Synthesis, structural investigation and computational modelling of water-binding aquafoldamers

Article abstract of DOI:10.1039/c1ob06609a

Detailed studies on water-binding aquafoldamers are presented that illustrate the potential use of the elongated larger aquafoldamers for recognizing larger water clusters of diverse topologies. A novel self-trapping dimerization mode involving two tetramer molecules is proposed, which is consistent with the obtained varying experimental evidences. The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c1ob06609a

Organic &
Biomolecular
Chemistry
www.rsc.org/obc
Volume 10 | Number 6 | 14 February 2012 | Pages 1125–1312
ISSN 1477-0520
PAPER
Zeng et al.
Synthesis, structural investigation and computational modelling
of water-binding aquafoldamers
1477-0520(2012)10:6;1-H

Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2012, 10, 1172
www.rsc.org/obc
PAPER
Synthesis, structural investigation and computational modelling of
water-binding aquafoldamers†
Huaiqing Zhao,a Wei Qiang Ong,a Xiao Fang,a Feng Zhou,b Meng Ni Hii,a Sam Fong Yau Li,a Haibin Sub and
Huaqiang Zeng*a
Received 21st September 2011, Accepted 31st October 2011
DOI: 10.1039/c1ob06609a
Detailed studies on water-binding aquafoldamers are presented that illustrate the potential use of the
elongated larger aquafoldamers for recognizing larger water clusters of diverse topologies. A novel
self-trapping dimerization mode involving two tetramer molecules is proposed, which is consistent with
the obtained varying experimental evidences.
bio-inspired water hosts had also been shown to be able to host a
1D helical chain of water molecules in their framework, reminis-
cent of aquaporins. Here we describe our approach using pyridine-
Introduction
Aquaporins are a group of specialized transmembrane proteins
that form a hydrophobic pore system across the cell membrane.1
based aquafoldamers 1–5 that fold into a crescent structure to en-
Four of these proteins, in the tetrameric form, form a water
close a defined water-binding cavity by intramolecular H-bonding
networks.9,10 The elongated aquafoldamers with sufficiently long
enough backbones may function as synthetic water-transporting
channels. In this article, detailed characterization of water-binding
aquafoldamers composed of up to five repeating units were
carried out using X-ray crystallography, variable-temperature 1H
NMR experiments, 2D NOESY studies and ab initio molecular
modelling.
channel across the lipid bilayer allowing for the transportation
of water molecules, in a 1D chain-like arrangement, across the
membrane.2 The possible use of water-transporting aquaporins for
wastewater reclamation and re-use as well as seawater desalination
is being investigated industrially by a company called Aquaporin.
However, significant challenges associated with manipulating
channel proteins in terms of complexity, stability, availability and
activity reconstitution still exist. An alternative strategy is to design
small molecule-based synthetic water channels that can mimic
aquaporins to a certain degree of related functionality for various
applications including water purification. Toward this goal, we
have been interested in designing water-binding aquafoldamers
as illustrated by oligoamides 1–53 with an ultimate aim to
realize synthetic water channels for rapid, efficient transportation
of water molecules while excluding all the other molecular
species.
There have been a few different approaches to mimic the aqua-
porin water channel, namely using supramolecular4 and metal–
organic framework approaches.5 These water hosts have usually
relied on conformationally more flexible organic or organometallic
molecules with respect to foldamers whose well defined backbones
are primarily stabilized by non-covalent forces such as p–p stack-
ing interactions, solvophobic forces and H-bonds. Interestingly,
despite their great diversities,6 only a few foldamer molecules of
similar type have been reported by Lehn and Huc.7 Some of these
aDepartment of Chemistry, National University of Singapore, 3 Science Drive
3, Singapore. E-mail: chmzh@nus.edu.sg; Fax: (+) 65-6779-1691; Tel: (+)
65-6516-2683
Results and discussion
bDivision of Materials Science, 50 Nanyang Avenue, Nanyang Technological
University, Singapore, 639798
Synthesis of aquafoldamers 1–5
† Electronic supplementary information (ESI) available. CCDC reference
number 834710. For ESI and crystallographic data in CIF or other
electronic format see DOI: 10.1039/c1ob06609a
We recently reported our studies on the pyridine-based foldamers
2–4 and 5a,3 demonstrating that with an increasing addition
1172 | Org. Biomol. Chem., 2012, 10, 1172–1180
This journal is
The Royal Society of Chemistry 2012
©

of the pyridine-based building blocks, the elongated backbone
becomes increasingly curved in one direction as a result of the
stabilizing forces from the progressively lengthened intramolecular
H-bonding network.10 This new class of aromatic foldamers
closely mimics the helicity requirement by p-helices where about
4.3 units are required for each helical turn. This value also
closely resembles the tetrameric arrangement of the aquaporin
required to furnish a water channel in the membrane. All the
aquafoldamers described in Scheme 2 were synthesized from
commercially available pyridine-2,6-dicarboxylic acid (1a) and
diethyl oxalate (1f) after up to 12 steps with an overall yield of
about 1% for pentamers 5a and 5b.
Scheme 2 Synthesis of aquafoldamers 2–5. Reagents and conditions: (a)
4-methylmorpholine, ethyl chloroformate, THF–DMF, -10 to 0 ◦C, then
1c; (b) NaOH, dioxane then 1 M HCl; (c) (COCl)2, DMF, CH2Cl2, then 1e
(for 3a and 4), 1l (for 3b), 2-aminopyridine (for 5a) or 2,6-diaminopyridine
(for 5b), Et3N, CH2Cl2.
the monohydrolyzed product 1c involves using one equivalent of
potassium hydroxide in methanol and allowing the reaction to stir
at room temperature overnight. The monoacid 1c then underwent
a Curtius rearrangement reaction to obtain monomer 1 in 65%
yield. Compound 1c first reacted with 4-methylmorpholine and
ethyl chloroformate to give an active ester intermediate, which
was then reacted with the nucleophile sodium azide to form the
acyl azide intermediate. The acyl azide intermediate underwent
Curtius rearrangement in the presence of benzyl alcohol to directly
form the carbobenzyloxy (Cbz) protected amine 1. Monomer acid
1d was then obtained by subjecting monomer 1 to the hydrolysis
by NaOH at room temperature overnight. The reaction must not
be heated as, instead of 1d, an amino acid product was obtained
where the Cbz protecting group was deprotected under the basic
condition at high temperatures. In order to obtain monomer amine
1e, monomer 1 underwent a catalytic hydrogenation reaction using
10% palladium on activated carbon as the catalyst. The reaction
was clean and efficient, giving a quantitative yield.
