Evaluation of the binding ability of tetraaza[2]arene[2]triazine receptors anchoring l-alanine units for aromatic carboxylate anions

Article abstract of DOI:10.1016/j.tet.2011.10.090

The binding affinities of three new tetraaza[2]arene[2]triazine based macrocycles anchoring one (AC1A) and two (AC2A and Me₄AC2A) l-alanine amino acid units for five aromatic carboxylate anions (bz^-, ph ^2-, iph^2-, tph^2- and btc^3-) were investigated in DMSO-d₆. ¹H NMR titrations revealed that the AC1A and AC2A receptors exhibit comparable anion affinities, suggesting that the two l-alanine binding units in AC2A are not simultaneously involved in the anion recognition as indicated by molecular dynamics simulations carried out for selected AC1A and AC2A associations using the AMBER force field (GAFF). New force field parameters were developed in order to mimic the structural singularity derived from the N-H macrocyclic bridges.

Full text of DOI:10.1016/j.tet.2011.10.090

Tetrahedron 68 (2012) 670e680
Contents lists available at SciVerse ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Evaluation of the binding ability of tetraaza[2]arene[2]triazine receptors
anchoring -alanine units for aromatic carboxylate anions
L
,
b
a
,
b
a
b
c
a,
*
Ana I. Vicente a , Joao M. Caio , Joao Sardinha , Cristina Moiteiro , Rita Delgado , Vítor Felix
~
~
ꢀ
ˇ
a
~
ꢀ
ꢀ
Departamento Química, CICECO and Secc¸ ao Autonoma das Ciencias da Saude, Universidade de Aveiro, 3810-193 Aveiro, ˇ Portugal
b Departamento de Química e Bioquímica and Centro de Química e Bioquímica, Universidade de Lisboa, Faculdade de Ciencias, Campo Grande, 1749-016 Lisboa, Portugal
c
ꢀ
ꢀ
Instituto de Tecnologia Química e Biologica, UNL, Av. da RepublicadEAN, 2780e157 Oeiras, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history:
The binding affinities of three new tetraaza[2]arene[2]triazine based macrocycles anchoring one (AC1A)
and two (AC2A and Me4AC2A)
iph2ꢀ, tph2ꢀ and btc3ꢀ) were investigated in DMSO-d6. 1H NMR titrations revealed that the AC1A and
AC2A receptors exhibit comparable anion affinities, suggesting that the two -alanine binding units in
L ,
-alanine amino acid units for five aromatic carboxylate anions (bzꢀ, ph2ꢀ
Received 14 September 2011
Received in revised form 23 October 2011
Accepted 27 October 2011
L
Available online 6 November 2011
AC2A are not simultaneously involved in the anion recognition as indicated by molecular dynamics
simulations carried out for selected AC1A and AC2A associations using the AMBER force field (GAFF).
New force field parameters were developed in order to mimic the structural singularity derived from the
NeH macrocyclic bridges.
Keywords:
Macrocycles
Azacalixarenes
Anion recognition
Carboxylate anions
Molecular dynamics
Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction
efficient convergent synthetic strategy to prepare in good yields
various azacalix[2]arene[2]triazine macrocycles by two successive
Along with the impressive developments in traditional calixar-
ene chemistry, the heterocalixarenes incorporating heteroaromatic
arene units into the calixarene skeleton (calixpyrroles and cal-
ixpyridines among others) have also attracted much interest be-
cause their distinct conformational preferences, cavity sizes and
hosteguest properties.1 In contrast, the heteroaromatic calixarenes
in which the methylene bridges are replaced by heteroatoms are
much less widespread, in spite of their singular structural and
binding properties derived from the electronic and steric grounds
of the heteroatom bridges.2e13 Furthermore, when the bridges are
composed of nitrogen atoms, extra potential binding and func-
tionalization sites are available for the design of novel synthetic
receptors. Among this growing variety of macrocyclic host mole-
cules, a representative number of azacalix[n]heteroarenes has been
reported over the last decade,2,3,5,8,14e19 but remain almost un-
explored regarding their binding capabilities.
coupling steps between the high reactive cyanuric chloride and
selected nucleophilic aromatic diamines, as building blocks. More
recently, an equivalent stepwise pathway was followed by Clayden
et al.17 in the synthesis of water soluble sulfonated azacalix[4]are-
nes composed of benzene and naphthalene diamine based linkers
with different macrocyclic cavity sizes and shapes.
Searching for novel azamacrocyclic receptors with selective
binding ability for carboxylate anions with biological and envi-
ronmental impact, we undertook the functionalization of the tet-
raaza[2]arene[2]triazine platform with amino acid binding units.
In this work, we present the synthesis of new tetraaza[2]arene
[2]triazine anchoring one and two L-alanine units in the phenyl
rings (Scheme 1) and the evaluation of their binding properties in
DMSO towards a whole set of carboxylate aromatic anions (Scheme
2).
To the best of our knowledge, it is the first time that two aza-
heteroaromatic calixarenes incorporating one (AC1A) and two al-
anine units (AC2A) are reported. These receptors exhibit two dif-
ferent binding regions. The first region comprises four NeH
linkages, and the second comprises the amide binding sites from
amino acid chains. The simultaneous interaction of both types of
binding sites on the anion recognition is also expected. Further-
more, in the AC2A, the amide sites will eventually bind the car-
boxylate anions in a cooperative fashion. The single crystal X-ray
Azacalix[n]heteroarenes can be obtained using
a Buch-
waldeHartwig cross coupling reaction in macrocyclization step8,20
or through a nucleophilic aromatic substitution (SNAr) between
diamine aromatic linkers and electrophilic activated species.18,21 In
this context, Wang and Yang15 devised a straightforward and
ꢀ
* Corresponding author. E-mail address: vitor.felix@ua.pt (V. Felix).
0040-4020/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2011.10.090

A.I. Vicente et al. / Tetrahedron 68 (2012) 670e680
671
H
N
structure of AC2A was also determined giving valuable structural
hints about the accessibility of the NeH linkages for the anion
binding in the competitive solvent used (DMSO). In order to un-
derstand, the role of the NeH binding sites on the binding behav-
iour, the AC2A was completely methylated in the nitrogen bridges
yielding the Me4AC2A molecule (Scheme 1). Further insights into
the molecular recognition mechanism were obtained by means of
molecular dynamic simulations carried out in solution with se-
lected AC2A and AC1A anion associations.
6
COOMe
O
H
7
COOMe
O
N
5
11
12
4
8
3
2
RN
NR
HN
HN
NH
14
N
N
N
N
13
1
N
N
N
N
N
Cl
N
Cl
Cl
N
N
Cl
RN
25
NR
18
24
21
NH
19
20
Moreover, the evaluation of the binding affinity of these re-
ceptors with aliphatic carboxylate anions including various amino
acid derivatives is currently in progress.
33
23
22
29
O
COOMe
N
31
H
30
2. Results and discussion
2.1. Synthesis
AC2A: R=H
AC1A
Me4AC2A: R=Me
Scheme 1.
The novel receptors, AC1A and AC2A, were prepared by SNAr
substitution of two labile chlorine atoms in two triazines by amine
groups from two m-phenylenediamine units following a stepwise
fragment coupling pathway as described earlier.15 The synthetic
-
-
-
CO2
CO2
CO2
-
CO2
strategy is depicted in Schemes 3 and 4, respectively. The
methyl ester of 3,5-dinitrobenzamide intermediate 3 was prepared
from the reaction of 3,5-dinitrobenzoyl chloride with -alanine
L-alanine
-
L
CO2
methyl ester 2 in tetrahydrofuran (THF) at room temperature (rt). In
order to prevent racemization of amino-acid motifs this reaction
was carried out in presence of propylene oxide. The 1H NMR
benzoate (bz-)
isophthalate (iph2-)
phthalate (ph2-)
-
CO2
spectrum of 3 showed a well-defined quintet at d 4.86 ppm,
-
CO2
assigned to the chiral proton, which is indicative of the existence of
a single (S)-enantiomeric form. Afterwards, the two nitro groups
were reduced by catalytic hydrogenation affording the 3,5-diamine
chiral synthon 4 required in quantitative yield. Subsequently, this
intermediate was coupled with two cyanuric chloride units in THF
at 0 ꢁC in the presence of N,N-diisopropylethylamine (DIPEA) to
afford the linear trimer 5, in 60% yield (Scheme 3a).
-O2C
CO2
-
-
CO2
terephthalate (tph2-)
benzenetricarboxylate (btc3-
)
The AC1A compound can be achieved through the cyclo-
condensation of trimer 5 with 1,3-phenylenediamine (route 3b1) or
Scheme 2.
