Thiourea isosteres as anion receptors and transmembrane transporters

Article abstract of DOI:10.1039/c1cc12439k

Compounds containing cyanoguanidine and 3-amino-1,2,4-benzothiadiazine-1,1- dioxide have been studied as anion receptors and transporters. Significant affinity for oxo-anions was observed in organic solution and the receptors were found to function as tra

Full text of DOI:10.1039/c1cc12439k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2011, 47, 7641–7643
www.rsc.org/chemcomm
COMMUNICATION
Thiourea isosteres as anion receptors and transmembrane transportersw
Marco Wenzel,a Mark E. Light,a Anthony P. Davisb and Philip A. Gale*a
Received 26th April 2011, Accepted 24th May 2011
DOI: 10.1039/c1cc12439k
Compounds containing cyanoguanidine and 3-amino-1,2,4-
benzothiadiazine-1,1-dioxide have been studied as anion receptors
and transporters. Significant affinity for oxo-anions was observed in
organic solution and the receptors were found to function as
transmembrane chloride/nitrate antiporters with transport rates
enhanced in the presence of valinomycin–K+ complex.
potential anion receptors 1–3 containing these groups and
study their lipid bilayer transport properties compared to
simple thiourea 4.
The receptors 1–4 were synthesised by modifications to
literature procedures (see ESIw).4,11,12 Crystal structures of
the both cyanoguanidine receptors 1 and 2 show the compounds
adopt a trans orientation of the two hydrogen atoms of the
guanidine subunit in the solid state (Fig. 1a and Fig. 1b).z
Strong intermolecular alternating N–Hꢀ ꢀ ꢀN hydrogen bonds
(NꢀꢀꢀN distances of 2.932(4) and 2.950(4) A in 1 and 2.857(3) A
in 2, respectively, Table S1 ESIw) between the cyanoguanidine
CRN group and the cis-orientated N–H group of two ligands
support the present arrangement. In addition weaker
N–Hꢀ ꢀ ꢀN hydrogen bonds (Nꢀ ꢀ ꢀN distance of 3.122(3) and
3.127(3) A in 1 and 3.013(3) A in 2, respectively) are formed
between the cyanoguanidine CRN– group and the trans-
orientated N–H group to link the dimer. This results in 2-D
network of hydrogen bonds in the crystal. In contradistinction
to these results, the cis-orientation is observed for the thiourea
group in the crystal structure of 4 (Fig. S1 ESIw).
The transport of anions across lipid bilayers is vital to a
number of important biological processes. Dysregulation of
passive chloride transport causes a number of diseases including
cystic fibrosis. There is therefore intense current interest in
developing synthetic transmembrane transporters to act as
potential future therapeutic substitutes for naturally occurring
ion channels (and as tools for the study of anion transport
processes in cells).1 Recent work within our group has high-
lighted the anion transport capabilities of a variety of molecules2
including a series of simple thioureas.3,4 However, some
thiocarbamates and thioureas have been found to be toxic,5
with the mechanism of toxicity not fully understood.6 Non-
classical isosteres such as substituted guanidine ligands 1 and 2,
offer a potentially non-toxic alternative to thioureas. In particular
cyanoguanidine has comparable physical properties to thiourea
(e.g. pKa: 15 (thiourea) and 14 (cyanoguanidine); hydrophobic
partition P: 0.09 (thiourea) and 0.07 (cyanoguanidine); C–N
bond lengths: 1.34 A both thiourea and cyanoguanidine;
N–C–N angle: 1191 (thiourea) and 1241 (cyanoguanidine),
respectively) and is present in the drug cimetidine.7 However,
N,N0,N0 0-trisubstituted guanidine groups often adopt a trans
orientation of the two hydrogen atoms of the guanidine
subunit and hence are not preorganised to bind anionic guests
(Fig. 1).8 Alternative conformationally locked systems containing
parallel hydrogen bond donors based upon 3-amino-1,2,4-
benzothiadiazine-1,1-dioxide (e.g. 3) have been synthesised
previously9 but have not yet applied to anion recognition.
