Concentration-Dependent Self-Assembly of an Unusually Large Hexameric Hydrogen-Bonded Molecular Cage

Article abstract of DOI:10.1002/chem.202005046

The sizes of available self-assembled hydrogen-bond-based supramolecular capsules and cages are rather limited. The largest systems have volumes of approximately 1400–2300 ?³. Herein, we report a large, hexameric cage based on intermolecular am

Full text of DOI:10.1002/chem.202005046

Accepted Article
Accepted Article
Title: Concentration-dependent Self-assembly of an Unusually Large
Hexameric Hydrogen-bonded Molecular Cage
Authors: Severin Merget, Lorenzo Catti, Shani Zev, Dan T. Major, Nils
Trapp, and Konrad Tiefenbacher
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202005046
Link to VoR: https://doi.org/10.1002/chem.202005046
01/2020

10.1002/chem.202005046
Chemistry - A European Journal
RESEARCH ARTICLE
Concentration-dependent Self-assembly of an Unusually Large
Hexameric Hydrogen-bonded Molecular Cage
Severin Merget[a], Lorenzo Catti[b], Shani Zev[c], Dan T. Major[c], Nils Trapp[d] and Konrad
Tiefenbacher*[a, e]
[a]
M. Sc. S. Merget., Prof. Dr. K. Tiefenbacher
Department of Chemistry,
University of Basel
Mattenstrasse 24a, CH-4058 Basel, Switzerland
E-mail: konrad.tiefenbacher@unibas.ch, tkonrad@ethz.ch
Dr. L. Catti,
WPI Nano Life Science Institute (WPI-NanoLSI),
Kanazawa University
Kakuma-machi, Kanazawa 920-1192, Japan
Prof. Dr. D. T. Major
Department,of Chemistry and Institute for Nanotechnology & Advanced Materials
Bar-Ilan University
Ramat-Gan 52900, Israel
Dr. N. Trapp
Laboratorium für Organische Chemie
ETH Zurich,
Vladimir-Prelog-Weg 3, CH-, 8093, Zürich, Switzerland
Prof. Dr. K. Tiefenbacher
[b]
[c]
[d]
[e]
Department of Biosystems Science and Engineering,
ETH Zürich
Mattenstrasse 26, CH-4058 Basel, Switzerland
Supporting information for this article is given via a link at the end of the document.
Abstract: The sizes of available self-assembled hydrogen bond Introduction
based supramolecular capsules and cages are rather limited. The
largest systems feature volumes of approx. 1400 − 2300 Å3. Here, we
The self-assembly of molecular capsules and cages using
2]
noncovalent interactions like hydrogen bonds,[1,
halogen
report a large, hexameric cage based on intermolecular amide−amide
dimerization. The unusual structure with openings, reminiscent of
covalently linked cages, is held together by 24 hydrogen bonds. With
a diameter of 2.3 nm and a cavity volume of ∼ 2800 Å3 the assembly
is larger than any previously known capsule/cage structures relying
exclusively on hydrogen bonds. The self-assembly process in
chlorinated, organic solvents was found to be strongly concentration
dependent with the monomeric form prevailing at low concentrations.
Additionally, the formation of host−guest complexes with fullerenes
(C60 and C70) was observed.
bonds[3] or hydrophobic interactions[4] has been studied
intensively over the last decades. Interestingly, in the case of
hydrogen bond based systems, the size of the assemblies formed
remained rather modest. The octameric capsule (Fig. 1a),
reported by the Mastalerz group in 2016 constitutes the largest
structure with a diameter of approx. 1.8 nm (V ∼ 2300 Å3).[2b]
1
This article is protected by copyright. All rights reserved.

10.1002/chem.202005046
Chemistry - A European Journal
RESEARCH ARTICLE
Mastalerz et al. (2016)
Warmuth et al. (2006)
a) 'Closed-shell' Capsule (Hydrogen-bonded):
b) Organic Cage (Covalently Linked):
O
O
O
NH
O
O
R
R
R
R
H2N-(CH2 2-NH2
TFA (cat.)
)
O
O
O
O
apolar
solvents
O
nPr
nPr
(CHCl3)
then
Reduction
Me
1.8 nm
2.5 nm
O
O
O
HN
nPr
O
O
N
O
H
O
O
R = (CH2 4CH3
)
c) This work: Largest Cage based on Hydrogen Bonds
O
NH2
OH
HO
R
R
HO
HO
OH
H2N
O
O
apolar
solvents
2.3 nm
NH2

2.1 nm
OH
R
R
HO
OH
H2N
O
1a R = (CH2 10CH3
)
≡
≡
II
16
I
16
1b R = CH
2CH(CH3)2
'Closed-shell' Capsule
Cage
(Hydrogen-bonded)
(not observed)
Figure 1. Molecular models of a) the octameric structure reported by Mastalerz, b) the covalently linked cage reported by Warmuth and c) the herein reported
structure I, assembled from six units of 1a (R residues are omitted for clarity).