(a) Synthesis of the chiral linear trimer
H
N
H
N
H
N
COOMe
Cl
O
O
Cl
O
COOMe
O
COOMe
Cl
N
Cl
N
DIPEA, THF
0 ºC, 60%
THF,C3H6O
70%
COOMe
H2, Pd/C, 10%
N
N
N
N
N
N
+
+
NO2
EtOH
H2N
N
H
N
H
N
Cl
Cl
O2N
H2N
NH2
Cl
Cl
O2N
NO2
4
1
2
5
3
(b) Synthetic routes towards AC1A macrocyle
H
1)
COOMe
O
N
H
N
COOMe
Cl
O
Cl
DIPEA
N
N
N
N
HN
NH
+
NH2 Cl
(CH3)2CO, reflux
Cl
32%
N
N
N
N
N
H
N
N
H2N
N
H
Cl
N
N
Cl
5
HN
NH
2)
AC1A
Cl
Cl
N
Cl
N
4, DIPEA
(CH
N
N
DIPEA, THF
0 ºC, 81%
N
N
N
N
3)2CO
reflux, 50%
+
H2N
NH2
Cl
N
H
N
H
N
Cl
Cl
Cl
6
Scheme 3.

672
A.I. Vicente et al. / Tetrahedron 68 (2012) 670e680
ꢁ
H
CeH/HeC distance of 2.959 A at lower rim. This cavity size di-
mension compares well with those found in the crystal structures of
other tetrazacalix[2]arene[2]triazine derivatives.23 Selected bond
distances and angles subtended at nitrogen bridges are given in
Table S1. The CeN bonds to the triazine rings with an average dis-
COOMe
O
N
H
N
COOMe
Cl
O
RN
NR
Cl
N
N
N
N
N
4, DIPEA
N
N
N
N
Cl
N
N
Cl
2
ꢁ
tance of 1.34(1) A are typical CeNsp double bonds, while those of
(CH ) CO, reflux
40%
N
H
N
N
Cl
Cl
H
RN
NR
5
the phenyl rings are systematically longer (average distance 1.42(1)
A) and close of the normal value of a CeNsp single bond. This
3
ꢁ
structural feature indicates a better electron delocalization from the
O
COOMe
N
nitrogen bridges to the
p deficient electron chlorotriazine rings than
H
Cs CO , CH CN
MeI, 50 ºC, 52%
to the phenyl rings, which is in agreement with average tilt angles
quoted above for both aromatic ring entities.
R=H, AC2A
R=Me, Me AC2A
Scheme 4.
alternatively by SNAr reaction between synthon 4 and the linear
trimer 6 (route 3b2), which is previously prepared by the addition
of two cyanuric chloride units to 1,3-phenylenediamine.15 Both
synthetic pathways led to comparable yields of 32% and 50% for 3b1
and 3b2, respectively; however, intermediate 5 was poorly soluble
in acetone. All cyclization reactions were performed in refluxing
acetone using DIPEA in order to favour the production of the
macrocyclic molecule and decreasing the formation of undesired
oligomer side products.22 The AC2A compound was obtained by
cyclization reaction between trimer 5 and synthon 4 in 40% yield
(Scheme 4). Afterwards, the Me4AC2A macrocycle was synthesized
in moderate yield by direct methylation of AC2A using MeI and
Cs2CO3 in acetonitrile at 50 ꢁC.
The structural elucidation of the AC1A, AC2A and Me4AC2A
macrocycles was preliminary established by means of high-
resolution mass spectrometry and 1H and 13C NMR spectroscopies.
The ESIeMS spectra of the three macrocycles exhibit Mþ, [Mþ2]þ
and [Mþ4]þ 35Cl/37Cl isotopic peaks with relative intensities con-
sistent with the existence of two chlorine atoms. The complete as-
signments of 1H and 13C resonances were done using HMBC and
HMQC NMR techniques, whose spectra are presented in
Supplementary data (ESD). The phenyl rings of AC1A receptor dis-
play two different sets of 1H and 13C NMR resonances at downfield
while for AC2A only a single set was found in agreement with the C2
symmetry witnessed at rt in the NMR time experiment. The ste-
Fig. 1. Insight depictions in the solid state structure of AC2A$4.5(DMSO)$2H2O: view
(a) shows the A and B molecules encapsulating a single DMSO molecule (drawn as
spheres) into the upper and the eventual
pep stacking between their phenyl rings;
reochemistry of the two L-alanine units and 1,3-conformation of the
calix[2]arene[2]triazine platform was confirmed with the structure
determination of AC2A by single crystal X-ray diffraction.
view (b) shows the 1-D network of S]O/HeN hydrogen bonds established by AC2A
molecules (illustrated with two B and one A molecules) with surrounding DMSO and
water solvent molecules. Carbon atoms are drawn in grey, hydrogen atoms in white,
nitrogens in blue, chlorides in green, sulfurs in yellow and oxygens in red. Hydrogen
bonds are drawn as green dashed lines.
2.2. Single crystal X-ray structure of AC2A
The AC2A exhibits further structural features derived from the
crystal packing as can be seen in Fig. 1: A and B accommodate in the
upper rim one DMSO molecule of crystallization a methyl group
completely inserted into the annulus; an almost parallel arrange-
ment of two phenyl rings from A and B molecules, separated by
The crystal structure of AC2A is built up from an asymmetric
unit composed of two independent molecules (A and B) and fifteen
mother liquor molecules comprising ten DMSO and five water
molecules. Furthermore, the two DMSO and two water solvent
molecules have occupancy factors of 0.5, which is consistent with
the molecular formula AC2A$4.5(DMSO)$2H2O.
ꢁ
3.464 A distance, is observed, which is consistent with the exis-
tence of
pep interactions. The AC2A molecules are assembled
The two independent AC2A molecules (A and B), shown in Fig. 1,
are almost identical apart from the spatial disposition of the methyl
ester groups of the two L-alanine units. In fact, in A they exhibit
through the surrounded DMSO solvent molecules into 1-D chains,
which are composed of four independent S]O/HeN hydrogen
bond bridges (eight hydrogen bonds) with O/N distances ranging
comparable NeCeC]O torsion angles of ꢀ55.6 and ꢀ47.4ꢁ and
NeCeCeO torsion angles of 147.6 and 136.9ꢁ, while in B they have
NeCeC]O and NeCeCeO torsion angles of 147.0 and ꢀ123.9ꢁ and
ꢀ36.6 and 56.8ꢁ, respectively. This conformational difference is due
to the different hydrogen bonds established between the NeH
amide binding groups and DMSO solvent molecules (see infra).
The AC2A displays a 1,3-alternated conformationwith the phenyl
andtriazine ringsinterceptingthe planedefinedbythefour nitrogen
bridges at average dihedral angles of 47.8 and 27.3ꢁ, respectively. A
well-defined concave cavity was formed with a short average
ꢁ
from 2.811 to 2.891 A and corresponding O/HeN angles from 171
to 155ꢁ. The A and B molecules are not equivalent regarding the
hydrogen bonds formed by the amide binding groups: the two NeH
amide binding sites of A are hydrogen bonded to two water mol-
ꢁ
ecules with N/O distances of 2.886 and 2.995 A; while one amide
binding site of B participates in a single S]O/HeN hydrogen bond
with a O/N distance of 2.768 A. These structural findings are en-
tirely consistent with conformational differences reported above
for two alanine motifs. Furthermore, the intermolecular distances
ꢁ

A.I. Vicente et al. / Tetrahedron 68 (2012) 670e680
673
found from the water molecules to the triazine nitrogen atoms and
to the DMSO solvent molecules also suggest the existence of
N/HeO and O/HeO hydrogen bonds, but the full analysis of
these hydrogen interactions is limited by the absence of the water
hydrogen atoms.
Finally, when the crystal packing is in view along the crystal-
lographic direction, channels composed of A and B molecules filled
with DMSO and water solvent molecules emerge (Fig. 2).
interactions between aromatic rings in the receptors and the sub-
strates can not be discarded.
1H NMR was used to follow the values of Dd¼(dfree ligand
ꢀ
dobs)
obtained by addition of known amounts of anionic substrate so-
lutions. Separated 1H NMR signals were found for the receptor and
the substrate protons upon formation of the receptoresubstrate
entity, indicating a fast exchange between the free and associated
receptor compared to the NMR time scale. Downfield shifts of the
free receptor proton resonances were observed by addition of bzꢀ,
iph2ꢀ, tph2ꢀ and btc3ꢀ anions, as expected. This is well illustrated in
Fig. 3 for the case of iph2ꢀ with AC1A and AC2A receptors. Sur-
prisingly, for the ph2ꢀ anion the reverse behaviour occurred and
upfield shifts were observed with simultaneous broadening and
then disappearance of the proton amine resonances. This point will
be discussed later.
Fig. 2. Crystal packing view of AC2A$4.5(DMSO)$2H2O showing the formation of
channels along the a crystallographic direction, which are occupied by DMSO and
water solvent molecules. A and B molecule arrangements are drawn as yellow and blue
spheres, respectively.
