Selective recognition of sulfate ions by tripodal cyclic peptides functionalised with (thio)urea binding sites

Article abstract of DOI:10.1039/c2ob06964d

A tripodal urea and tripodal thiourea with the same cyclic peptide core have been synthesised and their anion binding ability investigated. In CDCl ₃, the tripodal urea self-associates whereas the thiourea does not. Neither compound shows self-

Full text of DOI:10.1039/c2ob06964d

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 2664
www.rsc.org/obc
PAPER
Selective recognition of sulfate ions by tripodal cyclic peptides functionalised
with (thio)urea binding sites†
Philip G. Young and Katrina A. Jolliffe*
Received 23rd November 2011, Accepted 17th January 2012
DOI: 10.1039/c2ob06964d
A tripodal urea and tripodal thiourea with the same cyclic peptide core have been synthesised and their
anion binding ability investigated. In CDCl3, the tripodal urea self-associates whereas the thiourea does
not. Neither compound shows self-association in the more polar 10% v/v DMSO-d6/CDCl3. Both
compounds bind strongly and selectively to sulfate ions in CDCl3 and 10% v/v DMSO-d6/CDCl3. This
selectivity is attributed to a unique binding mode for sulfate, in which this tetrahedral anion forms nine
hydrogen bonds to the receptors, with three of these coming from the amide protons of the cyclic peptide.
anions. For example, an optimal coordination number of twelve
Introduction
has been predicted for sulfate binding.6 In an elegant example,
The design of host molecules capable of the selective recognition
of anionic species is an area of intense current interest. In par-
ticular, significant effort has been directed towards the develop-
ment of receptors capable of binding anions solely through
hydrogen bonding interactions, in efforts to mimic natural anion
receptors such as the sulfate and phosphate binding proteins.
Neutral receptors with amide,1 sulfonamide,2 pyrrole,3 indole4
and thio(urea)5 hydrogen bond donors (and combinations
thereof) have successfully been employed in anion recognition.
The thio(urea) class of receptors are particularly attractive
because the two hydrogen bond donors of the thio(urea) can
bind to either a single acceptor in a 6-membered chelate or to
two atoms of a tetrahedral anion (e.g. sulfate) in an 8-membered
chelate.6 Careful positioning of several of these hydrogen bond
donors allows receptors to be designed for specific anions. For
example, calculations by Hay et al.6 indicate that tripodal recep-
tors will exhibit selectivity for tetrahedral anions (e.g. sulfate,
phosphate) over trigonal planar anions (e.g. nitrate).
A number of tripodal (thio)ureas have been reported pre-
viously, with a variety of ‘cores’ including tren,7,8 trisubstituted
benzene derivatives,9,10 cyanuric acid,11 triindane12 and calix[6]
arenes13 employed to provide the tripodal geometry, providing
six hydrogen bond donors to the guest. Many of these have been
found to favour the binding of tetrahedral anions. A logical step
in the design of new tripodal receptors is the inclusion of
additional hydrogen bond donor sites to fully complement target
Jia et al. have recently demonstrated the incorporation of
additional hydrogen bond donors into the ‘arms’ of tripodal
receptors8 but with traditional scaffolds it is difficult to add
additional hydrogen bond donors to the ‘core’ of such tripodal
systems. Indeed, Gunnlaugsson and co-workers have recently
demonstrated that for tripodal benzene derivatives bearing both
amide and urea hydrogen bond donors in the arms, the amide
protons do not participate in hydrogen bonds with anions.10
Backbone modified cyclic hexapeptides based on the lissocli-
namide family of natural products14,15 provide a unique ‘core’
for the synthesis of tripodal receptors, incorporating three hydro-
gen bond donors in the core which are available for guest
binding. We have recently reported the anion binding abilities of
two cryptands with a Lissoclinum type cyclic peptide ‘core’.16,17
We found that these receptors displayed strong affinity for sulfate
anions with a unique recognition motif involving the cyclic
peptide amide hydrogen bond donors. In comparison, spherical
anions such as chloride appeared to bind solely through the
thiourea binding sites, with no interaction from the peptide
hydrogen bond donors. In efforts to further explore the anion
binding ability of the lissoclinamide cyclic peptides, we report
here an investigation of the related tripodal receptors, thiourea 1
and urea 2 (Fig. 1). These tripodal receptors are significantly
more flexible than the related cryptands; it was therefore of inter-
est to examine whether a similar binding motif would be
adopted.
School of Chemistry (F11), The University of Sydney, 2006 NSW,
Australia. E-mail: kate.jolliffe@sydney.edu.au; Fax: +61 2 9351 3329;
Tel: +61 2 9351 2297
Results and discussion
†Electronic supplementary information (ESI) available: 1H and 13C
spectra of 1 and 2; fitted curves for determination of Kas; Job plots;
selected 1H NMR titration spectra for dimerisation and anion binding
experiments. See DOI: 10.1039/c2ob06964d
Synthesis
The synthesis of the tris-(thio)urea receptors is outlined in
Scheme 1. The para-tbutylphenyl substituents were chosen to
2664 | Org. Biomol. Chem., 2012, 10, 2664–2672
This journal is © The Royal Society of Chemistry 2012

View Article Online
provide solubility in organic solvents. Under optimised con-
ditions, (thio)urea 1 was obtained by treating the known tripodal
cyclic peptide scaffold 3a15 with triethylamine and para-tbutyl-
phenyl isothiocyanate in chloroform at reflux. When 3a was
similarly treated with para-tbutylphenyl isocyanate, it was
difficult to separate the desired tris-urea 2 from the triethylam-
monium bromide byproduct, providing 2 in only a modest yield
of 61% after several attempts at column chromatography. There-
respectively, with increasing concentration of 2. In contrast to
the downfield shifts observed for NHa and NHb, all other proton
environments (Hc–Hi) were observed to undergo small to negli-
gible upfield shifts. Throughout the titration the molecule main-
tained C3-symmetry and no additional resonances were
observed, indicating the formation of well-defined aggregates.
