Spontaneous and catalytic fusion of supramolecules

Article abstract of DOI:10.1039/c2cc35036j

Using principles of completive and integrative self-sorting, a clean supramolecule-to-supramolecule transformation is realised that involves fusion of a 3-component rectangle and a 2-component equilateral triangle into a 5-component scalene triangle. While the spontaneous process takes 15 h at 25 °C, the catalytic transformation is completed within 1 h.

Full text of DOI:10.1039/c2cc35036j

View Online / Journal Homepage / Table of Contents for this issue
Chemical Communications
www.rsc.org/chemcomm
Volume 48 | Number 76 | 4 October 2012 | Pages 9429–9544
ISSN 1359-7345
COMMUNICATION
Michael Schmittel et al.
Spontaneous and catalytic fusion of supramolecules

Dynamic Article Links
ChemComm
View Online
Cite this: Chem. Commun., 2012, 48, 9459–9461
www.rsc.org/chemcomm
COMMUNICATION
Spontaneous and catalytic fusion of supramoleculesw
Manik Lal Saha, Susnata Pramanik and Michael Schmittel*
Received 13th July 2012, Accepted 1st August 2012
DOI: 10.1039/c2cc35036j
Using principles of completive and integrative self-sorting, a
clean supramolecule-to-supramolecule transformation is realised
that involves fusion of a 3-component rectangle and a 2-component
equilateral triangle into a 5-component scalene triangle. While the
spontaneous process takes 15 h at 25 8C, the catalytic transforma-
tion is completed within 1 h.
Gene shuffling, i.e. combination of dissimilar genes to create
novel and improved recombinant genes, is one of the key
Scheme 1 Clean supramolecular fusion of T1 and R1.
mechanisms through which new genes emerge in Nature.1 In
fact, gene shuffling is a highly effective mechanism for innova-
tion because it can generate new genes with structures and
functions drastically different from those of the parental
progenitors. Similarly, the use of an evolutionary mechanism
may allow us to design and fabricate intricate new aggregates
starting from simple self-assemblies, i.e. eventually from
libraries of supramolecular assemblies. However, such a
protocol inherently demands, in contrast to the typical
bottom-up approach, a clean supramolecule-to-supramolecule
transformation2 with at least two distinct self-assemblies
merging into a new assembly via shuffling of their components,
a process described only once to date. In an example reported
by Stang et al.3 two homoleptic 2-component architectures
merge into a heteroleptic 3-component system upon heating.
While this reaction may be considered somewhat analogous to
gene shuffling, biological processes usually occur under con-
ditions that are tightly regulated, often by enzymatic action. In
order to account for regulation, we report herein not only on
the spontaneous but also catalytic fusion of two supra-
molecular assemblies (Scheme 1). Explicitly, the two distinct
and dynamic supramolecules T1 and R1 endure spontaneous
and catalytic reshuffling of their components upon mixing and
evolve as the 5-component scalene triangle T2, a rare topology.4
Clearly, the diversity and complexity of T2 are much increased
due to the enlarged number of different components and
reversible orthogonal interactions.5
mononuclear cornerstones in ‘disfavoured’ combinations,
e.g. in T1 and R1. Upon mixing, both aggregates should be
strongly biased to reshuffle their components and reassemble
in their thermodynamically ‘favoured’ pair, as allocated in T2.
In order to set up the high-fidelity self-sorting required in
Scheme 1, we first interrogated path a in Scheme 2. To check
for alternative combinations, we prepared separately the pure
complexes C1 = [Cu(1)(4)](PF6) and C2 = [Zn(2)(5)](OTf)2
using the HETPHEN8 and HETTAP9 complexation approach.
Upon their mixing in a 1 : 1 ratio in MeCN at 25 1C, clean
formation of C3 = [Cu(4)(5)](PF6) and C4 = [Zn(1)(2)](OTf)2
is observed after ca. 30 min (Scheme 2, path a), as evidenced by
1H-NMR analysis (Fig. S5, ESIw). The ease of component
shuffling indicates that the kinetic barrier for such a process is
not too high. As expected, the thermodynamically driven self-
sorting is similar to one of the free components,4c as shown in
path b (Scheme 2). Component exchange of C1 and C2 is
equally successful in the presence of the preassembled pyridine–
zinc porphyrin complex C5 = [(3)(6)] (Scheme 2, path a).