In contrast, the self-assembly of cages via covalent bonds[5] and
especially metal−ligand interactions[6] has furnished much larger
structures with volumes of up to 157000 Å3 and more than a
hundred components.[7] What causes this size discrepancy?
Metal−ligand self-assembly is highly predictable concerning the
number of ligands binding to the metal center and the angle
between these ligands. In contrast, hydrogen bonding is less
predictable. While a linear binding mode is ideal, large deviations
are tolerated. Moreover, different binding motives can be
observed (f.i. bifurcation of the hydrogen bond) that can further
complicate a rational design. Most importantly, however, the
linearity of the hydrogen bond requires the design and synthesis
of curved building blocks, while in the case of metal−ligand self-
assembly the curvature can be imparted by the metal binding site.
These factors reduce the predictability of the self-assembly
process via hydrogen bonds, and in many cases the formation of
Furthermore, the pores potentially offer the possibility for further
modification of the assembly. Warmuth and co-workers have
pioneered the assembly of covalent organic cages based on
methylene bridged resorcinarene derivatives.[12] (Figure 1b). The
cages are formed using dynamic covalent imine chemistry and
feature openings of considerable size (d ∼ 8 Å). The approach[13]
has been extended towards chiral hydrazone cages by Szumna
and co-workers using resorcinarene derivatives with
unfunctionalized phenol groups.[14]
In order to construct a very large assembly based on hydrogen
bonds we decided to synthesize the tetraamide-functionalized
resorcinarene 1a (Fig. 1c). Interestingly, this building block would
be suited to self-assemble into two structurally very different
hexameric structures, according to our initial molecular modelling.
A direct amide−amide dimerization would form an unusually large
cage structure I, while an amide trimerization binding motive
would form a smaller closed-shell capsule II (Fig. 1c).
8]
extended networks is observed.[2b, These problems can be
overcome to a large extent by designing dimeric systems.[1, 9]
Unfortunately, the synthetic effort usually scales with the size of
the building blocks, rendering this approach challenging when
trying to obtain larger assemblies. Among the noteworthy
Herein, we present our results concerning the self-assembly of
the organic cage I with an internal cavity volume of ∼ 2800 Å3 −
to the best of our knowledge the largest cage structure that relies
solely on hydrogen bonding.
exceptions
are
the
hexameric
resorcinarene[10]
and
pyrogallolarene[11] capsules (d ∼ 1.8 nm; V ∼ 1400 Å3), the
Stefankiewicz[2c] capsule (d ∼ 1.9 nm; V ∼ 1700 Å3) and the largest
example by the Mastalerz group[2b] (d ∼ 1.8 nm; V ∼ 2300 Å3)
already mentioned before (Figure 1a).
Another striking feature of most multimeric assemblies based on
hydrogen bonding is the formation of a ‘closed-shell’ capsule,
effectively isolating the interior from the surrounding solvent. This
is in stark contrast to metal−ligand based[5] and covalently linked
cages[6] many of which have large openings. Such open
structures are of interest not only since the pores potentially offer
an additional binding site, but also because they allow small
molecules to diffuse in and out of the cavity without the need for
partial disassembly required for closed-shell capsules.
Results and Discussion
First, a reliable synthetic route to macrocycle 1a had to be
developed (Figure 2a). Starting from resorcinarene, the
tetrabrominated compound 2a was accessible on the decagram
scale via bromination and methylation following a literature
procedure.[15] Initial attempts to access tetraamide 3a from 2a via
cyanide coupling followed by hydrolysis proved futile, due to the
unreactive nature of the nitrile compound. Ultimately, the
introduction of the four amide moieties was achieved in 47% yield
by treatment of 2a with an excess of n-butyllithium, followed by
2
This article is protected by copyright. All rights reserved.

10.1002/chem.202005046
Chemistry - A European Journal
RESEARCH ARTICLE
addition of trimethylsilyl isocyanate (TMSNCO) as the electrophile. 10.8 ppm) forms a hydrogen bond with the phenol-group at the
Finally, the methyl protecting groups were removed with
trimethylsilyl iodide (TMSI) to yield macrocycle 1a in 74% yield.