Scheme 1 Synthesis of monomeric building blocks. Reagents and condi-
tions: (a) conc. H2SO4, MeOH, reflux; (b) KOH, MeOH; (c) 4-methylmor-
pholine, ethyl chloroformate, THF–DMF, -10 to 0 ◦C; (d) aq. NaN3; (e)
BnOH, toluene, reflux; (f) NaOH, dioxane; (g) Pd/C, H2, THF; (h) sodium
ethoxide, then acetone; (i) conc. HCl; (j) 10% aq. NH3; (k) SOCl2, MeOH;
(l) K2CO3, CH3I, DMF, reflux.
Monomer amine 1l was obtained in eight steps starting from
diethyl oxalate 1f. Diacid 1g was obtained after three steps:
diethyl oxalate first underwent a Claisen condensation reaction
with acetone using sodium ethoxide as the base, followed by
a cyclization reaction in the presence of concentrated HCl and
subsequent treatment with 10% aqueous ammonia solution.
Esterification of diacid 1g using thionyl chloride in methanol
afforded diester methyl 1h. The hydroxyl group, meta to the
ester functionality, then underwent an alkylation reaction using
potassium carbonate and methyl iodide to generate the alkylated
product 1i in 73% yield. Similar to the procedure to obtain
monomer amine 1e, the diester 1i was subjected to a series of
reactions consisting of monohydrolysis, Curtius rearrangement
and catalytic hydrogenation to produce amine building block 1l.
Following the elaboration of the synthetic routes for the
efficient preparation of the various monomeric building blocks,
Monomeric building blocks 1d, 1e and 1l were prepared accord-
ing to Scheme 1. Carboxylic acid 1d and amine 1e were synthesized
in four steps starting from pyridine-2,6-dicarboxylic acid 1a. As
shown in Scheme 1, pyridine-2,6-dicarboxylic acid underwent
esterification in methanol to afford the diester methyl 1b in a
high yield of 96%. The second step involved the monohydrolysis
of the diester to provide the monoacid 1c. Several methods, such as
using one equivalent of sodium hydroxide in cold or hot methanol
or in cold or hot dioxane, or using one equivalent of potassium
hydroxide in hot DMSO or in cold or hot methanol or in cold
or hot dioxane were tried to obtain the product 1c, however,
most of the methods gave a low yield of the desired product with
either the majority of starting material 1b remaining unreacted
or the dihydrolyzed product pyridine-2,6-dicarboxylic acid 1a
as the major product. The condition producing a 78% yield of
This journal is
The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 1172–1180 | 1173
©

aquafoldamers 2–5 was then prepared according to Scheme 2 using
a step-by-step approach with the various monomeric building
blocks. The synthesis of dimer 2 started with monomer acid 1d
and monomer amine 1e via an active ester intermediate.
Compound 1d first reacted with 4-methylmorpholine and
ethylchloroformate to form the active ester intermediate and
the above in situ generated intermediate was then coupled with
compound 1e to yield dimer 2 in 52% yield. Hydrolysis of dimer 2
using NaOH at room temperature overnight yielded the dimer acid
in 84% yield. An acid chloride was then generated from the dimer
acid by treating the dimer acid with oxalyl chloride and a few drops
of DMF in CH2Cl2 at room temperature. Trimers 3a and 3b were
then synthesized by adding monomer amine 1e and 1l, respectively,
to dimer acid chloride together with triethylamine. The higher
oligomers were synthesized in the same fashion by allowing trimer
3a to undergo a series of hydrolysis and coupling reactions to
produce tetramer 4 and pentamers 5a and 5b. A unidirectional
stepwise approach was used in the synthesis instead of a convergent
approach where a higher oligomer such as a dimer acid is coupled
to a dimer amine to synthesize a tetramer molecule, because (1) the
higher oligomer amines have very poor solubilities in most organic
solvents and (2) the H-bonding network in the higher oligomer
amine is very strong, possibly resulting in a much lower reactivity
of the amine in higher oligomers and hence giving a much lower
yield. Therefore, for this particular class of oligoamides, a stepwise
approach proves to be more efficient, in terms of both time and
materials.
Fig. 2 Computationally determined structures for 1 : 1 water complexes
Aq·H2O (Aq = 1–5b) at the B3LYP/6-311G+(2d,p) level in the gas phase.
The corresponding binding energies per water molecule are shown below
the structures. The crystal structure of 4 containing one CH2Cl2 molecule
in its cavity is shown in (g). Both water and CH2Cl2 molecules are shown
as CPK models. In the computed structures, all the water molecules
are H-bonded to the amide H-atoms or pyridine N-atoms from the
aquafoldamer hosts. Cbz groups are represented as the dummy atom in
yellow.
cavity, respectively. More specifically, the water-binding abilities
of this series of pyridine-derived cavity-enclosing aquapentamers
can be proven by the water-containing crystal structure of 3b, 5a
and 5b (Fig. 3).
Solid-state structure of aquapentamer 5b
The crystal structures of 2, 3, 4 and 5a have been recently
reported by us.3 Similar to these structures, the crystal struc-
ture of the pyridine-based aquapentamer 5b (Fig. 1) shows an
expected crescent-shaped helically folded structure as a result of
an efficient backbone rigidification by the stabilizing forces from
˚
the continuous intramolecular H-bonding network (2.13–2.41 A)
formed from the inward-pointing amide protons and N-atoms on
the pyridine rings. It could be observed that the helical backbone of
5b requires about 4.3 repeating units to make up one helical turn,
a helicity requirement also demanded by 5a.3a The cavity enclosed
Fig. 3 Crystal structures for water complexes of (a) 3b·H2O, (b) 5a·2H2O
and (c) 5b·2H2O, encapsulating both unconventional water dimer in
3b·H2O and conventional water dimers in 5a·2H2O and 5b·2H2O. Due
to a difference in geometry, planar 3 binds only one water while helically
folded 5a and 5b each trap two water molecules in their helical cavity. In (c),
the Cbz group in 5b is represented as the dummy atom in yellow. See Table
2 for the corresponding binding energies for forming water complexes and
water dimers in these aquafoldamers.