Re-assembly of the 8-component library to C3–C5 is thus
warranted irrespective of the preassembled state of ligands
1–5, two metals (Cu+ and Zn2+) and zinc porphyrin 6.
The flawless error correction between two heteroleptic
complexes, as represented above, encouraged us to engineer
a supramolecule-to-supramolecule fusion (Scheme 1). In order
At the start, using an approach based on completive4
and integrative self-sorting,6,7 we decided to preassemble the
Center of Micro and Nanochemistry and Engineering,
Organische Chemie I, Universitat Siegen, Adolf-Reichwein-Str. 2,
¨
D-57068 Siegen, Germany. E-mail: schmittel@chemie.uni-siegen.de;
Tel: +49 271740-4356
w Electronic supplementary information (ESI) available: Experimental
procedures and spectroscopic data for ligand 7 and all complex
(T1, T2, R1) assemblies. See DOI: 10.1039/c2cc35036j
Scheme 2 Reshuffling of 32,3,3-fold(8) completive library.4c
Chem. Commun., 2012, 48, 9459–9461 9459
c
This journal is The Royal Society of Chemistry 2012

View Online
the existence of two diastereomers in a 1 : 3 ratio, as a result of
three stereogenic cornerstones of the type [Cu(9phen)(9phenAr2)]+.10
Notably, the c0-H and c00-H (Scheme 3) appear as 4 singlets with a
ratio 1 : 1 : 1 : 1. Apparently, in one of the diastereomers all
3 ligands show up in a single set due to its C3 symmetry
(P*,P*,P*), while in the second isomer (P*,P*,M*) the protons
emerge as three distinct sets (1 : 1 : 1) (Fig. S27, ESIw). Other
valuable information arises from the 1H-NMR signals of the
OMe protons belonging to ligand 9. All four OMe protons of
ligand 9 in T1 appear as separate singlets due to their diastereo-
topicity caused by the stereogenic [Cu(9phen)(9phenAr2)]+ units.10
The collected spectroscopic evidence convincingly estab-
lishes the integrity of both T1 and R1 in solution. T1 and
R1 are thermodynamically favoured over any larger structures
because they constitute discrete entities at the lowest possible
entropic costs while realising maximum site occupancy.
Finally, in order to test our concept presented in Scheme 1,
T1 and R1 were mixed in a 2 : 3 ratio and kept under NMR
surveillance for 15 h at 25 1C. The progress of the reaction was
monitored in 0.5 h intervals (Fig. S9, ESIw) and may be best
judged by characteristic resonances of R1 (ligand 8, 4-H,
and x, x0) and T1 (ligand 9, b0, b00). They reveal that the
spectrum gradually becomes simpler, with full convergence
into a single set after 15 h. Additionally, a single diffusion
coefficient obtained in the DOSY provides unambiguous
evidence for the clean formation of a single product, identified
as T2 below.
Scheme 3 (a) Chemical structures of ligands 7–10. (b) Synthesis of
rectangle R1 and equilateral triangle T1.
to establish positional control, we installed the ligands 2 and 6
as termini of the new porphyrin–terpyridine hybrid 7, readily
accessible via Sonogashira cross-coupling. Along known
procedures, the coordination units of 3 and 5 were com-
bined within the phenanthroline–pyridine hybrid 84c and
the complexing properties of 1 and 4 in the unsymme-
trical bisphenanthroline 9.4b For all ligands, spacers were
chosen to render the building blocks 7–9 unequal in length
(Scheme 3a).