Derivative 1b featuring shorter iso-butyl chains was synthesized
along the same route and used for X-ray structure analysis (see
SI for details).
adjacent aromatic ring and experiences a weaker downfield shift
(see SI for 2D-NMR data). All major peaks correspond to a single
diffusing entity (D = 2.39 ± 0.005 × 10−10 m2/s in CDCl3 (5 mM)) as
indicated by DOSY-NMR in different solvents (see SI). The
hydrodynamic radius (rh) in apolar solvents was estimated to be
[2b,
1.81 – 1.96 nm, which indicates an assembly of higher order
17]
similar in size to previously reported hexameric structures.[18]
a)
O
NH2
OMe
Br
The spectrum of 1a (20 mM) in CDCl3 as wells as TCE-d2,
features additional small peaks and shoulders next to the OH
signals (A and B) and at 5.84 ppm (red dots in Figure 3), beside
the main peaks (green dots). We found the ratio between the main
peaks and these smaller signals to be highly dependent on the
concentration of the NMR sample (Figure 4b). Furthermore,
DOSY-NMR spectra revealed a significantly higher diffusion
coefficient of the signals marked with the red dots (D = 3.53 ±
0.021 × 10−10 m2/s (5 mM in CDCl3)) indicating a much smaller
species (Figure 4d). Upon varying the concentration, it became
evident that at low concentration the small species is preferred
while at high concentration the hexameric structure I is formed
predominantly (Figure 4c). The strong upfield shift from 8.98 to
5.78 ppm of NH-proton C’ (highlighted in red in Fig. 4b) indicates
the loss of a strong hydrogen bond in the smaller species as
compared to the larger one.[19] The loss of a strong hydrogen bond
at atom C’ indicates that the smaller species is the monomeric
species 1a. In the hexameric structure, one of the NH-protons (C)
forms an intermolecular hydrogen bond with the carbonyl moiety
of a second macrocycle leading to a significant deshielding. This
hydrogen bond is lost in the monomeric species and the NH-
proton (C’) experiences an upfield shift. [20]
MeO
R
OMe
R
MeO
R
R
MeO
Br
MeO
OMe
Br
OMe
nBuLi, THF
then TMSNCO
MeO
MeO
OMe
OMe
H2N
O
O
NH2
R
R
R
R
MeO
OMe
MeO
OMe
Br
H2N
O
R = (CH
R = (CH
(47%)
(47%)
2a
2b
3a
R = CH 2
10CH3
10CH3
)
)
R = CH 2
3b
2CH(CH3)2
2CH(CH3)2
b)
O
NH2
OH
HO
TMSI,
CHCl3
reflux
R
R
HO
HO
OH
OH
H2N
O
O
1.9 Å
NH2
2.0 Å
R
R
HO
OH
H2N
R = (CH
R = CH 2
O
(74%)
(14%)
1a
1b
10CH3
)
2CH(CH3)2
Figure 2. a) Synthesis of 1a and 1b starting from 2a and 2b, respectively.
b) Representation of the wave-like crystal structure of 1b obtained from
diethylether/nitrobenzene. Solvent molecules were omitted for clarity.
Crystals were obtained by slow evaporation of a saturated
solution of 1b in diethylether/nitrobenzene (10/1) at room
temperature. Two types of structures were identified, using this
protocol. Occasionally, 1b crystallized forming a wave-like
structure (Figure 2b) that could be solved using small molecule
methods and was also found when employing other solvents (i.e.
toluene) for crystallization. In this structure the intermolecular
amide−amide hydrogen bonding is evident with measured NH −
O=C distances of 1.9 − 2.0 Å.[16] In most cases however, the
crystals obtained showed a much larger cubic unit cell (a = 3.7
nm) with a maximum resolution of 1.2 Å, suggesting a large self-
assembled structure. Unfortunately, disordered solvent molecules
within the cavity prevented solving of this larger structure with
small molecule methods.
HC
O
N
HD
O
HA
O
HB
CDCl3
R
A
A
E
D
D
B
HF
C
C
F
HE
TCE-d2
CD2Cl2
E
E
B
F
A
C/D
F
B
C6D6
A
D
D
E
B
C
C
F
Toluene-d8
A
B
E
F
Acetone-d6
A
B
E
Although macrocycle 1a contains several polar groups on its
upper rim it was found to be well soluble (≥ 50 µmol/mL,
≥ 64 mg/mL) in apolar, chlorinated solvents (i.e. CDCl3, CD2Cl2,
tetrachloroethane-d2 (TCE-d2)) as well as aromatic solvents (C6D6,
toluene-d8). The relevant low-field regions of the 1H-NMR spectra
in different solvents, including acetone-d6 that prevents hydrogen
bond based self-assembly, are displayed in Figure 3. The spectra
in apolar solvents show largely similar features: (i) the two amide
NH-protons (C and D) resonate at a similar chemical shift in the
range of 8.50 – 9.50 ppm, indicating that they both participate in
the hydrogen bond network; (ii) the protons corresponding to the
phenol groups (A and B) differ quite significantly with proton A
(16.0 – 17.0 ppm) being highly deshielded due to the hydrogen
bond with the neighboring carbonyl group, while proton B (9.70 –
D
C
F
16.5
16.0
10.0
9.5
16
15
14
13
12
11
f1 (ppm)
10
9
8
7
6
5
Figure 3. Partial 1H-NMR spectra of 1a (20 mmol/L) in different solvents. Signals
corresponding to the main species I are marked with a green dot (●). Signals
corresponding to the monomeric species 1a are marked with a red dot (●).
