Supramolecular balance: Using cooperativity to amplify weak interactions

Article abstract of DOI:10.1021/ja105717u

Gathering precise knowledge on weak supramolecular interactions is difficult yet is of utmost importance for numerous scientific fields, including catalysis, crystal engineering, ligand binding, and protein folding. We report on a combined theoretical and

Full text of DOI:10.1021/ja105717u

Published on Web 11/04/2010
Supramolecular Balance: Using Cooperativity To Amplify
Weak Interactions
Mihaela Roman,† Caroline Cannizzo,‡ Thomas Pinault,‡ Benjamin Isare,‡
Bruno Andrioletti,† Paul van der Schoot,*,§ and Laurent Bouteiller*,‡
UniVersite´ Claude Bernard-Lyon 1, ICBMS-UMR 5246, 43 BouleVard du 11 NoVembre 1918,
F-69622 Villeurbanne cedex, France, Chimie des Polyme`res, UPMC UniVersite´ Paris 06 and
CNRS, UMR 7610, F-75005 Paris, France, and Group Theory of Polymers and Soft Matter,
Technische UniVersiteit EindhoVen, P.O. Box 513, 5600 MB EindhoVen, and Instituut Voor
Theoretische Fysica, UniVersiteit Utrecht, LeuVenlaan 4, 3584 CE Utrecht, The Netherlands
Received July 5, 2010; E-mail: laurent.bouteiller@upmc.fr; p.p.a.m.v.d.schoot@tue.nl
Abstract: Gathering precise knowledge on weak supramolecular interactions is difficult yet is of utmost
importance for numerous scientific fields, including catalysis, crystal engineering, ligand binding, and protein
folding. We report on a combined theoretical and experimental approach showing that it is possible to vastly
improve the sensitivity of current methods to probe weak supramolecular interactions in solution. The concept
consists of using a supramolecular platform involving a highly cooperative configurational transition, the
perturbation of which (by the modification of the molecular building blocks) can be monitored in a temperature
scanning experiment. We tested this concept with a particular bisurea platform, and our first results show that
it is possible to detect the presence of interaction differences as low as 60 J/mol, which may be due to steric
repulsion between vinyl and alkyl groups or may be the result of solvation effects.
mol or less has been recognized.4 Finally, predicting the
pathways of protein folding and quantifying the relative
Introduction
The precise knowledge of interactions that occur at the
molecular level is of the utmost importance for numerous
scientific fields. Indeed, manipulating noncovalent interactions
is at the root of the whole field of supramolecular chemistry1
and has consequences in a wide range of scientific domains and
practical applications. For instance, in crystal engineering, one
tries to obtain crystal structures that are related in obvious ways
to molecular structures and to achieve the ultimate goal of crystal
structure prediction. However, crystal packing is not simply
determined by a small number of strong interactions, but by
the cooperation and competition of a large number of strong
and weak interactions.2 In catalysis, the concept of manipulating
weak attractive ligand-substrate interactions has been demon-
strated. This has been implemented, for example, in the case of
bifunctional catalysis and in olefin polymerization, where this
approach has been shown to improve the control of the reactivity
and the microstructure of the polymer chain.3 In the field of
pharmaceuticals, a key step is to rationalize and optimize the
interactions between a potential drug and a relevant receptor.
In this context, the importance of interactions as weak as 1 kJ/
thermodynamic stability of intermediate and final states along
these pathways constitute two important challenges in modern
chemistry.5
These examples illustrate that even weak interactions such
as van der Waals interactions can collectively have a major
influence on the structure and thus on the properties of a given
assembly. It is therefore very important to be able to quantita-
tively measure weak molecular interactions. To this end, several
clever synthetic systems have been devised to isolate and
measure the energetic contribution of a particular interaction.
For instance, a molecular torsion balance, based on a molecule
with two restricted conformational states, allows one to probe
weak intramolecular interactions through direct NMR measure-
ment of the relative population of each state.6 Another possibility
(4) Mu¨ller, K.; Faeh, C.; Diederich, F. Science 2007, 317, 1881–1886.
(5) (a) Baltzer, L.; Nilsson, H.; Nilsson, J. Chem. ReV. 2001, 101, 3153–
3163. (b) Shakhnovich, E. Chem. ReV. 2006, 106, 1559–1588.