In a first experiment, ligands 7 and 8 as well as Zn(OTf)2
were mixed in a 1 : 1 : 1 ratio and refluxed for 1 h in
MeCN–DCM (1 : 8) (Scheme 3b). The reaction mixture was
characterised without any further purification by 1H-NMR,
1H–1H COSY, diffusion-ordered spectroscopy (DOSY),
elemental analysis, and electrospray ionisation mass spectro-
metry (ESI-MS). All spectroscopic data confirm clean for-
mation and integrity of rectangle R1 (Fig. S25, ESIw), a
discrete 3-component nanostructure. 1H-NMR shows a diagnostic
upfield shift of the mesityl protons in R1 (x, x0, d = 6.30 ppm)
as compared to those in free 8 (d = 6.92 ppm).4c In light of the
HETTAP concept, an intimate p–p stacking between the
mesityl groups of 8 and the terpyridine of 7 is responsible
for such a shift in the 1H-NMR of R1.9 Furthermore, the
pyridine protons (a, b) of 8 in R1 experience upfield shifts from
8.61 to 3.41 ppm and from 7.38 to 5.62 ppm, respectively, that
are diagnostic for pyridine protons upon axial coordination to
a zinc porphyrin.8b 1H-NMR peak assignments were corrobo-
rated by COSY experiments (Fig. S11, ESIw). The ESI-MS of
R1 further substantiates the structure by showing signals
m/z = 1011.7 and 1399.0 (Fig. S19, ESIw) corresponding to
[Zn2(7)2(8)2]4+ and [Zn2(7)2(8)2](OTf)3+, respectively. Finally,
exclusive formation of R1 was confirmed by a single diffusion
coefficient in the DOSY NMR (Fig. S14, ESIw).
In another experiment, we studied the coordination beha-
viour of ligand 9 and [Cu(CH3CN)4]PF6 when mixed in a 1 : 1
ratio and refluxed for 1 h in MeCN–DCM (4 : 1) to furnish a
clear red solution of T1 (Scheme 3b and Fig. S26 in ESIw). The
product was characterised by ESI-MS, 1H-NMR, DOSY, and
elemental analysis. The ESI-MS spectrum exhibits two major
peaks at m/z = 962.0, 1515.7 for [Cu3(9)3]3+, [Cu3(9)3](PF6)2+
in full accord with assembly T1 (Fig. S20, ESIw). The experi-
mental isotopic splitting pattern of the major peak is in full
agreement with its calculated splitting pattern. Importantly,
no other peak was noticed in the spectral region between 150
and 2000 Da. 1H-NMR and DOSY NMR further support the
assignment. Exclusive formation of T1 in solution was con-
firmed by a single diffusion coefficient in the DOSY NMR. To
substantiate the clean self-assembly process, we carefully
interrogated the 1H-NMR of T1. It is complicated due to
In order to examine the connectivity of the ligands in the
putative T2, we paid special attention to several characteristic
proton resonances in the 1H-NMR. For example, the 0.44 and
0.52 ppm downfield shifts of b0 and b00-H (from 6.52–6.38 ppm
in T1 to 6.96 and 6.90 ppm in T2) are indicative of a HETTAP
complex between 7 and 9 (Scheme 1 and Fig. 1). In contrast to
R1, the pyridine b-proton in T2 experiences a slight downfield
shift of 0.26 ppm (from 5.62 to 5.88 ppm), indicating the
survival of the pyridine–zinc porphyrin interaction in solution.
Similar to T1, the suggested structure requires that T2 is chiral,
due to the stereogenic [Cu(9phen)(8phenAr2)]+ unit.4c,10 As a
result, the mesityl protons x and x0-H of ligand 8 being
enantiotropic in R1 become diastereotopic in T2. This is
corroborated by the existence of two singlets at 6.15 and
5.97 ppm in T2. The 0.15 and 0.33 ppm upfield shift (from
6.30 ppm in R1) of the two mesityl protons (x and x0) in T2
also supports the presence of a [Cu(9phen)(8phenAr2)]+ HETPHEN
coordination center. As further support, four singlets for the
diastereotopic methoxy protons (OMe) appear at 2.73–2.83 ppm,
a range typical for HETTAP complexes.4a
Fig. 1 Comparison of partial 1H-NMR spectra of (a) R1 (CD2Cl2),
(b) T1 (CD3CN), (c) T1 : R1 (2 : 3) (CD3CN : CD2Cl2 = 4 : 1) after
16 h at 25 1C, and (d) T1 : R1 (2 : 3) (CD3CN : CD2Cl2 = 4 : 1) with
10 (10 mol% with respect to R1) after 1.5 h at 25 1C.
