Solvent induced selectivity switching in aromatic-anion binding molecular receptors

Article abstract of DOI:10.1007/s10847-009-9725-4

Selectivity reversal within a series of hydroxy-substituted benzoates can be induced simply by changing from an aqueous to a non-aqueous environment when using a cyclen derived molecular receptor offering two alternative guest binding mechanisms. Graphica

Full text of DOI:10.1007/s10847-009-9725-4

J Incl Phenom Macrocycl Chem (2010) 67:483–487
DOI 10.1007/s10847-009-9725-4
SHORT COMMUNICATION
Solvent induced selectivity switching in aromatic-anion binding
molecular receptors
•
Jozef A. Z. Hodyl Stephen F. Lincoln
•
Kevin P. Wainwright
Received: 10 September 2009 / Accepted: 10 December 2009 / Published online: 10 January 2010
ꢀ Springer Science+Business Media B.V. 2010
Abstract Selectivity reversal within a series of hydroxy-
substituted benzoates can be induced simply by changing
from an aqueous to a non-aqueous environment when using
a cyclen derived molecular receptor offering two alterna-
tive guest binding mechanisms.
o-hydroxybenzoate anions [6]. This became evident when
anion association constant measurements were made using
a series of structurally isomeric hydroxybenzoates, by
monitoring fluorescence perturbations during titrations
conducted in 20% aqueous 1,4-dioxane. It was found that
as the hydroxy substituent moved sequentially from the
para- to the ortho-position of the benzoate the association
constant increased markedly and, if the meta- and ortho-
dihydroxybenzoates are considered, by up to three orders
of magnitude. [Cd(1)]2? and related receptors are known
from X-ray crystallographic studies to bind to benzoate
guests primarily through classical hydrogen bonding from
the O–H and N–H donors that are convergent into the base
of the binding pocket [7, 8]. The remarkable thing about
this selectivity pattern is that o-hydroxybenzoate (salicy-
late) is markedly less basic than p-hydroxybenzoate and
consequently, if this binding mechanism alone were to
operate, should form much weaker hydrogen bonds. On the
basis of observed fluorescence quenching and geometrical
considerations we proposed that the enhanced affinity of
the o-hydroxybenzoates, and to a lesser extent the m-hy-
droxybenzoates, arises from a second effect whereby
O–H…p hydrogen bonding from the guest to one or other
of the aromatic rings that collectively define the binding
pocket in the host occurs, as shown in Fig. 1b. The
sequence of stability ortho [ meta [ para matches the
degree to which the hydroxy group of the guest is directed
towards the interior of the binding pocket and hence
towards an aromatic ring. A variation of this proposal is
that a water molecule may interpose in the O–H…p
hydrogen bonding and bridge between the hydroxy group
of the guest and an aromatic ring (see, for example, the
structure shown as Fig. 1c) in a manner somewhat akin to
that seen when aromatic p hydrogen bonding to water was
first reported [9].
Keywords Inclusion complex Á Hydroxybenzoates Á
Cyclen derivatives Á Association constants Á Solvent effects
Introduction
The availability of selective sensors for the simple and
immediate detection of anionic species is of great signifi-
cance to a number of different areas of human endeavor,
for example, in the rapid and accurate diagnosis of various
medical and environmental conditions. Because of this
there is enormous global effort in pursuit of this objective
[1–5]. In earlier work we discovered that a cyclen (cy-
clen = 1,4,7,10-tetraazacyclododecane) derived molecular
receptor for aromatic anions, [Cd(1)]2?, shown in Fig. 1,
shows significant affinity and selectivity towards
Electronic supplementary material The online version of this
article (doi:10.1007/s10847-009-9725-4) contains supplementary
material, which is available to authorized users.
J. A. Z. Hodyl Á K. P. Wainwright (&)
School of Chemistry, Physics and Earth Sciences,
Flinders University, GPO Box 2100, Adelaide,
SA 5001, Australia
e-mail: Kevin.Wainwright@flinders.edu.au
S. F. Lincoln
School of Chemistry and Physics, University of Adelaide,
Adelaide, SA 5005, Australia
123

484
J Incl Phenom Macrocycl Chem (2010) 67:483–487
Fig. 1 [Cd(1)]2? binding of a p-hydroxybenzoate, b and c o-hydroxybenzoate, without and with the interposition of water. Log Kassoc values
determined in 20% aqueous 1,4-dioxane [6]
To search for further evidence of O–H…p hydrogen
bonding between hydroxybenzoates with favourably posi-
tioned hydroxy moieties and this class of receptor we have
now synthesised [Cd(3)]2? (Scheme 1) in which the pen-
dant anthracene of [Cd(1)]2? has been replaced by a pen-
dant methyl group. Whilst the main objective of this work
was to look for reductions in the association constants of
guests suspected of forming O–H…p hydrogen bonds,
concomitant with loss of the principal aromatic moiety
from [Cd(1)]2?, we also took the opportunity to investigate
the role of water as a possible stabilizing intermediary in
the suspected O–H…p hydrogen bonding. We are now able
to report not only further evidence for O–H…p hydrogen
bonds in favourable cases, but also that the presence or
absence of water has the quite remarkable effect of being
able to reverse the ordering of association constants for the
series of o-, m- and p-hydroxy-substituted benzoates with
these hosts.