Fitting the data to a monomer–dimer equilibrium model (Fig. 3)
was carried out by non-linear least squares regression,21 giving
dimerisation constants (Kdimer) of < 10 M−1 for (thio)urea 1 and
90 19 M−1 for urea 2.
t
fore, the free base tris-amine 3b was treated with butylphenyl
isocyanate in chloroform, in the absence of triethylamine, to
provide 2 in an improved yield of 80%.
Relative to the high dimerisation propensity often observed
with urea-functionalised calixarene systems (up to 109 M−1),19
the Kdimer-value obtained for 2 suggests only a small tendency of
the receptor to dimerise in solution. In analogy with these pre-
viously described systems, the large downfield shift of the urea
Dimerisation studies
Given the propensity of urea and thiourea derivatives to self-
assemble in solution,18–20 we first investigated the dimerisation
behaviour of 1 and 2 by obtaining their 1H NMR spectra in
CDCl3 at 300 K at a range of concentrations between 0.174 and
22.9 mM. During the course of the titration with thiourea 1, NHa
experienced a downfield shift of 0.25 ppm, whereas the change
in chemical shift of NHb was negligible (Δδ = 0.02 ppm). (See
Scheme 1 for proton identification.) The changes in chemical
shifts of all other proton environments of 1 (Hc–Hi) were also
negligible. In contrast, a significant concentration dependence of
the chemical shifts of the urea protons (NHa and NHb) of 2 was
observed (Fig. 2). Signals attributable to both NHa and NHb
underwent substantial downfield shifts from 6.78 to 7.41 ppm
(Δδ = 0.63 ppm) and from 5.36 to 5.81 ppm (Δδ = 0.45 ppm),
Fig. 2 1H NMR (400 MHz, 300 K) spectra of receptor 2 in CDCl3
Fig. 1 Structures of tripodal receptors 1 and 2.
recorded at a range of concentrations.
Scheme 1 Synthesis of tripodal tris-(thio)urea receptors. Reagents and conditions: (i) 3a, para-tbutylphenyl isothiocyanate (for 1), anhydrous
CHCl3, triethylamine, reflux. (ii) 3b, para-tButylphenyl isocyanate (for 2), anhydrous CHCl3, reflux.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2664–2672 | 2665

View Article Online
range to that above, indicating that in this more polar environ-
ment the equilibria can be completely shifted to the monomer.
The difference in the dimerisation behaviours of receptors 1
and 2 can be explained by the differences in the strengths of the
intermolecular hydrogen bonding interactions between the
thiourea and urea moieties. In general, the thiocarbonyl of a
thiourea is a poor hydrogen bond acceptor owing to the lower
electronegativity of sulfur in relation to the oxygen of a urea.20
As such, the oxygen of a urea will more readily accept hydrogen
bonds than the sulfur of a thiourea, thus reflecting the trend
observed with 1 and 2.
Anion binding studies
The anion binding properties of 1 and 2 were quantitatively
1
determined by H NMR spectroscopic titrations with a range of
anions as their tetrabutylammonium salts in CDCl3. As a repre-
1
sentative example, H NMR spectra of receptor 1 recorded over
the course of the titration with a solution of tetrabutylammonium
chloride are shown in Fig. 4. Throughout the titration the mol-
ecule maintained C3-symmetry and fast exchange processes were
observed. Significant downfield shifts of the (thio)urea protons,
NHa and NHb, were observed over the titration range (Δδ = 2.39
and 1.87 ppm, respectively) which was a strong indication that
all six (thio)urea protons were cooperatively involved in hydro-
gen bonding to the anions in solution.
The apparent stability constants (Ka) of receptors 1 and 2 with
a range of anions in CDCl3 are presented in Table 1. In all cases,
except for the titration with receptor 1 and sulfate, fast exchange
processes were observed. For thiourea 1, data were fitted to a
1 : 1 binding model by non-linear curve fitting (see ESI† for
titration curves).21 The 1 : 1 stoichiometry of the complexes was
confirmed by Job plots (see ESI†). In the case of the urea 2, the
dimerisation behaviour (Kdimer = 90 M−1), was incorporated into
the binding model to account for its ability to self-associate in
CDCl3. In contrast to all other systems investigated, when sulfate
ions were added to receptor 1, slow exchange processes were
observed with the appearance of new signals corresponding to a
host–guest complex in the 1H NMR spectra.
Fig. 3 Concentration dependence of the chemical shift of the (thio)
urea protons (NHa and NHb) of (a) 2 and (b) 1 in CDCl3. Solid lines are
the calculated chemical shifts for 1 : 1 dimerisation. The chemical shift
for NHb of 1 could not be fitted to a 1 : 1 dimerisation model.
The data obtained suggests that both 1 and 2 are selective for
sulfate as they both bind the tetrahedral oxoanion extremely
strongly with stability constants (Ka) > 104 M−1. Receptor 1
favours benzoate over acetate, chloride and bromide while 2
favours chloride over bromide, benzoate and nitrate, indicating
that the relative acidities22 of the two receptors influence the
binding selectivities. Both receptors showed selectivity towards
smaller halides, with both 1 and 2 displaying the lowest affinity
for iodide. The experimental data obtained from the titration of
hydrogen sulfate and 2 suggests strong binding, however, the
data did not fit to a suitable binding model which suggests a
proton transfer reaction may be occurring.23 Peak broadening
during the titration of 1 with dihydrogenphosphate prevented an
association constant from being determined. Notably, the data
presented in Table 1 suggests that the receptors have selectivity
for tetrahedral oxoanions, and sulfate in particular.
