] A. Echavarren, A. Galan, J.-M. Lehn, J. De Mendoza, J. Am. Chem.
Soc. 1989, 111, 4994–4995.
80.0964(6), g=73.5887(10)8; V=756.30(2) Š3; Z=2; 1calcd
=
1.622 gcmÀ3; F000 =388; m=0.147 mmÀ1; Rint =0.036; 2766 (2494: Io >
2s(Io)) independent reflections; 303 parameters; R1=0.0301 (Io >
2s(Io)); wR2=0.0792 (all data); GOF=1.043. d) CCDC 673725 (3
(free base)), 673724 (3 (bromide, acetonitrile adduct)), and 673723
(23 (trihydrate)) contain the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif. For more details, see the Supporting Information.
[38] F. P. Schmidtchen, Top. Curr. Chem. 2005, 255, 1–29.
[39] V. D. Jadhav, F. P. Schmidtchen, Org. Lett. 2006, 8, 2329–2332.
[40] J. Barthel, L. Iberl, J. Rossmaier, H. J. Gores, B. Kaukal, J. Solution
Chem. 1990, 19, 321–337.
[20] A. Echavarren, A. Galan, J. De Mendoza, A. Salmeron, J.-M. Lehn,
Helv. Chim. Acta 1988, 71, 685–693.
[21] K. Kobiro, Y. Inoue, J. Am. Chem. Soc. 2003, 125, 421–427.
[22] A. P. Davis, K. J. Dempsey, Tetrahedron: Asymmetry 1995, 6, 2829–
2840.
[23] M. J. Chmielewski, T. Zielinski, J. Jurczak, Pure Appl. Chem. 2007,
79, 1087–1096.
[24] K. Choi, A. D. Hamilton, Coord. Chem. Rev. 2003, 240, 101–110.
[25] R. J. Fitzmaurice, F. Gaggini, N. Srinivasan, J. D. Kilburn, Org.
Biomol. Chem. 2007, 5, 1706–1714.
[26] E. A. Katayev, J. L. Sessler, V. N. Khrustalev, Y. A. Ustynyuk, J.
Org. Chem. 2007, 72, 7244–7252.
[27] Z. Rodriguez-Docampo, S. I. Pascu, S. Kubik, S. Otto, J. Am. Chem.
Soc. 2006, 128, 11206–11210.
[41] O. Tiba, C. G. Overberger, J. Polym. Sci. Part A: Polym. Chem.
1987, 25, 3437–3458.
[28] a) V. D. Jadhav, F. P. Schmidtchen, Org. Lett. 2005, 7, 3311–3314;
b) F. P. Schmidtchen, Coord. Chem. Rev. 2006, 250, 2918–2928.
[29] V. Amendola, M. Bonizzoni, D. Esteban-Gomez, L. Fabbrizzi, M.
Licchelli, F. Sancenon, A. Taglietti, Coord. Chem. Rev. 2006, 250,
1451–1470.
Received: December 25, 2007
Published online: June 3, 2008
Chem. Eur. J. 2008, 14, 6098 – 6107
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6107