Triple helix formation in amphiphilic discotics: Demystifying solvent effects in supramolecular self-assembly

Article abstract of DOI:10.1021/ja4104183

A set of chiral, amphiphilic, self-assembling discotic molecules based on the 3,3′-bis(acylamino)-2,2′-bipyridine-substituted benzene-1,3,5-tricarboxamide motif (BiPy-BTA) was prepared. Amphiphilicity was induced into the discotic molecules by an asymmetrical distribution of alkyl and oligo(ethylene oxide) groups in the periphery of the molecules. Small-angle X-ray scattering, cryogenic transmission electron microscopy, and circular dichroism spectroscopy measurements were performed on the discotic amphiphiles in mixtures of water and alcohol at temperatures between 0 C an 90 C. The combined results show that these amphiphilic discotic molecules self-assemble into supramolecular fibers consisting of either one or three discotic molecules in the fiber cross-section and that the presence of water induces the bundling of the supramolecular fibers. The rich phase behavior observed for these molecules proves to be intimately connected to the mixing thermodynamics of the water-alcohol mixtures.

Full text of DOI:10.1021/ja4104183

Article
pubs.acs.org/JACS
Triple Helix Formation in Amphiphilic Discotics: Demystifying
Solvent Effects in Supramolecular Self-Assembly
Martijn A. J. Gillissen, Marcel M. E. Koenigs, Jolanda J. H. Spiering, Jef A. J. M. Vekemans,
Anja R. A. Palmans,* Ilja K. Voets, and E. W. Meijer*
Institute of Complex Molecular Systems, Laboratory of Macromolecular and Organic Chemistry, Eindhoven University of
Technology, P.O. Box 513, 5600 MB, Eindhoven, The Netherlands
S
* Supporting Information
ABSTRACT: A set of chiral, amphiphilic, self-assembling
discotic molecules based on the 3,3′-bis(acylamino)-2,2′-
bipyridine-substituted benzene-1,3,5-tricarboxamide motif
(BiPy-BTA) was prepared. Amphiphilicity was induced into
the discotic molecules by an asymmetrical distribution of alkyl
and oligo(ethylene oxide) groups in the periphery of the
molecules. Small-angle X-ray scattering, cryogenic transmission
electron microscopy, and circular dichroism spectroscopy
measurements were performed on the discotic amphiphiles in
mixtures of water and alcohol at temperatures between 0 °C an 90 °C. The combined results show that these amphiphilic
discotic molecules self-assemble into supramolecular fibers consisting of either one or three discotic molecules in the fiber cross-
section and that the presence of water induces the bundling of the supramolecular fibers. The rich phase behavior observed for
these molecules proves to be intimately connected to the mixing thermodynamics of the water−alcohol mixtures.
We here illustrate how the self-assembly behavior in water of
an amphiphilic discotic molecule, encoded to form bundled
structures by incorporating two orthogonal interactions, is
affected by mixing water and an alcohol. We select the 3,3′-
bis(acylamino)-2,2′-bipyridine-substituted benzene-1,3,5-tricar-
boxamide discotic (BiPy-BTA), which is known to self-
assemble into 1D helical fibers in apolar and polar organic
solvents and even in water, due to π−π interactions.43−49 In
addition, the BiPy-BTA discotic is robust to modification of its
periphery. The amphiphilicity of the molecule is induced by
placing alkoxy groups on the three R1 positions and
oligo(ethylene oxide) groups on the six R2 positions, thus
incorporating the hydrophobic effect as an orthogonal
interaction (Figure 1). Breaking the C3 symmetry of the
BiPy-BTA discotic facilitates the hierarchical self-assembly into
bundled structures. To introduce a preference for a single
helical handedness, we incorporate chiral alkyl or oligo-
(ethylene oxide) side chains. This strategy results in the
formation of 1D helical fibers in pure isopropanol, when only
the π−π interactions are operative, while triple helical bundles
are formed in isopropanol−water (IPA−H2O) mixtures, in
which the hydrophobic effect starts playing a role. A detailed
study of the self-assembly as a function of solvent composition
and temperature reveals that the solvent drives the observed
morphological transitions, as suggested by Trappe and co-
workers for poly(N-isopropylacrylamide) in water−alcohol
mixtures.50 We observe three distinct structural regimes with
INTRODUCTION
■
The self-assembly of molecular components is an intriguing,
low-energy pathway toward materials with nanoscopic order.1 A
widely studied class of self-assembling molecules consists of
disc-shaped molecules that form one-dimensional (1D)
fibers.2−4 The creation of more complex superstructures with
advanced functionality requires not only control over the
assembly of molecules into 1D fibers but also control over the
next step, assembly of fibers into bundles or fibrils. Such
bundled structures occur frequently in natural proteins and
have also been realized in synthetic proteins and peptides.5 The
leucine zipper6 structural motif and the tropomyosin coiled
coil7 are stabilized to a large extent by hydrophobic
interactions,8 while other naturally occurring structures such
as the collagen triple helix9 are held together mainly by
hydrogen-bonding interactions. Inspired by the coiled-coils,
double and triple helices found in nature, chemists successfully
developed several elegant approaches toward these architec-
tures from fully synthetic building blocks in organic media.10−31
Driven by the potential of self-assembled materials in
biomedical applications, much current research is focused on
developing fully synthetic systems that form fibrillar structures
in aqueous media.32−36 Translating the design rules for self-
assembly in apolar organic solvents to water is not trivial, as
additional factors, such as the hydrophobic effect and
interference of water with hydrogen bonding interactions,
start playing a role. In addition, to help compatibilize synthetic
self-assembling systems with water, a cosolvent is frequently
added, resulting in self-assembly behavior that is not necessarily
predictable.37−42
Received: October 10, 2013
Published: December 7, 2013
© 2013 American Chemical Society
336
dx.doi.org/10.1021/ja4104183 | J. Am. Chem. Soc. 2014, 136, 336−343

Journal of the American Chemical Society
Article
proposed structures (see Supporting Information section 3 for
details of the synthesis).