Scheme 2 Proposed dimerisation mode of tris-urea 2.
protons of 2 implies that they participate in a band of hydrogen
bonds which act to hold two molecules of 2 together
(Scheme 2).19 The relatively small upfield shifts of the other
proton environments (Hc–Hi) are attributable to conformational
changes upon dimerisation.
Additional concentration dependent studies with 2 were
carried out in a more competitive solvent mixture (10% v/v
DMSO-d6/CDCl3). Only negligible changes in chemical shift
(Δδ = 0.02 ppm) were observed over a similar concentration
In all titrations, the addition of anions resulted in restricted
rotation of the receptor arms, as evidenced by changes to the
c
signals attributed to the CH2 environments of 1 and 2 (see
2666 | Org. Biomol. Chem., 2012, 10, 2664–2672
This journal is © The Royal Society of Chemistry 2012

View Article Online
Fig. 4 1H NMR (400 MHz, 300 K) spectra from the titration of 1 with tetrabutylammonium chloride in CDCl3. S: solvent residual. See Scheme 1
for proton assignments (Ha–i).
Table 1 Apparent stability constants (Ka, M−1) of 1 and 2 towards
various anions as determined by monitoring the urea or (thio)urea
proton resonances, NHa and NHb, during 1H NMR titrations in CDCl3
a
Anionb
Receptor 1
Receptor 2
Cl−
Br−
I−
298
104
36
74
486
1920
1210
619
944
919
1170
1500
−
NO3
AcO−
BzO−
579
Scheme 3 Proposed conformational changes to the tripodal receptors
incurred during anion binding. Such changes are expected to lead to
changes in the nature of the hydrogen-bond network of the macrocycle.
−
c
H2PO4
TsO−
HSO4
189
1250
−
d
1700
2−
SO4
>104e,f
>104
a Determined at 300 K. Data was fitted to a 1 : 1 binding model as
confirmed by Job plot titrations. Ka values are an average obtained from
monitoring NHa and NHb. Errors < 15%. b Anions added as their
tetrabutylammonium salts. c Peak broadening prevented an association
constant from being determined. d Titration data suggests strong binding
however it could not be fitted to a suitable binding model. e Titration
displayed slow exchange on the NMR timescale. f K1 value only.
ESI†). Based on these observations, it is hypothesised that co-
operative anion binding by the three (thio)urea moieties of the
receptors results in restricted conformational freedom of the
binding arms (Scheme 3). The three binding arms of the recep-
tors become fixed in place due to the formation of six hydrogen-
bonds from the (thio)urea groups which converge on a central
anion. Notably, larger splittings of signals for these diastereoto-
pic protons were found to correlate with larger Ka values.
The binding titrations for sulfate with each receptor exhibited
distinct behaviour which merits comment. The chemical shift
data for the urea protons (NHa and NHb) of 2 with increasing
equivalents of sulfate is shown in Fig. 5. Addition of 0.2 to 0.6
equivalents of anion caused both proton signals to broaden sig-
nificantly into the baseline of the spectra so that they were
Fig. 5 Change in chemical shift (ppm) of the urea protons, NHa and
NHb, of 2 in CDCl3 with the addition of bis(tetrabutylammonium)
sulfate.
hardly observed (see ESI†). However, upon further addition of
sulfate both urea signals were observed significantly downfield
from their original positions and continued to shift downfield
until one equivalent of anion had been added.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2664–2672 | 2667

View Article Online
Table 2 Maximum changes in the chemical shifts (Δδmax, ppm) of the
thiourea or urea NHa protons and the amide protons (NHe) of 1 and 2
over the course of titrations with a variety of anions in CDCl3
NHa [thiourea]
NHe [amide]
Aniona
Receptor 1
Receptor 2
Receptor 1
Receptor 2
Cl−
Br−
I−
2.385
1.696
0.680
1.364
3.472
3.369
2.173
0.618
1.411
2.565
3.235
2.997
0.804
1.029
3.263
0.076
0.044
0.002
0.012
0.092
0.019
0.285
0.052
0.150
0.765
0.109
0.079
0.049
0.032
0.115
0.050
0.260
0.053
0.219
0.703
−
NO3
AcO−
BzO−
3.169
−
b
H2PO4
TsO−
HSO4
1.149
1.430
3.271
−
2−
SO4
a Anions added as their tetrabutylammonium salts. b Titration resulted in
broadening of the thiourea protons.
Fig. 6 1H NMR (400 MHz, 300 K) spectra from the titration of 1 with
bis(tetrabutylammonium) sulfate in CDCl3. The titration resulted in slow
exchange up to approximately 2 equivalents of sulfate. See Scheme 1
for proton assignments (Ha–i).
The titration of sulfate with receptor 1 resulted in dissimilar
behaviour indicative of slow exchange on the NMR timescale
(Fig. 6). New signals for each proton environment appeared sig-
nificantly downfield from their original positions upon the
addition of anion; these were attributed to the formation of a new
host–guest complex in solution. A steady increase in the popu-
lation of the new signals was observed as the titration progressed
until, after one equivalent of anion had been added, the original
signals could no longer be observed indicating very strong 1 : 1
binding (Ka > 104 M−1). Beyond the addition of two equivalents
of sulfate, the (thio)urea (NHa and NHb) and amide (NHe)
signals broadened to such an extent that they could no longer be
observed. In contrast, the signal attributable to the ornithine α-
protons (CHd) underwent a gradual upfield shift with fast
exchange. This indicated that the receptor was undergoing
further conformational changes in the presence of higher equiva-
lents of anion. These observations suggest that 1 initially forms
strong 1 : 1 complexes at low concentrations of sulfate (i.e. less
than two equivalents of sulfate), which display slow exchange
on the NMR timescale, while at higher concentrations of anion,
more complex binding equilibria exist which result in fast
exchange processes. Similar behaviour has recently been
observed by Gale and co-workers with acyclic indole and carba-
zole-based receptors.23,24
In order to gain further insight into the complex behaviour 1
exhibited at high concentrations of sulfate, and to further investi-
gate the specific binding interactions involved in anion recog-
nition, the chemical shifts of the thiourea or urea protons (NHa)
and the amide protons (NHe) of 1 and 2 were examined in closer
detail. Table 2 summarises the maximum changes in the chemi-
cal shift (Δδmax) of these protons over the course of titrations
carried out with a variety of anions.