(6) (a) Paliwal, S.; Geib, S.; Wilcox, C. S. J. Am. Chem. Soc. 1994,
116, 4497–4498. (b) Nakamura, K.; Houk, K. N. Org. Lett. 1999,
1, 2049–2051. (c) Cockroft, S. L.; Hunter, C. A. Chem. Commun.
2006, 3806–3808. (d) Gung, B. W.; Xue, X.; Zou, Y. J. Org. Chem.
2007, 72, 2469–2475. (e) Bhayana, B.; Wilcox, C. S. Angew. Chem.,
Int. Ed. 2007, 46, 6833–6836. (f) Fischer, F. R.; Schweizer, W. B.;
Diederich, F. Angew. Chem., Int. Ed. 2007, 46, 8270–8273. (g)
Fischer, F. R.; Schweizer, W. B.; Diederich, F. Chem. Commun.
2008, 4031–4033. (h) Fischer, F. R.; Wood, P. A.; Allen, F. H.;
Diederich, F. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 17290–17294.
(i) Carroll, W. R.; Pellechia, P.; Shimizu, K. D. Org. Lett. 2008,
10, 3547–3550. (j) Cornago, P.; Claramunt, R. M.; Bouissane, L.;
Elguero, J. Tetrahedron 2008, 64, 3667–3673. (k) Aliev, A. E.;
Mo¨ıse, J.; Motherwell, W. B. Phys. Chem. Chem. Phys. 2009, 11,
97–100. (l) Cockroft, S. L.; Hunter, C. A. Chem. Commun. 2009,
3961–3963. For a review, see: (m) Mati, I. K.; Cockroft, S. L. Chem.
Soc. ReV. 2010, 39, 4195-4205.
† Universite´ Claude Bernard-Lyon 1.
‡ UPMC Universite´ Paris 06 and CNRS.
§ Technische Universiteit Eindhoven and Universiteit Utrecht.
(1) (a) Lehn, J.-M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4763–4768.
(b) Ling, X. Y.; Reinhoudt, D. N.; Huskens, J. Pure Appl. Chem. 2009,
81, 2225–2233. (c) Sijbesma, R. P.; Beijer, F. H.; Brunsveld, L.;
Folmer, B. J. B.; Hirschberg, J. H. K. K.; Lange, R. F. M.; Lowe,
J. K. L.; Meijer, E. W. Science 1997, 278, 1601–1604.
(2) Desiraju, G. R.; Steiner, T. The Weak Hydrogen Bond; Oxford
University Press: Oxford, U.K., 1999.
(3) (a) Ikariya, T.; Blacker, A. J. Acc. Chem. Res. 2007, 40, 1300. (b)
Chan, M. C. W. Chem. Asian J. 2008, 3, 18–27.
9
16818 J. AM. CHEM. SOC. 2010, 132, 16818–16824
10.1021/ja105717u  2010 American Chemical Society

Supramolecular Balance
A R T I C L E S
is to measure the association constant of a well-defined dimeric
assembly. Incremental modification of the substituents can be
used to probe a particular interaction, especially if a double
mutant cycle is performed to factor out unwanted secondary
interactions.7 These approaches have been very successful when
it comes to the measurement of free energies for interactions
such as hydrogen bonds, dipolar interactions, or π-π interac-
tions. Unfortunately, an intrinsic limitation of these strategies
is their limited sensitivity. In fact, the smallest uncertainty ever
reported for these kinds of studies is 0.12 kJ/mol,6f,h and there
is little hope to push this limit with the current methodology,
even with an increase in the NMR magnetic field. The limited
sensitivity is due to the fact that the energetic information derives
from the evaluation of concentrations (either of the two
conformational states or of the intermolecular complexes), which
are difficult to measure precisely. For example, biasing a
molecular torsion balance with an interaction of 0.1 kJ/mol
means that the equilibrium population of the two conformations
is shifted from a 50/50 to a 51/49 ratio. Measuring such a small
shift by the integration of NMR signals is clearly difficult.
Therefore, weaker interactions such as van der Waals interac-
tions, isotope effects, or chiral biases are currently out of reach
by these approaches.
forward that any platform10 based on the same physical
principles should be as effective as the one that we present here
by way of proof of principle.