Characterization of Compounds 1−3. A combination of
spectroscopic, microscopic, and scattering techniques was
applied to investigate how solvent composition, temperature,
and concentration affect the supramolecular structures formed
upon self-assembly of compound 1. Circular dichroism (CD)
experiments were performed as a function of the volume
fraction of isopropanol (ϕIPA), temperature, and concentration
to investigate under which conditions 1 self-assembles. The
shape and size of the supramolecular structures obtained were
subsequently investigated by cryogenic transmission electron
microscopy (cryo-TEM) and small-angle X-ray scattering
(SAXS). This combination of experimental techniques reveals
that solvent composition and temperature determine the nature
of the aggregates, and three distinct regimes can be
distinguished. The results are discussed in detail below.
Circular Dichroism Spectroscopy. CD spectra of 1 were
measured in IPA−H2O mixtures at 0 °C. The volume fraction
of isopropanol (ϕIPA) was 1, 0.5, or 0.25 and the concentration
of 1 was constant at 3 × 10−4 M. In all solvent mixtures, Cotton
effects were observed, indicating the presence of helical objects
of one predominant handedness (Figure 2). However, the
shape of the CD spectra differs considerably between the three
solvent compositions.
At ϕIPA = 1, the CD spectrum shows minima at λ = 235, 286,
367, and 385 nm, with a molar circular dichroism (Δε) at 286
nm of −39 L mol−1 cm−1 (Figure 2F). The Cotton effect has a
shape and magnitude similar to those previously observed for
symmetrical, aliphatic BiPy-BTA compounds in apolar organic
solvents.46 At ϕIPA = 0.5, the Cotton effect shows a maximum at
λ = 335 nm (Δε = 119 L mol−1 cm−1) and minima at λ = 263
and 302 nm (Figure 2E). Finally at ϕIPA = 0.25 we find a
Cotton effect with a maximum at λ = 310 nm (Δε = 39 L mol−1
cm−1) and minima at λ = 235, 283, and 385 nm shown in
Figure 2D. These observations show that whereas 1 forms
Figure 1. Chemical structure of amphiphilic BiPy-BTA discotics 1−3
and their hierarchical self-assembly, first into a supramolecular fiber
and then into a triple helical bundle.
boundaries that coincide with the extrema in the excess
enthalpy of mixing (ΔHmixE) of IPA−H2O mixtures.
RESULTS AND DISCUSSION
■
Synthesis of Compounds 1−3. Compounds 1−3 were
synthesized in a stepwise approach (Scheme 1) starting from 5-
(methoxycarbonyl)benzene-1,3-dicarboxylic acid (4).51 In the
initial step, the diacid was converted to the diacid dichloride
using oxalyl chloride, after which it was coupled to the
bipyridine monoamines 5a or 5b that contain water-soluble
oligo(ethylene oxide) chains. The obtained monoesters 6a and
6b were then saponified using lithium hydroxide at 85 °C and
subsequently neutralized with oxalic acid. In the final step, the
monoacids 7a and 7b were converted into the acid chloride
using Ghosez’s reagent52 and coupled to the alkyl-decorated
bipyridines 8a,b to obtain the final products 1−3. All end-
1
products were analyzed by H and 13C NMR, MALDI−TOF
MS, and elemental analysis; the results were consistent with the
Scheme 1. Synthesis of Amphiphilic BiPy-BTA Discotics 1−3
337
dx.doi.org/10.1021/ja4104183 | J. Am. Chem. Soc. 2014, 136, 336−343

Journal of the American Chemical Society
Article
Figure 2. (A−C) Schematic depiction of the three regimes in self-assembly of 1 in IPA−H2O. CD-spectra of 1 at c = 3 × 10−4 M and T = 0 °C for
(D) ϕIPA = 0.25, (E) ϕIPA = 0.5, and (F) ϕIPA = 1. Cryo-TEM images of 1 at c = 3 × 10−4 M for (G) ϕIPA = 0.25, (H) ϕIPA = 0.75, and (I) ϕIPA = 1.
SAXS profiles (symbols) and form factor fits (lines) for 1 in IPA−H2O mixtures at c = 3 × 10−4 M and T = 20 °C for (J) ϕIPA = 0.25, (K) ϕIPA = 0.5,
and (L) ϕIPA = 1.
chiral aggregates in all solvent mixtures, the arrangement of the
discotic molecules inside the aggregates is strongly determined
by the solvent composition.
allows quantification of the melting temperature (Tm) and the
enthalpy release for bond formation (ΔH) summarized in
Figure 4 and Table S1 (see Supporting Information section 4
To study the influence of temperature on the formation of
self-assembled structures of 1, we performed temperature-
dependent CD experiments. In all cases, the solutions of 1 were
cooled from 80 °C to 0 °C and the CD effect was monitored at
one specific wavelength (Figure S1, Supporting Information).
In pure isopropanol (ϕIPA = 1), a sigmoidal temperature-
dependence of the Cotton effect is observed when monitoring
Δε at 286 nm (Figure 3A). Such sigmoidal curves are indicative
for an isodesmic self-assembly mechanism.53 The temperature-
dependent cooling curves were fit to an isodesmic model, which
Figure 4. (A) Transition temperature (Tm or LCST) for compound 1
in regimes I, II, and III in IPA−H2O mixtures. (B) Enthalpy release
upon bond formation (ΔH) for compound 1 in regimes I, II, and III in
IPA−H2O mixtures.
for details). At a concentration of 3 × 10−4 M and at ϕIPA = 1,
ΔH = −142 kJ mol−1 and Tm = 328 K. Upon decreasing the
concentration of 1 in IPA, Tm shifts to lower temperatures
while ΔH remains rather constant.