For both 1 and 2 the more basic anions, such as acetate, ben-
zoate, dihydrogenphosphate and sulfate, were found to induce
the largest downfield shifts in the thiourea or urea protons
(Table 2). However, the same trend was not observed in the case
Fig. 7 Plot of the change in chemical shift (Δδ, ppm) of the cyclic
peptide amide protons (NHe) of 2 in CDCl3 with the addition of various
anions. ◇ sulfate; ◆ hydrogensulfate; ■ dihydrogenphosphate; ×
bromide; ◯ tosylate; 4 nitrate.
of the amide protons. For instance, for both 1 and 2, the
maximum downfield shift of the amide protons incurred during
the titration with benzoate was small (0.019 and 0.050 ppm,
respectively). Small downfield shifts for the amide protons were
also observed for the titrations performed with chloride,
bromide, iodide, nitrate, acetate and tosylate indicating that the
predominant hydrogen bonding interactions to these guests in
solution were from the (thio)urea or urea binding sites only. In
sharp contrast to these observations, relatively large Δδmax
-
values for the amide protons of both 1 and 2 were observed for
the titrations involving dihydrogenphosphate, hydrogensulfate
and in particular for sulfate indicating that the peptide backbone
is involved with binding to these anions to varying extents.
The titration profiles of the amide protons (NHe) of 2 with a
variety of anions are shown in Fig. 7. Although the magnitude
of the downfield shift was different for each anion, the signal in
each case first underwent a downfield shift with the addition of
anion which was then followed by an upfield shift for the
remainder of the titration. All titrations resulted in similar
profiles, irrespective of the binding strength to the anions. With
the exception of dihydrogenphosphate, the downfield shifts reach
a maximum value upon the addition of approximately one
2668 | Org. Biomol. Chem., 2012, 10, 2664–2672
This journal is © The Royal Society of Chemistry 2012

View Article Online
Table 3 Apparent stability constants (Ka, M−1) of 1 and 2 towards
various anions as determined by monitoring the urea or thiourea proton
resonances, NHa and NHb, during 1H NMR titrations in 10% v/v
a
DMSO-d6/CDCl3
Anionb
Cl−
Receptor 1
Receptor 2
477
89
593
129
521
>104
−
NO3
−
HSO4
289
2−
SO4
>104c,d
a Determined at 300 K. Data was fitted to a 1 : 1 binding model. Ka
values are an average obtained from monitoring NHa and NHb. Error
<15%. b Anions added as their tetrabutylammonium salts. c Titration
displayed slow exchange on the NMR timescale. d K1 value only.
Fig. 8 Proposed binding modes of receptors 1 (X = S) and 2 (X = O)
for (a) sulfate anions involving nine hydrogen bonding interactions and
(b) spherical anions involving six hydrogen bonding interactions.
Notably, with the exception of sulfate, the stability constants
obtained for tris-urea 2 were several times larger than those
found for tris-thiourea 1. In the most extreme case with iodide,
binding with 2 is more than 17 times stronger than with 1. While
thiourea is more acidic than urea (pKa = 21.1 and 26.9, respect-
ively, in DMSO)22 and therefore expected to establish stronger
hydrogen bonding interactions with anions, it has recently been
shown that the conformational preferences of urea and thiourea
groups can be an overriding factor which dictates binding behav-
iour. While ureas frequently adopt the cis–cis conformation
required for binding, thioureas have been found to favour a cis–
trans conformation and require rotational reorganisation to adopt
the required geometry for anion binding.27 In this regard, 2 has
an added degree of pre-organisation compared to 1, and this is
reflected in the lower binding constants observed for 1.
equivalent of anion, which further suggests a 1 : 1 binding stoi-
chiometry. In the case of dihydrogenphosphate, the curve
reached a maximum value with the addition of three equivalents
of anion, however, Job’s plot titrations of the urea protons indi-
cate 1 : 1 binding (see ESI†).
Small changes in chemical shift of the amide hydrogens are
considered to be a result of conformational changes of the scaf-
folds. However, the large changes observed upon sulfate addition
strongly suggest the formation of hydrogen-bond interactions
with this anion.25 Similar shifts previously observed for the
amide proton signals of related cryptand-like receptors upon
binding sulfate ions have been attributed to the formation of
hydrogen bonds between the receptors and this anion.17 Due to a
combination of the hydrogen bonding and conformational
effects, the chemical shift data for the amide proton signals of 1
and 2 could not be fitted to an appropriate binding model, pre-
venting the determination of stability constants using these
protons. Indeed, the titration profiles observed drastically differ
from those reported previously during anion titrations with other
Lissoclinum type receptors.26
The significant and rapid downfield shift of the amide proton
signals of 1 and 2 observed in the titration with sulfate (Table 2,
Fig. 7), in contrast to the smaller shifts observed for all other
anions, strongly suggests that the amide hydrogens contribute
significantly to binding this particular anion. Based on the titra-
tion data obtained, a proposed binding mode for sulfate with
receptors 1 and 2 is shown in Fig. 8a. It is postulated that a total
of nine hydrogen bonding interactions (six from the (thio)urea
groups and three from the amide protons) act to selectively and
strongly hold a central sulfate anion in place. In contrast, the pro-
posed binding mode for spherical anions features only six
binding contributions from the three urea or (thio)urea groups
(Fig. 8b). The observed selectivity of the receptors for sulfate
may be attributed to these additional hydrogen bonding inter-
actions. This is in agreement with the proposed binding modes
we have previously reported for related cryptands.17 The smaller
changes to the amide protons observed in the titrations with
Anion binding in more competitive solvent mixtures
In order to investigate the effect of solvent polarity on the
binding affinities of 1 and 2, further binding studies with
selected anions were carried out in a more competitive medium,
namely 10% v/v DMSO-d6/CDCl3 (Table 3). It was previously
established that 2 does not self-associate in this solvent system;
as such a pre-dimerisation equilibrium was not taken into
account. In all cases, except for titration of 1 with sulfate, fast
exchange processes were observed.