3. Binding affinity of AC1A, AC2A and Me4AC2A receptors for
aromatic carboxylate substrates in DMSO
3.1. 1H NMR studies
The binding constants of the supramolecular entities formed by
azacalix[2]arene[2]triazine derivative receptors, AC1A and AC2A,
with aromatic mono-, di- and tricarboxylate anions were de-
termined by 1H NMR titrations at 298 K in DMSO-d6. The studied
anions, benzoate (bzꢀ), phthalate (ph2ꢀ), isophthalate (iph2ꢀ),
terephthalate (tph2ꢀ) and benzenetricarboxylate (btc3ꢀ), were used
as tetrabutylammonium salts. The constants were determined
taking into account the maximum number of shifted proton reso-
nances of the receptors titrated with the anions using HypNMR
program24 and the obtained values are listed in Table 1. The AC1A,
AC2A receptors and anionic substrates showed scarce solubility in
non-polar solvents and consequently the binding experiments
were carried out in the high solvating medium DMSO-d6, as well
illustrated in the X-ray structure of AC2A (see supra), in which 4.5
DMSO molecules interact with the receptor (Fig. 1). Therefore, the
affinities of AC1A and AC2A receptors for anions would be lower in
this solvent than in non-polar ones. The expected interactions
should be mainly governed by NeH/O hydrogen bonds between
the receptor anchoring sites (amide protons and amine bridges)
Fig. 3. 1H NMR titration of AC1A (solid triangles) and AC2A (open triangles) with iph2ꢀ
anion, Dd of the receptor proton resonances in function of the number of equivalents of
substrate added in DMSO-d6, following resonances H6, H1,14, H3,11 and H8, see Scheme 1
for labelling scheme.
Only one constant, corresponding to the formation of RS species
with 1:1 stoichiometry was found, as confirmed by the Job’s plot
presented in Fig. 4 for the AC1A association with btc3ꢀ and in Figs.
S13 and S14 for the associations between AC2A and iph2ꢀ or tph2ꢀ
and carboxylate groups of the substrates, although
pep stacking
Table 1
Binding constants (log KRS
a
)
of AC1A, AC2A and Me4AC2A receptors with the aro-
Receptors (R)
matic carboxylate anions
Substrate (S)
AC1A
AC2A
Me4AC2A
bzꢀ
1.39(1)
d
1.43(1)
d
d
ph2ꢀ
iph2ꢀ
tph2ꢀ
btc3ꢀ
<1
1.88(1)
d
2.59(2)
1.87(1)
2.63(1)
2.82(1)
1.89(1)
2.77(1)
d
a
Errors in log KRS values are the standard deviations given directly by the pro-
gram HypNMR24 by fitting the experimental points derived from two to five NMR
signals measured. T¼298 K; solvent: DMSO-d6. The self-assembly of the receptor is
insignificant in the concentrations used as shown in Experimental section.
Fig. 4. Job’s plot in function of the receptor for the AC1A/btc3ꢀ system in DMSO-d6
solutions.

674
A.I. Vicente et al. / Tetrahedron 68 (2012) 670e680
anions.25 For each substrate, both receptors exhibited identical
binding constants and similar pattern of NMR shifts (Fig. 3). This
indicates that in spite of the presence of two-amide binding units in
AC2A with spatial disposition suitable for the simultaneous binding
of both carboxylate groups, only a single binding unit is involved in
the anion recognition process. Furthermore, the binding is stronger
for iph2ꢀ and btc3ꢀ anions than for tph2ꢀ and bzꢀ associations.
These data suggest that tph2ꢀ should interact mainly through
a single carboxylate group, revealing that its binding geometry is
not appropriate for the simultaneous interaction of both carbox-
ylates with the NeH receptor binding sites.
However, independently of the number of simultaneous an-
chorage points, the btc3ꢀ, iph2ꢀ, tph2ꢀ and bzꢀ anions interact with
AC1A and AC2A through NeH/O hydrogen bonds between the
receptor NeH binding sites (NamineseH macrocyclic bridges and/or
the NamideseH) and anion oxygen atoms from carboxylate groups,
as indicated by the main downfield shifts of H1,14 of AC1A (and also
H18,25 of AC2A) and H6 resonances. The 1H NMR titrations of both
receptors are similar as shown in Figs. 5 and 6 for titrations of AC1A
and AC2A with btc3ꢀ anion, respectively; and in Figs. S15e20 for
the remaining three anions (excluding ph2ꢀ). Henceforth, the fol-
lowing analysis will be focused on the AC1A receptor titrations.
Fig. 6. 1H NMR spectra of AC2A receptor with different amounts of btc3ꢀ anion added:
(a) 0; (b) 0.2; (c) 0.6; (d) 1.1 and (e) 3.2 equiv. The proton labelling adopted is given in
Scheme 1.
broadening of almost all proton resonances of the receptors, with
especial relevance to the amine protons, H1,14 and H18,25. For in-
stance, these two AC1A singlets coalesce in a large singlet upon
addition of the first 0.1 equiv of anion as well as the H3,11 benzene
protons of the macrocycle (Fig. 7). Moreover, the proton resonances
of the receptor are upfield shifted. Other resonances also move,
from these the most significant shift is verified for H3,11 proton
resonance, followed successively by the shifts of H20, H21, H6 and
H24 resonances.
Fig. 5. 1H NMR spectra of AC1A receptor with different amounts of btc3ꢀ anion added:
(a) 0, (b) 0.2, (c) 0.6, (d) 1.1 and (e) 3.2 equiv. The proton labelling adopted is given in
Scheme 1.
Titrating AC1A with btc3ꢀ, iph2ꢀ, tph2ꢀ and bzꢀ led only to
downfield shifts of the receptor proton resonances, being the larger
shift verified for H6, followed by H1,14 (H18,25 does not move), and
then H3,11, while H8 proton resonances have small downfield shifts
(Figs. 3 and 5). For tph2ꢀ system a small upfield shift of H18,25 is also
observed. In addition, the signals of some proton resonances
broaden along the titration, especially H1,14 and in minor extent
H3,11 and H13, evidencing varied protons contributing to the signal.
These data indicate that the amide group together with the close
amine hydrogen of the bridges are the main hydrogen-donor sites
in the interaction of AC1A with the carboxylate anions. The shifts or
enlargement of signals are more significant for btc3ꢀ and less im-
portant for bzꢀ, following the trend: btc3ꢀ>iph2ꢀ>tph2ꢀ>bzꢀ, in
agreement with the values obtained for the corresponding associ-
ation constants.
Fig. 7. 1H NMR spectra of AC1A receptor with different amounts of ph2ꢀ anion added:
(a) 0; (b) 0.2; (c) 0.4 and (d) 0.6 equiv. The proton labelling adopted is given in
Scheme 1.
As well known the chemical shifts of the protons are far
downfield when they interact with strong electronegative atoms,
although exceptions were reported.26 Our results seem to indicate
that the main effect caused by ph2ꢀ anions in both receptors is not
the association, but the deprotonation of the amine protons, as also
found in other cases.27e33 The upfield shifts of the resonances de-
rives from the increase of the electron density in the phenyl rings,
which caused through-bond propagation of the electronic density
generated in the amine centres upon deprotonation.
The titrations of both receptors with ph2ꢀ anion follow a reverse
behaviour compared to the already described for the other anions.
The addition of this anion to the receptor results in severe

A.I. Vicente et al. / Tetrahedron 68 (2012) 670e680
675
The same effect was observed when titrating AC1A with a strong
base, such as TBAOH (Fig. S22), indicating definitely that ph2ꢀ is
a stronger base than the amine receptor bridges. Therefore the equi-
librium exemplified for AC1A: H4AC1A þ ph2ꢀ%H3AC1Aꢀ þ Hphꢀ
occurs with ph2ꢀ anion but not with the other studied anions. The pKa
values in DMSO of H2ph are consistent with this equilibrium
(pKa1¼6.04e6.434e36and pKa2¼16.0e17.434,36). Unfortunately values
for all the other anions were not foundin the literature, except for Hbz
(pKa¼11.1,37 or 11.0835), and H2tph (pKa1¼9.8035).
scenario S1 is illustrated in Fig. 8 with tph2ꢀ association (left) while
the binding arrangement S2 (right) is depicted for the iph2ꢀ asso-
ciation. It is noteworthy that among the minimized structures, no
binding arrangement with the carboxylate groups interacting ex-
clusively with the NeH macrocyclic linkages was found; which is
understandable taking into account the mismatch between the
stereochemistry of AC2A binding sites and the anion’s carboxylate
binding geometries. For the AC1A anion associations, only S2 type
binding scenarios wereobviously found. Subsequently the AC1A and
AC2A carboxylate associations were immersed in cubic boxes of
DMSO (as described in the Experimental section) and their struc-
tural behaviours investigated through molecular dynamics simula-
tions runs ranging from 10 to 15 ns.