Results and Discussion
Platform Description. Bisureas consisting of a toluene central
core bearing two urea moieties in the 2 and 4 positions have
been shown to self-assemble through hydrogen bonding into
two distinct supramolecular structures of high molar masses.11
Remarkably, in low-polarity solvents both structures have similar
stabilities and are in fast dynamic exchange.12 Therefore, it is
possible to switch the assembly from one structure to the other
by changing the temperature. The high-temperature structure
is a long filament with a single bisurea in the cross-section
(Figure 1a),13 while the low-temperature structure is a very long
and rigid tube with three bisureas in the cross-section (Figure
1b).14 Both structures are characterized by a strongly cooperative
growth, meaning that the formation of short oligomers is
disfavored, compared to longer oligomers.12,15 Because of this
high level of cooperativity, the transition between tubes and
filaments is very sharp and can conveniently be detected by an
endothermic peak in a differential scanning calorimetry (DSC)
experiment.12 These features have been demonstrated for several
bisureas bearing the same associating core and seem to be
common to this family of compounds.16 Moreover, because of
the much tighter packing of the monomers in the case of the
tube structure, intermolecular interactions within the tube can
be expected to be significantly different from those within the
filament. This system therefore constitutes an ideal platform to
test our concept of a supramolecular balance, the theoretical
principles of which we outline next.
In biochemistry, interactions between DNA bases are com-
monly quantified by the measurement of the melting temperature
of suitable DNA duplexes.8 Similarly, interactions between
peptide residues are probed by comparing the denaturation
temperature of related protein mutants,9 rather than the relative
concentrations of folded and unfolded proteins, which would
be much more difficult to quantify. In these cases, measuring a
transition temperature is possible because transitions in these
biochemical assemblies are cooperative; i.e., they assemble or
disassemble within a narrow temperature range.
Theory. The key ingredient in the supramolecular balance is
the ability of the monomer units to self-assemble in two
competing types of supramolecular structures. At the (absolute)
transition temperature T ) T0** both types are equally stable
and hence equally prevalent in the solution. Above (and below)
this temperature one of them is more (less) stable than the other.
As is shown explicitly in the Supporting Information, despite
the presence of assemblies of all degrees of polymerization, the
relatiVe stability of these two distinct supramolecular structures
turns out to be characterized by a single equilibrium constant,
We propose to apply this successful approach to synthetic
supramolecular systems, and the crucial idea for obtaining
ultrahigh sensitivity to differences in interaction free energies
is to use a cooperatively self-associating system as a platform.
Incremental modification of the substituents will be used to
probe a particular type of interaction, through variation of a
suitable transition temperature of the platform that makes use
of the competition between free monomers and two types of
supramolecular polymer present in the solution. The achieved
sensitivity of the measurement is due on one hand to the high
level of cooperativity of supramolecular polymerization of the
chosen platform and on the other to the competition between
two types of assemblies, i.e., the involvement of two equilibrium
constants rather than one. The platform that we make use of is
based on the self-organization of bisureas in solution and enables
us to distinguish the strength of van der Waals interactions
between alkyl or alkene groups in an organic solvent. We put
(10) For example, see: (a) Brunsveld, L.; Zhang, H.; Glasbeek, M.;
Vekemans, J. A. J. M.; Meijer, E. W. J. Am. Chem. Soc. 2000, 122,
6175–6182. (b) van der Schoot, P.; Michels, M. A. J.; Brunsveld, L.;
Sijbesma, R. P.; Ramzi, A. Langmuir 2000, 16, 10076–10083.
(11) Bouteiller, L.; Colombani, O.; Lortie, F.; Terech, P. J. Am. Chem.
Soc. 2005, 127, 8893–8898.
(12) Bellot, M.; Bouteiller, L. Langmuir 2008, 24, 14176–14182.
(13) (a) Vonau, F.; Suhr, D.; Aubel, D.; Bouteiller, L.; Reiter, G.; Simon,
L. Phys. ReV. Lett. 2005, 94, 066103. (b) Vonau, F.; Aubel, D.;
Bouteiller, L.; Reiter, G.; Simon, L. Phys. ReV. Lett. 2007, 99, 086103.