Solvent peaks are marked in grey.
In contrast, the chemical environment of the protons involved in
intramolecular hydrogen bonds (A’, B’, D’HA’, HB’, HD’) remains
rather similar and only a small shift is observed at different
concentrations.
3
This article is protected by copyright. All rights reserved.

10.1002/chem.202005046
Chemistry - A European Journal
RESEARCH ARTICLE
a)
c)
N
O
HC'
HC
O
N
HC
N
high
concentration
HD'
O
HA'
O
O
HD
O
HA
O
HB'
R
low
HB
HF'
concentration
R
HF
HE'
1a
HE
I
b)
A
B
D
E
C
F
50 mM in CDCl3
20 mM
10 mM
d)
7.5 mM
5.0 mM
= 2.39 x 10-10
m
DHexamer
2/s
2.5 mM
1.0 mM
0.5 mM
= 3.53 x 10-10
m
DMonomer
2/s
E'
B'
A'
F'
C'
D'
0.1 mM
16
15
14
13
12
11
10
9
8
7
6
5
f1 (ppm)
Figure 4. a) Concentration dependent assembly process between monomer 1a (red ●) and hexamer I (green ●) in apolar solvents. b) Partial 1H-NMR spectra of
1a/I in CDCl3 at different concentrations (0.1 – 50 mM based on 1a). c) Ratio between monomer 1a and hexamer I as a function of the monomer concentration.
d) DOSY-NMR spectrum of 1a/I (5.0 mM) showing the different diffusion coefficients for 1a and I.
Similar trends are observed in the acetone-d6 solution of 1a
(Figure 3), as the self-assembly process is suppressed by the
solvent. The spectrum is in fact comparable to the one of 1a in
CDCl3 at lower concentration (Figure 4b), although the difference
for the two NH-protons (C’ and D’) is not as pronounced in
acetone-d6; presumably because proton C’ still experiences some
deshielding due to hydrogen bonding with the solvent. Varying the
concentration of 1a in TCE-d2 and CD2Cl2 (see SI) as well as in
aromatic solvents (C6D6, toluene-d8) led to similar results
to have a diffusion coefficient of 4.86 × 10−10 m2/s (5.0 mM in
CDCl3), which is in good agreement with the value obtained for 1a
at low concentration. In contrast, at high concentrations the
assembled species I prevails and the diffusion coefficient
decreases to reach 1.71 × 10−10 m2/s at 50 mM in CDCl3 indicating
the formation of a large, discrete assembly. As the decreasing
diffusion coefficient might be simply caused by an increase in
viscosity at higher macrocycle concentrations, we compared the
behavior
to
the
well-known
resorcinarene[10]
and
concerning
a
second smaller species prevailing at low
pyrogallolarene[11] hexamers assembling from macrocycles 6 and
7, respectively (Figure 5c). In all cases the diffusion coefficient
decreases at higher concentration, however, the effect is much
more pronounced for macrocycle 1a which indicates that indeed
the hexamer I is significantly larger than the resorcinarene and
pyrogallarene based hexamers. Assuming that for all assemblies
the contribution of the alkyl feet is similar, the hydrodynamic
radius determined from the respective diffusion coefficient
(measured at 50 mM) correlates very well with the molecular
models (Table S4). For the resorcinarene hexamer (rh = 1.99 nm;
estimated r = 2.01 nm) and the pyrogallolarene hexamer (rh =
1.91 nm; estimated r = 1.97 nm), an excellent correlation was
observed. Also the proposed structure of cage-like hexamer I (rh
= 2.36 nm) displays a very good fit to the molecular model
concentration.
Interestingly, a study concerning the self-assembly of the
pyridinearene macrocycle 4 (Figure 5a) by Cohen and co-
workers[18c] revealed analogous behavior. Also in this case two
species were distinguished by DOSY-NMR in CDCl3: a large
hexameric capsule (D = 2.36 × 10−10 m2/s) and a smaller species
(D = 3.50 × 10−10 m2/s in CDCl3) assigned as a dimer. Addition of
TFA (trifluoroacetic acid) led to disassembly and allowed the
observation of the monomer in solution (D = 4.31 × 10−10 m2/s).