The binding data obtained reveals that moving to this more
polar solvent mixture had a significant influence on the binding
behaviour of receptor 2. For example, the stability constants
obtained for chloride and nitrate with 2 are 3 and 7 times
smaller, respectively, than those values obtained previously in
CDCl3. In addition, in this more polar solvent, the binding of 2
and hydrogen sulfate was simplified with good fits obtained to a
1 : 1 binding model. In contrast, the stability constants found for
the tris-thiourea 1 are not as dramatically affected upon moving
to the more competitive solvent system. In fact, surprisingly the
stability constant obtained for chloride and 1 increased in
moving to this more polar solvent system (Ka = 477 M−1 in 10%
v/v DMSO-d6/CDCl3 versus 298 M−1 in CDCl3). Such behav-
iour suggests that the rotational barriers of the thiourea and urea
binding groups are influenced to different extents by the polarity
of the solvent system employed and that in more polar solvents,
dihydrogenphosphate and hydrogensulfate also suggest
a
binding association. However, these other tetrahedral anions are
presumably not able to adopt a binding orientation like that rep-
resented in Fig. 8a due to their higher protonation states and, as
such, are not bound as strongly.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2664–2672 | 2669

View Article Online
1
the higher acidity of the thiourea protons outweighs the confor-
mational preferences of these groups.28
polarimeter using the indicated spectroscopic grade solvents. H
NMR spectra were recorded using a Bruker Avance DPX 400 at
a frequency of 400.13 MHz or a Bruker Avance DPX 300 at a
frequency of 300.13 MHz and are reported as parts per million
(ppm) with CDCl3 (δH 7.26 ppm) as an internal reference. The
data are reported as chemical shift (δ), multiplicity (br = broad,
s = singlet, d = doublet, t = triplet, m = multiplet), coupling con-
stant (J Hz) and relative integral. 13C NMR spectra were
recorded using a Bruker Avance DPX 400 at a frequency of
100.61 MHz or a Bruker Avance DPX 300 at a frequency of
75.47 MHz and are reported as parts per million (ppm) with
CDCl3 (δC 77.16 ppm) as an internal reference. High resolution
ESI spectra were recorded on a Bruker BioApex Fourier Trans-
form Ion Cyclotron Resonance mass spectrometer (FTICR) with
an Analytica ESI source, operating at 4.7 T or a Bruker Daltonics
Apex Ultra FTICR with an Apollo Dual source, operating at 7
T. Analytical TLC was performed using precoated silica gel
plates (Merck Kieselgel 60 F254). Preparative column chromato-
graphy was carried out using either Merck Kieselgel 60 silica
gel (SiO2; 0.040–0.063 mm) or Davisil® Chromatographic Silica
Media (SiO2; 0.040–0.063 mm), with the indicated solvents
which were mixed v/v as specified. Triethylamine was distilled
from calcium hydride prior to use. Chloroform was distilled
from calcium chloride and passed through a column of basic
alumina prior to use. Tetrabutylammonium salts were used as
supplied and were stored in a vacuum desiccator over silica
drying beads and phosphorous pentoxide. Compound 3 was syn-
thesised according to our previously reported procedures.15,16
All other reagents were commercially available and were used as
supplied.
Both 1 and 2 bind sulfate selectively and strongly in this more
competitive medium (Table 3). The results clearly reflect the
importance of the highly complementary binding network on
binding affinity and selectivity of this tripodal system. As with
the data obtained in CDCl3, titrations with sulfate and tris-
thiourea 1 resulted in slow exchange on the NMR timescale and
at high concentrations of anion the NMR data suggested that
more complex equilibria existed. The titration with tris-urea 2,
on the other hand, resulted in fast exchange processes and fitted
well to a 1 : 1 binding model.
In comparison to the cryptand-like thiourea systems pre-
viously reported,16,17 1 and 2 show similar selectivity for sulfate
ions over all other anions tested, attributable to a binding mode
involving the amide hydrogen bond donors of the cyclic peptide
scaffold. Notably, binding affinity for other anions (in particular
for chloride) is much lower in these more flexible tripodal
systems providing greater selectivity for sulfate. However, direct
comparisons of the binding constants of the tripodal and cryp-
tand-like systems can not be made as measurements were made
in different solvents for solubility reasons.
Conclusions
Quantitative binding studies carried out with the tripodal recep-
tors indicate that both 1 and 2 bind sulfate ions strongly and
selectively in CDCl3 and 10% v/v DMSO-d6/CDCl3 (Ka values
> 104 M−1 in both solvents). Notably, in CDCl3 the tris-urea 2
exhibited higher association constants for all anions than the tris-
thiourea 1. This can be attributed to the relative reorganisational
energies required for the thiourea and urea groups of 1 and 2,
respectively, to adopt the cis–cis conformation necessary for
anion binding. In the more polar solvent, 10% v/v DMSO-d6/
CDCl3, both receptors bound anions with similar affinities
suggesting that the higher acidity of the thiourea protons out-
weighs the reorganisational energies in more polar media.