In order to elucidate the preferential anchorage sites of the
substrates to the receptors, the associations of the methylated re-
ceptor Me4AC2A with iph2ꢀ and ph2ꢀ anions were also studied by
1H NMR titrations, in the same experimental conditions (at 298 K in
DMSO-d6). The addition of iph2ꢀ anion to Me4AC2A receptor led to
downfield shifts of the H6 and H3,11 proton resonances, with the last
one presenting only a slight move. The value of log KRS¼1.88(2) for
the binding constant was obtained for the formation of
[Me4AC2A(iph)]2ꢀ (Fig. S23). However, the addition of ph2ꢀ anion
to the same receptor results in only very small shifts of H6 reso-
nance and concomitantly the doublet signals coalesce in a broad
single signal (Fig. S24). The binding constant is very small. These
results emphasize the importance of the four amine bridges in
AC2A compared to Me4AC2A in the recognition of the studied an-
ions. The presence of these binding sites increases the number of
hydrogen bonds established among partners, however, due to the
significant acidic character of these amines, the deprotonation can
occur depending on the basicity of the substrate, as happened with
Fig. 8. Representative binding scenarios of the AC2A associations with tph2ꢀ (left, S1)
and iph2ꢀ (right, S2) anions. The O/HeN hydrogen bonds are drawn as green dashed
lines.
ph2ꢀ
.
The Oc/HeN intermolecular distances between the anion car-
boxylate groups (Oc is the centre of mass of a single carboxylate
group) and receptor NeH binding sites (selected L-alanine amide
3.2. Molecular dynamics simulations
binding sites and nitrogen bridges) plotted in Fig. 9 for the iph2ꢀ
and tph2ꢀ AC2A associations, show how anion recognition pro-
cesses advance during the course of the corresponding MD simu-
lations. A more perceptible depiction of the anion recognition at
molecular level is given by the surface grids presented in Fig. 10,
which were constructed with the positions occupied by the in-
dividual carboxylate groups over the entire MD simulation time.
The NMR solution studies showed that L-alanine functionalized
calix[2]arene[2]triazine receptors are able to recognize aromatic
mono-, di- and tricarboxylate anions mainly through 1:1 associ-
ations. In order to obtain further insights into the anion recogni-
tion, molecular mechanics (MM) and molecular dynamics (MD)
simulations were carried out for the associations between AC1A
and AC2A with iph2ꢀ and tph2ꢀ anions. These theoretical in-
vestigations were undertaken using the general AMBER force field
(GAFF)38 and RESP charges39 with the AMBER 10 software suite.40
Primarily, the X-ray molecular structures of AC2A (individual A
and B molecules) and Me4AC2A together with eight calix[2]arene
[2]triazine derivative structures obtained from Cambridge Struc-
tural Database23 were energy minimized in the gas phase using
default force field parameters. All MM optimized structures
showed an almost planar macrocyclic shape, eventually favoured
by the
p electron current ring through the nitrogen bridges, in-
stead of the 1,3-conformation. Furthermore, the CeN distances
from nitrogen bridging atoms to chlorotriazine and phenyl rings
are equivalent contrary to the experimental structural data. In
other words, the default GAFF parameters are not suitable to
model the structural singularity of the calix[2]arene[2]triazine
framework derived from the electronic nature of the nitrogen
bridges, and hence, it was necessary to develop specific parame-
ters as described in ESD.
The association structures between AC1A or AC2A macrocycles
with iph2ꢀ and tph2ꢀ anions (stoichiometry 1:1) were primarily
obtained through MD runs carried out in gas phase at 2000 K for 2 ns
(see below). Afterwards, all the recorded structures were minimized
byMMand energysorted. The lowestenergystructure (S1)found for
the AC2A anion associations, displays the anion located in the upper
rim with the carboxylate groups hydrogen bonded to both NeH
Fig. 9. Evolution of Oc/HeN intermolecular distances for the AC2A associated with
iph2ꢀ and tph2ꢀ anions along the MD simulations.
The AC2A$iph2ꢀ S1 binding arrangement is almost stable during
the first 8.4 ns of simulation (see Fig. 9, top left), with exception of
a relatively short period, ca. 1 ns, where this scenario is interrupted.
Afterwards, the O/HeN hydrogen bond with the second carbox-
ylate group is definitively broken and concomitantly only a single
hydrogen bond interaction, between the NeH amide binding site
amide L-alanine binding sites. The next structure on the energy
ranking (S2) appears ca. 17 kcal/mol above. In this S2 binding sce-
nario, the anion is O/HeN hydrogen bonded to single NeH amide
and NeH bridgebindingsitesvia one carboxylate group. Thebinding

676
A.I. Vicente et al. / Tetrahedron 68 (2012) 670e680
the number of oxygen atoms of each carboxylate group involved in
the main hydrogen bond interactions with AC2A NeH binding sites.
The association of AC1A with the iph2ꢀ and tph2ꢀ anions
(binding scenario S2) has shown to be almost stable as revealed by
the evolution of Oc/HeN intermolecular distances for 10 ns of MD
simulation runs (see Fig. S27). As would be expected, the dynamic
behaviours of these binding arrangements are equivalent to those
found for the corresponding AC2A anion associations (see Fig. S28
and Table S5).
4. Conclusion
Three novel calix[2]arene[2]triazine based receptors (AC1A,
AC2A and Me4AC2A) anchoring L-alanine units were prepared in
moderate yields. The first two macrocycles were soluble only in
DMSO, in which the binding affinities of the three receptors to-
wards the aromatic anions bzꢀ, ph2ꢀ, iph2ꢀ, tph2ꢀ and btc3ꢀ were
evaluated. The solvent being a hydrogen bond acceptor severely
competes with carboxylate groups of the anions for the NeH re-
ceptor binding sites, as suggested by the crystal structure of AC2A,
in which most of the anchorage sites, in particular the four amine
NeH bridges, are NeH/O hydrogen bonded to DMSO solvent
molecules. Simulations carried out with free AC2A and AC1A
revealed identical structural findings. In fact, the first DMSO solvent
shells calculated for all possible NeH/O]S interactions showed
apparently sharp peaks at N/O intermolecular distances consis-
tent with the existence of hydrogen bonds, and that the HeN
bridges are solvated in more extension than the NeH amide groups.
In this solvation conditions, the binding constants found for bzꢀ,
iph2ꢀ, tph2 and btc3ꢀ anions are unsurprisingly small.
Fig. 10. Surface areas representing the histograms of the positions occupied by the two
carboxylate groups of iph2ꢀ (top) and tph2ꢀ (bottom) around the AC2A in two alter-
native scenarios. A density contour value of 5 was used with surfaces of first and
second carboxylate groups drawn in green chartreuse and yellow, respectively.
and the first carboxylate, is maintained until the end of the MD
simulation (15 ns), leading to a relatively narrow green grid surface
for this carboxylate group; whereas the second one displays two
well-defined yellow grid surfaces corresponding to bound and
unbound states (Fig. 10, top left).
The 1H NMR binding investigations showed that AC1A and AC2A
receptors exhibit comparable anion binding affinities, which in-
dicates that the studied anions are not simultaneously recognized
by both arms of the AC2A receptor. Furthermore, the 1H NMR
studies also show unequivocally that the NeH bridges of AC1A and
AC2A interact with anions, in agreement with the bulky solvation of
AC2A by DMSO solvent molecules, and further confirmed by sub-
sequent molecular dynamics simulations undertaken for AC1A and
AC2A associations with the iph2ꢀ and tph2ꢀ anions. Unlikely, the
AC2A/AC1A/ph2ꢀ systems are primarily involved in a Brønsted
acid-base equilibrium rather than an association processes. In other
words, this study showed that the AC1A and AC2A NeH bridges
behave either as binding sites or Brønsted acids depending on the
base strength of the carboxylate anion.
This simulation was repeated for 10 ns using different starting
system coordinates, and an equivalent sequence of events was
observed. In contrast, the S2 arrangement is retained for 10 ns
simulation length (see Fig. 9, top right) with first carboxylate group
strongly hydrogen bonded to AC2A, which is consistent with the
tight green surface grid area (Fig. 10, top right). Furthermore, the
widespread yellow surface grid indicates that second carboxylate
group oscillates extensively around the NeH binding sites region,
as would be expected for a free carboxylate group.