In our case the diffusion coefficients of both species 1a and I were
found to be concentration dependent (Figure 5b). At low
concentrations, where the assembled species I is effectively
absent, the diffusion coefficient of the small species increases to
reach values of up to 4.83 × 10−10 m2/s (1.0 mM in CDCl3) with
decreasing concentration, which is in good agreement with values
previously reported for pyridinearene 4 in its monomeric form. To
obtain an independent reference for the monomeric form the
globally methyl protected resorcinarene derivative 5 was used,
which cannot self-assemble. Macrocycle 5 was accordingly found
(estimated
r = 2.39 nm). As the alternative closed-shell
structure II does not match the observed hydrodynamic radius,
we excluded this option. In conclusion, the DOSY-NMR
investigations provided evidence that strongly support the
formation of the cage-like hexamer I.
4
This article is protected by copyright. All rights reserved.

10.1002/chem.202005046
Chemistry - A European Journal
RESEARCH ARTICLE
I
a) Related Macrocycles
b) Diffusion Coefficients of 1a and Dependent on Concentration
HO
N
OH
MeO
R
OMe
R
R
R
HO
N
OH
N
MeO
MeO
OMe
OMe
I
HO
OH
R
4
R
R
R
HO
N
OH
MeO
OMe
R = (CH
R = (CH
5
10CH3
10CH3
2
)
2
)
−
−
= 2.37 10 10 m
= 4.86 10 10 m
×
×
DHexamer
2/s
2/s
DMonomer
2/s
−
= 4.31 10 10 m
×
DMonomer
c) Diffusion Coefficients of Different Hexameric Assemblies Dependent
on Concentration
OH
I
HO
OH
R
HO
OH
R
R
R
HO
OH
HO
OH
OH
HO
OH
HO
R
OH
HO
R
R
R
HO
OH
HO
OH
OH
R = (CH
R = (CH
6
7
10CH3
10CH3
2
)
2
)
−
−
= 2.34 10 10 m
= 2.34 10 10 m
×
×
DHexamer
2/s
DHexamer
2/s
Figure 5. a) Related macrocycles: pyridinearene 4, octamethylated resorcinarene derivative 5, resorcinarene 6 and pyrogallarene 7 and their respective diffusion
coefficients in CDCl3. b) Concentration dependence of the diffusion coefficients of the monomeric species 1a and the hexameric species I (1 mM − 50 mM in CDCl3).
c) Comparison of the diffusion coefficients of the related hexameric assemblies based on macrocycles 1a, 6 and 7 at different concentrations(1 mM − 50 mM in
CDCl3).
EXSY-spectra of a 5.0 mM sample of 1a in CDCl3, corresponding
to a monomer:hexamer ratio of 29:71, were obtained using
different mixing times.[21] The results revealed an exchange
barrier of 16 – 17 kcal/mol between the two species.[17b]
Assembly I was also studied at different temperatures (20 mM in
CDCl3) utilizing variable temperature (VT) NMR spectroscopy.
While lowering the temperature to −45 °C resulted in considerable
broadening and decreasing quality of the spectra, at higher
temperatures (up to 55 °C) coalescence of the peaks associated
with monomer 1a and hexamer I was observed. This indicates
that the peaks are indeed two different species of the same
compound, which are in equilibrium.
The results presented above are in agreement with the hypothesis
that in chlorinated solvents a dynamic, concentration dependent
self-assembly process between monomer and hexamer is
observed (Figure 4a). The self-assembly of I exhibits a critical
aggregation concentration of about 3 mM in CDCl3 (Figure 4c).
Once this concentration is reached hexamer I is formed in a
cooperative fashion while other multimeric species remain
undetectable at least by NMR spectroscopy. Based on this
conclusion the association constant Ka between monomer 1a and
hexamer I can be formulated as:
to selectively bind fullerenes for separation purposes and to
22]
modulate the electronic properties of fullerenes.[2c,
To our
delight the spherical C60- and C70-fullernes proved to be suitable
guest molecules for cage I (Table 1, Entry 1 and 2) most likely
due to favorable dispersive interactions and potentially π−π-
interactions with the cage walls. The encapsulation of C60 within
the cavity of cage I was indicated by a downfield shift of the C60-
carbon signal in the 13C-NMR spectra in CDCl3. Due to the low
solubility of C60 in CDCl3 or potentially as a result of fast guest
exchange on the time scale of 13C-NMR we were unable to
observe a second signal corresponding to nonencapsulated C60,
even in the presence of excess guest. Using C70 in toluene-d8 in
the presence of I similar results were obtained. The low solubility
and the absence of slow exchange prevented the determination
of binding constants by 13C-NMR. Due to the low concentrations
required for UV/Vis-spectroscopy that prevent the self-assembly
of I, this technique was also not applicable. To still gain some
insight into the encapsulation process by NMR spectroscopy, the
soluble ethyl and tert-butyl malonyl derivatives of C60 and C70
were synthesized. The encapsulation of these derivatives was
indicated by the appearance of a second set of upfield shifted
1
signals (slow exchange) in the H-NMR in presence of I. Since
[
]
Hexamer
only one set of well-defined guest signals was observed, a 1:1
binding mode was assumed. Additional evidence for
encapsulation was provided by DOSY-NMR experiments showing
similar diffusion coefficients for host I and the encapsulated guest
molecules (see SI). Association constants were accordingly
Ka=
6
[
]
Monomer
and was found to be 6.12 ± 1.35 × 1013 mol-5•L5 in CDCl3.