Experimental
General
All reactions were performed under an atmosphere of nitro-
gen. Solvents were pre-dried and purified using known liter-
ature methods. 1,4,7-Tris((2S)-2-hydroxy-3-phenoxypropyl)
cyclen, 2, was prepared by the literature procedure [6].
Microanalyses were carried out at the University of Otago,
New Zealand. 1H and 13C NMR spectra were obtained using a
Bruker Avance 400 MHz spectrometer. Referencing of the
13C NMR chemical shifts was to the central resonance of the
solvent multiplet: d 77.00 for CDCl3 and d 49.00 for CD3OD.
1H NMR chemical shifts were referenced to the central
resonance of the residual non-deuterated solvent peak: d 7.26
for CDCl3, d 2.50 for DMSO-d6 and d 1.94 for CD3CN.
Molar conductivity measurements (K) were made in
10-3 mol dm-3 solutions in anhydrous DMF at ambient
Scheme 1
123

J Incl Phenom Macrocycl Chem (2010) 67:483–487
485
temperatureusingaModelAqua-CConductivity-TDS-Temp.
Meter.
dried (MgSO4) and the solvent then evaporated leaving
spectroscopically pure 3 as a viscous, brown oil, 630 mg,
1
96%. H NMR (CDCl3): d 7.3 (6H, m, ArH); 6.9 (9H, m,
Association constant measurements
ArH); 4.8–2.8 (38H, br m, –CH2, –OH, –CH); 1.3 (3H, s,
(–CH3)). 13C NMR (CDCl3): d 158.35 (2 C, Ar, ipso); 158.32
(1 C, Ar, ipso); 129.25 (6 C, Ar); 120.92 (2 C, Ar); 120.73
(1 C, Ar); 114.57 (2 C, Ar); 114.49 (2 C, Ar); 114.39 (2 C,
Ar); 70.13 (1 C, OCH2); 70.04 (2 C, OCH2); 67.84 (1 C,
CH); 65.89 (1 C, CH); 65.67 (1 C, CH); 65.34 (1 C, CH);
63.26 (1 C, –CH2N) 58.04 (1 C, CH2N); 56.92 (2 C,
CH2N); 52.33 (2 C, cyclen CH2); 51.04 (2 C, cyclen CH2);
50.36 (2 C, cyclen CH2); 49.71 (2 C, cyclen CH2); 21.87 (1
C, –CH3). To obtain a sample suitable for microanalysis 3
was converted to its tetrahydrochloride by treating a solu-
tion of it (534 mg, 0.79 mmol), in ice-cold ethanol (5 cm3)
with 37% aqueous HCl (10 cm3, 121 mmol). The resultant
solution was then evaporated to dryness and redissolved in
methanol (0.5 cm3). Addition of diethyl ether then pre-
cipitated the product, which was filtered off, washed with
diethyl ether (10 cm3) and dried in vacuo to give
3Á4HClÁ1.5H2O as an off-white powder, 583 mg, 89%.
(Found: C, 53.46; H, 7.23; N, 6.57. C38H63Cl4N4O8.5
requires C, 53.46; H, 7.44; N, 6.56%).
NMR monitored titrations, conducted at 298 K, in which
1
the changing chemical shift of the H resonance from one
of the guest protons was tracked were used. The procedure
involved preparation of individual samples containing
guest alone or host and guest in 0.7 cm3 of deuterated
solvent in 5 9 180 mm NMR tubes. Samples for NMR
titrations were added to the NMR tubes with Gilson
Pipetman micropipettes. Concentration of the guest species
in each tube was kept constant at 1 mM while varying the
host for each sample from 0 to 10 mM. The samples in
each tube were prepared by addition of 50 lL of guest
from a 14 mM stock solution made up in the appropriate
deuterated solvent. A host stock solution of 14 mM was
prepared, in the same solvent, from which aliquots of
between 5 and 500 lL were added to each sample, each in
a different NMR tube, to give a series of samples with host
: guest ratios between 0.1 and 10 that were topped up to
0.7 cm3 with deuterated solvent. A locally written, non-
linear regression procedure was used to process the
chemical shift data and hence obtain the association con-
stants and theoretical titration curves.