The AC2A$tph2ꢀ in S1 binding arrangement displays a different
dynamic behaviour as found for the equivalent iph2ꢀ scenario. In-
deed, this scenario is kept over the entire time of simulation (Fig. 9,
bottom left) as well illustrated by the two separated green and
yellow surface areas of both carboxylate groups (Fig. 10, bottom
left). The S2 binding geometry is retained for 10 ns of MD simula-
tion with the anion using a single carboxylate group to interact with
AC2A, (Figs. 9 and 10, bottom right) as also observed for the S2
iph2ꢀ association. The average O/HeN hydrogen bond dimensions
calculated using a cut-off of 120ꢁ for the O/HeN angles are
gathered in Table 2. Moreover, the hydrogen bonds with occupan-
cies lesser than 40% were discarded and so the values quoted give
5. Experimental section
5.1. General
Melting points were determined with a Reicher Model Ther-
movar melting-point apparatus and are uncorrected. Optical rota-
tions were recorded with a PerkineElmer 241 polarimeter. IR
spectra were obtained with a PerkineElmer FT-IR spectropho-
tometer. HRMS (ESI) and MS (ESI) measurements were performed
on an ApexQe FTICR Mass Spectrometer from Bruker Daltonics
equipped with a combined Apollo II electrospray/MALDI ion source
and a 7 T actively shielded superconducting magnet. 1H, 13C (APT)
and 2D NMR (HMQC and HMBC) spectra were obtained on Bruker
spectrometers MX300 and CXP400 operating at 300 and 400 MHz,
respectively.
Table 2
a
ꢁ
Average O/N distances (A) with their standard deviations for the O/HeN hy-
drogen bonds between AC2A and iph2ꢀ and tph2ꢀ anions
NeH bindings sites Amide groups
Nitrogen bridges
1
eCO22ꢀ groups
1
2
iph2ꢀ
S1
S2
3.04(27); 3.29(35) 3.07(25)
3.14(25)
2.85(13)
Reagents: 1,3-phenylenediamine, 2,4,6-trichloro-1,3,5-triazine,
3,5-dinitrobenzoyl chloride, propylene oxide, diisopropylethyl-
tph2ꢀ
S1
S2
amine (DIPEA), thionyl chloride (SOCl2), L-alanine, 10% Pd/C cata-
3.03(24); 3.06(28) 3.04(24); 3.11(28)
3.10(23); 3.13(25)
ꢁ
2.93(28); 3.17(48)
lyst, silica gel 60 A (230e400 mesh ASTM), aromatic carboxylic
acids, methyl iodide, Cs2CO3 and NH4Cl salts were purchased from
Sigma Aldrich. All these reagents were used as supplied without
a
Standard deviations were calculated using all trajectory frames using all O/N
distances consistent with the existence of hydrogen bonds.

A.I. Vicente et al. / Tetrahedron 68 (2012) 670e680
677
further purification. The 1H and 13C NMR spectra were recorded
using one of the following deuterated solvents, DMSO-d6 (99.9%),
CDCl3 or C3D6O, which were also purchased from Sigma Aldrich. All
solvents, dimethylformamide (DMF), methanol (MeOH), ethanol
(EtOH), tetrahydrofuran (THF), acetone and acetonitrile were pu-
rified and dried before use according to standard methods.41 The
tetrabutylammonium hydroxide in methanol solution (1 M) was
also purchased from Aldrich. However, the potentiometric titration
of this reagent gave a concentration of 1.27 M.
NMR (400 MHz, C3D6O, 25 ꢁC):
d
173.0, 169.8, 169.0, 165.8, 164.1,
137.4, 135.5, 117.8, 52.0, 48.3, 16.7; ESIeMS m/z (%): 532.0 (79)
[MþH]þ, 533.0 (18) [MþHþ1]þ, 534.0 (100) [MþHþ2]þ, 535.0 (22)
[MþHþ3]þ, 536.0 (46) [MþHþ4]þ, 537.0 (10) [MþHþ5]þ, 538.0 (8)
[MþHþ6]þ; HReESIeMS (m/z): calcd for [C17H13Cl4N9O3þH]þ,
531.99682; found, 531.99455.
5.2.5. Linear trimer (6). Following the experimental procedure
used for the preparation of 5, the reaction of 1,3-phenylenediamine
(0.36 g, 3.3 mmol) with 2,4,6-trichloro-1,3,5-triazine (1.23 g,
6.7 mmol) in presence of DIPEA (1.08 g, 8.3 mmol) gave the trimer
6, which was further purified by silica gel column chromatography
using a mixture of ethyl acetate/petroleum ether, 1:4 (v/v) as elu-
ent. The compound 6 was obtained (1.09 g) in 81% yield as white
5.2. Synthetic procedures
5.2.1. -Alanine methyl ester (2). To a cooled solution of the L-ala-
L
nine (2.00 g, 11 mmol) in dry methanol (12.0 mL), SOCl2 (2.2 mL,
15.41 mmol) was slowly added. The mixture was refluxed for 3 h
and left at rt under stirring for 45 h. Methanol was then removed
and the crude product was washed with toluene three times. The
toluene was removed under reduce pressure and 2 was obtained as
white powder (2.31 g) in a quantitative yield: 1H NMR (400 MHz,
solid: 1H NMR (400 MHz, C3D6O, 25 ꢁC):
d 10.09 (s, 2NH), 8.18 (t,
J¼2.2 Hz, 1H), 7.63 (d, J¼8 Hz, 2H), 7.48 (t, J¼8 Hz, 1H); 13C NMR
(400 MHz, C3D6O, 25 ꢁC):
d 171.5, 170.6, 165.5, 138.6, 130.3, 119.2,
115.4. These structural data are comparable to those previously
reported for this compound.15
CDCl3 and a drop of MeOD, 25 ꢁC, TMS):
J¼7.2 Hz, 1H), 3.84 (s, 3H), 1.65 (d, J¼7.2 Hz, 3H).
d 8.65 (br, NH2), 4.10 (q,
5.2.6. Tetraazacalix[2]arene[2]triazine AC1A. The macrocycle AC1A
was prepared through two alternative synthetic pathways as
depicted in Scheme 3. Following the route b1, a solution of trimer 5
(267 mg, 0.5 mmol) and 1,3-phenylenediamine (54 mg, 0.5 mmol),
both in acetone, were added dropwise and simultaneously to a so-
lution of DIPEA (0.21 mL, 1.2 mmol) in acetone at reflux during 4 h
approximately. The reaction mixture was left stirring at reflux
during 48 h until all starting materials were consumed. The solvent
was removed and the crude solid was successively precipitated
(DMSO/acetone) to give a white power in yield of 32% of the
macrocycle AC1A (91 mg). The AC1A was also prepared via route b2
by simultaneous addition of a solution of trimer 6 (403 mg,1 mmol)
5.2.2. -Alanine methyl ester 3,5-dinitrobenzamide (3). A mixture of
L
3,5-dinitrobenzoyl chloride (1.15 mg, 5 mmol) and propylene ox-
ide (1.05 mL, 15 mmol) was added to a solution of -alanine methyl
L
ester 2 (515.0 mg, 5 mmol) in dry THF under nitrogen. The reaction
mixture was stirred at rt for 2 h. Subsequently, the solvent was
removed under reduced pressure and the crude product was pu-
rified by column chromatography using as eluent a mixture of
acetone/petroleum ether, 1:3 (v/v). The compound 3 was obtained
as white solid (465.5 mg) in a yield of 70%: mp 125e126 ꢁC; ½a D20
ꢂ
þ10.37 (c 0.27, acetone); IR (KBr) nmax 3315 (NH amide), 3109,
2958, 1739 (C]O ester), 1645 (C]O amide), 1542 (N]O), 1640,
1540, 1550 (CC Ar), 1345 cmꢀ1 (CN); 1H NMR (300 MHz, CDCl3,
and
L-alanine methyl ester 3,5-diaminobenzamide 4 (237 mg,
1 mmol) in acetone to a solution of DIPEA (0.42 mL, 2.4 mmol) in
acetone to give a white powder (284 mg, 50% yield) after pre-
cipitation in a mixture of DMSO/acetone solution; mp >350 ꢁC;
25 ꢁC, TMS):
d
9.08 (br s, 1H); 8.86 (d, J¼2 Hz, 2H), 7.65 (d, J¼7.2 Hz,
NH), 4.81 (quint, J¼7.2 Hz, 1H), 3.81 (s, 3H), 1.54 (d, J¼7.6 Hz, 3H);
13C NMR (400 MHz CDCl3, 25 ꢁC, TMS):
173.9, 162.3, 148.7, 137.1,
d
½
a 2D0
ꢂ
þ2.86 (c 0.022, DMSO); IR (KBr) nmax 3281 (NH), 3070, 1731
127.5, 121.4, 53.2, 49.2, 18.3 ppm; ESIeMS m/z (%): 320.1 (100)
[MþNa]þ, 321.1 (11) [MþNaþ1]þ, 322.1 (1.3) [MþNaþ2]þ;
HReESIeMS (m/z): calcd for [C11H11N3O7þNa]þ, 320.04947; found,
320.05021.