The observed selectivity for sulfate is attributed to a network
of hydrogen bonds which is highly complementary for the tetra-
hedral oxoanion. 1H NMR evidence indicates that sulfate accepts
a total of nine hydrogen bonds, six from the (thio)urea groups
and three from the amide NH groups of the peptide backbone,
while spherical anions, e.g. chloride, only form six hydrogen
bonds to the (thio)urea groups. Presumably, the selectivity for
sulfate over the other tetrahedral anions examined (hydrogensul-
fate and dihydrogenphosphate) is due to the inability of these
anions to adopt a similar orientation to sulfate due to their higher
protonation states. In this regard, the selectivity mechanism of
the receptors mimics that of biological systems. We are currently
extending our investigations of the anion binding affinities and
selectivities of these and structurally similar receptors.
Receptor 1
Under an atmosphere of nitrogen, anhydrous CHCl3 (6.5 mL)
was added to tris-hydrobromide salt 3a (261 mg, 0.315 mmol)
to provide a suspension. Triethylamine (0.141 mL, 1.01 mmol)
was then added and the resulting mixture was allowed to reflux
for 30 min followed by cooling to room temperature. para-tButyl-
isothiocyanate (193 mg, 1.01 mmol) was then added and the
reaction mixture was returned to reflux for a further 19 h. The
mixture was cooled to room temperature, diluted with CHCl3
(10 mL), quenched with 0.5 M HCl (5 mL) and the layers separ-
ated. Further extraction of the aqueous layer with CHCl3 (2 ×
20 mL) was followed by washing the combined organic layers
with water (15 mL) and saturated brine (15 mL). The organic
layers were then dried (MgSO4) and concentrated under reduced
pressure to give the crude product as a brown oil. Subsequent
purification by flash chromatography (silica gel; EtOAc:hexanes
[1 : 4] then CH2Cl2:MeOH [20 : 1]) gave the desired tris-(thio)
urea 1 as an off-white solid (325 mg, 89%). m.p. 160–165 °C;
[α]2D0 −33.6 (c 0.4, CHCl3); 1H NMR (300 MHz, CDCl3) δ
(ppm) 8.33 (d, J = 6.5 Hz, 3 H), 8.12 (br s, 3 H), 7.44–7.31 (m,
6 H), 7.17–7.06 (m, 6 H), 6.27 (br s, 3 H), 5.16–5.03 (m, 3 H),
3.75–3.51 (m, 6 H), 2.57 (s, 9 H), 2.20–2.02 (m, 3 H),
2.00–1.54 (m, 9 H), 1.27 (s, 9 H); 13C NMR (75 MHz, CDCl3)
δ (ppm) 180.7, 161.1, 160.7, 154.1, 150.3, 133.6, 128.4, 127.0,
125.1, 47.6, 44.8, 34.7, 32.1, 31.3, 24.5, 11.7; HRMS (ESI) m/z
calcd for C60H79N12O6S3 [M + H]+: 1159.5408, found:
1159.5433.
Experimental
General remarks
Melting points were measured using a Stanford Research
Systems Optimelt apparatus and are uncorrected. Optical
rotations were performed using a Perkin Elmer Model 341
2670 | Org. Biomol. Chem., 2012, 10, 2664–2672
This journal is © The Royal Society of Chemistry 2012

View Article Online
Receptor 2
Job’s plot titrations
Under an atmosphere of nitrogen, para-tbutylisocyanate
(48.6 μL, 0.273 mmol) was added to a solution of tris-amine 3b
(50.0 mg, 85.4 μmol) in anhydrous CHCl3 (5 mL). The reaction
mixture was refluxed for 15 h after which time it was concen-
trated under reduced pressure to give a pale yellow oil. Sub-
sequent purification by flash chromatography (silica gel;
CH2Cl2:MeOH [12 : 1]) provided the desired tris-urea 2 as an
off-white solid (74.5 mg, 80%). m.p. 186 °C (decomp.); [α]D20
Separate stock solutions of receptor (2.00 mM) and an anion
guest (2.00 mM) were prepared in CDCl3 using volumetric
flasks. H NMR spectra were recorded for eight different sol-
1
utions containing a total volume of 500 μL in the following
receptor : anion ratios: 500 : 0, 450 : 50, 375 : 125, 325 : 175,
250 : 250, 175 : 325, 125 : 375 and 50 : 450. The molar fraction
of the receptor (XR) was then plotted as a function of Δδ × XR,
where Δδ = δobs − δfree (δobs is the observed chemical shift and
δfree is the chemical shift of the free receptor). In each case the
chemical shifts of both the (thio)urea protons, NHa and NHb,
were examined.
1
−12.4 (c 0.2, CHCl3); H NMR (400 MHz, CDCl3) δ (ppm)
8.33 (br s, 3 H), 7.59 (br s, 3 H), 7.23–6.99 (m, 12 H), 5.97 (br
s, 3 H), 5.12 (br s, 3 H), 3.25–2.97 (m, 6 H), 2.55 (s, 9 H),
2.15–1.99 (m, 3 H), 1.97–1.78 (m, 3 H), 1.69–1.51 (m, 3 H),
1.49–1.33 (m, 3 H), 1.21 (s, 9 H); 13C NMR (100 MHz, CDCl3)
δ (ppm) 160.9, 156.9, 154.1, 146.0, 136.3, 128.4, 125.9, 120.1,
47.8, 39.6, 34.3, 32.0, 31.5, 25.4, 11.7 (one quaternary carbon
signal obscured or overlapping); HRMS (ESI) m/z calcd for
C60H79N12O9 [M + H]+: 1111.6093, found: 1111.6072.