In the range of 0.4 ≤ ϕIPA ≤ 0.9, the CD spectra are similar
to those observed at ϕIPA = 0.5. In addition, monitoring Δε at
340 nm as a function of temperature in the range of 0.4 ≤ ϕIPA
≤ 0.9 shows a sigmoidal temperature-dependence of the CD-
Figure 3. (A) Temperature-dependent CD spectra for 1 at c = 3 ×
10−4 M and T = 80 to −10 °C in ϕIPA = 1. (B) UV heating curve
monitored at λ = 310 nm for 1 at c = 3 × 10−5 M in ϕIPA = 0.1.
338
dx.doi.org/10.1021/ja4104183 | J. Am. Chem. Soc. 2014, 136, 336−343

Journal of the American Chemical Society
Article
effect (Figure 3B). The melting temperatures (Tm) for these
solvent compositions are substantially lower than that observed
in pure isopropanol. For example, at a concentration of 3 ×
10−4 M and at ϕIPA = 0.5, Tm is 280 K with a corresponding ΔH
of −166 kJ mol−1. The Tm does not change strongly with ϕIPA
or the concentration of 1. The ΔH, on the other hand, has a
with a length within the experimentally probed q range. At high
q values, form factor oscillations are observed, indicating the
presence of a well-defined cross-section. The position of the
first minimum is significantly different for high and low ϕIPA
,
indicating a significant difference in cross-section dimensions in
these two regions. We applied two methods to collect
information on the dimensions of the self-assembled objects,
the indirect Fourier transform (IFT), and form factor analysis.
The IFT is a model-independent method to obtain
information on the scattering object by generating a pair
distance distribution function (PDDF, p(r)) of the scattering
particle.54 This PDDF gives the probability of finding a point at
a distance r starting from a random point within the particle,
weighed by the difference in electron density between these
two points if it is not constant over the whole particle.
Therefore, the maximal dimension of a particle corresponds to
the maximal distance (r) of the PDDF, where p(r) ≠ 0. The
PDDF typically exhibits one or several peaks at distances
characteristic for the object studied. For example, in the case of
a sphere, this maximum occurs approximately at the radius. In
the case of objects that are elongated in one dimension, a
PDDF of the cross-section can be determined (pc(r)). Figure
5B and Figure S4 (Supporting Information) show the obtained
pc(r) and p(r). The pc(r) of the aggregate in pure isopropanol is
a single Gaussian-like peak with a maximum at r = 1.3 nm,
extending to r = 4 nm (Figure 5B). This shape is typical for an
object having a circular cross-section with one fixed scattering
length density (ρ) over the entire cross-section.54 However, at
ϕIPA = 0.25 and 0.3, the pc(r) of the aggregate show two
maxima at r = 0.9 and 4.8 nm, extending to r = 7.5 nm. This
shape is typical for an object having two distinct scattering
length densities along the cross-section, i.e., a core−shell
object.54 Finally, at the intermediate solvent compositions, the
p(r) show two maxima at r = 2.1 and 5.5 nm, extending to r =
8, 12, and 18 nm for ϕIPA = 0.75, 0.5 and 0.4, respectively
(Figures 5B and Figure S4, Supporting Information). These
PDDFs are typical for objects having a core−shell structure
where the length of the objects becomes larger going from ϕIPA
= 0.75 to 0.4.
strong nonlinear dependence on ϕIPA
.
At ϕIPA ≤ 0.3, the CD spectrum changes shape, and
temperature-dependent CD experiments reveal that the
magnitude of the CD-effects hardly changes with increasing
temperature. The UV trace of the cooling curve shows an
abrupt increase in intensity above ca. 340 K, indicative of a
lower critical solution temperature (LCST) as shown in Figure
3B. The temperature of the LCST transition is virtually
independent of the BiPy-BTA concentration and ϕIPA as shown
in Figure 4A.
Cryogenic Transmission Electron Microscopy. The
structures of the aggregates formed by 1 in IPA−H2O mixtures
at a concentration of 3 × 10−4 M were visualized by cryo-TEM
(Figure S3, Supporting Information). These experiments reveal
a remarkable difference in aggregate structure upon changing
the solvent composition from pure IPA to predominantly
water. In pure isopropanol (Figure 2I), long, fibrous objects are
present. The length of these aggregates ranges from 100 to
several 100 nm. In contrast, the cross-section of the fibers is
constant with a diameter of approximately 7 nm. At
intermediate isopropanol content, ϕIPA = 0.75 (Figure 2H),
low aspect ratio objects are observed. The length of these
aggregates ranges from 10 to 50 nm, whereas the cross-section
is constant at approximately 10 nm. Finally, at low isopropanol
content, ϕIPA = 0.25 (Figure 2G), we observe very long fibrous
objects with lengths of several micrometers. The cross-section
is constant at approximately 11 nm. It should be noted that
contrast, resolution, and focusing issues hamper a more exact
determination of the dimensions of the cross-section.
Small Angle X-ray Scattering. The aggregate structure of
1 in IPA−H2O mixtures at a concentration of 3 × 10−4 M was
studied in more detail by SAXS. The scattering profiles are
shown in Figures 5A, and Figures S4 and S5 (Supporting
Information). At high and low ϕIPA the scattering profiles at low
q values exhibit power law scaling behavior, I ∝ q−1, indicative
of rigid objects elongated in one dimension with a length
beyond the experimentally probed q range. At the intermediate
solvent compositions we see a I ∝ q0 dependence of the
scattered intensity at low q, typical for low aspect ratio objects
The PDDFs show that aggregates of 1 can be described as
objects elongated along one direction, which at ϕIPA < 1 have a
core−shell cross-section and at ϕIPA = 1 have a cross-section
with only a core (without a shell). This can be depicted as a
(core−shell) cylinder having a length (L), a (core) radius (R),
and a shell thickness (T), each having their own scattering
length density (ρ) depicted in Figure 6.