Finally, assembly I was investigated concerning its host−guest
chemistry. Supramolecular structures have been used in the past
5
This article is protected by copyright. All rights reserved.

10.1002/chem.202005046
Chemistry - A European Journal
RESEARCH ARTICLE
determined from the integrals of free and encapsulated guest. The
binding constants were found to range from 750 to 2220 M−1
(Entry 3 – 6) with C60-derivatives binding slightly stronger than the
corresponding C70-derivatives. Contrastingly, in presence of the
hexameric resorcinarene and pyrogallolarene capsules no
indication for the encapsulation of C60((CH(CO2Et)2) was
observed. These results are interpreted as additional evidence for
the formation of a large, discrete cage I, which is able to
accommodate large, spherical molecules.
class feature a much more complex hydrogen bond network.
Furthermore, it features large openings commonly associated
with covalently linked cages but unusual for hydrogen-bonded
assemblies. We believe that porous structures such as the one
presented could potentially be advantageous as they offer an
additional handle for further modifications and/or alternative
binding sites to obtain heteroassemblies. In general, the
simplification of the binding pattern achieved here is expected to
aid rational development of future systems and we are confident
that the results presented here will help to design new, more
sophisticated assemblies in order to overcome the current
limitations concerning size and encapsulation of large molecules
within hydrogen bonded structures.
Table 1. Encapsulation of guest molecules within assembly I.
[M-1
Guest[a]
]
Entry
Ka
[b]
[b]
n. d.
n. d.
1
2
3
4
5
6
C60
C70
Acknowledgements
C
Et))
2220[c]
790[c]
910[c]
750[c]
60(CH(CO2
tBu))
This work was supported by funding from the European Research
Council Horizon 2020 Programme (ERC Starting Grant 714620 –
TERPENECAT) and the Swiss National Science Foundation as
part of the NCCR Molecular Systems Engineering. We thank PD
Dr. Daniel Häussinger and Fabian Bissegger for assistance with
the DOSY-NMR and VT-NMR measurements as well as Dr.
Michael Pfeffer for HR-MS analysis. L. C. thanks JSPS and the
Alexander von Humboldt Foundation for a postdoctoral fellowship.
This work was also supported by a grant from the Israeli Science
Foundation (Grant 1683/18) (DTM)
C
60(CH(CO2
C
Et))
70(CH(CO2
tBu))
C
70(CH(CO2
[a] 1a:guest ratio = 6:1 or 12:1, 1a 20 mM in CDCl
[b] n. d. = not determined due to low solubility;
[c] determined by
3, 16h @ 50 °C;
1H-NMR integration.
Although the results of the DOSY-NMR measurements and the
binding motif observed in the solid state (Figure 2b) indicate the
formation of the larger cage-like structure I, we decided also to
investigate the energy differences between the two possible
hexameric assemblies I and II with computational methods.
According to our gas-phase calculations, the cage structure I is
lower in energy by 4.5 − 17.2 kcal/mol than the alternative closed-
shell structure II, depending on the DFT method and the basis set
used (Table S8). Also in chloroform, an energy preference for
structure I in the range of 1.8 – 14.0 kcal/mol was observed (Table
S9). This energetic difference may be a result of stronger interunit
hydrogen bonds formed in the cage structure I. Both the cage I
and closed-shell structure II have 24 interunit hydrogen bonds,
but these are shorter in the former by 0.04 − 0.05 Å, depending
on the DFT method used (Table S10). These results are fully in
line with the experimental findings of this study.
Keywords: Supramolecular Chemistry • Self-assembly • Host-Guest Complex
References
[1]
[2]
For reviews see: a) L. Avram, Y. Cohen, J. Rebek, Jr., Chem. Commun.
2011, 47, 5368-5375; b) D. Ajami, J. Rebek, Jr., Acc. Chem. Res. 2013,
46, 990-999; c) L. Adriaenssens, P. Ballester, Chem Soc Rev 2013, 42,
3261-3277; d) L. J. Liu, J. Rebek, in Hydrogen Bonded Supramolecular
Structures (Eds.: Z.-T. Li, L.-Z. Wu), Springer Berlin Heidelberg, Berlin,
Heidelberg, 2015, pp. 227-248.
For recent examples see: a) D. Rackauskaite, K. E. Bergquist, Q. Shi, A.
Sundin, E. Butkus, K. Warnmark, E. Orentas, J. Am. Chem. Soc. 2015,
137, 10536-10546; b) D. Beaudoin, F. Rominger, M. Mastalerz, Angew.