(CO ester), 1664 (CO amide), 1600, 1589, 1458 (CC Ar), 1545 (CN Ar),
1396 (CN amide), 1293, 1179, 806 cmꢀ1 (CCl); 1H NMR (400 MHz,
DMSO-d6, 25 ꢁC):
d
10.05 (s, 2NH), 9.98 (s, 2NH), 8.72 (d, J¼6.8 Hz,
NH), 7.83 (br s, 1H), 7.72 (br s, 1H), 7.30 (d, J¼2.4 Hz, 2H), 7.21 (t,
J¼8.0 Hz, 1H), 6.78 (dd, J¼8.0 and 2.4 Hz, 2H), 4.41 (quint, J¼7.2 Hz,
1H), 3.61 (s, 3H), 1.34 (d, J¼7.2 Hz, 3H); 13C NMR (400 MHz, DMSO-
5.2.3. -Alanine methyl ester 3,5-diaminobenzamide (4). A mixture
L
of 3 (465.5 mg, 1.57 mmol) with 10% wt Pd(C) (47 mg) catalyst in
ethanol (16 mL) was stirred in a Parr shaker at a 50 psi H2 atmo-
sphere for 1 h. The reaction evolution was monitored by thin layer
chromatography. The crude mixture was filtered through a pad of
Celite and the solvent was removed under reduced pressure to
afford the compound 4 (368 mg) as a brown solid in quantitative
yield. The product was directly engaged in the next step without
further purification.
d6, 25 ꢁC):
d 172.9,168.2,165.4, 164.7,164.3,137.8,134.5,129.1, 123.2,
118.9, 118.8, 118.3, 51.7, 48.2, 17.0; ESIeMS m/z (%): 568.1 (100)
[MþH]þ, 569.1 (26) [MþHþ1]þ, 570.1 (65) [MþHþ2]þ, 571.1 (17)
[MþHþ3]þ, 572.1 (11) [MþHþ4]þ, 573.1 (3) [MþHþ5]þ;
HReESIeMS (m/z): calcd for [C23H19Cl2N11O3þH]þ, 568.11222;
found, 568.11662.
5.2.7. Tetraazacalix[2]arene[2]triazine AC2A. A solution of
4
(237 mg, 1 mmol) in acetone (90 mL) and trimer 5 (533 mg,
1 mmol) in acetone (90 mL) were added dropwise and simulta-
neously to a solution of DIPEA (0.42 mL, 2.4 mmol) in acetone
(200 mL) at reflux during 8 h approximately. The reaction mixture
was left stirring at reflux during 48 h until all starting materials
were consumed. The solvent was removed and the crude solid was
successively precipitated (DMSO/acetone) to give AC2A as a white
powder (279 mg) in 40% yield. Colourless needles suitable for single
crystal X-ray diffraction determination were grown up from DMSO:
5.2.4. Trimer (5). A mixture of diamine 4 (368 mg, 1.6 mmol) and
DIPEA (0.7 mL, 4.0 mmol) in THF was dropped wise into a THF so-
lution of triazine (570 mg, 3.1 mmol), previously cooled under an
ice bath. The reaction mixture was stirred at 0 ꢁC for 2 h followed by
half hour at rt. The DIPEA hydrochloride salt was removed by fil-
tration and the crude product was then concentrated in vacuo. The
trimer was obtained as a white powder (497 mg), in 60% yield after
purification by silica gel column chromatography using a mixture of
acetone/petroleum ether, 1:1 (v/v) as eluent; mp 294 ꢁC with de-
mp >350 ꢁC; ½a 2D0
þ94.55 (c 0.022, DMSO); IR (KBr) nmax 3263 (NH),
ꢂ
composition, ½a D20
ꢂ
þ1.05 (c 0.017, DMSO); IR (KBr) nmax 3267 (NH),
3070, 2958, 1733 (CO ester), 1654 (CO amide), 1600, 1534, 1458 (CC
Ar), 1545 (CN Ar), 1395 (CN amide), 1288, 1224, 1177, 807 cmꢀ1
3109, 2958, 1725 (C]O ester), 1647 (C]O amide), 1600, 1534, 1459
(CC Ar), 1241, 1184, 852 (CCl), 796 cmꢀ1; 1H NMR (400 MHz, C3D6O,
(CCl); 1H NMR (400 MHz, DMSO-d6, 25 ꢁC):
d 10.06 (s, 4NH), 8.73 (d,
25 ꢁC):
d
11.37 (s, 2NH), 8.84 (s,1NH), 8.06 (br s, H), 7.86 (d, J¼1.6 Hz,
J¼6.8 Hz, 2H), 7.80 (br s, 2H), 7.30 (d, J¼2.4 Hz, 4H), 4.41 (quint,
2H), 4.48 (quint, J¼6 Hz, 1H), 3.66 (s, 3H), 1.39 (d, J¼6 Hz, 3H); 13C
J¼7.2 Hz, 2H), 3.61 (s, 6H), 1.34 (d, J¼7.2 Hz, 6H); 13C NMR

678
A.I. Vicente et al. / Tetrahedron 68 (2012) 670e680
(400 MHz, DMSO-d6, 25 ꢁC):
d
172.9, 168.2, 165.4, 164.6, 137.8, 134.8,
The association constants of the species formed in solution were
determined from the experimental titration data using HypNMR,24
which required as input the concentration of each component and
the observed chemical shift. The fitting included the binding con-
stants and the chemical shiftof each Rr/Si species. Initial calculations
provided overall stabilityconstants, bR S values, bR S ¼[RrSs]/[R]r[S]s.
123.3, 119.3, 52.1, 48.5, 16.8; ESIeMS m/z (%): 697.2 (100) [MþH]þ,
698.2 (44) [MþHþ1]þ, 699.1 (68) [MþHþ2]þ, 700.1 (25)
[MþHþ3]þ, 701.2 (9) [MþHþ4]þ, 702.1 (3) [MþHþ5]þ;
HReESIeMS (m/z): calcd for [C28H26Cl2N12O6þH]þ, 697.15481;
found, 697.15259.
r
s
r s
Only RS species were found for all systems. The errors quoted are the
standard deviations of the overall stability constants given directly
by the program from the input data, including the experimental
points for all resonances of the titration curves.
5.2.8. N-Tetramethyltetraazacalix[2]arene [2]triazine Me4AC2A. To
a solution of AC2A (100 mg, 0.14 mmol) with Cs2CO3 (190 mg,
1.4 mmol) in acetonitrile at 50 ꢁC was added MeI (70
mL,1.12 mmol)
and the reaction was left stirring for 2 h until total consumption of
starting macrocycle. The solvent was removed, a saturated solution
of NH4Cl was added and the crude product was obtained by suc-
cessive extractions with ethyl acetate (3ꢃ50 mL). The Me4AC2Awas
isolated as white powder (57 mg), in yield of 52% after purification
by silica gel column chromatography using a mixture of acetone/
petroleum ether, 2:1 (v/v) as eluent and subsequent precipitation
The AC1A receptor self-assembling was investigated through 1H
NMR experiments carried out with 17 receptor solutions in con-
centrations ranging between 7.1ꢃ10ꢀ4 and 3.09ꢃ10ꢀ2 M in DMSO-
d6. No significant changes of the chemical shift were found in this
range of concentrations.
5.4. Crystallography
with chloroform and petroleum ether; mp >350 ꢁC; ½a D20
þ7.88 (c
ꢂ
0.022, acetone); IR (KBr) nmax 3320 (NH amide), 1738 (CO ester),
The X-ray data of AC2A$4.5DMSO$2H2O were collected on
1656 (CO amide), 1572, 1495, 1457 (CC Ar), 1551 (CN Ar), 1402 (CN
a CCD Bruker APEX II at 150(2) K using graphite monochromatized
amide),1282,1216, 802 (CCl) cmꢀ1; 1H NMR (400 MHz, CDCl3)
d
1.51
Mo K
a
radiation (
l
¼0.71073 A). The crystal was positioned at
ꢁ
(d, J¼7.1 Hz, 6H), 3.37 (s, 12H), 3.78 (s, 6H), 4.64 (quint, J¼6 Hz, 2H),
35 mm from the CCD and the spots were measured using a count-
ing time of 60 s. Data reduction including a multi-scan absorption
correction was carried out using the SAINT-NT software package
from Bruker AXS. The structure was solved by a combination of
direct methods with subsequent difference Fourier syntheses and
refined by full matrix least squares on F2 using the SHELX-97
suite.42 Anisotropic thermal displacements were used for all non-
hydrogen atoms. In addition, two waters and two DMSO mother
liquor molecules were refined with occupancy factors of 0.5 and
with ISOR restraints. Hydrogen atoms bonded to carbon and to
nitrogen atoms were placed at geometrical positions and refined
with Uiso¼1.2Ueq of the atom they are attached. The atomic posi-
tions of the water hydrogen atoms were not discernible from dif-
ference Fourier maps and therefore they were not included in the
structure refinement. Molecular diagrams were drawn with
PYMOL software.44
6.59 (d, J¼6.7 Hz, 2NH), 6.87 (s, 2H), 7.23 (s, 2H), 7.28 (s, 2H); 13C
NMR (400 MHz, CDCl3)
d 18.4, 38.2, 48.9, 52.8, 124.9, 125.3, 130.6,
137.1, 145.6, 165.5, 165.9, 170.0, 173.6; MS (ESI) m/z (%): 753.2 (100)
[MþHþ1]þ, 754.2 (27) [MþHþ2]þ, 755.2 (66.6) [MþHþ3]þ, 756.2
(15.5) [MþHþ4]þ, 757.2 (5.8) [MþHþ5]þ, 758.0 (0.8) [MþHþ6]þ;
HReESIeMS (m/z): calcd for [C32H34Cl2N12O6þH]þ, 753.21013;
obsd, 753.21393.