To further quantify the dimensions of the aggregates, the
SAXS profiles were fit to a model for either a cylinder (cyl) or a
core−shell cylinder (cs−cyl). The SAXS profiles and form
factor fits are shown in Figures 5A and Figures S4 and S5
(Supporting Information), and the extracted aggregate
dimensions are given in Table 1 and Table S3 (Supporting
Information). At high and low isopropanol contents, where the
q−1 power law scaling of the scattered intensity is observed, the
aggregates are too long to determine their length by the SAXS
experiments, and these are therefore set at 1000 nm. In pure
isopropanol the aggregates have a radius (R) of ca. 1.65 nm,
having a homogeneous scattering length density (ρ) over the
whole cross-section. Experiments at ϕIPA = 0.3 and 0.25 reveal
that the aggregates have an inhomogeneous ρ over the cross-
section, i.e., a core−shell structure, with an inner radius (R) of
ca. 1.5 nm and an outer radius (R+T) of ca. 3.5 nm. At
intermediate isopropanol contents, the length of the aggregates
Figure 5. SAXS data for 1 in IPA−H2O mixtures at c = 3 × 10−4
and T = 20 °C: (A) SAXS profiles (symbols) and form factor fits
(lines) (data is offset by a factor of 102 (ϕIPA = 0.5) and 104 (ϕIPA
0.25) for clarity). (B) Cross-sectional PDDFs and PDDFs for ϕIPA = 1,
0.25, and 0.5.
M
=
339
dx.doi.org/10.1021/ja4104183 | J. Am. Chem. Soc. 2014, 136, 336−343

Journal of the American Chemical Society
Article
superstructure, in which the aliphatic substituents cluster in the
center of the fiber. This separation of aliphatic and ethylene
oxide chains results in a contrast difference along the fiber
cross-section. It is likely that the addition of water will cause the
helical aggregates to assemble in such a way that the
hydrophobic parts are shielded from the hydrophilic outside.
Because the discotics have a flexible propeller shape but are
never complete flat, the most likely conformation is a triple
helix of helices. In fact, MD calculation performed previously by
Amabilino and co-workers43 on bipyridine-based discotics
showed that three primary fibers in an M conformation wrap
spontaneously one around the other to converge into a coiled
superhelix. In our case, with the hydrophobic effect present,
such an organization is even more likely.
Figure 6. (A) Schematic depiction of a cylinder and core−shell
cylinder model. (B) Depiction of scattering length density (ρ) along
the cylinder cross-section and contrast (Δρ) of core and shell for ϕIPA
= 0.5 (arrows).
Three Regimes of Self-Assembly. The above experiments
show that we can distinguish three regimes for the self-assembly
of compound 1 and in each regime, the structural and
mechanistic properties of aggregates formed by 1 differ (Figure
2).
Table 1. Fit Parameters Obtained from Form Factor Fits to
SAXS Profiles of Compound 1 in IPA−H2O Mixtures,
Where L = Cylinder Length, R = Core Cross-Section, and T
= Shell Thickness
Regime I. In the low isopropanol content regime, at ϕIPA
≤
0.3, we observe long fibrous aggregates having a core−shell
cross-section. The aggregates have a preferential chirality and
consist of three BiPy-BTA molecules in their cross-section: a
triple helical bundle.
dimensions (nm)
ϕIPA
regime
L
a
R
T
1
III
II
II
II
I
1000
1.7 0.1
1.4 0.1
1.6 0.2
1.6 0.1
1.4 0.1
1.6 0.1
−
−
Regime II. In the intermediate isopropanol content regime,
at 0.4 ≤ϕIPA ≤0.9, we observe low aspect ratio objects having a
core−shell cross-section and a preferential chirality.
0.75
0.5
8.5 1.7
5.1 0.7
1.7 0.3
1.8 0.2
2.3 0.2
2.1 0.2
0.4
13.8 2.0
a
0.3
1000
Regime III. In the high isopropanol content regime, at ϕIPA =
a
0.25
I
1000
1, we observe long fibrous aggregates having a cross-section
with a single, homogeneous contrast. The aggregates have a
preferential chirality and consist of one BiPy-BTA molecule in
their cross-section: a one−dimensional helical fiber.
a
This value was set as fixed in the fitting procedure.
falls within the experimentally probed range. Here, we observe
the presence of low aspect ratio objects, with an average length
of 5 to 15 nm. In all cases an inner radius of ca. 1.5 nm is
determined. Where for ϕIPA = 0.4 and 0.5 an outer radius of ca.
3.2 nm is observed, no shell contrast is seen for ϕIPA = 0.75.
The latter observation could be caused by the change in shell
contrast (Δρshell), which is the scattering length density
difference between the shell and the surrounding solvent,
gradually diminishing the “visibility” of the shell (Figure 6B).
Overall we observe that the fiber radius in the regime where
core−shell structures are formed is approximately twice that
observed for fibers in pure isopropanol. The latter is consistent
with bundle formation of the smaller isopropanol regime fibers
into bundles of 3 to 4 fibers. This observation is further
substantiated by calculation of the mass per unit length (ML) of
the aggregates from the SAXS profiles. With this information
we can estimate the number of discotic molecules in the cross-
section of the aggregate (Table S5, Supporting Information).
Whereas the cross-section at ϕIPA = 1 contains one discotic
molecule, the ϕIPA = 0.25 aggregate consists of three discotic
molecules.
To summarize, the SAXS experiments corroborate the ϕIPA
dependence of aggregate length apparent in the cryo-TEM
images. Whereas at high and low isopropanol content long
fibers are present, in the intermediate regime low aspect ratio
objects are observed. The SAXS experiments give additional
information on the cross-section of the observed fibers. The
increased diameter of the fibers upon addition of water is
pointing to the formation of a core−shell structure of the cross-
section. We propose that the observed core−shell cross-section
is the result of bundling of three 1D fibers into a coiled
The findings discussed above unambiguously demonstrate
that the combination of solvent composition and temperature
dictate the structure of the supramolecular polymers formed by
1 and their formation mechanism. Inspired by recent work by
Trappe and co-workers, who showed that the LCST transition
in poly(N-isopropylacrylamide) in mixed solvents is correlated
to the solvent mixing thermodynamics, we decided to
investigate whether such correlation exists in the self-assembly
process of compounds 1−3.50
Enthalpy of Mixing of IPA and H2O. We first focus on the
(excess) enthalpy of mixing (ΔHmixE) of IPA and H2O, which is
known to be strongly asymmetric in its composition depend-
ence.55−57 The excess enthalpy of mixing is one of the several
properties that reflect the well-documented anomalous proper-
ties of many water−alcohol mixtures.55 The ΔHmix shows a
E
temperature-dependent minimum at low isopropanol content
(ϕmin) and maximum at high isopropanol content (ϕmax). Both
ϕmin and ϕmax shift to higher ϕIPA upon decreasing temperature
(Figure 7A).