Chem. Int. Ed. 2016, 55, 15599-15603; c) G. Markiewicz, A. Jenczak, M.
Kolodziejski, J. J. Holstein, J. K. M. Sanders, A. R. Stefankiewicz, Nat.
Commun. 2017, 8, 15109; d) A. Kiesila, N. K. Beyeh, J. O. Moilanen, R.
Puttreddy, S. Gotz, K. Rissanen, P. Barran, A. Lutzen, E. Kalenius, Org.
Biomol. Chem. 2019, 17, 6980-6984; e) K. Eichstaedt, K. Szpotkowski,
M. Grajda, M. Gilski, S. Wosicki, M. Jaskolski, A. Szumna, Chem. Eur. J.
2019, 25, 3091-3097; f) H. Jedrzejewska, A. Szumna, Chem. Sci. 2019,
10, 4412-4421; g) Y. S. Park, Y. Kim, K. Paek, Org. Lett. 2019, 21, 8300-
8303.
Conclusion
In summary, we have presented the rational design and synthesis
of a new supramolecular assembly based on hydrogen bonding.
Cage I spontaneously self-assembles in chloroform and other
chlorinated solvents based on intermolecular amide−amide
hydrogen bonding as evidenced by detailed NMR studies and
indicated by solid state structures. With an internal cavity volume
of approximately 2800 Å3 cage I represents, to the best of our
knowledge, the largest capsule/cage structure to date solely
based on hydrogen bonding. Comparative DFT studies revealed
a significant energy difference between two possible hexameric
structures clearly favoring the self-assembly of I. The
investigation of the properties of assembly I in apolar solvents
revealed a strong concentration dependence of the self-assembly
process and an affinity for fullerenes, which are encapsulated with
moderate binding constants. Remarkably, cage I is held together
by only 24 intermolecular hydrogen bonds all based on simple
amide−amide dimerization, while most other assemblies of this
[3]
a) N. K. Beyeh, F. Pan, K. Rissanen, Angew. Chem. Int. Ed. 2015, 54,
7303-7307; b) O. Dumele, N. Trapp, F. Diederich, Angew. Chem. Int. Ed.
2015, 54, 12339-12344; c) L. Turunen, U. Warzok, R. Puttreddy, N. K.
Beyeh, C. A. Schalley, K. Rissanen, Angew. Chem. Int. Ed. 2016, 55,
14033-14036; d) O. Dumele, B. Schreib, U. Warzok, N. Trapp, C. A.
Schalley, F. Diederich, Angew. Chem. Int. Ed. 2017, 56, 1152-1157; e)
L. Turunen, A. Peuronen, S. Forsblom, E. Kalenius, M. Lahtinen, K.
Rissanen, Chem. Eur. J. 2017, 23, 11714-11718.
[4]
a) C. L. Gibb, B. C. Gibb, J. Am. Chem. Soc. 2004, 126, 11408-11409;
b) A. Suzuki, K. Kondo, M. Akita, M. Yoshizawa, Angew. Chem. Int. Ed.
2013, 52, 8120-8123; c) J. H. Jordan, B. C. Gibb, Chem Soc Rev 2015,
44, 547-585; d) A. Suzuki, M. Akita, M. Yoshizawa, Chem. Commun.
2016, 52, 10024-10027.
6
This article is protected by copyright. All rights reserved.

10.1002/chem.202005046
Chemistry - A European Journal
RESEARCH ARTICLE
[5]
a) G. Zhang, M. Mastalerz, Chem Soc Rev 2014, 43, 1934-1947; b) M.
Mastalerz, Acc. Chem. Res. 2018, 51, 2411-2422; c) M. A. Little, A. I.
Cooper, Adv. Funct. Mater. 2020, 1909842; d) T. Hasell, A. I. Cooper,
Nat. Rev. Mater. 2016, 1, 16053.
[6]
a) M. J. Wiester, P. A. Ulmann, C. A. Mirkin, Angew. Chem. Int. Ed. 2011,
50, 114-137; b) T. R. Cook, Y. R. Zheng, P. J. Stang, Chem. Rev. 2013,
113, 734-777; c) S. H. Leenders, R. Gramage-Doria, B. de Bruin, J. N.
Reek, Chem Soc Rev 2015, 44, 433-448; d) D. Zhang, T. K. Ronson, J.
R. Nitschke, Acc. Chem. Res. 2018, 51, 2423-2436; e) C. M. Hong, R.
G. Bergman, K. N. Raymond, F. D. Toste, Acc. Chem. Res. 2018, 51,
2447-2455; f) M. Yoshizawa, L. Catti, Acc. Chem. Res. 2019, 52, 2392-
2404.