(1-((2S)-2-hydroxypropyl)-4,7,10-tris-((2S)-2-hydroxy-3-
phenoxypropyl)-1,4,7,10-
All guest species were used as their sodium salts
(although Na(I) can bind to free host ligand, 3, since 3
already has the much more strongly binding Cd(II) ion
coordinated to it when used as the host the possibility of
any interference by Na(I) is remote), which were prepared
by treatment of an aqueous or ethanolic solution of the
conjugate acid with NaOH until the pH of the solution was
at least two pH units above the pKa of the carboxyl moiety.
The salt was then isolated either by filtration or removal of
the solvent in vacuo and recrystallised from ethanol or
methanol.
tetraazacyclododecane)cadmium(II) diperchlorate
hemihydrate, [Cd(3)](ClO4)2Á0.5H2O
To a refluxing solution of 3 (248 mg, 0.36 mmol) in dry
ethanol (10 cm3), cadmium perchlorate hexahydrate
(168 mg, 0.40 mmol) dissolved in dry ethanol (5 cm3) was
added dropwise over 5 min. A white precipitate formed
and redissolved immediately. The reaction was heated at
reflux temperature for a further 1 h and then cooled to
room temperature. The solvent was evaporated leaving a
viscous, light-brown oil which was dissolved in minimal
methanol then precipitated and triturated with diethyl ether
to yield [Cd(3)](ClO4)2Á0.5H2O as an off-white powder
that was dried in vacuo, 281 mg, 79%. KM = 128 X-1
cm2 mol-1 (1 3 10-3 mol dm-3 in DMF). (Found: C,
45.62; H, 5.54; N, 5.42. C38H59CdCl2N4O16.5 requires C,
45.59; H, 5.74; N, 5.60%.) 13C NMR (CD3OD): d 159.92
(1 C, Ar, ipso); 159.86 (1 C, Ar, ipso); 159.79 (1 C, Ar,
ipso); 130.67 (2 C, Ar); 130.63 (2 C, Ar); 130.56 (2 C, Ar);
122.40 (3 C, Ar); 115.81 (2 C, Ar); 115.74 (2 C, Ar);
115.66 (2 C, Ar); 70.66 (1 C, –OCH2); 70.42 (1 C,
–OCH2); 70.04 (1 C, –OCH2); 66.65 (2 C, –CH); 66.50 (1 C,
–CH); 65.79 (1 C, –CH); 63.26 (1 C, –NCH2); 56.24 (2 C,
–NCH2); 56.15 (1 C, –NCH2); 54.56 (2 C, cyclen –CH2);
53.05 (2 C, cyclen –CH2); 51.53 (2 C, cyclen –CH2); 50.49
(2 C, cyclen –CH2); 22.13 (1 C, –CH3).
Syntheses
1-((2S)-2-hydroxypropyl)-4,7,10-tris-((2S)-2-hydroxy-3-
phenoxypropyl)-1,4,7,10-tetraazacyclododecane (3)
To a pressure vessel containing a solution of 2 (600 mg,
0.963 mmol) in dry acetonitrile (1 cm3) and LiBr
(26.2 mg, 0.302 mmol), (S)-propylene oxide (88 mg,
1.51 mmol) was added. The vessel was sealed and the
contents then heated to 80 ꢁC and stirred at that tempera-
ture for 3 days. Following this the vessel was cooled,
opened and the reaction mixture dissolved in water
(20 cm3). The aqueous solution was then extracted with
chloroform (3 9 10 cm3). The extracts were combined,
123

486
J Incl Phenom Macrocycl Chem (2010) 67:483–487
Results and discussion
to the ortho-position(s) can be seen. It is also seen that this
selectivity pattern is mirrored with [Cd(3)]2?, when in the
similar solvent mixture, 20% aqueous CD3CN (deuterated
1,4-dioxane would have been prohibitively expensive,
CD3CN is closer in dielectric constant than DMSO). It is
noticeable, however, that the substitution of the anthracene
containing arm by a methyl containing arm, lowers the
affinity for all the hydroxybenzoates, but not for benzoate
itself where O–H…p hydrogen bonding is not possible.