There is a remarkable correlation between the dependence of
the ΔHmixE and regimes I, II, and III on ϕIPA. At 20 °C, we find
the transition from regime I to II (ϕI→II) at ϕIPA ∼ 0.35, a
solvent composition almost identical to the minimum ϕmin
∼
0.35 in Figure 7A. In addition, the transition from regime II to
III (ϕII→III) is situated at ϕIPA ∼ 0.95, which is close to the
maximum at ϕmax ∼ 0.93 in Figure 7A. To test whether the
boundary between regime II and III in the self-assembly of
compound 1 is indeed coupled to the excess enthalpy of
mixing, we performed a temperature-dependent experiment at
ϕIPA = 0.95.
340
dx.doi.org/10.1021/ja4104183 | J. Am. Chem. Soc. 2014, 136, 336−343

Journal of the American Chemical Society
Article
LCSTs are at slightly higher temperatures (340−350 K) as
compared to compound 1 in IPA−H2O (335 K). Alternatively,
if the solvent composition is between ϕmin and ϕmax, we observe
a sigmoidal temperature-dependence of the CD effect,
indicative of regime II behavior. Whereas in a 1:1 mixture the
Tm of compound 1 in EtOH−H2O (276 K) is similar to that in
IPA−H2O (278 K), the Tm for compounds 2 and 3 in IPA−
H2O are shifted to higher temperatures of 285 K and 295 K,
respectively (Figure 8A and Figure S1B, Supporting Informa-
tion).
In pure ethanol and isopropanol, all compounds show a
sigmoidal temperature-dependence, indicative for regime III
behavior. The Tm in regime III is similar for 2 at 326 K as
compared to compound 1 in IPA where it was 328 K. For
compound 3, Tm was not within the experimentally accessible
temperature range and exceeds 333 K. Next to this isodesmic
self-assembly, at lower temperatures a second transition (T2) at
ca. 300−310 K is observed which coincides with the appearance
of linear dichroism in the sample (Figure 8A). This indicates
that at these temperatures aggregates are formed, which are
sufficiently long to align within the sample.
The shape and size of the structures obtained for 1 in
EtOH−H2O, 2 in IPA−H2O, and 3 in IPA−H2O were
characterized by SAXS. The results are very similar to those
obtained for 1 in IPA−H2O. Form factor fits reveal core−shell
cylinder structures under all conditions at 20 °C (Figure 8B).
The aggregate cross-section dimensions are similar to those of
compound 1 in IPA−H2O mixtures, having a core radius of ca.
1.5 nm and an outer radius of ca. 3.5 nm (Table S4, Supporting
Information). This can be readily explained, as the cross-
sectional size of the self-assembled structures is determined by
the size of the amphiphilic BiPy-BTA discotic molecules, which
is similar for all compounds. The length of the aggregates is not
attainable in most scattering profiles.
Figure 7. (A) Temperature- and ϕIPA-dependence of ΔHmixE for IPA−
H2O (figure based on data from refs 38 and 39). (B) Temperature-
dependent CD spectra for 1 at c = 3 × 10−5 M in ϕIPA = 0.95.
Figure 7A shows that at 80 °C, ϕmax is well below 0.95 while
at 0 °C, ϕmax shifts to values larger than 0.95. This would
suggest that cooling a solution of 1 in ϕIPA = 0.95 should induce
a change in the nature of the aggregates typical for regime III to
an aggregate typical for regime II. To test this, the CD spectra
of a solution of 1 in ϕIPA = 0.95 were measured at different
temperatures. Figure 7B shows that at 10 °C, a Cotton effect
typical for regime III is observed. Upon cooling the solution to
0 °C, the CD-effect changes significantly and is typical for
regime II. This clearly shows that the self-assembly of 1 in
water−IPA mixtures is intricately linked to the mixing behavior
of this solvent mixture.
Changing Cosolvent and Peripheral Substitution. The
positions of ϕmin and ϕmax depend not only on temperature but
also on the alcohol used as cosolvent. In EtOH−H2O at 25 °C,
ϕmin is ca. 0.4, and there is no maximum in ΔHmixE. At elevated
temperatures, there is a maximum at ϕmax ∼ 0.77 at T = 75
°C.55 To further substantiate the relation between the phase
E
behavior of amphiphilic BiPy-BTAs and the extrema of ΔHmix
,
A Phase Diagram. The previous observations can be
summarized into the generalized phase diagram depicted in
Figure 9. In this phase diagram, there are ϕcosolvent-dependent
we performed experiments on compound 1 in ethanol−water
mixtures (EtOH−H2O). Additionally, we investigated the
behavior of BiPy-BTA derivatives 2 and 3 in isopropanol−
water mixtures to evaluate to what extent the nature of the side
chains influences the self-assembly properties of the BiPy-BTA
derivatives. The results are summarized in Figure 8 and Tables
S2 and S4 (Supporting Information).
In all three studies, the general features are similar to those
described for 1 in IPA−H2O. An LCST transition is observed
when the cosolvent content is below ϕmin in all samples. The
Figure 9. Generalized phase diagram of amphiphilic BiPy-BTA
derivatives in alcohol−water mixtures.
boundaries between different aggregate structures linked to the
positions of ϕmin and ϕmax. Below ϕmax, hydrophobic
interactions lead to bundle formation of the amphiphilic
discotics while above ϕmax, there is no driving force for this
lateral association. In the intermediate regime the hydrophobic
domains aggregate, but elongation into long, triple helical fibers
occurs only below a certain temperature.