(c) Vonau, F.; Linares, M.; Isare, B.; Aubel, D.; Habar, M.; Bouteiller,
L.; Reiter, G.; Geskin, V.; Zerbetto, F.; Lazzaroni, R.; Simon, L. J.
Phys. Chem. C 2009, 113, 4955–4959.
(14) (a) Pinault, T.; Isare, B.; Bouteiller, L. ChemPhysChem. 2006, 7, 816–
819. (b) Shikata, T.; Nishida, T.; Lazzaroni, R.; Bouteiller, L. J. Phys.
Chem. B 2008, 112, 8459–8465. (c) Isare, B.; Linares, M.; Lazzaroni,
R.; Bouteiller, L. J. Phys. Chem. B 2009, 113, 3360–3364. (d) Isare,
B.; Linares, M.; Zargarian, L.; Fermandjian, S.; Miura, M.; Motohashi,
S.; Vanthuyne, N.; Lazzaroni, R.; Bouteiller, L. Chem.sEur. J. 2010,
16, 173–177.
(7) (a) Adams, H.; Carver, F. J.; Hunter, C. A.; Morales, J. C.; Seward,
E. M. Angew. Chem., Int. Ed. 1996, 35, 1542–1544. For a review,
see: (b) Cockroft, S. L.; Hunter, C. A. Chem. Soc. ReV. 2007, 36,
172–188.
(8) (a) Petersheim, M.; Turner, D. H. Biochemistry 1983, 22, 256–263.
(b) Guckian, K. M.; Schweitzer, B. A.; Ren, R. X.-F.; Sheils, C. J.;
Paris, P. L.; Tahmassebi, D. C.; Kool, E. T. J. Am. Chem. Soc. 1996,
118, 8182–8183.
(15) Simic, V.; Bouteiller, L.; Jalabert, M. J. Am. Chem. Soc. 2003, 125,
13148–13154.
(9) (a) Fernandez, A. M.; Villegas, V.; Martinez, J. C.; van Nuland,
N. A. J.; Conejero-Lara, F.; Aviles, F. X.; Serrano, L.; Filimonov,
V. V.; Mateo, P. L. Eur. J. Biochem. 2000, 267, 5891–5899. (b)
Vassall, K. A.; Stathopulos, P. B.; Rumfeldt, J. A. O.; Lepock, J. R.;
Meiering, E. M. Biochemistry 2006, 45, 7366–7379.
(16) (a) Colombani, O.; Bouteiller, L. New J. Chem. 2004, 28, 1373–1382.
(b) Isare, B.; Bouteiller, L.; Ducouret, G.; Lequeux, F. Supramol.
Chem. 2009, 21, 416–421. (c) Pensec, S.; Nouvel, N.; Guilleman, A.;
Creton, C.; Boue´, F.; Bouteiller, L. Macromolecules 2010, 43, 2529–
2534.
9
J. AM. CHEM. SOC. VOL. 132, NO. 47, 2010 16819

A R T I C L E S
Roman et al.
Figure 1. Schematic representations (a, b) and optimized geometries14b (c, d) for bisurea Br6 in filament (a, c) or tube (b, d) form. In each assembly, a
single bisurea is highlighted with O in red, N in blue, C in light blue, and H in white.
K0 ) K0(T), provided (i) the solution is dilute and (ii) their mean
2
2
K0(T) ) K0(T0**)e(∆h /k T ** )(T-T **) ) e(∆h /k T ** )(T-T **)
0
B
0
0
0
B
0
0
aggregation number is sufficiently large near the transition
temperature T0**. This happens to be the case for sufficiently
cooperative equilibrium polymerizations, even for conditions
not too remote from the polymerization transition of either
supramolecular structure. The (dimensionless) equilibrium
constant K0 ≡ exp(-∆G0/kBT), with kB Boltzmann’s constant,
is a function of the free energy difference ∆G0 ) ∆G0(T) of
binding a single monomer in the two types of assemblies. If K0
< 1or K0 > 1, then one of the two states of aggregation
predominates over the other. At the transition temperature we
find that K0(T0**) ) 1, again provided the assemblies are in
the mean of sufficiently large degree of polymerization.