5.2.9. Tetrabutylammonium carboxylate salts. The carboxylate an-
ions (bzꢀ, ph2ꢀ, iph2ꢀ, tph2ꢀ and btc3ꢀ) were used as tetrabuty-
lammonium salts and prepared as follows: 1e3 equiv of
a methanolic solution of tetrabutylammonium hydroxide (1.27 M),
depending on the number of carboxylic groups, was added to
a solution of carboxylic acid in methanol. The mixture was left
stirring at rt for 2 h. Subsequently, the solvent was evaporated and
the salts were recrystallized from acetone/ethanol and dried under
high vacuum for 24 h. The purity of the salts was confirmed by the
area ratio between 1H NMR resonances of the tetrabutyl cation and
the carboxylate anions.
The crystal data and refinement details are summarized in Table
S2 given in the Supplementary data.
5.5. Molecular modelling
MM calculations and MD simulations were carried out with the
AMBER 10 suite of programs40 using GAFF parameters.38 Atomic
partial charges were calculated from previous optimized structures
at HF/6-31G* level for AC1A, AC2A, tetrabutylammonium and at HF/
6-31G*(þþ) level for ph2ꢀ, iph2ꢀ and tph2ꢀ anions using the
Gaussian 03.44 Subsequently, RESP charges39 were obtained from an
HF/6-31G* single point calculation with Gaussian IOP (6/33¼2, 6/
41¼4, 6/42¼6) via RESP fitting using a single conformation for an-
ions and tetrabutylammonium cation and five different conforma-
tions for AC1A and AC2A macrocycles. Default force field parameters
were used apart from those involving directly the CeN macrocyclic
bridges, which were handled as follows: one nitrogen and two car-
bon atom types were created in order to mimic the different values
of the NeC bridging distances. The nitrogen-bridging atom was
called NH while carbon bridges were assigned as CI (triazine rings)
and CX (phenyl rings) as shown in Fig. S25. The NHeCX and NHeCI
ideal distances and the corresponding CXeNHeCI bending angles
terms used were the average values calculated from the X-ray single
crystal data of eight calix[2]arene[2]triazine derivatives currently
available from CSD23 plus the AC2A and Me4AC2A (unpublished
results) structures. The force constants for NHeCX and NHeCI bond
stretching and CXeNHeCI bending angle terms were calculated
using the corresponding force field equations.38,45 In addition, in
order to reproduce well the 1,3-conformational shape, the terms of
5.3. NMR studies
Reagents and solutions: the tetrabutylammonium anion salts
used as substrates were prepared as described above. Stock solu-
tions of the AC1A, AC2A and Me4AC2A (4.0e4.3ꢃ10ꢀ3 M) receptors
and anions (2.0e3.0ꢃ10ꢀ2 M) were prepared in DMSO-d6 and
stored in a glove box for further use.
Titrations of the receptors with anions followed by 1H NMR
started with the free receptor solution (400
aliquots of the anion salt solution were added successively using
a Hamilton syringe (Microlitre 700 series, 10 L) until no further
change in chemical shift has been observed. All measurements
were carried out at 298 K and the solutions were left to equilibrate
for 15e30 min before new addition of anion solution.
mL) to which 15 to 30
m
For the Job’s plots, ten solutions were prepared (500 mL) by the
mixture of the receptor (AC1A or AC2A) and anion (iph2ꢀ, tph2ꢀ or
btc3ꢀ) DMSO-d6 solutions at the same concentration (2e3 mM).
The following R/S (R is the receptor and S the anion) volume ratios
were used: 500:0, 450:50, 400:100, 350:150, 300:200, 250:250,
200:300, 150:350, 100:400 and 50:450.
The 1H NMR titrations were carried out on a Bruker CXP 300
spectrometer. No effort was made to maintain the ionic strength
constant in order to avoid the competition of other anions.

A.I. Vicente et al. / Tetrahedron 68 (2012) 670e680
679
the dihedral angles around the CeN bridges were also changed. The
Supplementary data
Electronic Supplementary data (ESD) available: 1H NMR and 13
remaining atoms of the calix[2]arene[2]triazine skeleton were
assigned with default force field atom types. The force field pa-
rameters for this platform are given in Table S4 (ESD) together with
van der Waals used for CX, CI and NH atom types. With this specific
parameterization the molecular structures of ten calix[2]arene[2]
triazine derivatives were well reproduced in gas phase by MM as
shown by the fitting and corresponding rms values (root mean
square deviation) between the MM minimized structures and solid
state structures (see Table S3). In addition, the mimetic structural
improvement obtained with this specific parameterization is illus-
trated with three representative examples in Fig. S26.
The starting coordinates of AC1A and AC2A were generated from
the crystal structure of AC2A. The selected anion (iph2ꢀor tph2ꢀ) was
positioned on the upper rim and the 1:1 model association was
subsequently energy-minimized by MM through the conjugate
gradient algorithm until a convergence criteria of 0.0001 kcal molꢀ1
was achieved. The minimized structures were then subjected to
a molecular dynamic quenching run at 2000 K using a time step of
1 fs. 20,000 Gas phase conformations were generated at 0.1 ps in-
tervals and subsequently minimized by MM using a house script.
MD simulations in DMSO were performed for the S1 and S2 low
energy anion associations (see supra). Single solute molecules were
solvated with a DMSO full atom solvent model46 leading to periodic
cubic boxes containing 2080 DMSO molecules. The electrostatic
neutrality of each systemwas achieved with the replacement of two
DMSO molecules by an equivalent number of tetrabutylammonium
cations.
C
spectra of all compounds as well as the COSY and HMQC spectra of
the new macrocycles are included. The 1H NMR titration data in-
clude the Job’s plots for AC2A associations with iph2ꢀ and tph2ꢀ
anions and the 1H NMR spectra for titrations of ACA with bzꢀ, iph2ꢀ
and tph2ꢀ, AC2A with bzꢀ, iph2ꢀ, tph2ꢀ and ph2ꢀ, ACA with TBAOH
and Me4AC2A with iph2ꢀ and ph2ꢀ anions. Crystallographic details
are presented. Computational data and protocol are also included.
Crystallographic data for AC2A$4.5(DMSO)$2H2O have been de-
posited with the Cambridge Crystallographic Data Centre allocated
with the deposit number CCDC 836953. Copy of the data can be
obtained free of charge on application to CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK fax (þ44) 1223 336033, e-mail: depos-
it@ccdc.cam.ac.uk. Supplementary data related to this article can be
found online at doi:10.1016/j.tet.2011.10.090.
References and notes
€
1. Calixarenes 2001, 1st ed; Asfari, M. Z., Bohmer, V., Harrowfield, J., Vicens, J., Eds.;
Springer: Dordrecht, 2010.
2. Xue, M.; Chen, C.-F. Chem. Commun. 2011, 2318e2320.
3. Wang, L.-X.; Zhao, L.; Wang, D.-X.; Wang, M.-X. Chem. Commun. 2011,
9690e9692.
4. Maes, W.; Dehaen, W. Chem. Soc. Rev. 2008, 37, 2393e2402.
5. Wang, D.-X.; Zheng, Q.-Y.; Wang, Q.-Q.; Wang, M.-X. Angew. Chem., Int. Ed. 2008,
47, 7485e7488.
6. Wang, Q.-Q.; Wang, D.-X.; Yang, H.-B.; Huang, Z.-T.; Wang, M.-X. Chem.dEur. J.
2010, 16, 7265e7275.
7. Hou, B.-Y.; Zheng, Q.-Y.; Wang, D.-X.; Huang, Z.-T.; Wang, M.-X. Chem. Commun.
2008, 3864e3866.
All MD simulations started with equilibration of each system
under periodic boundary conditions using the following multistage
protocol. The potential bad contacts between the solvent molecules
and the solute were removed via MM minimization with 1000 steps
by the steepest descent method followed by 10,000 steps of con-
8. Zhang, E.-X.; Wang, D.-X.; Zheng, Q.-Y.; Wang, M.-X. Org. Lett. 2008, 10,
2565e2568.