Acknowledgements
We thank Dr Ian Luck (USyd) for technical assistance with
NMR and the Australian Research Council (DP0877726) for
financial support.
Determination of dimerisation constants (Kdimer, M−1) by
References
1H NMR spectroscopic titrations
1 For reviews see: J. L. Sessler, P. A. Gale and W.-S. Cho, Anion receptor
chemistry, Royal Society of Chemistry, Cambridge, 2006, ch. 4;
S. Kubik, Chem. Soc. Rev., 2009, 38, 585; S. O. Kang, R. A. Begum and
K. Bowman-James, Angew. Chem., Int. Ed., 2006, 45, 7882; C. R. Bondy
and S. J. Loeb, Coord. Chem. Rev., 2003, 240, 77.
2 For recent examples see: Á. L. Fuentes de Arriba, M. G. Turiel,
L. Simón, F. Sanz, J. F. Boyero, F. M. Muñiz, J. R. Morán and
V. Alcázar, Org. Biomol. Chem., 2011, 9, 8321; V. Amendola,
L. Fabbrizzi, L. Mosca and F.-P. Schmidtchen, Chem.–Eur. J., 2011, 17,
5972; M. V. López, M. R. Bermejo, M. E. Vásquez, A. Taglietti,
G. Zaragoza, R. Pedrido and M. Martínez-Calvo, Org. Biomol. Chem.,
2010, 8, 357; C. Caltagirone, G. W. Bates, P. A. Gale and M. A. Light,
Chem. Commun., 2008, 44, 61; O. B. Berryman, C. A. Johnson II, L.
N. Zakharov, M. M. Haley and D. W. Johnson, Angew. Chem., Int. Ed.,
2008, 47, 117.
In a typical dimerisation experiment a solution of receptor 1 or 2
was prepared in the stated deuterated solvents (v/v) of known
concentration (of the order of 40 mM) using a volumetric flask.
In each case, aliquots of these stock solutions were added to
500 μL of the same deuterated solvent in an NMR tube. After
each addition the solution was thoroughly mixed and the 1H
NMR (400 MHz, 300 K) spectrum was recorded. Titrations were
performed in duplicate. Non-linear curve fitting21 of the exper-
imentally obtained titration isotherms (concentration of receptor
versus chemical shift of (thio)urea NH protons) enabled the cal-
culation of dimerisation constants (Kdimer, M−1).
3 For a review see: J. L. Sessler, P. A. Gale and W.-S. Cho, Anion receptor
chemistry, Royal Society of Chemistry, Cambridge, 2006, ch. 5.
4 For a review see: P. A. Gale, Chem. Commun., 2008, 4525.
5 For reviews see: A.-F. Li, J.-H. Wang, F. Wang and Y.-B. Jiang, Chem.
Soc. Rev., 2010, 39, 3729; V. Amendola, L. Fabbrizzi and L. Mosca,
Chem. Soc. Rev., 2010, 39, 3889.
Determination of association constants (Ka, M−1) for various
anions by 1H NMR spectroscopic titrations
6 B. P. Hay, T. K. Firman and B. A. Moyer, J. Am. Chem. Soc., 2005, 127,
1810.
In a typical anion titration experiment a 2–5 mM stock solution
of receptor was accurately prepared in the stated deuterated sol-
vents (v/v) using a volumetric flask. Solutions of anions (as their
tetrabutylammonium salts) to be titrated were then prepared in
separate volumetric flasks using the same host solution so that
the concentration of the host remained constant throughout a
given titration experiment. The concentration of anion solutions
were made 50 times that of the host (i.e. 0.1–0.25 M). In each
case, 500 μL of host solution in an NMR tube was titrated with
7 R. Custelcean, B. A. Moyer and B. P. Hay, Chem. Commun., 2005, 5971;
R. Custelcean, P. Remy, P. V. Bonnesen, D.-E. Jiang and B. A. Moyer,
Angew. Chem., Int. Ed., 2008, 47, 1866; B. Wu, J. Liang, J. Yang, C. Jia,
X.-J. Yang, H. Zhang, N. Tang and C. Janiak, Chem. Commun., 2008,
1762; Y. Li, K. M. Mullen, T. D. W. Claridge, P. J. Costa, V. Felix and P.
D. Beer, Chem. Commun., 2009, 7134; I. Ravikumar, P.
S. Lakshminarayanan, M. Arunachalam, E. Suresh and P. Ghosh, Dalton
Trans., 2009, 4160; R. Custelcean, A. Bock and B. A. Moyer, J. Am.
Chem. Soc., 2010, 132, 7177; N. Busschaert, P. A. Gale, C. J. E. Haynes,
M. E. Light, S. J. Moore, C. C. Tong, J. T. Davis and W. A. Harrell Jr.,
Chem. Commun., 2010, 46, 6252; S. K. Dey and G. Das, Dalton Trans.,
2011, 40, 12048; M. Li, Y. Hao, B. Wu, C. Jia, X. Huang and X.-J. Yang,
Org. Biomol. Chem., 2011, 9, 5637; N. Busschaert, M. Wenzel, M.
E. Light, P. Iglesias-Hernández, R. Pérez-Tomás and P. A. Gale, J. Am.
Chem. Soc., 2011, 133, 14136; Y. Hao, C. Jia, S. Li, X. Huang,
X.-J. Yang, C. Janiak and B. Wu, Supramol. Chem., DOI: 10.1081/
10610278.2011.622389.
1
aliquots of anion stock solution and after each addition the H
NMR (400 MHz, 300 K) spectrum was recorded after thorough
mixing. Typically this was performed in the following order:
10 × 2 μL, 3 × 10 μL, 20 μL then 30 μL (total 100 μL). Titrations
were performed in duplicate. Non-linear curve fitting21 of the
experimentally obtained titration isotherms (equivalents of anion
versus chemical shift of the (thio)urea NH protons) enabled the
calculation of association constants (Ka, M−1). In all cases com-
plete dissociation of the tetrabutylammonium salts was assumed
and the data was fitted to a 1 : 1 binding model as confirmed by
Job’s plot titrations (see below).