Although nontrivial and based (in essence) on the law of mass
action, where we again refer to the Supporting Information for
details, this result is actually quite intuitive. The reason is that
at the transition from one to the other aggregate type, like in a
true phase transition, the chemical potentials of the monomers
in the two types of assemblies must be equal. These chemical
potentials are dominated by the respective binding free energies
as the assemblies the monomers are part of have very little
translational entropy.17 This implies that ∆G0(T0**) ) 0, which
is equivalent to the statement that K0(T0**) ) 1. Strictly
speaking, ∆G0(T0**) is not quite zero but very small on the
scale of the thermal energy kBT0** due to finite-size effects.
The temperature dependence of this equilibrium constant near
its transition temperature is given by
(1)
where ∆h0 ≡ ∆h0(T0**) is the enthalpy difference of a monomer
bound to the two types of supramolecular structure, measured
at the transition temperature. Equation 1 follows from a Taylor
expansion and is valid for temperatures |T - T0**| , T0** and
presumes the solution to be incompressible. See the Supporting
Information. If we now consider a similar compound that can
self-assemble into the same two competing supramolecular
structures, then their relative stability can be expressed in a
different equilibrium constant that we denote K1. For this
association constant we can write
2
K1(T) ) K1(T0**)e(∆h /k T ** )(T-T **)
(2)
1
B
0
0
for temperatures near the transition temperature of the other
component, so |T - T0**| , T0**, with ∆h1 ≡ ∆h1(T0**) again
an enthalpy difference measured at the temperature T0**. Here,
we have taken the transition of component 0 as the reference
point. As we have seen, at the respective transition temperatures
of the supramolecular structures for the two compounds, T0**
and T1**, the equilibrium constants are, to leading order in the
reciprocal aggregation numbers and at equal overall concentra-
tion of monomers, equal. Hence, equating eqs 1 and 2 gives
2
K0(T0**) ) K1(T1**) ) K1(T0**)e(∆h /k T ** )(T **-T **)
1
B
0
1
0
(17) Israelachvili, J. Intermolecular and Surface Forces, 2nd ed.; Academic
(3)
Press: New York, 1992.
9
16820 J. AM. CHEM. SOC. VOL. 132, NO. 47, 2010

Supramolecular Balance
A R T I C L E S
Chart 1. Structures of the Monomers
and therefore
∆h1
kBT0**2
(T1** - T0**) ) ln K0(T0**) - ln K1(T0**) )
∆G0(T0**) - ∆G1(T0**)
kBT0**
-
(4)
(5)
which leads to the central result of our analysis
T1** - T0**
∆∆G ≡ ∆G1(T0**) - ∆G0(T0**) ) ∆h1
T0**
According to eq 5, a structural change from compound 0 to
compound 1 can be associated with a free energy difference
that is the product of the transition enthalpy and the relative
change in transition temperature, both of which can be measured
as already indicated, e.g., in a DSC experiment. Because the
temperature is in Kelvin, a free energy change ∆∆G 2 orders
of magnitudes lower than the transition enthalpy ∆h can easily
be detected. Therefore, a very small free energy change can be
derived from a precise transition temperature measurement on
a platform having a suitably low (but measurable) transition
enthalpy.
Interactions between Alkyl Chains. As a first test of our
concept of the supramolecular balance, we decided to investigate
the quantitative influence of the length of the alkyl substituents
of the bisurea monomers, which should inform us on the strength
of van der Waals interactions between alkyl chains. Thus, a
range of monomers was synthesized (DL2 to DL18, Chart 1)
by reaction of tolyl 2,4-diisocyanate on n-alkylamines of various
lengths. Unfortunately, none of these bisureas with linear alkyl
substituents are soluble in nonpolar solvents (such as cyclo-
hexane, toluene, or chloroform), so that the relative stability of
the filament and tube forms cannot be probed. To improve
solubility, branching was introduced either in one substituent
or in both substituents. In the former case, bisurea monomers
L1 to L16 bear a 2-ethylhexyl group on one side and a linear
alkyl of given length on the other side. The dissymmetry in the
monomers is derived from 2-amino-4-nitrotoluene (Scheme 1).
In the latter case, bisurea monomers Br5 to Br8 bear a
2-ethylalkyl group of given length on both sides of the molecule.
In this case, the branched alkylamines were not commercially
available and were synthesized by alkylation and reduction of
butyronitrile (Scheme 2). Satisfyingly, all branched bisureas
synthesized are soluble in toluene.