˚
by this set of aquapentamers 3–5 has a radius of about 2.6–3.3 A,
˚
values that are slightly larger than ~ 2 A pore radius found in the
aquaporin water channel, suggesting that every crescent-shaped
(e.g., 3 and 4) or helically folded (e.g., 5) aquafoldamer molecule is
able to accommodate one or more than one water molecule in its
Water complexes
Ab initio calculations performed on 1–5 at the level of B3LYP/6-
311G+(2d,p) show that the 1 : 1 water complexes Aq·H2O (Aq =
1–5b) have a respective stability of 9.07 (1), 10.18 (2), 11.81
(3a), 12.18 (3b), 11.97 (4), 9.80 (5a), 9.58 (5b) kcal mol-1 with
respect to its individual components (Fig. 2). Although the water
complex 4·H2O is found to be energetically very favored (11.97
kcal mol-1, Fig. 2d), 4 crystallographically encapsulates only
CH2Cl2 molecules, rather than water molecules, in its cavity with a
calculated binding energy of 2.28 kcal mol-1 per CH2Cl2 molecule
Fig. 1 Top and side views of crystal structure of the helically folded
aquapentamer 5b, illustrating the pyridine N-atoms (blue balls) and
amide protons (gray balls) that participate in forming an intramolecular
H-bonding network that induces the molecular backbone of 1 into a helical
structure and encloses a water-binding helical cavity of about 2.7 A in
radius as measured from the cavity center to the center of the amide
H-atom.
˚
1174 | Org. Biomol. Chem., 2012, 10, 1172–1180
This journal is
The Royal Society of Chemistry 2012
©

Table 1 Crystal growth conditions for oligomers 2–5a
Table 2 Binding energies for water complexes and water dimers found in
3b·H2O, 5a·2H2O and 5b·2H2O in the gas phase
Slow diffusion with varying
solvent pairs (1 : 1)
Slow
evaporation
Binding energy/kcal mol-1
b
3b·H2O
5a·2H2O
5b·2H2O
2
CH2Cl2 : MeOH
5a CH2Cl2 : C6H12
4
5b
CH2Cl2 : C6H12
CHCl3 : EtOH
3b
CH2Cl2
b
Aq + mH2O → Aq·mH2Oa
7.61
2.22
10.73
3.88
7.97
3.58
b
a Crystals of 3a could not be obtained under any the conditions tested,
and oligomers have good solubilities in either CH2Cl2 or CHCl3.
b Cyclohexane.
2H2O → (H2O)2
a For instance, Aq = 3b and m = 1 for 3b·H2O. b For comparison, the
computationally derived binding energy for the most stable water dimer is
5.21 kcal mol-1.
(Fig. 2g). Further, the computationally determined water position
in 3b (Fig. 2c) differs from that found in its crystal structure
(Fig. 3a). These discrepancies suggest that the computationally
determined energetic favorability order can be possibly overridden
by crystal packing effects.
3.58 kcal mol-1 at the level of M062X/aug-cc-pVTZ. These water
dimers are destabilized respectively by 2.99, 1.33 and 1.63 kcal
mol-1 with respect to the binding energy of 5.21 kcal mol-1 for
the most stable water dimer. This points to the instability of the
trapped water dimers, formation of which is greatly facilitated by
the stronger H-bonding network (Table 2) and a restricted cavity
provided by these H-bond-rigidified aquafoldamers.
Despite of repetitive efforts made to obtain the water complexes
for these aquafoldamers, experimentally, it was not as straight-
forward. After screening numerous conditions involving various
conditions of all the common organic solvents by method of either
slow evaporation or diffusion, X-ray quality crystals of 2–5 were
all obtained by their respective methods outlined in Table 1. Of
further note is that both water-containing crystals of 3b and 5a
were crystallized from water immiscible solvents, and only trace
amounts of water molecules can be found in these crystallization
conditions. And in the case of 4, despite of having tried many
conditions, even with using solvents that were saturated with water,
water complexes of 4 could not be obtained. Except for crystals
of dimer 2 containing no solvent or water molecules in its crystal
lattice and tetramer 4 containing CH2Cl2 in its cavity, trimer 3b,
and pentamers 5a and 5b were all found to be capable of trapping
water molecules in their interior cavities.
One-dimensional 1H NMR studies of the water complexes
The bound water molecules in the cavity of 2–5b was further
1
investigated by H NMR in deuterated chloroform with varying
water contents. More specifically, 5 mM of each of the oligoamide
was prepared using (a) “normal” CDCl3 directly taken from the
bottle, (b) “wet” CDCl3 saturated with water and (c) “dry” CDCl3
1
dried by activated A4 molecular sieves. The corresponding H
NMR of the samples were recorded to examine the dependence
of the chemical shifts of the amide protons on the water contents,
which serves as an indicator on the water-binding ability of these
molecules.
For 2, it was observed that varying water contents essentially
produces no changes among the NMR spectra recorded in
“normal”, “wet” and “dry” CDCl3 (Fig. 4), indicating an inability
for 2 to trap water molecules due to the open nature of its
cavity. Accordingly, the amide protons in 2 do not H-bond to the
water molecules, and thus no appreciable changes in the chemical
shifts of the amide protons can be observed. This observation is
consistent with the fact that no solvent or water molecules can be
found in the crystal lattice of 2.
Similar to 2, the water content exhibits little effect on the
chemical shifts of the amide protons in 4 (Fig. 4), which is
crystallographically verified to be incapable of binding water
molecules. But we found it difficult to explain this unusual
behaviour by 4 as compared to other oligomers such as 3b and
5 of the same series.
In contrast to 2 and 4 and consistent with their water-binding
ability, significant differences were observed for 1H NMR spectra
of both aquafoldamers 3b and 5b in CDCl3 containing varying
water contents (Fig. 4). In general, the amide proton signals larger
than 10.2 ppm shifted most downfield in the “wet” CDCl3, while
in “dry” CDCl3, the signals shifted most upfield as compared to
the “normal” and “wet” CDCl3. The Cbz amide protons around
7.8 ppm for 3b and 7.5 ppm for 5b follow the same trend. These
results were expected for aquafoldamers capable of binding water
molecules because as the water contents in the solvent increases,
the percentage of water complexes vs. “free” oligomer increases
too, making the amide protons in average more deshielded and
resulting in a downfield shift in the amide proton signals.
