A water soluble donor-acceptor [2]catenane that can switch between a coplanar and a gemini-sign conformation

Article abstract of DOI:10.1002/anie.201000807

When the moon Is in the seventh house: A conformationally switchable donor-acceptor [2]catenane was synthesized from a dynamic combinatorial library in water. The arrangement of the π units in one of the observed conformations features an unprecedented Gemini shape, which is reminiscent of the astrological Gemini sign (see picture). The catenane can be switched between the parallel and nonparallel conformations upon thermal, chemical, or hydrophobic stimuli. (Figure Presented).

Full text of DOI:10.1002/anie.201000807

Angewandte
Chemie
DOI: 10.1002/anie.201000807
Catenanes
A Water Soluble Donor–Acceptor [2]Catenane that Can Switch
between a Coplanar and a Gemini-Sign Conformation**
Ho Yu Au-Yeung, G. Dan Pantos¸, and Jeremy K. M. Sanders*
We describe herein the discovery of a conformationally
switchable donor–acceptor (D–A) [2]catenane 1, which was
obtained from an aqueous dynamic combinatorial library
(DCL; Scheme 1). One of the two observed conformations of
the D–A units,[1] in which the flat p cores are vertically
stacked in parallel with each other. The T-shape geometry,
although favorable for two interacting p systems, has, to the
best of our knowledge, not previously been observed in D–A
interlocked systems.[2] Previously, we reported the discovery
of D–A catenanes 2 and 3 that contain two interlocked rings
each formed from a naphthalenediimide (NDI; acceptor) and
a dioxynaphthalene (DN; donor).[3] As a result, the D–Aunits
cannot stack in the conventional alternating fashion but adopt
a D–A–A–D stacking geometry, which is one of the three
possible coplanar structures (conformer I, Figure 1). We also
Figure 1. Possible conformers of 1. The NDI and DN cores are
represented by the green and purple shapes.
showed that a [3]pseudorotaxane is formed from 2 and DN 4,
in which the guest is located in the center to create a fully
alternating
D–A–D–A–D sequence. The new catenane 1 was formed in
approximately 30% yield in a high-salt-concentration (1m
NaNO3) aqueous disulfide DCL of the corresponding build-
ing blocks.[4] The interlocked nature of 1 was shown by mass
spectrometry (Figure 2): the molecular ion of 1 has a mass
that corresponds to two of each of the building blocks, and the
largest daughter ion is derived from the D–A dimer. No
homodimeric fragments were observed, thus indicating that 1
is a [2]catenane formed from two interlocked D–A dimers,
analogous to 2 and 3. The difference between 1 and 2 lies in
the substitution pattern of the DN core (2,6- instead of 1,5-
substitution), while 1 is different from 3 because the DN core
is linked with the cysteine through four rather than two
carbon atoms.
Scheme 1. Structures of the two building blocks NDI and DN, cate-
nanes 1–3, and template 4. 1 is presented in conformation I. 1,5-long-
DN and 2-6-short-DN refer to changes in the substitution and carbon
chain lengths, respectively (see text).
catenane 1 is an unprecedented arrangement of the p surfa-
ces, which form the shape of c, the astrological Gemini sign.
Switching between the conformers of this catenane can be
achieved by changing the temperature or solvent polarity, or
by introducing an appropriate guest.
Supramolecular assemblies based on p donors and p ac-
ceptors are often designed with an alternating arrangement of
In contrast to 2 and 3, the 1H NMR spectrum of 1 contains
two sets of NDI and DN signals. Variable-temperature (VT)
1H NMR spectroscopy showed that these two sets of signals
arise from two separate species (designated here as 1a and
1b) that are in slow exchange; the relative ratio of these peaks
varies as a function of temperature (Figure 3).[5] A DH value
of À17 kJmolÀ1 and a DS value of À65 JmolÀ1 KÀ1 for the
equilibrium between 1a and 1b were obtained from the VT
NMR data. The presence of two exchanging species is further
supported by exchange cross-peaks in the NOESY spectrum.
The associated NDI and DN units are assigned from their
integration. The variation in the 1a/1b ratio excludes the
[*] H. Y. Au-Yeung, Dr. G. D. Panto¸s, Prof. J. K. M. Sanders
Department of Chemistry, University of Cambridge
Lensfield Road, Cambridge, CB2 1 EW (UK)
Fax: (+44)1223-336-017
E-mail: jkms@cam.ac.uk
[**] We thank the Croucher Foundation, Pembroke College, and EPSRC
for financial support, and Dr. Ana Belenguer for maintaining the
HPLC laboratory.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201000807.
Angew. Chem. Int. Ed. 2010, 49, 5331 –5334
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5331