Scheme 1. Synthesis of Monomers L1 to L16a
First, the supramolecular structures formed by these
compounds in toluene were characterized by the combination
of small-angle neutron scattering (SANS) and FTIR spec-
troscopy (see Figures S1 and S2 and Table S1 in the
Supporting Information). All the bisureas studied were found
to self-assemble into a quasi-one-dimensional structure
containing three molecules in the cross-section (and previ-
ously identified as the tube form) at low temperatures and
into filaments containing a single molecule in the cross-
section at high temperatures. As expected, the minor changes
introduced in the alkyl side chains do not change the structure
of the assemblies: the platform thus appears to be sufficiently
robust. The relative stability of the two supramolecular
structures was then probed by high-sensitivity DSC experi-
ments. All solutions yield a narrow and reversible endother-
mal peak characteristic for the transition between tube and
filament (see Figure S3 and Table S1 in the Supporting
a Reagents and conditions: (i) 2-ethylhexyl isocyanate, CH2Cl2, room
temperature, 56%; (ii) cyclohexene, Pd/C, 2-propanol, 85%; (iii)
CmH2m+1NCO or CmH2m+1NH2, triphosgene, diisopropylethylamine, THF,
room temperature, 11-64%.
Information). Remarkably, the minor changes introduced in the
alkyl side chains have a very strong effect on the value of the
transition temperature, which spans the range 27-50 °C (Figure
2a). The transition temperatures and enthalpies measured by
DSC can be converted into free energy differences, according
9
J. AM. CHEM. SOC. VOL. 132, NO. 47, 2010 16821

A R T I C L E S
Roman et al.
Scheme 2. Synthesis of Monomers Br5 to Br8a
that they cancel each other when the chains are six carbon atoms
long. In principle, an increase of the tube to filament transition
temperature can be due either to stabilizing interactions in the
tube form or to destabilizing interactions in the filament form.19
However, in the filament form, a large distance between bisureas
is imposed by the hydrogen-bonded core (4.6 Å), so that it is
reasonable to assume that alkyl chains interact less in the
filament form than in the more compact tube form (Figure 1).
Therefore, in the first approximation, we plausibly presume the
filament to be unaffected by the alkyl chain length,21 and we
attempt to correlate the stability of the tube to filament transition
to interactions occurring in the tube form. Accordingly, the data
of Figure 2 can be interpreted by the predominance of attractive
van der Waals interactions among alkyl chains, if they are less
than six carbon atoms long, and by that of steric repulsions
among alkyl chains, if they are more than six carbon atoms
long. Indeed, steric repulsion leads to reduced numbers of
conformations and hence reduced conformational entropy, an
effect that becomes stronger with increasing alkyl chain length.
Furthermore, from the first part of the curve, it can be estimated
that each additional methylene group contributes 70 J/mol to
the stabilization of the tube form. This value may seem
surprisingly small compared to known values of van der Waals
interaction, derived for instance from the Gibbs free energy of
micellization of surfactants (∆G ) 3.5 kJ/mol per methylene).22
However, one has to realize that the reference state is not the
same. In the present case, we measure the free energy gain for
a methylene group in the tube structure compared to the filament
structure. Thus, in a first approximation, we measure the free
energy gain for a methylene group interacting with alkyl groups
compared to the situation where it is solvated by toluene (instead
of being solvated by water, as is the case in micellization).
a Reagents and conditions: (i) LDA, -78 °C, then room temperature,
41-88%; (ii) LiAlH4, Et2O, 0 °C, then reflux, 66-85%; (iii) tolyl 2,4-
diisocyanate, CH2Cl2, room temperature, 41-93%.
Interactions Involving Alkene Chains. Aromatic groups are
well-known to interact by π-stacking or weak hydrogen
bonding. However, much less is known about interactions
involving vinyl groups, certainly because of their weak
intensity. Therefore, bisurea DV6 (Chart 1) was synthesized
to probe these interactions. The vinyl groups were introduced
in the terminal position of the alkyl substituents, because it
was anticipated that, at these positions, they would not induce
any conformational bias in the monomer. DSC experiments
(Figures 3 and S3, Supporting Information) show that the
tube to filament transition occurs at 31.5 °C, as compared to
42.7 °C for bisurea Br6. This 11 °C difference is much larger
(18) The reference temperature was chosen as T0** ) 27.4°C, which is
the transition temperature of compound Br5 (any other compound can
be chosen as a reference without affecting the shape of the curve).