Figure 8. (A) Transition temperature (Tm or LCST) for 1 in EtOH−
H2O at c = 3 × 10−5 M and 2 and 3 in IPA−H2O mixtures at c = 3 ×
10−4 M. (B) SAXS profiles for 1 in EtOH−H2O and 2 and 3 in IPA−
H2O mixtures at c = 3 × 10−4 M and T = 20 °C (data is offset by a
factor of 102 (2) and 104 (3) for clarity).
341
dx.doi.org/10.1021/ja4104183 | J. Am. Chem. Soc. 2014, 136, 336−343

Journal of the American Chemical Society
Article
There are also temperature-dependent boundaries between
different aggregate structures. The LCST transition in regime I
is weakly dependent on the nature of the cosolvent and the
side-chain substitution. The temperature-dependent boundaries
in regimes II and III are determined to a large extent by the
nature of the side chains of the compound (R1 and R2). The
transition temperature (Tm) for compound 1 in regime II is
virtually the same in isopropanol and ethanol. On the other
hand, Tm is higher for compound 3 than for compound 1,
indicating that the potential for packing/crystallization of the
aliphatic side chains is important. This effect is also observed in
mixing experiments of compounds 1 and 3 in pure isopropanol.
Here we find a gradual, linear decrease of the T2 transition
temperature upon addition of 1 to a solution of 2 (Figure S2,
Supporting Information). This indicates that disrupting the
side-chain packing by mixing in a branched aliphatic side chain
reduces the tendency of the aggregates to bundle.
It is noteworthy that the phase behavior of amphiphilic
synthetic polymers and amphiphilic self-assembled structures is
so similar in water−cosolvent mixtures. Whereas a full
explanation of the observed phase behavior is not trivial, it is
clearly related to the properties of the solvent mixture itself.
Recent progress in the study of these complex water−cosolvent
mixtures has revealed that even though they are macroscopi-
cally mixed, they phase separate at the nanoscopic level.58−60
This could be the underlying mechanism for the phase behavior
observed in these systems. The amphiphilic (small) molecules
present in these complex mixtures “feel” the phase separation of
the amphiphilic cosolvents and adapt their structure to
accommodate this.
structures could help us understand the pathways by which
the final structures in water are formed. Therefore, further
investigation of this solvent effect and elucidating the
mechanism behind it are crucially important to enable full
control over hierarchical aggregation processes.
ASSOCIATED CONTENT
* Supporting Information
■
S
Synthetic procedures, experimental details, spectroscopic SAXS
and cryo-TEM data, equations used, and derived parameters.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Authors
a.palmans@tue.nl;
■
e.w.meijer@tue.nl
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work is financed by The Netherlands Organisation for
Scientific Research (NWO - TOP grant: 10007851), the Dutch
Ministry of Education, Culture and Science (Gravity program
024.001.035), and NRSC-C. I.K.V. is grateful for financial
support from The Netherlands Organization for Scientific
research (NWO - VENI grant: 700.10.406) and the European
Union through the Marie Curie 5 Fellowship program FP7-
PEOPLE-2011-CIG (contract no. 293788). The authors thank
Koen Pieterse and Pol Besenius for their initial efforts on
coiled-coil formation and Veronique Trappe for the helpful
discussions. Part of this work is based on experiments
performed at the ID2 High Brilliance Beamline at the European
Synchrotron Radiation Facility (ESRF) in Grenoble. We also
thank Dr. Narayanan Theyencheri for his assistance.
CONCLUSIONS
■
We are able to rationally design molecules that by means of
π−π stacking and hydrophobic interactions self-assemble into
triple helical fibers. The morphology of aggregates of 1−3 can
be tuned by the properties of the solvent mixture itself. The
correlation between solvent composition and LCST behavior of
temperature-sensitive polymers and supramolecular aggregates
has been observed previously by various groups,61−64 but
Trappe and co-workers first noted the correlation with solvent
REFERENCES
(1) Aida, T.; Meijer, E. W.; Stupp, S. I. Science 2012, 335, 813.
(2) Kumar, S. Chem. Soc. Rev. 2006, 35, 83.
■
E
thermodynamics, that is, the excess enthalpy of mixing, ΔHmix
(3) Kato, T.; Mizoshita, N.; Kishimoto, K. Angew. Chem., Int. Ed.
2006, 45, 38.
(4) Elemans, J. A. A. W.; van Hameren, R.; Nolte, R. J. M.; Rowan, A.
E. Adv. Mater. 2006, 18, 1251.
(5) Mart, R. J.; Osborne, R. D.; Stevens, M. M.; Ulijn, R. V. Soft
Matter 2006, 2, 822.
and ϕmin for thermo-sensitive polymers and proteins.50 Herein
we demonstrate that supramolecular polymers behave in a
highly similar fashion. We report two morphological transitions
that are coincident with the minima and maxima in ΔHmixE with
ϕI→II ∼ ϕmin and ϕII→III ∼ ϕmax. To the best of our knowledge,
no such correlation has previously been reported for (supra-
molecular) polymers. A possible explanation could be the
persistence of π−π stacking up to ϕIPA = 1 whereas the
hydrophobic effect gradually disappears.
(6) Hodges, R. S. Curr. Biol. 1992, 2, 122.
(7) Ebashi, S.; Ohtsuki, I.; Nitanai, Y.; Minakata, S.; Maeda, K.; Oda,
N.; Maeda, Y. In Regulatory Mechanisms of Striated Muscle Contraction;
Springer: Japan, 2007; Vol. 592, p 137.
(8) Houston, M. E.; Campbell, A. P.; Lix, B.; Kay, C. M.; Sykes, B.
D.; Hodges, R. S. Biochemistry 1996, 35, 10041.