Communications
bound and unbound states on the chemical-shift time
scale. In the NOESY spectrum of the host–guest
complex, correlations between the NDI and DN of
1a, and the NDI of 1a and the DN of 4 indicate the
formation of a [3]pseudorotaxane with 4 located in
the center of the catenane (see the Supporting
Information). When less than two equivalents of 4
were present, no change in chemical shifts of 1b was
observed, thus indicating that 1b does not bind to 4.
The binding behavior of 1a to 4 is similar to that
of 2 to 4, with the NDI and DN signals of both
catenanes being shifted upfield upon addition of 4.
This result suggests that the binding of 4 to 1a or 2
involves the same mechanism.[3a] The chemical shifts
of the NDI unit in 1a, 2, and 3 also suggest that these
NDIs are in a similar chemical environment, which is
Figure 2. a) ESI-MS and b) MS–MS spectra of 1.
Figure 3. Partial 1H NMR spectra (500 MHz, D2O) of 1. NDI and DN
signals are labeled as A (green) and D (purple). Peaks corresponding
to a small amount of the D–A dimer are labeled with a blue star.[6]
Figure 4. Partial 1H NMR spectra (500 MHz, D2O, 300 K) of 1 in the
presence of varying amounts of 4. Signals from 4 are labeled as T
(blue).[6]
presence of the unsymmetrical conformation II. HPLC anal-
ysis of the sample showed that it was pure and contained only
one compound, thus confirming the notion of two intercon-
verting conformers on the HPLC timescale.
The NMR data showed that 1 does not adopt exclusively
conformation I, which is expected to be the most favorable.
While it is tempting to assign 1a and 1b respectively as
conformers III (A–D–D–A) and I (D–A–A–D) on the basis
of chemical shifts, this assignment is not consistent with the
observed spectral changes of 1 upon binding with 4 and the
chemical shift values of the equivalent protons for the
structurally similar 2 and 3.
Table 1: Chemical shifts comparison of catenanes 1, 2, and 3.
Catenane d NDI [ppm]
d DN [ppm]
1a
1b
2
8.02 (s)[a]
6.35, 5.97, 5.71 (d, br, br)
7.80, 7.67 (d, d) average=7.74 6.93, 6.44, 6.27 (d, d, s)
8.09, 7.79 (d, d) average=7.94 6.46, 6.11, 5.50 (d, t, d)
3
7.98 (s)
6.32, 6.12, 5.83 (d, br, br)
[a] Letters in parentheses represent multiplicity; br: broad, d: doublet, s:
singlet, t: triplet.
different from that of 1b (Table 1). The DN signals follow the
same trend, thus indicating an analogy between the structures
of 1a, 2, and 3. The NDI protons of the homologues 1a and 3
both appear as a singlet, which shows that both sets of protons
experience the same local symmetry. All these observations
indicate that 1a, 2, and 3 are all in conformation I (D–A–A–D
p stack).
In a 1H NMR titration experiment, addition of 4 resulted
in a decrease in the concentration of 1b relative to that of 1a
(Figure 4). Conversion of 1b to 1a was completed after
addition of more than two equivalents of 4. Upfield shifts
were observed for the NDI and DN resonances of 1a as the
amount of 4 increased, thus indicating fast exchange between
5332
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5331 –5334