The transition enthalpy measured for the other compounds is
∆h1(T1**). We assume that ∆h1(T1**) ) ∆h1(T0**), which is justified
by the fact that there is no heat capacity jump at the tube-filament
transition (Figure S3, Supporting Information) and therefore that the
temperature variation of the enthalpy is negligible.12
(19) We could also envisage that the transition is affected by a change of
conformation of the alkyl chains.20 However, FTIR spectroscopy shows
that the proportion of gauche conformations is not affected by the
tube to filament transition, whatever the chain length of the bisurea
(see Figure S4 in the Supporting Information).
Figure 2. (a) Transition temperature for bisureas Br5 to Br8 (red) and
bisureas L1 to L16 (black), measured in toluene (10 mM). (b) Relative
free energy of the tube to filament transition versus alkyl chain length (the
reference is Br5, and the reference temperature is 27.4 °C).
(20) Zweep, N.; Hopkinson, A.; Meetsma, A.; Browne, W. R.; Feringa,
B. L.; van Esch, J. H. Langmuir 2009, 25, 8802–8809.
to eq 5 (Figure 2b).18 The first striking result is that both series
of compounds display a nonmonotonic variation of the free
energy as a function of the alkyl chain length. Moreover, in
both cases, the maximum of stability of the tube form is reached
when the longest branch of the variable alkyl chain is composed
of six carbon atoms. This indicates that two competing effects
are involved when the length of the alkyl chains is altered and
(21) To test this assumption, the stability of the monomer to filament
transition was studied by isothermal titration calorimetry (ITC) (see
Figure S5 in the Supporting Information). No influence of the alkyl
chain length on the filament stability could be detected. Unfortunately,
this test is not conclusive, because the uncertainty of the ITC
experiment, although small, is larger than the effect detected by DSC.
(22) Tong, W.; Zheng, Q.; Shao, S.; Lei, Q.; Fang, W. J. Chem. Eng. Data
2010, 55, 3766–3771.
9
16822 J. AM. CHEM. SOC. VOL. 132, NO. 47, 2010

Supramolecular Balance
A R T I C L E S
a single vinyl group (Chart 1), was synthesized and charac-
terized. Because of the lack of symmetry of bisurea V6, two
orientations are possible for each monomer in the tube: the
vinyl group can be either on the left or on the right. However,
given the low value of the interaction considered, we can
safely assume that the statistics are not biased by the
interaction, so that the orientation of the monomer in the
tube is random. In this case, the number of vinyl/vinyl
contacts for V6 should be a quarter of the value for DV6. In
contrast, in the tube structure, the vinyl group of V6 is always
in the vicinity of methylene or methyl groups from its
neighbors; therefore, the number of vinyl/alkyl contacts for
V6 should be half the value for DV6.23 The DSC experiment
(Figures 3 and S3, Supporting Information) shows that the
tube to filament transition occurs at 38.0 °C for bisurea V6,
as compared to 31.5 and 42.7 °C for bisureas DV6 and Br6,
respectively. This median value means that replacing a single
terminal ethyl group per bisurea by a terminal vinyl group
destabilizes the tube structure by 60 ( 10 J/mol, i.e., about
half the value for DV6. It can therefore be concluded that
the destabilization is consistent with repulsive interactions
between vinyl and methyl or methylene groups. Of course,
it is difficult to attribute more precisely the nature of this
interaction, but it may plausibly be related to an unfavorable
packing of the vinyl and alkyl groups, by analogy to the
melting point depression of unsaturated fatty acids.
Figure 3. (a) Transition temperature measured in toluene (10 mM). (b)
Relative free energy of the tube to filament transition (the reference is Br6,
and the reference temperature is 42.7 °C).
than the experimental error and unambiguously shows that
replacing the terminal ethyl groups by terminal vinyl groups
destabilizes the bisureas in the tube structure (or stabilizes
the bisureas in the filament structure) by 110 ( 15 J/mol.