In both the water-containing crystal structures of 3b and 5a
recently reported by us,3b the water molecules are stabilized in the
cavities by intermolecular H-bonds of varying strengths with H-
˚
bond distances ranging from 1.92 to 2.91 A. Due to its planar
geometry, molecules of 3b stack on top of each other to form
a 1D columnar structure where every two molecules of 3b trap
an unconventional water dimer mediated by a H ◊ ◊ ◊ H interaction
˚
˚
(dH–H = 2.253 A, a distance that is ~ 0.15 A less than twice the van
˚
der Waals radius of 1.20 A for a hydrogen atom, Fig. 3a). Helically
folded 5a having a 3D-shaped cavity is able to accommodate two
water molecules or one conventional water dimer mediated by a
8
˚
strong H-bond of 1.849 A (Fig. 3b).
˚
In 5b, ten intramolecular H-bonds (2.130–2.491 A) are similarly
found among pyridine N-atoms and amide H-atoms. These H-
bonding forces lead to a helical conformation in 5b that encloses
˚
a small cavity of ~2.7 A in radius, measured from the centre of
the cavity to the amide proton. A water dimer is thus trapped in
the helical cavity of 5b with intermolecular H-bonding distances
between the water molecules and 5b ranging from 1.93 to 2.76
˚
A. The water molecules further serve as exo-bidentate ligands,
bridging the neighbouring two molecules of 5b in a stair-like
formation.
Computationally at the level of B3LYP/g-31G*, the H-bonding
networks around the water molecules in 3b, 5a and 5b provide a
stabilizing energy of 7.61, 10.73 and 7.97 kcal mol-1, respectively,
for forming the water complexes (Table 2). The trapped water
dimers in them have a respective binding energy of 2.22, 3.88 and
This journal is
The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 1172–1180 | 1175
©

Table 3 The chemical shifts of amide protons in 3a and 3b in CDCl3 of
varying water contents
d/ppm
Oligoamide
Solvent
Ha
Hb
Hc
3a
Dry
Normal
Wet
10.37
10.39
10.45
10.23
10.25
10.31
7.83
7.87
> 7.90a
3b
Dry
Normal
Wet
10.34
10.37
10.40
10.22
10.26
10.29
7.83
7.91
> 7.92a
a Signals overlap with other protons.
Fig. 5 1H NMR spectra of 4, 5a and 5b at 5 mM in “normal” CDCl3 at
(a) 300 K and (b) 223 K, illustrating significant aggregations in 4 and 5b
while aggregation in 5a is barely noticeable at 223 K.
Fig. 4 Expanded 1H NMR spectra of aromatic regions for 2–5 at 5 mM at
300 K in “dry”, “normal” and “wet” CDCl3 respectively shown from top
to bottom. The shaded gray regions highlight differences in the chemical
shifts of amide protons observed for 3b and 5b, and possibly also for 5a,
which are indicative of their water-binding ability.
2D NOESY studies of the water complexes
Taking 3b as an example, the proton signal for amide proton Ha
shifted downfield from d 10.34 ppm in “dry” CDCl3 to 10.37 ppm
in “normal” CDCl3 and to 10.40 ppm in “wet” CDCl3. The other
two amide protons Hb and Hc display a similar trend (Table 3).
Being similar in structure, the behaviour of the amide protons in
3a was the same as those in 3b in CDCl3 containing varying water
contents (Table 3).
Given that the NOE contacts become detectable if the inter-
˚
atomic distance is less than 5 A, a distance that is apparently
longer that most of interatomic distances among H-atoms in
water and water-containing aquafoldamers, observable NOE
contacts among these H-atoms in close proximity are anticipated.
Accordingly, these water complexes were further probed using
2D NOESY experiments. We further thought “freezing” all the
atoms at low temperatures such as 223 K may make it possible
to separate the “bound” water from “free” water,7e and to allow
for an observation of NOE contacts between aquafoldamers and
“bound” water molecules to provide direct evidences for the
presence of bound water in the cavity of these aquafoldamers
in solution.
1
As to 5a, a small difference in H NMR spectrum between
“wet” and “dry’ CDCl3 does exist, but comparison of all the three
spectrum including that in “normal” CDCl3 gives inconclusive
information on whether 5a entraps water molecules in its cavity or
not in solution, even though a water dimer is bound in its cavity
in the solid state (Fig. 3b).
Cooling the samples from 300 to 223 K causes the most
peak broadening in 4 and 5b (Fig. 5), second most in 5a and
essentially no broadening in 2, 3a and 3b, suggesting that aromatic
p–p stacking interactions in 4 and 5b are somewhat stronger
that those in 5a. As will be presented later in the paper, 2D
NOESY experiments, ab initio calculations and 1H NMR at
varying concentrations were performed to probe for this unusual
abnormality, and findings obtained are consistent with each other.
2D NOESY experiments were then performed on trimer 3b and
pentamer 5a at 223 K in “normal” CDCl3. Expectedly, in addition
to the signal from “free” water molecules, a new downfield-shifted
1
peak ascribable to “bound” water molecules appeared in the H
NMR spectrum of 3b (Fig. 6a). NOE contacts between the amide
protons including that in Cbz group and protons from the water
molecule bound in the cavity of 3b were clearly seen (Fig. 6a).
These NOE contacts are consistent with the results obtained from
1176 | Org. Biomol. Chem., 2012, 10, 1172–1180
This journal is
The Royal Society of Chemistry 2012
©

˚
are shorter than 5 A, accounting for the observed NOE pattern
in Fig. 6c. We therefore thought that at 223 K, 4 probably exists
in a form of 4C in solution where the ester methyl protons point
into its cavity, thus preventing the water molecule from entering
the cavity.
Ab initio studies of the conformers of 4 and dimeric structures
Presumably due to the predominant occurrence of conformer 4C
in solution, water molecules cannot enter the cavity in 4. Ab initio
computations were then carried out to investigate the likelihood of
4C as the major conformer with respect to the crystallographically
found 4A and 4B by comparing their relative stabilities on the basis
of the computationally derived relative energies among them.