8 C. Jia, B. Wu, S. Li, X. Huang, Q. Zhao, Q.-S. Li and X.-J. Yang,
Angew. Chem., Int. Ed., 2011, 50, 486.
9 H. R. Seong, D.-S. Kim, S.-G. Kim, H.-J. Choi and K. H. Ahn, Tetrahe-
dron Lett., 2004, 45, 723; I. Hisaki, S. Sasaki, K. Hirose and Y. Tobe,
Eur. J. Org. Chem., 2007, 2007, 607; S. Sasaki, D. Citterio, S. Ozawa
and K. Suzuki, J. Chem. Soc., Perkin Trans. 2, 2001, 2309; D. R. Turner,
M. J. Paterson and J. W. Steed, J. Org. Chem., 2006, 71, 1598.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2664–2672 | 2671

View Article Online
10 C. M. Gomes dos Santos, E. M. Boyle, S. De Solis, P. E. Kruger and
T. Gunnlaugsson, Chem. Commun., 2011, 47, 12176.
11 I. Ravikumar and P. Ghosh, Chem. Commun., 2010, 46, 6741.
12 H.-J. Choi, Y. S. Park, S. H. Yun, H.-S. Kim, C. S. Cho, K. Ko and K.
H. Ahn, Org. Lett., 2002, 4, 795.
12403; P. Ballester and G. Gil-Ramirez, Proc. Natl. Acad. Sci. U. S. A.,
2009, 106, 10455; M. Alajarín, A. Pastor, R.-A. Orenes and J. W. Steed,
J. Org. Chem., 2002, 67, 7091.
20 A. Asadi, D. Ajami and J. Rebek Jr., J. Am. Chem. Soc., 2011, 133,
10682.
13 M. Hamon, M. Ménand, S. Le Gac, M. Luhmer, V. Dalla and I. Jabin, J.
Org. Chem., 2008, 73, 7067; J. Scheerder, J. F. J. Engbersen, A. Casnati,
R. Ungaro and D. N. Reinhoudt, J. Org. Chem., 1995, 60, 6448.
14 D. Mink, S. Mecozzi and J. Rebek Jr., Tetrahedron Lett., 1998, 39, 5709;
A. J. Lucke, J. D. A. Tyndall, Y. Singh and D. P. Fairlie, J. Mol. Graphics
Modell., 2003, 21, 341; Y. Singh, M. J. Stoermer, A. J. Lucke, T. Guthrie
and D. P. Fairlie, J. Am. Chem. Soc., 2005, 127, 6563; K. A. Jolliffe,
Supramol. Chem., 2005, 17, 81; A. Pintér, G. Haberhauer, I. Hyla-
Kryspin and S. Grimme, Chem. Commun., 2007, 3711; A. Pintér and
G. Haberhauer, Synlett, 2009, 3082.
21 All curve fitting was performed using the Equilibria program,
C. E. Marjo, University of New South Wales Analytical Centre, Sydney,
Australia, http://www.sseau.unsw.edu.au/Index.htm, 2009.
22 F. G. Bordwell, Acc. Chem. Res., 1988, 21, 456.
23 P. A. Gale, J. R. Hiscock, S. J. Moore, C. Caltagirone, M. B. Hursthouse
and M. E. Light, Chem.–Asian J., 2010, 5, 555.
24 P. A. Gale, J. R. Hiscock, C. Z. Jie, M. B. Hursthouse and M. E. Light,
Chem. Sci., 2010, 1, 215.
25 G. Wagner, A. Pardi and K. Wüthrich, J. Am. Chem. Soc., 1983, 105,
5948.
15 R. J. G. Black, V. J. Dungan, R. Y. T. Li, P. G. Young and K. A. Jolliffe,
Synlett, 2010, 551.
16 P. G. Young, J. K. Clegg, M. Bhadbhade and K. A. Jolliffe, Chem.
Commun., 2011, 47, 463–465.
17 P. G. Young, J. K. Clegg and K. A. Jolliffe, Supramol. Chem., DOI:
10.1081/10610278.2011.622388.
18 M. George, G. Tan, V. T. John and R. G. Weiss, Chem.–Eur. J., 2005, 11,
3243 and references therein.
26 M. Schnopp, S. Ernst and G. Haberhauer, Eur. J. Org. Chem., 2009, 213;
L. Molina, E. Moreno-Clavijo, A. J. Moreno-Vargas, A. T. Carmona and
I. Robina, Eur. J. Org. Chem., 2010, 4049.
27 J. R. Hiscock, P. A. Gale, C. Caltagirone, M. B. Hursthouse and M.
E. Light, Supramol. Chem., 2010, 22, 647; V. S. Bryantsev and B. P. Hay,
J. Phys. Chem. A, 2006, 110, 4678–4688; B. P. Hay, Chem. Soc. Rev.,
2010, 39, 3700; V. S. Bryantsev, T. K. Firman and B. P. Hay, J. Phys.
Chem. A, 2005, 109, 832.
19 V. Rudzevich, Y. Rudzevich and V. Boehmer, Synlett, 2009, 1887; K.
D. Shimizu and J. Rebek Jr., Proc. Natl. Acad. Sci. U. S. A., 1995, 92,
28 K. A. Haushalter, J. Lau and J. D. Roberts, J. Am. Chem. Soc., 1996,
118, 8891; J. B. Gerken and S. Ross, J. Phys. Org. Chem., 2010, 23, 806.
2672 | Org. Biomol. Chem., 2012, 10, 2664–2672
This journal is © The Royal Society of Chemistry 2012