Not surprisingly, this value is low; however, it is not
negligible, and a better knowledge of this interaction could
have an impact, for instance, on the design of polymerization
catalysts.3
Our supramolecular balance is clearly of unprecedented
sensitivity, but at the same time unfortunately hampered by
our poor knowledge of the relative positions of the bisureas
in the tube assembly. At this point, no crystal of the tube
structure has been obtained, and the very broad NMR spectra
of bisureas in toluene preclude NOE experiments. This is in
fact the exactly symmetrical situation compared to crystal-
lographic data analysis, which yields an unambiguous position
for the atoms, but no direct information about the energetics
of a particular interaction. In the case of crystallography, a
careful statistical analysis of a large number of crystal
structures makes it possible to isolate the energetic contribu-
tion of a particular interaction from the pervading crystal
packing forces. In the present case, the buildup of a large
data set on related compounds should make it possible to
deduce also precise geometrical information.
Let us first assume that no differential solvation effect is
present between the tube and filament forms. Then, as in the
previous section, we can assume that the decrease of the tube
to filament transition is due to destabilizing interactions in
the tube form, rather than stabilizing interactions in the
filament form. It is clear from the molecular simulations of
the tube structure (Figure 1)14b-d that the vinyl groups can
only come into contact with other vinyl groups or with alkyl
groups. To discriminate whether the 110 J/mol bias is due
to vinyl/vinyl interactions or to vinyl/alkyl interactions, it is
possible to design a suitable blank experiment. For instance,
“diluting” the vinyl groups in the tube structure should lead
to a reduction of both vinyl/vinyl and vinyl/alkyl contacts,
but not in the same proportion. Accordingly, bisurea V6, with
Let us now examine possible solvation effects. The filament
form is less compact than the tube form; therefore, it is likely
that the vinyl groups in the filament form are more exposed to
the solvent than in the tube form. If this is the case, then the
decrease in the tube to filament transition can be due to favorable
interactions between the vinyl groups and the solvent (toluene).
The linearity of the effect on going from Br6 to V6 and to DV6
(Figure 3b) is compatible with this interpretation.
We have thus identified two plausible interpretations: steric
repulsion between vinyl and alkyl groups in the tube form or
favorable solvation of the vinyl groups by toluene in the filament
form. Reproducing the experiment in a range of solvents may
help to discriminate between the two interpretations.
Conclusion
In conclusion, we report a theoretical analysis showing
that it is possible to vastly improve the sensitiVity of current
methods to measure weak supramolecular interactions in
solution. The concept consists in using a supramolecular
platform, which presents a cooperative transition, so that a
perturbation of this transition can be monitored by a
temperature scanning experiment. This concept has been
tested with a particular bisurea platform. Our first results
show that it is possible to detect the presence of interactions
as low as 60 J/mol, which may occur between vinyl and alkyl
groups or may be the result of solvation effects. At present,
the weakness of our proposed supramolecular balance is the
poor knowledge of the relative position of the interacting
groups, but it is our opinion that the buildup of a large data
set on related compounds will make it possible to deduce
also precise geometrical information. Moreover, the excep-
(23) We count as alkyl groups all the methylene and methyl groups coming
either from the alkyl or from the alkylene side chains.
9
J. AM. CHEM. SOC. VOL. 132, NO. 47, 2010 16823

A R T I C L E S
Roman et al.
synthesis of bisureas. We thank Franc¸ois Boue´ (LLB, Saclay)
for assistance with the SANS experiments.
tional sensitivity of this platform can also be used to probe
weak interactions between solvent molecules in solvent
mixtures.24
Supporting Information Available: Theory of competing
supramolecular assembly, Experimental Section (synthesis and
characterization of bisureas), and SANS, FTIR, DSC, and ITC
data. This material is available free of charge via the Internet
at http://pubs.acs.org/.
Acknowledgment. Financial support from the French ANR
program (Grant ANR-06-BLAN-0159 zeoliplastic) is acknowl-
edged. Guillaume Zagury is thanked for his contribution to the
(24) van der Schoot, P.; Bouteiller, L.; Brocorens, P.; Lazzaroni, R.
Manuscript in preparation.
JA105717U
9
16824 J. AM. CHEM. SOC. VOL. 132, NO. 47, 2010