(9) Bhattacharjee, A.; Bansal, M. IUBMB Life 2005, 57, 161.
(10) Ferrand, Y.; Gan, Q.; Kauffmann, B.; Jiang, H.; Huc, I. Angew.
Chem., Int Ed. 2011, 50, 7572.
(11) Gan, Q.; Ferrand, Y.; Chandramouli, N.; Kauffmann, B.; Aube,
C.; Dubreuil, D.; Huc, I. J. Am. Chem. Soc. 2012, 134, 15656.
(12) Yamada, H.; Furusho, Y.; Yashima, E. J. Am. Chem. Soc. 2012,
134, 7250.
(13) Hasenknopf, B.; Lehn, J. M.; Baum, G.; Fenske, D. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 1397.
Solvent-driven morphological transitions may dominate
supramolecular self-assembly in many aqueous cosolvent
mixtures.65−67 For example, George and co-workers studied
the self-assembly of a coronene bisimide amphiphile in water−
tetrahydrofuran mixtures and observed three self-assembly
regimes.68 The boundaries between these seem to correlate well
E
with the ΔHmix of the water−THF mixture.
In summary, the nanostructures obtained upon self-assembly
in water can be greatly influenced by the preparation method.
Frequently such samples are prepared by direct injection into
water of solutions of small self-assembling units in a “good”
solvent such as an alcohol or THF.38,66,69−71 Understanding
how the solvent composition influences the aggregate
(14) Tanaka, Y.; Katagiri, H.; Furusho, Y.; Yashima, E. Angew. Chem.,
Int Ed. 2005, 44, 3867.
(15) Peterca, M.; Percec, V.; Imam, M. R.; Leowanawat, P.;
Morimitsu, K.; Heiney, P. A. J. Am. Chem. Soc. 2008, 130, 14840.
342
dx.doi.org/10.1021/ja4104183 | J. Am. Chem. Soc. 2014, 136, 336−343

Journal of the American Chemical Society
Article
(16) Goto, H.; Furusho, Y.; Yashima, E. J. Am. Chem. Soc. 2007, 129,
9168.
(17) Goto, H.; Furusho, Y.; Miwa, K.; Yashima, E. J. Am. Chem. Soc.
(48) van Houtem, M. H. C. J.; Benaskar, F.; Fitie, C. F. C.; Martin-
Rapun, R.; Vekemans, J. A. J. M.; Meijer, E. W. Org. Biomol. Chem.
2012, 10, 5898.
(49) van Houtem, M. H. C. J.; Martín-Rapun, R.; Vekemans, J. A. J.
́
2009, 131, 4710.
M.; Meijer, E. W. Chem.Eur. J. 2010, 16, 2258.
(50) Trappe, V. Hydrophobic hydration: a matter of the mean
energetics of water. Personal communication.
(51) Roosma, J.; Mes, T.; Leclere, P.; Palmans, A. R. A.; Meijer, E. W.
̀
J. Am. Chem. Soc. 2008, 130, 1120.
(18) Gan, Q.; Bao, C. Y.; Kauffmann, B.; Grelard, A.; Xiang, J. F.; Liu,
S. H.; Huc, I.; Jiang, H. Angew. Chem., Int Ed. 2008, 47, 1715.
(19) Ferrand, Y.; Kendhale, A. M.; Garric, J.; Kauffmann, B.; Huc, I.
Angew. Chem., Int Ed. 2010, 49, 1778.
(20) Berl, V.; Huc, I.; Khoury, R. G.; Krische, M. J.; Lehn, J. M.
(52) Devos, A.; Remion, J.; Frisque-Hesbain, A.-M.; Colens, A.;
Ghosez, L. J. Chem. Soc., Chem. Commun. 1979, 1180.
(53) Smulders, M. M. J.; Nieuwenhuizen, M. M. L.; de Greef, T. F.
A.; van der Schoot, P.; Schenning, A. P. H. J.; Meijer, E. W. Chem.
Eur. J. 2010, 16, 362.
(54) Glatter, O. In Neutrons, X-rays and Light: Scattering Methods
Applied to Soft Condensed Matter; The Inverse Scattering Problem in
Small-Angle Scattering; Lindner, P., Zemb, T., Eds.; Elsevier:
Amsterdam, 2002; p 73.
(55) Franks, F.; Ives, D. J. G. Q. Rev. Chem. Soc. 1966, 20, 1.
(56) Katayama, T. Kagaku Kogaku Ronbun. 1962, 26, 361.
(57) Lama, R. F.; Lu, B. C. Y. J. Chem. Eng. Data 1965, 10, 216.
(58) Dixit, S.; Crain, J.; Poon, W. C. K.; Finney, J. L.; Soper, A. K.
Nature 2002, 416, 829.
(59) Nakada, M.; Maruyama, K.; Yamamuro, O.; Misawa, M. J. Chem.
Phys. 2009, 130, 074503.
(60) Takamuku, T.; Kumai, T.; Yoshida, K.; Otomo, T.; Yamaguchi,
T. J. Phys. Chem. A 2005, 109, 7667.
(61) Betancourt, J. E.; Rivera, J. M. J. Am. Chem. Soc. 2009, 131,
16666.
(62) Richards, G. J.; Labuta, J.; Hill, J. P.; Mori, T.; Ariga, K. J. Phys.
Chem. Lett. 2010, 1, 1336.
(63) Costa, R. O. R.; Freitas, R. F. S. Polymer 2002, 43, 5879.
(64) Schild, H. G.; Muthukumar, M.; Tirrell, D. A. Macromolecules
1991, 24, 948.
(65) Boekhoven, J.; Brizard, A. M.; van Rijn, P.; Stuart, M. C. A.;
Eelkema, R.; van Esch, J. H. Angew. Chem., Int. Ed. 2011, 50, 12285.
(66) Che, Y.; Datar, A.; Balakrishnan, K.; Zang, L. J. Am. Chem. Soc.
2007, 129, 7234.
Nature 2000, 407, 720.