As both cyanoguanidines and derivatives of 3-amino-1,2,4-
benzothiadiazine-1,1-dioxide are known to be compatible
with biological systems7,9–11 we decided to synthesise some
Proton NMR titration techniques were used to measure the
stability of the receptor : anion complexes. Job plot analyses
a Chemistry, University of Southampton, Southampton, SO17 1BJ,
UK. E-mail: philip.gale@soton.ac.uk; Fax: +44 (0)23 8059 6805;
Tel: + 44 (0)23 8059 3332
Fig. 1 Schematic draw of the crystal structure (a) of 1 and (b) of 2,
selected atom labels, dotted lines represent hydrogen bonds, non-
interacting hydrogen atoms omitted for clarity, for 2 major occupied
species only, Symmetry codes: i = 1 + x,y,z; ii = 1/2 + x,3/2 ꢁ y,ꢁz;
iii = ꢁ1/2 + x,1/2 ꢁ y,ꢁz; iv = 2 ꢁ x,1 ꢁ y,1 ꢁ z; v = ꢁ2 ꢁ x,ꢁ1/
2 + y,1/2 ꢁ z; vi = 2 ꢁ x,ꢁ1/2 + y,1/2 ꢁ z.
b School of Chemistry, University of Bristol, Cantock’s Close, Bristol,
BS8 1TS, UK. E-mail: anthony.davis@bristol.ac.uk
w Electronic supplementary information (ESI) available: Ligand
synthesis, additional crystallography, transport and NMR studies.
CCDC 821666–821668. For ESI and crystallographic data in CIF or
other electronic format see DOI: 10.1039/c1cc12439k
c
This journal is The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 7641–7643 7641

View Article Online
Table 1 Stability constants log K of 1 : 1 complexes in DMSO-d6/
H2O 0.5%, 25 1C. Errors are estimated to be o15%
series of unilamellar 1-palmitoyl-2-oleoylphosphatidylcholine
(POPC) vesicles loaded with KCl (489 mM) and suspended
them in an external KNO3 (489 mM) solution using previously
reported procedures.6,7 A sample of receptor 1–4 (2% molar
carrier to lipid) was added as a DMSO solution and the
resultant Clꢁ efflux monitored using a chloride selective
electrode. After 300 s, the vesicles were lysed by addition of
detergent and the final reading of the electrode was used to
calibrate 100% release of chloride. The results are shown in
Fig. 3 and reveal chloride transport of compounds 1 and 3
(both 5% after 270 s). The pentafluorophenyl substituted
receptor 2 was found to be more active and approaching
29% release of chloride over the timescale of the experiment.
This is comparable to the transport ability of the thiourea
receptor 4. The latter shows a chloride efflux of approximately
30%, which is similar to the activity of a previously reported
phenyl thiourea receptor.4
Anion
1
2
3
4
2ꢁ
SO4
H2PO4
Clꢁ
3.59
2.37
0.56a
2.03
2.51
44
—
44
3.61a
2.42
—
2.32
2.64
ꢁ
b
3.74
1.56
3.70
3.91
c
c
—
—
PhCOOꢁ
b
CH3COOꢁ
3.23
a
b
c
Error 20%. Broadened NMR resonances. Small shifts—could
not be fitted to a 1 : 1 or 1 : 2 binding model.
indicate a 1 : 1 binding mode for all the receptors (see ESIw).
Stability constants for the complexes were determined using
the EQNMR computer program.13 Proton NMR titrations
were conducted in DMSO-d6/0.5% water solutions with tetra-
butylammonium anion salts. The results show that in all cases
the receptors bind sulfate strongly with log Ka of 3.59 and 3.62
for receptor 1 and 4 and log Ka, 44 for receptors 2 and 3.
Stability constants with chloride, dihydrogen phosphate,
benzoate and acetate were also determined (Table 1). Addition
of tetrabutylammonium nitrate under these conditions did not
cause a shift of the receptors’ proton resonances.
There are a number of potential mechanisms that could be
responsible for releasing chloride under the conditions of the
experiment. We repeated the transport experiments with sodium
rather than potassium salts and found no change in the
chloride transport rate (evidence against a K+/Clꢁ co-transport
mechanism). The transport experiments were repeated under
similar conditions but with sulfate as the extravesicular anion.
Under these conditions no chloride transport was observed.
This is evidence that leads us to suggest that Clꢁ/NO3ꢁ antiport is
the predominant mechanism responsible for the release of
chloride from the vesicles. Further studies in POPC/cholesterol
70 : 30 vesicles showed a reduction in chloride transport rate
consistent with the compounds 1–4 functioning as discrete
molecular carriers and not as channels (Fig. S2–S5 ESIw).