c
9460 Chem. Commun., 2012, 48, 9459–9461
This journal is The Royal Society of Chemistry 2012

View Online
The fusion was additionally followed by ESI-MS. Upon
mixing T1 and R1, initially only the starting materials are
observable in the ESI-MS spectra (Fig. S21, ESIw). However,
over time they convert to T2. Finally, after 15 h, the ESI-MS
spectrum exhibits no more signals corresponding to T1 or R1,
but only peaks being in full agreement with the newly formed
triangle T2 (Fig. S22, ESIw). For example, the signals at
m/z = 995.7 and 1565.6 Da represent [ZnCu(7)(8)(9)]3+ and
[ZnCu(7)(8)(9)](PF6)2+, respectively. The experimental iso-
topic splitting of all major peaks agrees with the calculated
one.
in appropriate stoichiometry is without precedence. Further-
more, the supramolecular fusion is readily catalysed.
In our view, a supramolecular fusion is more valuable than
a bottom-up self-assembly of all constituents, as both pre-
cursor supramolecules already represent sophisticated
chemical information. While T1 is fully defined by the length
of one side, description of R1 requires two inputs and that of
T2 at least three inputs, i.e. three different lengths. The fusion
thus not only involves a first step toward evolution of supra-
molecular architectures but equally to higher information
content.5c
Existence of the [Cu(9phen)(8phenAr2)]+ unit in one of the
metal corners of T2 was also interrogated by differential pulse
voltammetry (DPV) probing the Cu+ oxidation wave. A single
oxidation wave at 0.70 VSCE in T2 (Fig. S24, ESIw) confirms
the presence of only one type of copper(I) complex, pointing
persuasively to the formation of [Cu(9phen)(8phenAr2)]+.4c
A combination of ESI-MS, 1H-NMR, DPV, DOSY, and
elemental analysis thus unambiguously provides evidence for
the clean formation of scalene triangle T2.
Looking more on molecular details, the reported example
has clearly some analogy to gene-shuffling, chiefly to indel
mutations (Fig. S29, ESIw)1b because T2 forms by insertion
and deletion of subunits that are delivered by T1 and R1.
Efforts to extend this strategy to multi-stage adaptive assem-
blies are underway in our laboratory.
We are indebted to the DFG and the University of Siegen
for financial support. We thank Dr T. Paululat/University of
Siegen for 600 MHz NMR measurements.
All attempts to obtain single crystals of T2 met with failure.
Fortunately, MM+ force field computations on T2 provide
some insight into the scalene triangular structure. Taking the
metal–metal distance as a measure, the three metal corners of
T2 are separated by 1.70, 1.86, and 1.95 nm in the energy
minimised structure (Fig. S28, ESIw), nicely illustrating the
geometrically scalene arrangement of T2.
Notes and references
1 (a) G. C. Conant and A. Wagner, Genome Biol., 2005, 6, R50;
´
(b) S. Pasek, J.-L. Risler and P. Brezellec, Bioinformatics, 2006,
22, 1418.
2 (a) S.-S. Sun, J. A. Anspach and A. J. Lees, Inorg. Chem., 2002,
41, 1862; (b) J. Heo, Y.-M. Jeon and C. A. Mirkin, J. Am. Chem.
Soc., 2007, 129, 7712; (c) L. Zhao, B. H. Northrop and P. J. Stang,
J. Am. Chem. Soc., 2008, 130, 11886; (d) S. Hiraoka, Y. Sakata and
M. Shionoya, J. Am. Chem. Soc., 2008, 130, 10058;
(e) V. E. Campbell, X. D. Hatten, N. Delsuc, B. Kauffmann,
I. Huc and J. R. Nitschke, Nat. Chem., 2010, 2, 684; (f) K. Parimal,
E. H. Witlicki and A. H. Flood, Angew. Chem., Int. Ed., 2010,
49, 4628; (g) L.-P. Cao, J.-G. Wang, J.-Y. Ding, A.-X. Wu and
L. Isaacs, Chem. Commun., 2011, 47, 8548; (h) B. Kilbas,
S. Mirtschin, R. Scopelliti and K. Severin, Chem. Sci., 2012, 3,
701.