[7]
a) D. Fujita, Y. Ueda, S. Sato, H. Yokoyama, N. Mizuno, T. Kumasaka,
M. Fujita, Chem 2016, 1, 91-101; b) D. Fujita, Y. Ueda, S. Sato, N.
Mizuno, T. Kumasaka, M. Fujita, Nature 2016, 540, 563-566.
a) K. Muller-Dethlefs, P. Hobza, Chem. Rev. 2000, 100, 143-168; b) P.
A. Wood, F. H. Allen, E. Pidcock, CrystEngComm 2009, 11, 1563-1571.
C. Gaeta, C. Talotta, M. De Rosa, P. La Manna, A. Soriente, P. Neri,
Chem. Eur. J. 2019, 25, 4899-4913.
[8]
[9]
[10] L. R. MacGillivray, J. L. Atwood, Nature 1997, 389, 469-472.
[11] T. Gerkensmeier, W. Iwanek, C. Agena, R. Frohlich, S. Kotila, C. Nather,
J. Mattay, Eur. J. Org. Chem. 1999, 1999, 2257-2262.
[12] a) X. Liu, Y. Liu, G. Li, R. Warmuth, Angew. Chem. Int. Ed. 2006, 45,
901-904; b) X. Liu, R. Warmuth, J. Am. Chem. Soc. 2006, 128, 14120-
14127; c) J. Sun, R. Warmuth, Chem. Commun. 2011, 47, 9351-9353.
[13] K. Kobayashi, M. Yamanaka, Chem Soc Rev 2015, 44, 449-466.
[14] M. Wierzbicki, A. A. Glowacka, M. P. Szymanski, A. Szumna, Chem.
Commun. 2017, 53, 5200-5203.
[15] C. Schäfer, J. Mattay, Photochem. Photobiol. Sci. 2004, 3, 331-333.
[16] R. Taylor, O. Kennard, W. Versichel, Acta. Crystallogr. B 1984, 40, 280-
288.
[17] a) H. C. Chen, S. H. Chen, J. Phys. Chem. 1984, 88, 5118-5121; b) A.
Macchioni, G. Ciancaleoni, C. Zuccaccia, D. Zuccaccia, Chem Soc Rev
2008, 37, 479-489.
[18] a) L. Avram, Y. Cohen, Org. Lett. 2002, 4, 4365-4368; b) L. Avram, Y.
Cohen, Org. Lett. 2003, 5, 3329-3332; c) T. Evan-Salem, Y. Cohen,
Chem. Eur. J. 2007, 13, 7659-7663.
[19] M. H. Abraham, R. J. Abraham, W. E. Acree, Jr., A. E. Aliev, A. J. Leo,
W. L. Whaley, J. Org. Chem. 2014, 79, 11075-11083.
[20] a) J. R. Neale, N. B. Richter, K. E. Merten, K. G. Taylor, S. Singh, L. C.
Waite, N. K. Emery, N. B. Smith, J. Cai, W. M. Pierce, Jr., Bioorg. Med.
Chem. Lett. 2009, 19, 680-683; b) D. Joarder, S. Gayen, R. Sarkar, R.
Bhattacharya, S. Roy, D. K. Maiti, J. Org. Chem. 2019, 84, 8468-8480.
[21] S. Y. Chang, H. Y. Jang, K. S. Jeong, Chem. Eur. J. 2003, 9, 1535-1541.
[22] a) S. J. Park, O. H. Kwon, K. S. Lee, K. Yamaguchi, D. J. Jang, J. I. Hong,
Chem. Eur. J. 2008, 14, 5353-5359; b) N. Kishi, M. Akita, M. Kamiya, S.
Hayashi, H. F. Hsu, M. Yoshizawa, J. Am. Chem. Soc. 2013, 135, 12976-
12979; c) J. Camacho Gonzalez, S. Mondal, F. Ocayo, R. Guajardo‐
Maturana, A. Muñoz ‐ Castro, Int. J. Quantum Chem. 2019, 120,
e26080; d) X. Chang, S. Lin, G. Wang, C. Shang, Z. Wang, K. Liu, Y.
Fang, P. J. Stang, J. Am. Chem. Soc. 2020.
7
This article is protected by copyright. All rights reserved.

10.1002/chem.202005046
Chemistry - A European Journal
RESEARCH ARTICLE
Entry for the Table of Contents
2800 Å
≈
3
•
•
Large Cage Structure (V
)
Concentration-dependent Self-assembly
via Dimerization of Amides
A large, hexameric cage based on amide dimerization self-assembles from a designed macrocyclic building block in apolar solvents.
The unusually large structure (V ~ 2800 Å3) features openings commonly associated with covalently linked cages and forms
host−guest complexes with fullerenes (C60 and C70). The formation of the hexameric structure was found to be strongly concentration
dependent.
Institute and/or researcher Twitter usernames: @TiefenbacherLab
8
This article is protected by copyright. All rights reserved.