Contrary to what we hypothesized above, calculations using
density functional theory at the level of B3LYP/6-31G* with a
polarized continuum model (PCM) in chloroform at 223 K on
different conformers of 4A–4C with its ester methyl protons in
different positions (Fig. 7) reveals conformer 4A1 found in the
solid state to be computationally the most stable, marginally more
stable than 4B that is also found in the solid state by 0.55 kcal mol-1,
and significantly much more stable than the “desired” conformers
4C (Fig. 7d–e) by at least 7.45 kcal mol-1. With these differences
in energy, either 4C1 or 4C2 is less than 10-millionth of either 4A1
or 4B. In all the modeled structures, the Cbz groups are roughly
perpendicular to the macrocyclic plane. Further, it can be seen that
methyl group in the more stable states tends to stay away from the
cavity by pointing outward (4A1/4B vs. 4C1) and slightly upward
(4A1 vs. 4A2, and 4C1 vs. 4C2), rather than inward and down to
the cavity.
In order to account for the observations noted in the variable-
temperature 1H NMR (Fig. 5a) and 2D NOESY experiments (Fig.
6c) that signify a strong aggregation in 4 and close proximity
between its ester methyl protons and amide protons, the next
most possible logical explanation would be that 4 exists as
a self-trapping dimer with having its methyl protons pointing
or protruding into the cavity of another molecule of 4. The
extensive interactions involving feeble intermolecular H-bonds
and aromatic p–p stackings possibly would bring two molecules
of 4 much closer together, occupying mutual cavities and thus
preventing water molecules from entering the cavities. This dimeric
structure may constitute a good model for explaining the obtained
abnormal findings. Our calculation using a Dreiding force field11
indeed demonstrates a possibility of dimerization involving two
molecules of 4 with a very substantial binding energy of 80.69 kcal
mol-1 per dimer (Fig. 8a–8c). This is a result of three types of
stabilizing intermolecular forces provided largely by (1) extensive
H-bonds (Fig. 8b) and (2) aromatic p–p stacking, and (3) by van
der Waals interactions to a good extent as suggested by a nearly
perfect match in size between the ester group and the cavity (Fig.
8c). From the computed dimeric structures, it can be seen that the
Fig. 6 Expanded 2D NOESY (223 K, “normal” CDCl3, 500 MHz,
mixing time = 500 ms) spectra of (a) 3b at 10 mM, showing the NOE
contacts between the bound water molecule and the amide protons of 3b,
(b) 5a at 5 mM, showing the NOE contacts between the bound water
molecule and the amide protons of 5a, (c) 4 at 10 mM, showing the NOE
contacts between the amide and ester methyl protons and (d) 5b at 5 mM,
showing the NOE contacts between the amide and amine protons.
1
the H NMR studies carried out using CDCl3 of varying water
contents (Fig. 4) where the chemical shifts of the amide protons
progressively downfield shift upon increasing water content in
CDCl3. For 5a, similar NOE contacts between “bound” water
and amide protons of 5a were observed while it appears that no
“free” water molecules are present in solution.
However, it still remains unclear to us as to why the 2D NOESY
study of 5b at 223 K in “normal” CDCl3 did not give rise to any
NOE contacts between water and 5b. Nevertheless, two important
interactions between the amide protons and the amine protons in
5b were observed (Fig. 6d). One of these two NOE contacts is in
˚
good accord with the short distance of 2.768 A between the end
amine group and the amide proton involved in the three-center H-
bonds at the other end, suggesting that the helical structure found
in the solid state also persists in solution.
Surprisingly, strong NOE contacts between the amide protons
and the end ester methyl protons in 4 were observed at 223 K
(Fig. 6c). Crystallographically, conformers 4A and 4B are found
with the methyl group staying away from the center of the cavity,
resulting in the interatomic distances between methyl protons and
˚
amide protons excluding Cbz amide proton much larger than 5 A.
Only in 4C, there exist at least two interatomic distances among
methyl protons and the three amide protons Ha, Hb and Hc that
This journal is
The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 1172–1180 | 1177
©

Fig. 7 Top and side views of the computationally optimized geometries of varying conformers for 4 at BYLYP/6-31G* level in CHCl3 at 223 K. The
shortest distances among methyl protons and amide protons Ha, Hb and Hc are shown below the structures. Shown below the distances are the relative
energies in kcal mol-1 among the five conformers normalized against the most stable conformer 4A1. Methyl protons and methyl carbon are in gray and
green balls, respectively, with the Cbz group represented as a yellow dummy atom.
Fig. 8 Top and side views of the computationally optimized geometries for dimeric structures of (a–c) (4C)2, (d) 4A·4B, (e) (5a)2 and (f) (5b)2 using the
Dreiding field force in the gas phase. In (a) a skeleton representation of (4C)2 where only the inner atoms are shown with carbon atoms in green and
yellow, respectively. In (b), the fragment at the ester end of the second molecule is shown with carbon atoms in yellow, and with only one of two identical
˚
sets of H-bonds of <2.8 A. In (c), front and back views of the fragmented ester group inside the cavity (COOCH3, H, C and O as black, yellow and red
balls) from the second molecule are presented as a CPK model, illustrating an excellent fit of the ester group into the cavity enclosed by 4C. In (f), the
two H-atoms and N-atoms involved in H-bonds at the amine end are shown in gray and blue balls, respectively. The binding energy for dimerization is
shown in kcal mol-1 below the structures. Clearly, 4C exhibits the highest energetic gain by self-dimerizing.
1178 | Org. Biomol. Chem., 2012, 10, 1172–1180
This journal is
The Royal Society of Chemistry 2012
©

Table 4 Computationally calculated chemical shifts in ppm with TMS as
the reference for the ester methyl protons from monomer 4C and dimer
(4C)2 in both gas phase and chloroform
overlap involving helical aromatic backbones is the best in dimer
4A·4B. Due to the existence of extensive H-bonds in (4C)2, two
strong H-bonds in (5b)2 and longer backbones in 5a and 5b, the
energetic gain among the four dimers increase in the order of (5a)2
< 4A·4B < (5b)2 < (4C)2. This order explains well the aggregation
behaviors displayed by 4, 5a and 5b (Fig. 5). Further, since 4C
clearly exhibits a much higher energetic gain (80.69 kcal mol-1) by
self-dimerizing with respect to the heterodimer 4A·4B (47.59 kcal
mol-1) found in the solid state, it is likely that 4C, not 4A or 4B,
is the predominant conformation adopted by 4 in solution, and
increasingly dimerizes upon decreasing the temperature. On the
basis of the computed structures, the shortest distances between
ester methyl protons and amides protons a, b and c are 3.65, 4.76
Dreiding force fielda
Gas phase
Chloroform
4C
(4C)2
4C
(4C)2
d/ppm
4.15
1.93
4.23
2.21
a Structures are optimized using the Dreiding force field, followed by
calculating chemical shifts using B3LYP/6-31G*.