(21) Mateos-Timoneda, M. A.; Kerckhoffs, J.; Reinhoudt, D. N.;
Crego-Calama, M. Org. Biomol. Chem. 2007, 5, 447.
(22) Aggeli, A.; Bell, M.; Carrick, L. M.; Fishwick, C. W. G.; Harding,
R.; Mawer, P. J.; Radford, S. E.; Strong, A. E.; Boden, N. J. Am. Chem.
Soc. 2003, 125, 9619.
(23) Chen, X.-M.; Liu, G.-F. Chem.Eur. J. 2002, 8, 4811.
(24) Zubarev, E. R.; Pralle, M. U.; Sone, E. D.; Stupp, S. I. J. Am.
Chem. Soc. 2001, 123, 4105.
(25) Taylor, P. N.; Anderson, H. L. J. Am. Chem. Soc. 1999, 121,
11538.
(26) Mamula, O.; von Zelewsky, A.; Bark, T.; Bernardinelli, G.
Angew. Chem., Int. Ed. 1999, 38, 2945.
(27) Engelkamp, H.; Middelbeek, S.; J. M., R.; Nolte. Science 1999,
284, 785.
(28) Tikhomirov, G.; Yamazaki, T.; Kovalenko, A.; Fenniri, H.
Langmuir 2008, 24, 4447.
(29) Li, J. X.; Wisner, J. A.; Jennings, M. C. Org. Lett. 2007, 9, 3267.
(30) Goto, H.; Furusho, Y.; Yashima, E. J. Am. Chem. Soc. 2007, 129,
109.
(31) Goto, H.; Katagiri, H.; Furusho, Y.; Yashima, E. J. Am. Chem.
Soc. 2006, 128, 7176.
(32) Besenius, P.; Portale, G.; Bomans, P. H. H.; Janssen, H. M.;
Palmans, A. R. A.; Meijer, E. W. Proc. Natl. Acad. Sci. U.S.A. 2010, 107,
17888.
(33) Leenders, C. M. A.; Albertazzi, L.; Mes, T.; Koenigs, M. M. E.;
Palmans, A. R. A.; Meijer, E. W. Chem. Commun. 2012, 49, 1963.
(34) Brizard, A.; Stuart, M.; van Bommel, K.; Friggeri, A.; de Jong,
M.; van Esch, J. Angew. Chem., Int. Ed. 2008, 47, 2063.
(35) von Groning, M.; de Feijter, I.; Stuart, M. C. A.; Voets, I. K.;
Besenius, P. J. Mater. Chem. B 2013, 1, 2008.
(67) Yang, Y.; Suzuki, M.; Owa, S.; Shirai, H.; Hanabusa, K. J. Mater.
Chem. 2006, 16, 1644.
(68) Rao, K. V.; George, S. J. Org. Lett. 2010, 12, 2656.
(69) Huang, J.; Bonduelle, C.; Thevenot, J.; Lecommandoux, S.;
́
Heise, A. J. Am. Chem. Soc. 2011, 134, 119.
(36) Bae, J.; Choi, J.-H.; Yoo, Y.-S.; Oh, N.-K.; Kim, B.-S.; Lee, M. J.
Am. Chem. Soc. 2005, 127, 9668.
(37) Hermans, T. M.; BroerenMaarten, A. C.; Gomopoulos, N.; van
der, S.; van GenderenMarcel, H. P.; SommerdijkNico, A. J. M.; Fytas,
G.; Meijer, E. W. Nat. Nano 2009, 4, 721.
(70) Chaudhary, N.; Singh, S.; Nagaraj, R. J. Pept. Sci. 2009, 15, 675.
(71) Stevens, A. L.; Kaeser, A.; Schenning, A. P. H. J.; Herz, L. M.
ACS Nano 2012, 6, 4777.
(38) Hill, J. P.; Jin, W.; Kosaka, A.; Fukushima, T.; Ichihara, H.;
Shimomura, T.; Ito, K.; Hashizume, T.; Ishii, N.; Aida, T. Science 2004,
304, 1481.
(39) Boekhoven, J.; van Rijn, P.; Brizard, A. M.; Stuart, M. C. A.; van
Esch, J. H. Chem. Commun. 2010, 46, 3490.
(40) Hartgerink, J. D.; Beniash, E.; Stupp, S. I. Science 2001, 294,
1684.
(41) Shao, H.; Gao, M.; Kim, S. H.; Jaroniec, C. P.; Parquette, J. R.
Chem.Eur. J. 2011, 17, 12882.
(42) Marsden, H. R.; Korobko, A. V.; van Leeuwen, E. N. M.; Pouget,
E. M.; Veen, S. J.; Sommerdijk, N. A. J. M.; Kros, A. J. Am. Chem. Soc.
2008, 130, 9386.
́
(43) Danila, I.; Riobe, F.; Piron, F.; Puigmartí-Luis, J.; Wallis, J. D.;
Linares, M.; Ågren, H.; Beljonne, D.; Amabilino, D. B.; Avarvari, N. J.
Am. Chem. Soc. 2011, 133, 8344.
(44) Brunsveld, L.; Lohmeijer, B. G. G.; Vekemans, J. A. J. M.;
Meijer, E. W. Chem. Commun. 2000, 2305.
(45) Brunsveld, L.; Zhang, H.; Glasbeek, M.; Vekemans, J. A. J. M.;
Meijer, E. W. J. Am. Chem. Soc. 2000, 122, 6175.
(46) Palmans, A. R. A.; Vekemans, J. A. J. M.; Havinga, E. E.; Meijer,
E. W. Angew. Chem., Int. Ed. 1997, 36, 2648.
(47) Gillissen, M. A. J.; Hoeben, T. T.; Spiering, A. J. H.; Vekemans,
J. A. J. M.; Palmans, A. R. A.; Meijer, E. W. Isr. J. Chem. 2011, 51, 1118.
343
dx.doi.org/10.1021/ja4104183 | J. Am. Chem. Soc. 2014, 136, 336−343