We then repeated these experiments in the presence of the
potassium ionophore valinomycin. The same experimental
conditions were used, namely KCl (489 mM) loaded vesicles
suspended in an external KNO3 (489 mM) solution. The
results are shown in Fig. 4 and reveal an enhancement
Detailed examinations of the titration spectra of receptor 1
show a significant difference in the NH-proton shift upon the
titration with sulfate and chloride (Fig. 2). The stack plot
shows the simultaneous shift of the NH resonances by 3.3 and
3.5 ppm upon the addition of sulfate. This is evidence that
leads us to suggest that the receptor is conformationally
flexible in solution and in the presence of sulfate both NH
groups are oriented towards the anion. On the other hand, the
weaker binding of chloride results in a significant shift of only
one proton signal by 0.3 ppm in the presence of excess
chloride, evidence to suggest that this anion interacts with
only a single NH group in this case.
In order to study the chloride transport properties of
compounds 1–3 and the control compound 4 we prepared a
Fig. 3 Chloride efflux promoted by 0.02 molar equiv. of receptors
1–4 from unilamellar POPC vesicles loaded with 489 mM KCl
buffered to pH 7.2 with sodium phosphate salts. The vesicles were
dispersed in 489 mM KNO3 buffered to pH 7.2 with 5 mM sodium
phosphate salts. At the end of the experiment, detergent was added to
lyse the vesicles and calibrate the ISE to 100% chloride release. Each
point represents the average of three trials.
Fig. 2 NMR titrations of compound 1 with (a) TBA2ꢀSO4 (top) and
(b) TBAꢀCl (bottom) in DMSO-d6/0.5% water.
c
7642 Chem. Commun., 2011, 47, 7641–7643
This journal is The Royal Society of Chemistry 2011

View Article Online
activity which is similar to thiourea derivative 4 whilst the
transport properties of all the compounds are enhanced in the
presence of valinomycin-K+. These initial studies with simple
systems suggest that these hydrogen-bonding motifs may be
optimised for anion complexation and transport. We are
currently working to incorporate these new anion transport
hydrogen bonding motifs into more efficient transporters for
chloride and other anions.
We thank the EPSRC for funding and for use of the
crystallographic facilities at the University of Southampton.
Notes and references
z Crystal data for 1: C14H20N4,
M = 244.34, orthorhombic,
a = 10.2071(2) A, b = 11.2105(3) A, c = 24.2602(6) A, V =
2776.01(11) A3, T = 120 K, P212121 (no. 19), Z = 8, 27 023 reflections
measured, 3585 unique (Rint = 0.0775) of which 3585 were used in the
calculations, R1 = 0.0546 (2806 with F 4 2s(F)), for this light atom
structure Friedel opposites were merged and delta f’’ value for all
elements set to zero.
Fig. 4 Chloride efflux promoted by 0.02 molar equiv. of receptors
1–4 in the presence and absence of valinomycin (0.02 molar equiv.)
from unilamellar POPC vesicles loaded with 489 mM KCl buffered to
pH 7.2 with sodium phosphate salts. The vesicles were dispersed in
489 mM KNO3 buffered to pH 7.2 with 5 mM sodium phosphate salts.
At the end of the experiment, detergent was added to lyse the vesicles
and calibrate the ISE to 100% chloride release. Each point represents
the average of three trials.
2: Cystal data for C14H15N4F5,
M = 334.30, monoclinic,
a = 13.6862(9) A, b = 10.6669(7) A, c = 11.3955(7) A, b =
109.57(4), V = 1567.52(17) A3, T = 120 K, P21/c (no. 14), Z = 4,
25 328 reflections measured, 3614 unique (Rint = 0.0916) of which
3614 were used in the calculations, R1 = 0.0706 (2773 with
F 4 2s(F)), carbon 12 and 13 of the alkyl chain in 2 are disordered
over 2 positions with an occupancy of 0.5 each.
1 P. A. Gale, Acc. Chem. Res., 2011, 44, 216–226; C. J. E. Haynes
and P. A. Gale, Chem. Commun., 2011, DOI: 10.1039/c1cc12061a;
A. P. Davis, D. N. Sheppard and B. D. Smith, Chem. Soc. Rev.,
2007, 36, 348–357; J. T. Davis, O. Okunola and R. Quesada, Chem.
Soc. Rev., 2010, 39, 3843–3862.
2 J. T. Davis, P. A. Gale, O. A. Okunola, P. Prados, J. C. Iglesias-
Sanchez, T. Torroba and R. Quesada, Nat. Chem., 2009, 1,
138–144; P. A. Gale, C. C. Tong, C. J. E. Haynes, O. Adeosun,
D. E. Gross, E. Karnas, E. M. Sedenberg, R. Quesada and
J. L. Sessler, J. Am. Chem. Soc., 2010, 132, 3240–3241.
3 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, 6252–6254.