the Dreiding force field as shown in Fig. 8a–c with its NMR
chemical shift values calculated at the B3LYP/6-31G* level in
both gas phase and chloroform (Table 4). Monomer 4C was
taken out of the optimized dimer structure and subjected to
the same treatment. Our calculations indeed show that 4C in
100% dimer state experiences a very substantial shielding effect
by the nearby aromatic ring currents, leading to a NMR chemical
shift of 2.21 ppm in chloroform as compared to 4C in 100%
monomer state with a chemical shift of 4.23 ppm. Although
a difference in chemical shift of methyl group between 4C and
(4C)2 may not be as large as 2.02 ppm in chloroform, a general
trend where dimerization involving two molecules of 4C causes
their methyl protons to experience more shielding effects and
thus to have a smaller chemical shift value with respect to
the monomeric 4C largely should hold true, at least computa-
tionally.
˚
and 2.76 A, respectively, accounting for both the observation of
strong NOEs between methyl protons and amide protons (Fig. 6c)
and why tetramer 4, seemingly possessing a water-binding cavity,
does not bind water molecules in solution (Fig. 4).
In the computed dimeric structure of (4C)2, the two ester methyl
groups are surrounded by aromatic benzene rings, and their
chemical shift will be affected by these aromatic ring currents.
1
A concentration-dependant analysis of H NMR of 4 that may
be composed of a mixture of monomer and dimer in varying
ratios in CDCl3 demonstrates an upfield shift for the two methyl
groups from 3.883 to 3.846 upon increasing the concentration
from 1 to 20 mM (Fig. 9a). For comparison, the same analysis
of trimer 3a shows a very small upfield shift from 3.916 to 3.904
ppm for its methyl group within the same concentration range
(Fig. 9b). Since higher concentrations favor the formation of more
dimer (4C)2, a decrease in chemical shift at higher concentrations
suggests the methyl groups to be more shielded from the NMR
magnetic field. If the self-trapping dimerization model proposed
in Fig. 8a–c is essentially correct or resembles its real structure
to a very good extent, the benzene rings surrounding the methyl
groups should produce a shielding rather than deshielding effect
in order to explain the observed 1H NMR shift for methyl groups.
In this regard, we carried out the computational investigations
of the chemical shifts of the methyl groups in monomer 4C and
dimer (4C)2. The structure of dimer (4C)2 was optimized using
Conclusions
In this article, we describe our detailed studies on the various
aspects of water-binding aquafoldamers, particularly, aquatrimer
3b and aquapentamers 5a and 5b. While a short pyridine-based
3b is found to encapsulate a water molecule in its nearly planar
cavity and subsequently traps a unconventional water dimer, a 3D-
shaped helical cavity as in longer oligomers such as 5a and 5b is able
to accommodate two water molecules that dimerize via a strong
intermolecular H-bond of different strengths. As a result of a
structural rigidity imposed by these aquafoldamers onto the water
molecules, these water dimers are destabilized by 1.33–2.99 kcal
mol-1 with regard with the most stable “unbound” water dimer.
Consistent with these solid-state structures, investigations by both
1D 1H NMR using CDCl3 containing varying water contents and
2D NOESY studies at low temperature provide a good support
to the ability of these aquafoldamers to bind water molecules in
solution. We envisioned that by elongating the helical backbones
either covalently or non-covalently, enlarged or longer 3D-shaped
cavities can be created for encapsulating larger water clusters of
diverse topographies in their interiors. On the other hand, aided by
computational molecular modelling, a self-trapping dimerization
model involving two molecules of 4 protruding into mutual cavities
is proposed, which is consistent with the observed (1) inability
of 4 to bind water in solution state, (2) abnormal aggregation
in 4, (3) strong NOE contacts that suggest a close proximity
between ester methyl protons and amide protons in 4, and (4)
a progressive downfield shifting of ester methyl protons in 4 upon
increasing the concentration. Whether this dimeric structure can
Fig. 9 1H NMR spectra (CDCl3, 500 MHz, 298 K) for ester methyl
protons from (a) 4 and (b) 3a, illustrating comparably different changes in
concentration-dependant chemical shift between 4 and 3a within the same
concentration range of 1–20 mM.
This journal is
The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 1172–1180 | 1179
©

be found in the solid state or not is a subject worth of further
investigation.
Eur. J., 2001, 7, 2810; (c) I. Huc, V. Maurizot, H. Gornitzka and J.-M.
Leger, Chem. Commun., 2002, 578; (d) J. Garric, J.-M. Le´ger and I.
Huc, Angew. Chem., Int. Ed., 2005, 44, 1954; (e) J. Garric, J.-M. Le´ger
and I. Huc, Chem.–Eur. J., 2007, 13, 8454.
8 M. C. Bheda, J. S. Merola, W. A. Woodward, V. J. Vasudevan and H.
W. Gibson, J. Org. Chem., 1994, 59, 1694.
Acknowledgements
9 For some selected examples of aromatic foldamers, see: (a) Y. Hamuro,
S. J. Geib and A. D. Hamilton, Angew. Chem., Int. Ed. Engl., 1994,
33, 446; (b) Y. Hamuro, S. J. Geib and A. D. Hamilton, J. Am. Chem.
Soc., 1996, 118, 7529; (c) Y. Hamuro, S. J. Geib and A. D. Hamilton, J.
Am. Chem. Soc., 1997, 119, 10587; (d) V. Berl, I. Huc, R. G. Khoury,
M. J. Krische and J. M. Lehn, Nature, 2000, 407, 720; (e) V. Berl, I.
Huc, R. Khoury and J.-M. Lehn, Chem.–Eur. J., 2001, 7, 2810; (f) E.
Kolomiets, V. Berl, I. Odriozola, A. M. Stadler, N. Kyritsakas and J.