9. Wang, M.-X. Chem. Commun. 2008, 4541e4551.
10. Thomas, J.; Maes, W.; Robeyns, K.; Ovaere, M.; Meervelt, L. V.; Smet, M.; De-
haen, W. Org. Lett. 2009, 11, 3040e3043.
jugate gradient keeping the solute rigid with positional restraints of
A . Subsequently, the restraint was removed and
the entire system was relaxed by a further MM minimization. The
11. Van Rossom, W.; Maes, W.; Kishore, L.; Ovaere, M.; Van Meervelt, L.; Dehaen, W.
Org. Lett. 2008, 10, 585e588.
12. Van Rossom, W.; Robeyns, K.; Ovaere, M.; Van Meervelt, L.; Dehaen, W.; Maes,
W. Org. Lett. 2011, 13, 126e129.
ꢁꢀ2
500 kcal molꢀ1
system was then heated up to 300 K for 50 ps using an NVT en-
13. Van Rossom, W.; Ovaere, M.; Van Meervelt, L.; Dehaen, W.; Maes, W. Org. Lett.
ꢁꢀ2
semble and Lengvin thermostat. A 10 kcal molꢀ1
A
positional re-
2009, 11, 1681e1684.
14. Wang, M.-X.; Zhang, X.-H.; Zheng, Q.-Y. Angew. Chem., Int. Ed. 2004, 43, 838e842.
15. Wang, M.-X.; Yang, H.-B. J. Am. Chem. Soc. 2004, 126, 15412e15422.
16. Clayden, J.; Rowbottom, S. J. M.; Hutchings, M. G.; Ebenezer, W. J. Tetrahedron
Lett. 2009, 50, 3923e3925.
straint on solute was applied. Afterwards, the solute was released
and an MD run in NPT ensemble under average pressure of 1 atm
with isotropic scaling was carried out for 500 ps. At the end of this
simulation, the density of the cubic box was within the experi-
mental value for a DMSO solution at 300 K.
17. Clayden, J.; Rowbottom, S. J. M.; Ebenezer, W. J.; Hutchings, M. G. Org. Biomol.
Chem. 2009, 7, 4871e4880.
18. Wang, L.-X.; Wang, D.-X.; Huang, Z.-T.; Wang, M.-X. J. Org. Chem. 2010, 75,
741e747.
MD data collection runs were produced for 10e15 ns in an NPT
ensemble. Frames were saved every 0.2 ps leading to a trajectory files
containing between 50,000 and 75,000 frames. The SHAKE algo-
rithm47 was applied to constrain all bonds involving hydrogen atoms
and a time step of 2 fs was used. The Particle Mesh Ewald method was
used to describe the long-range electrostatic interactions and the
19. Katz, J. L.; Tschaen, B. A. Org. Lett. 2010, 12, 4300e4303.
20. Miyazaki, Y.; Kanbara, T.; Yamamoto, T. Tetrahedron Lett. 2002, 43, 7945e7948.
21. Yang, X.; Lowe, C. R. Tetrahedron Lett. 2003, 44, 1359e1362.
22. Wang, Q.-Q.; Wang, D.-X.; Ma, H.-W.; Wang, M.-X. Org. Lett. 2006, 8,
5967e5970.
23. Allen, F. H. Acta Cryst. Sect. B 2002, 58, 380e388.
24. Frassineti, C.; Ghelli, S.; Gans, P.; Sabatini, A.; Moruzzi, M. S.; Vacca, A. Anal.
Biochem. 1995, 231, 374e382.
ꢁ
non-bonded van der Waals interactions were subject to a 12 A cut-off.
All molecular diagrams were drawn with PyMOL.43
25. Hirose, K. Determination of Binding Constants In Analytical Methods in Supra-
molecular Chemistry; Schalley, C. A., Ed.; GmbH & KGaA: Weinheim, 2007.
26. Mahoney, J. M.; Stucker, K. A.; Jiang, H.; Carmichael, I.; Brinkmann, N. R.; Beatty,
A. M.; Noll, B. C.; Smith, B. D. J. Am. Chem. Soc. 2005, 127, 2922e2928.
Acknowledgements
~
27. Costero, A. M.; Banuls, M. J.; Aurell, M. J.; Ward, M. D.; Argent, S. Tetrahedron
2004, 60, 9471e9478.
28. Esteban-Gomez, D.; Fabbrizzi, L.; Licchelli, M. J. Org. Chem. 2005, 70,
This work was funded through the Operational Program Com-
ꢀ
petitiveness FactorsdCOMPETE and National Funds through
ˇ
5717e5720.
ꢀ
29. Boiocchi, M.; Del Boca, L.; Esteban-Gomez, D.; Fabbrizzi, L.; Licchelli, M.;
~
FCTdFundac¸ aoparaaCienciaeaTecnologia underprojectPTDC/QUI/
Monzani, E. Chem.dEur. J. 2005, 11, 3097e3104.
30. Amendola, V.; Boiocchi, M.; Fabbrizzi, L.; Palchetti, A. Chem.dEur. J. 2005, 11,
5648e5660.
68582/2006. Authors thank to Helena Santos (ITQB) and Pedro
Lamosa (ITQB) for the NMR interpretation and technical assistance
and Pedro Vaz by the mass spectrometry data acquisition. Acknowl-
edgements are also due to the National NMR (contract REDE/1517/
RMN/2005) and Mass Spectrometry (REDE/1501/REM/2005) net-
works supported by National Program for Scientific Re-equipment,
with funds from POCI 2010 (FEDER) and FCT by use of these facili-
ties. A.I.V. and J.M.C. also thank FCT for the Ph.D. scholarships SFRH/
BD/32487/2006 and SFRH/BD/66789/2009, respectively.
ꢀ
31. Esteban-Gomez, D.; Fabbrizzi, L.; Licchelli, M.; Sacchi, D. J. Mater. Chem. 2005,
15, 2670e2675.
ꢀ~
ꢀ
32. Ros-Lis, J. V.; Martínez-Manez, R.; Sancenon, F.; Soto, J.; Rurack, K.; Weibhoff, H.
Eur. J. Org. Chem. 2007, 2449e2458.
33. Amendola, V.; Fabbrizzi, L.; Mosca, L.; Schmidtchen, F.-P. Chem.dEur. J. 2011, 21,
5792e5981.
34. Choi, P. J.; Petterson, K. A.; Roberts, J. D. J. Phys. Org. Chem. 2002, 15, 278e286.
35. Shan, S.; Herschlag, D. J. Am. Chem. Soc. 1996, 118, 5515e5518.
36. Chantooni, M. K.; Kolthoff, I. M. J. Phys. Chem. 1975, 79, 1176e1182.

680
A.I. Vicente et al. / Tetrahedron 68 (2012) 670e680
37. Kolthoff, I. M.; Reddy, T. B. Inorg. Chem. 1962, 1, 189e194.
38. Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A. J. Comput. Chem.
2004, 25, 1157e1174.
39. Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Kollmann, P. A. J. Am. Chem. Soc. 1993, 115,
9620e9631.
Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.;
Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.;
Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao,
O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken,
V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A.
J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G.
A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.;
Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Fores-
man, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov,
B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.;
Keith, T.; Al Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P.
M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian
03, Revision C.02; 2008.
40. Case, D. A.; Darden, T.; Cheatham, T. E., III; Simmerling, C.; Wang, J.; Duke, R. E.;
Luo, R.; Crowley, M.; Walker, R.; Zhang, W.; Merz, K. M.; Wang, B.; Hayik, S.;
ꢀ
Roitberg, A.; Seabra, G.; Kolossvary, I.; Wong, K. F.; Paesani, F.; Vanicek, J.; Wu,
X.; Brozell, S. R.; Steinbrecher, T.; Gohlke, H.; Yang, L.; Tan, C.; Mongan, J.;
Hornak, V.; Cui, G.; Matthews, D. H.; Seetin, M. G.; Sagui, C.; Babin, V.; Kollman,
P. A. AMBER 10; University of California: 2008.
41. Armarego, W. L. F.; Chai, C. L. L. Purification of Laboratory Chemicals, 6th ed.;
Elsevier: Burlinghton, 2009.
42. Sheldrick, G. M. Acta Cryst. Sect. A 2007, 64, 112e122.
43. Delano, W. L. The PyMOL Molecular Graphics System Version 1.2r2; Schrodinger,
45. Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A. J. Comput. Chem.
2005, 26, 114.
€
LLC: 2002.
46. Fox, T.; Kollman, P. A. J. Phys. Chem. B 1998, 102, 8070e8079.
47. Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. C. J. Comput. Phys. 1977, 23,
327e341.
44. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;
Cheeseman, J. R.; Montgomery, J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.;