Furthermore, the reduced affinity of the hydroxybenzoates
is most marked for those having a hydroxy group in an
ortho-position, where it is best positioned to engage in
O–H…p hydrogen bonding, and progressively less so as the
hydroxy group moves further towards the upper extremity
of the binding pocket. Even more interesting is the total
reversal of the association constant ordering for [Cd(3)]2?
when non-aqueous conditions (DMSO-d6) are used, which
strongly supports the idea that water molecule interposition
between the hydroxy group and the aromatic wall of the
pocket, as shown in Fig. 1c, is an essential component of
O–H…p hydrogen bonding stabilization. The stability
ordering seen in DMSO-d6 is that which would be expected
simply from the trend in the pKa of the benzoate group,
with the most acidic benzoate (2,6-dihydroxybenzoic acid)
displaying the weakest binding, suggesting that O–H…p
hydrogen bonding is weak or non-existent in the absence of
water. The ordering of the two association constants
measured in CD3OD suggests that methanol is not able to
act in a similar way to water in facilitating O–H…p
hydrogen bonding, although with the limited data available
this conclusion should be treated with caution.
The homochiral receptor ligand 3 was prepared in 96%
yield from the known compound 1,4,7-tris((2S)-2-hydroxy-
3-phenoxypropyl)cyclen, 2 [6], by reacting it with (S)-
propylene oxide for 72 h in acetonitrile at 80 ꢁC in a sealed
pressure vessel (to avoid loss of the epoxide), using LiBr as
a catalyst. As we have noted previously [6], N-alkylation
reactions of 2 are exceedingly slow, and in the absence of
LiBr, which is a recognized catalyst for epoxide ring
openings [10, 11], no reaction could be induced. Com-
plexation with Cd(II) was straightforward leading to the
receptor complex [Cd(3)]2?, isolated as its diperchlorate
salt in 79% yield, that is presumed on the basis of
numerous crystal structures of related complexes to have
the eight-coordinate structure shown [7, 8, 12]. In DMF
[Cd(3)]2? behaves as a 1:2 electrolyte (KM = 128 X-1
cm2 mol-1) indicating [13], as expected [14], little or no
-
retention of ClO4 in the aromatic-anion binding pocket.
To establish the relative anion binding ability of
[Cd(3)]2? compared to [Cd(1)]2?, and the influence of
water on the stabilization of hydroxybenzoate guests within
the binding pocket, a series of association constant mea-
1
surements were made using H NMR monitored titrations
in different solvents. Prior to doing this Job’s Method of
1
Continuous Variation [15], was employed, using H NMR
monitoring in DMSO-d6 and a variety of different aromatic
anions, to deduce the stoichiometry of the host:guest
interactions. Through this it was verified that except in the
case of p-nitrophenolate, which has been shown previously
to give 1:2 host:guest inclusion complexes [7], only 1:1
inclusion complexes are formed. The association constants
determined are tabulated in Table 1 alongside the previ-
ously determined data for [Cd(1)]2? for comparison.
From Table 1 the progressive increase in the association
constant for a benzoate with [Cd(1)]2?, in 20% aqueous
1,4-dioxane, as a hydroxy group is first introduced to the
para-position and then moved sequentially to the meta- and
Conclusion
Although we have not yet been able to demonstrate by
X-ray crystallography the O–H…p hydrogen bonding that
we suspect to be stabilising some of these host–guest
Table 1 Association constants expressed as log(K/mol-1 dm3) for the inclusion of hydroxybenzoate anions within the receptor complexes
1
[Cd(1)]2? and [Cd(3)]2? determined by H NMR monitored titration, unless otherwise noted, following the previously reported procedure [6]
Guest anion
Solvent:
Receptor complex
pKca
[Cd(1)]2?a
20% H2O/dioxane
[Cd(3)]2?
[Cd(3)]2?
DMSO-d6
[Cd(3)]2?