M. Lehn, Chem. Commun., 2003, 2868; (g) J. Zhu, R. D. Parra, H. Q.
Zeng, E. Skrzypczak-Jankun, X. C. Zeng and B. Gong, J. Am. Chem.
Soc., 2000, 122, 4219; (h) B. Gong, H. Q. Zeng, J. Zhu, L. H. Yuan, Y.
H. Han, S. Z. Cheng, M. Furukawa, R. D. Parra, A. Y. Kovalevsky, J.
L. Mills, E. Skrzypczak-Jankun, S. Martinovic, R. D. Smith, C. Zheng,
T. Szyperski and X. C. Zeng, Proc. Natl. Acad. Sci. U. S. A., 2002,
99, 11583; (i) L. H. Yuan, H. Q. Zeng, K. Yamato, A. R. Sanford, W.
Feng, H. S. Atreya, D. K. Sukumaran, T. Szyperski and B. Gong, J.
Am. Chem. Soc., 2004, 126, 16528; (j) X. W. Yang, L. H. Yuan, K.
Yamamoto, A. L. Brown, W. Feng, M. Furukawa, X. C. Zeng and B.
Gong, J. Am. Chem. Soc., 2004, 126, 3148; (k) A. M. Zhang, Y. H. Han,
K. Yamato, X. C. Zeng and B. Gong, Org. Lett., 2006, 8, 803; (l) W.
Feng, K. Yamato, L. Q. Yang, J. Ferguson, L. J. Zhong, S. L. Zou, L.
H. Yuan, X. C. Zeng and B. Gong, J. Am. Chem. Soc., 2009, 131, 2629;
(m) P. S. Corbin, S. C. Zimmerman, P. A. Thiessen, N. A. Hawryluk
and T. J. Murray, J. Am. Chem. Soc., 2001, 123, 10475; (n) H. Jiang,
J.-M. Leger and I. Huc, J. Am. Chem. Soc., 2003, 125, 3448; (o) D.
Haldar, H. Jiang, J. M. Leger and I. Huc, Angew. Chem., Int. Ed., 2006,
45, 5483; (p) Q. Gan, C. Y. Bao, B. Kauffmann, A. Grelard, J. F. Xiang,
S. H. Liu, I. Huc and H. Jiang, Angew. Chem., Int. Ed., 2008, 47, 1715;
(q) J. L. Hou, X. B. Shao, G. J. Chen, Y. X. Zhou, X. K. Jiang and Z.
T. Li, J. Am. Chem. Soc., 2004, 126, 12386; (r) C. Li, S.-F. Ren, J.-L.
Hou, H.-P. Yi, S.-Z. Zhu, X.-K. Jiang and Z.-T. Li, Angew. Chem., Int.
Ed., 2005, 44, 5725; (s) J.-L. Hou, H.-P. Yi, X.-B. Sha, C. Li, Z.-Q.
Wu, X.-K. Jian, L.-Z. Wu, C.-H. Tung and Z.-T. Li, Angew. Chem.,
Int. Ed., 2006, 45, 796; (t) W. Cai, G. T. Wang, Y. X. Xu, X. K. Jiang
and Z. T. Li, J. Am. Chem. Soc., 2008, 130, 6936; (u) D. Kanamori, T.
A. Okamura, H. Yamamoto and N. Ueyama, Angew. Chem., Int. Ed.,
2005, 44, 969; (v) X. Li, C. L. Zhan, Y. B. Wang and J. N. Yao, Chem.
Commun., 2008, 2444; (w) J. Rebek, Acc. Chem. Res., 2009, 42, 1660;
(x) Q. Gan, Y. Ferrand, C. Bao, B. Kauffmann, A. Gre´lard, H. Jiang
and I. Huc, Science, 2011, 331, 1172.
10 For some selected examples of aromatic foldamers, see: (a) B. Qin, X.
Y. Chen, X. Fang, Y. Y. Shu, Y. K. Yip, Y. Yan, S. Y. Pan, W. Q. Ong,
C. L. Ren, H. B. Su and H. Q. Zeng, Org. Lett., 2008, 10, 5127; (b) Y.
Yan, B. Qin, Y. Y. Shu, X. Y. Chen, Y. K. Yip, D. W. Zhang, H. B. Su
and H. Q. Zeng, Org. Lett., 2009, 11, 1201; (c) B. Qin, C. L. Ren, R. J.
Ye, C. Sun, K. Chiad, X. Y. Chen, Z. Li, F. Xue, H. B. Su, G. A. Chass
and H. Q. Zeng, J. Am. Chem. Soc., 2010, 132, 9564; (d) Y. Yan, B. Qin,
C. L. Ren, X. Y. Chen, Y. K. Yip, R. J. Ye, D. W. Zhang, H. B. Su and
H. Q. Zeng, J. Am. Chem. Soc., 2010, 132, 5869; (e) B. Qin, C. Sun, Y.
Liu, J. Shen, R. J. Ye, J. Zhu, X.-F. Duan and H. Q. Zeng, Org. Lett.,
2011, 13, 2270; (f) B. Qin, W. Q. Ong, R. J. Ye, Z. Y. Du, X. Y. Chen, Y.
Yan, K. Zhang, H. B. Su and H. Q. Zeng, Chem. Commun., 2011, 47,
5419; (g) B. Qin, S. Shen, C. Sun, Z. Y. Du, K. Zhang and H. Q. Zeng,
Chem.–Asian J., 2011, 6, 3298; (h) C. L. Ren, S. Y. Xu, J. Xu, H. Y.
Chen and H. Q. Zeng, Org. Lett., 2011, 13, 3840; (i) C. L. Ren, F. Zhou,
B. Qin, R. J. Ye, S. Shen, H. B. Su and H. Q. Zeng, Angew. Chem., Int.
Ed., 2011, 50, 10612; (j) C. L. Ren, V. Maurizot, H. Q. Zhao, J. Shen,
F. Zhou, W. Q. Ong, Z. Y. Du, K. Zhang, H. B. Su and H. Q. Zeng, J.
Am. Chem. Soc., 2011, 133, 13930; (k) Z. Y. Du, C. L. Ren, R. J. Ye,
J. Shen, Y. J. Lu, J. Wang and H. Q. Zeng, Chem. Commun., 2011, 47,
12488.
Financial support by A*STAR BMRC research consortia (R-143-
000-388-305 to H. Z.) and by Environment and Water Industry De-
velopment Council and Economic Development Board (SPORE,
COY-15-EWI-RCFSA/N197-1 to H. Z.).