3 Y.-R. Zheng, Z. Zhao, M. Wang, K. Ghosh, J. B. Pollock,
T. R. Cook and P. J. Stang, J. Am. Chem. Soc., 2010, 132, 16873.
4 (a) K. Mahata and M. Schmittel, J. Am. Chem. Soc., 2009,
131, 16544; (b) M. Schmittel and K. Mahata, Chem. Commun.,
2010, 46, 4163; (c) K. Mahata, M. L. Saha and M. Schmittel,
J. Am. Chem. Soc., 2010, 132, 15933; (d) M. L. Saha and
M. Schmittel, Org. Biomol. Chem., 2012, 10, 4651.
5 (a) M. Schmittel and K. Mahata, Angew. Chem., Int. Ed., 2008,
47, 5284; (b) N. Christinat, R. Scopelliti and K. Severin, Angew.
Chem., Int. Ed., 2008, 47, 1848; (c) S. De, K. Mahata and
M. Schmittel, Chem. Soc. Rev., 2010, 39, 1555.
6 (a) M. M. Safont-Sempere, G. Ferna
Rev., 2011, 111, 5784; (b) K. Osowska and O. S. Miljanic
2011, 1643; (c) R. Chakrabarty, P. S. Mukherjee and P. J. Stang,
Chem. Rev., 2011, 111, 6810.
7 (a) W. Jiang, H. D. F. Winkler and C. A. Schalley, J. Am. Chem.
Soc., 2008, 130, 13852; (b) W. Jiang and C. A. Schalley, Proc. Natl.
Acad. Sci. U. S. A., 2009, 106, 10425.
In an effort to accelerate the process we anticipated that
labilisation of metal–ligand bonds11 may shorten the time of
the fusion. We envisaged 2-methylpyridine (10) to be a suitable
candidate because the methyl group should prevent any strong
binding to the zinc porphyrin unit of 7.12 Indeed, in a control
experiment, 10 mol% of 10 (related to the initial amount
of R1) was added to a 2 : 3 mixture of T1 and R1 at 25 1C and
the reaction monitored by NMR. To our delight, the trans-
formation was effected in ca. 1 h, as suggested by diagnostic
shifts in the NMR signals (Fig. 1). Clearly, the spectrum in
Fig. 1c resembles very much the spectrum in Fig. 1d. To
further prove the integrity of T2 generated in the catalytic
process, we measured its ESI-MS. We only observe peaks for
T2, indicating that 10 does not lead to destruction of the
triangular assembly (Fig. S23, ESIw).
The above fusion of supramolecules comprises several
distinct chemical events, including (i) self-correction under
thermodynamic control; (ii) favoured pair-selection due to
high-fidelity self-sorting; and (iii) acceleration of a supra-
molecular fusion reaction via labilisation of the metal–ligand
bonds.
´
ndez and F. Wurthner, Chem.
´
¨
, Synlett,
In conclusion, the present strategy based on completive and
integrative self-sorting describes a viable means for cons-
tructing topologically demanding supramolecules starting
from simpler supramolecular aggregates. As a demonstration
we describe herein the shuffling and recombination of com-
ponents from the 2-component equilateral triangle T1 and
3-component rectangle R1 to the 5-component scalene triangle
T2. To the best of our knowledge, the fabrication of a clean
multicomponent (n > 3) assembly by just mixing two assemblies
8 (a) M. Schmittel and A. Ganz, Chem. Commun., 1997, 999;
(b) R. S. K. Kishore, V. Kalsani and M. Schmittel, Chem. Commun.,
2006, 3690.
9 M. Schmittel, R. S. K. Kishore, H. Colfen and J. W. Bats, J. Am.
Chem. Soc., 2005, 127, 11544.
¨
10 phen
= [1,10]-phenanthroline; phenAr2 = 2,9-diaryl[1,10]-
phenanthroline.
11 D. Fujita, A. Takahashi, S. Sato and M. Fujita, J. Am. Chem. Soc.,
2011, 133, 13317.
12 C. H. Kirkesy, P. Hambright and C. B. Storm, Inorg. Chem., 1969,
8, 2141.
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 9459–9461 9461