CD3OD
20% D2O/CD3CNb
Benzoate
p-OH
2.3 0.1
4.5 0.3
5.3 0.5
7.1 0.5
7.5 0.9
3.10 0.08
3.18 0.28
3.49 0.17
3.70 0.15
4.64 0.36
4.11 0.06
4.09 0.07
4.16 0.09
3.08 0.08
2.05 0.03
3.96 0.09
3.23 0.03
4.19
4.54
4.30
2.97
1.05
m-OH
o-OH
2,6-diOH
Measured at 298 1 K with [guest anion] = 10-3 M. Uncertainties are one SD
a
Data taken from [6], determined by fluorescence perturbation, buffered at pH 7.0 (0.02 mol dm-3 lutidine). bBuffered at pH 7.0
(0.01 mol dm-3 HEPES). Taken from [16]
c
123

J Incl Phenom Macrocycl Chem (2010) 67:483–487
487
7. Smith, C.B., Stephens, A.K.W., Wallwork, K.S., Lincoln, S.F.,
Taylor, M.R., Wainwright, K.P.: Metal ion-dependent molecular
inclusion chemistry: inclusion of aromatic anions by coordinated
1, 4, 7, 10-tetrakis((S)-2-hydroxy-3-phenoxypropyl)-1, 4, 7, 10-
tetraazacyclododecane. Inorg. Chem. 41, 1093–1100 (2002)
8. Robinson, T.S., Wyness, O., Lincoln, S.F., Taylor, M.R., Tiekink,
E.R.T., Wainwright, K.P.: A structural study of the synergic
envelopment of acetonitrile by a Cd(II) activated molecular
receptor formed from cyclen with appended 2-hydroxy-3-phen-
ylpropyl moieties. Inorg. Chim. Acta 359, 1413–1420 (2006)
9. Atwood, J.L., Hamada, F., Robinson, K.D., Orr, G.W., Vincent,
R.L.: X-ray diffraction evidence for aromatic p hydrogen bonding
to water. Nature 349, 683–684 (1991)
10. Chakraborti, A.K., Basak, A., Grover, V.: Chemoselective pro-
tection of carboxylic acid as methyl ester: a practical alternative
to diazomethane protocol. J. Org. Chem. 64, 8014–8017 (1999)
11. Chakraborti, A.K., Rudrawar, S., Kondaskar, A.: Lithium bro-
mide, an inexpensive and efficient catalyst for opening of epoxide
rings by amines at room temperature under solvent-free condi-
tion. Eur. J. Org. Chem. 3597–3600 (2004)
complexes, circumstantial evidence for its existence is
mounting. Through this work we see manifestation of it not
only by the reduction of the o-hydroxybenzoate association
constants when the aromaticity of the binding cavity is
diminished by removal of the anthracene moiety, but also
by the effect of added water, which, instead of attracting
the water soluble o-hydroxybenzoates away from the cav-
ity, actually enhances their association, presumably by
facilitating intracavity O–H…p hydrogen bonding.
Acknowledgment We are grateful to the Australian Research
Council for the support of this work.
References
12. Wainwright, K.P.: Applications for polyaza macrocycles with
nitrogen-attached pendant arms. Adv. Inorg. Chem. 52, 293–334
(2001)
1. Kubik, S., Reyheller, C., Stuwe, S.: Recognition of anions by
synthetic receptors in aqueous solution. J. Incl. Phenom. Mac-
rocycl. Chem. 52, 137–187 (2005)
13. Geary, W.J.: The use of conductivity measurements in organic
solvents for the characterization of coordination compounds.
Coord. Chem. Rev. 7, 81–122 (1971)
14. Smith, C.B., Wallwork, K.S., Weeks, J.M., Buntine, M.A., Lin-
coln, S.F., Taylor, M.R., Wainwright, K.P.: Metal ion dependent
molecular inclusion chemistry: inclusion of p-toluenesulfonate
and p-nitrophenolate within the structure of coordinated 1,4,7,10-
tetrakis((S)-2-hydroxy-3-phenoxypropyl)-1, 4, 7, 10-tetraazacy-
clododecane. Inorg. Chem. 38, 4986–4992 (1999)
15. Hirose, K.: In: Schally, C.A. (ed.) Analytical Methods in
Supramolecular Chemistry, pp. 17–54. Weinheim, Wiley-VCH
(2007)
16. Howard, P., Meylan, W.: Physical Properties Database. Syracuse
Research Corporation, Environmental Science Centre, North
Syracuse (1999)
´
2. Amendola, V., Bonizzoni, M., Esteban-Gomez, D., Fabbrizzi, L.,
Licchelli, M., Sancenon, F., Taglietti, A.: Some guidelines for the
´
design of anion receptors. Coord. Chem. Rev. 250, 1451–1470
(2006)
3. Gale, P.A., Quesada, R.: Anion coordination and anion-templated
assembly: highlights from 2002 to 2004. Coord. Chem. Rev. 250,
3219–3244 (2006)
´
4. Gale, P.A., Garcıa-Garrido, S.E., Garric, J.: Anion receptors
based on organic frameworks: highlights from 2005 and 2006.
Chem. Soc. Rev. 37, 151–190 (2008)
5. Schneider, H.-J., Yatsimirsky, A.K.: Selectivity in supramolecu-
lar host–guest complexes. Chem. Soc. Rev. 37, 263–277 (2008)
6. Bradbury, A.J., Lincoln, S.F., Wainwright, K.P.: Fluorescent
signaling provides deeper insight into aromatic anion uptake by
metal-ion activated molecular receptors. New J. Chem. 32, 1500–
1508 (2008)
123