Notes and references
1 (a) A. S. Verkman, J. Cell Sci., 2011, 124, 2107; (b) P. Agre, M. Bonhivers
and M. J. Borgnia, J. Biol. Chem., 1998, 273, 14659; (c) M. Borgnia,
S. Nielsen, A. Engel and P. Agre, Annu. Rev. Biochem., 1999, 68, 425;
(d) G. M. Preston, T. P. Carroll, W. B. Guggino and P. Agre, Science,
1992, 256, 385.
2 (a) L. Strand, S. E. Moe, T. T. Solbu, M. Vaadal and T. Holen,
Biochemistry, 2009, 48, 5785; (b) B. Otto, N. Uehlein, S. Sdorra,
M. Fischer, M. Ayaz, X. Belastegui-Macadam, M. Heckwolf, M.
Lachnit, N. Pede, N. Priem, A. Reinhard, S. Siegfart, M. Urban and
R. Kaldenhoff, J. Biol. Chem., 2010, 285, 31253; (c) J. B. Heymann, P.
Agre and A. Engel, J. Struct. Biol., 1998, 121, 191.
3 (a) W. Q. Ong, H. Q. Zhao, Z. Y. Du, J. Z. Y. Yeh, C. L. Ren, L. Z.
W. Tan, K. Zhang and H. Q. Zeng, Chem. Commun., 2011, 47, 6416;
(b) W. Q. Ong, H. Q. Zhao, X. Fang, S. Woen, F. Zhou, W. L. Yap, H.
B. Su, S. F. Y. Li and H. Q. Zeng, Org. Lett., 2011, 13, 3194.
4 (a) L. X. Song, L. Bai, X. M. Xu, J. He and S. Z. Pan, Coord. Chem.
Rev., 2009, 253, 1276; (b) K. Ono, K. Tsukamoto, R. Hosokawa, M.
Kato, M. Suganuma, M. Tomura, K. Sako, K. Taga and K. Saito, Nano
Lett., 2009, 9, 122; (c) W. Si, X.-B. Hu, X.-H. Liu, R. Fan, Z. Chen, L.
Weng and J.-L. Hou, Tetrahedron Lett., 2011, 52, 2484; (d) Z. Zhang,
Y. Luo, J. Chen, S. Dong, Y. Yu, Z. Ma and F. Huang, Angew. Chem.,
Int. Ed., 2011, 50, 1397; (e) Z. Zhang, Y. Luo, B. Xia, C. Han, Y. Yu,
X. Chen and F. Huang, Chem. Commun., 2011, 47, 2417; (f) Z. Zhang,
B. Xia, C. Han, Y. Yu and F. Huang, Org. Lett., 2010, 12, 3285.
5 (a) J. R. Moon, A. J. Lough, Y. T. Yoon, Y. I. Kim and J. C. Kim, Inorg.
Chim. Acta, 2010, 363, 2682; (b) M. H. Mir, L. Wang, M. W. Wong and
J. J. Vittal, Chem. Commun., 2009, 4539.
6 For some recent reviews on foldamers, see: (a) S. H. Gellman, Acc.
Chem. Res., 1998, 31, 173; (b) K. D. Stigers, M. J. Soth and J. S.
Nowick, Curr. Opin. Chem. Biol., 1999, 3, 714; (c) B. Gong, Chem.–Eur.
J., 2001, 7, 4336; (d) D. J. Hill, M. J. Mio, R. B. Prince, T. S. Hughes
and J. S. Moore, Chem. Rev., 2001, 101, 3893; (e) R. P. Cheng, S. H.
Gellman and W. F. DeGrado, Chem. Rev., 2001, 101, 3219; (f) M. S.
Cubberley and B. L. Iverson, Curr. Opin. Chem. Biol., 2001, 5, 650;
(g) A. R. Sanford and B. Gong, Curr. Org. Chem., 2003, 7, 1649; (h) A.
R. Sanford, K. Yamato, X. Yang, L. Yuan, Y. Han and B. Gong, Eur.
J. Biochem., 2004, 271, 1416; (i) C. Schmuck, Angew. Chem., Int. Ed.,
2003, 42, 2448; (j) I. Huc, Eur. J. Org. Chem., 2004, 17; (k) R. P. Cheng,
Curr. Opin. Struct. Biol., 2004, 14, 512; (l) G. Licini, L. J. Prins and
P. Scrimin, Eur. J. Org. Chem., 2005, 969; (m) Z. T. Li, J. L. Hou, C.
Li and H. P. Yi, Chem.–Asian J., 2006, 1, 766; (n) X. Li and D. Yang,
Chem. Commun., 2006, 3367; (o) C. Z. Li, X. K. Jiang, Z. T. Li, X.
Gao and Q. R. Wang, Chin. J. Org. Chem., 2007, 27, 188; (p) C. M.
Goodman, S. Choi, S. Shandler and W. F. DeGrado, Nat. Chem. Biol.,
2007, 3, 252; (q) B. Gong, Acc. Chem. Res., 2008, 41, 1376; (r) D.
Haldar, Curr. Org. Synth., 2008, 5, 61; (s) Z. T. Li, J. L. Hou and C. Li,
Acc. Chem. Res., 2008, 41, 1343; (t) B. Baptiste, F. Godde and I. Huc,
ChemBioChem, 2009, 10, 1765; (u) W. S. Horne and S. H. Gellman, Acc.
Chem. Res., 2008, 41, 1399; (v) I. Saraogi and A. D. Hamilton, Chem.
Soc. Rev., 2009, 38, 1726; (w) D. Haldar and C. Schmuck, Chem. Soc.
Rev., 2009, 38, 363; (x) X. Zhao and Z. T. Li, Chem. Commun., 2010, 46,
1601.
7 (a) V. Berl, I. Huc, R. Khoury and J.-M. Lehn, Chem.–Eur. J., 2001,
7, 2798; (b) V. Berl, I. Huc, R. G. Khoury and J.-M. Lehn, Chem.–
11 S. L. Mayo, B. D. Olafson and W. A. Goddard III, J. Phys. Chem., 2009,
94, 8897.
1180 | Org. Biomol. Chem., 2012, 10, 1172–1180
This journal is
The Royal Society of Chemistry 2012
©