4 N. J. Andrews, C. J. E. Haynes, M. E. Light, S. J. Moore,
C. C. Tong, J. T. Davis, W. A. Harrell Jr. and P. A. Gale, Chem.
Sci., 2011, 2, 256–260.
5 P. B. Smith and C. Crespi, Biochem. Pharmacol., 2002, 63,
1941–1948; R. C. A. Onderwater, J. N. M. Commandeur,
E. J. Groot, A. Sitters, W. M. P. B. Menge and N. P. E.
Vermeulen, Toxicology, 1998, 125, 117–129.
6 S. A. Svarovsky, R. H. Simoyi and S. V. Makarov, J. Phys. Chem. B,
2001, 105, 12634–12643.
7 G. J. Durant, J. C. Emmett, C. R. Ganellin, P. D. Miles,
M. E. Parsons, H. D. Prain and G. R. White, J. Med. Chem.,
1977, 20, 901–906.
8 C. G. McCarty, R. V. Chastain and D. M. Wieland, J. Chem. Soc. D,
1971, 198–199.
9 S. Boverie, M.-H. Antoine, F. Somers, B. Becker, S. Sebille,
R. Ouedraogo, S. Counerotte, B. Pirotte, P. Lebrun and P. de
Tullio, J. Med. Chem., 2005, 48, 3492–3503.
in chloride transport when both valinomycin and one of the
receptors 1–4 is present. For example a total chloride efflux of
58% was observed for 2 and valinomycin present together
compared to 29% and 5% in the presence of either receptor
alone in the previous experiment. A similar result was obtained
with compound 4, namely an enhancement from 30% in
absence of valinomycin to 56% in the present of both, 4 and
valinomycin. The chloride transport for 1 and 3 is increased
from 5% to 23% if valinomycin is added in an equimolar
amount. The chloride efflux for the receptors 1–4 at 270 s in
the absence and presence of valinomycin are given in Table 2.
As no potassium gradient is present, the enhancement observed
may be caused by the presence of valinomycin bound potassium
cations in the membrane possibly either enhancing extraction
of the anions from the aqueous phase into the membrane or
stabilizing the receptor bound anion complexes in the lipid
bilayer.
This work has demonstrated that thiourea isosteres such as
functionalised cyanoguanidines and the hexyl derivative of
3-amino-1,2,4-benzothiadiazine-1,1-dioxide are capable of
binding and transporting anions. Compound 2 shows transport
10 P. Lebrun, B. Becker, N. Morel, P. Ghisdal, M.-H. Antoine, P. de
Tullio and B. Pirotte, Biochem. Pharmacol., 2008, 75, 468–475;
P. de Tullio, L. Dupont, P. Francotte, S. Counerotte, P. Lebrun and
B. Pirotte, J. Med. Chem., 2006, 49, 6779–6788; B. Pirotte, P. de Tullio,
Q.-A. Nguyen, F. Somers, P. Fraikin, X. Florence, P. Wahl,
J. B. Hansen and P. Lebrun, J. Med. Chem., 2010, 53, 147–154.
11 P. de Tullio, S. Boverie, B. Becker, M.-H. Antoine, Q.-A. Nguyen,
P. Francotte, S. Counerotte, S. Sebille, B. Pirotte and P. Lebrun,
J. Med. Chem., 2005, 48, 4990–5000.
Table 2 Chloride efflux (%) promoted by 0.02 molar equiv. of
receptors 1–4 in the absence and presence of valinomycin (0.02 molar
equiv.) from unilamellar POPC vesicles loaded with 489 mM KCl
buffered to pH 7.2 with sodium phosphate salts after 270 s. The
vesicles were dispersed in 489mM KNO3 buffered to pH 7.2 with
5 mM sodium phosphate salts
Chloride efflux at 270 s (%)
12 L. Novak, M. Hanania, P. Kovacs, C. E. Kovacs, P. Kolonits and
C. Szantay, Synth. Commun., 1999, 29, 1757–1766; P. de Tullio,
B. Pirotte, F. Somers, S. Boverie, F. Lacan and J. Delarge,
Tetrahedron, 1998, 54, 4935–4942.
1
2
3
4
Absence of valinomycin
Presence of valinomycin
5
23
29
58
5
23
30
56
13 M. J. Hynes, J. Chem. Soc., Dalton Trans., 1993, 311–312.
c
This journal is The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 7641–7643 7643