Angewandte
Chemie
The observed spectral changes of 1a during the titration
experiment are also consistent with this assignment. The
upfield shifts of both the NDI and DN resonances of 1a upon
binding the guest can be explained by a fast threading–
dethreading process. When threaded with the guest, the
catenane is tightened and the p units are closer. Therefore,
both the NDI and DN units experience a stronger shielding
effect and are shifted upfield. These changes cannot be
accounted for by the formation of conformer III. Since the
NOESY experiment showed that the final catenane complex
has an alternate D–A stack with 4 located in the center, the
binding of 4 to 1 in conformer III implies a binding
mechanism that is accompanied by a conformational change
of the host. Although this change is consistent with the upfield
shifts of the NDI as they move from the outermost to the
inner part of the stack, it contradicts the upfield shifts
observed for the DN. By moving from the inner to the
outermost position, the DNs are influenced by fewer neigh-
boring aromatic units, and should be less shielded by ring
currents and therefore downfield-shifted.[7]
Figure 5. Partial 1H NMR spectra (500 MHz, D2O/[D6]acetone, 300 K)
of 1 with varying [D6]acetone content.[6]
While all the available experimental evidence supports
the assignment of 1a as conformer I (D–A–A–D p stack), the
conformation of 1b is enigmatic: conformer II has been
eliminated as it requires the presence of two sets of NDI and
DN signals in a 1:1 ratio, while conformation III cannot
explain the relative chemical shift values of 1b. In conformer
III (A–D–D–A p-stack), the NDI and DN would be on the
outermost and inner positions, respectively, of the parallel
stack. Therefore, the corresponding signals should show
larger downfield and upfield shifts, respectively, compared
with those of conformer I. However, when compared to 1a,
which is assigned as conformer I, the NDI and DN signals of
1b show larger upfield and downfield shifts, which are
contrary to the expected behavior (see above).
Therefore, a fourth conformation where the four p units
do not stack in parallel is required in order to explain these
observations. In conformation IV, the two NDI cores are
parallel to each other, perhaps offset, and this NDI stack is
placed in between the DN moieties in an arrangement that is
best represented as c, the astrological Gemini sign. In this
conformation, the chemical shifts of the NDI protons will be
shifted upfield relative to conformer I as they are deeper in
the shielding region of the DNs; on the other hand, the DN
signals will be shifted downfield as the DNs are in the
deshielding region of the NDIs. These observations are
consistent with a T-shape type of interaction that occurs
between the NDI and the DN cores of catenane 1b, and
operates in tandem with face-to-face NDI contacts.
on the D–A stacking order, the overall area of exposed
hydrophobic surfaces is almost identical for each conformer,
so the hydrophobic effect should not have a significant
influence on the relative stability and ratio of 1a/1b if
conformers I–III were the only possibilities. The existence of
a fourth conformer is therefore essential. Further support for
the structural differences of 1a and 1b comes from the
NOESY spectrum, in which NOE cross-peaks between the
NDI and DN protons were observed for a sample in which 1b
was the dominant conformer (D2O with ca. 40% acetone).
This result shows that the NDI and DN are in proximity in 1b,
and is consistent with the tilted NDI arrangement that brings
the NDI and DN protons closer. On the other hand, no NOE
cross-peaks were observed in the NOESY spectrum of a
sample in which 1a was the main conformer in solution (0.5m
NaNO3, see the Supporting Information). Despite the lack of
detailed structural parameters, the presence of a nonparallel
conformation is necessary based on all these observations.
In summary, we have shown that a new flexible D–A
[2]catenane 1, which was synthesized from a dynamic
combinatorial library, can adopt different conformations in
solution. NMR studies showed that one of the observed
conformers of 1 has a conformation, which we believe to be
previously unknown, where the electronically complementary
p units do not stack in parallel. In addition, we can
manipulate the conformational equilibrium by thermal,
solvophobic, or chemical stimuli, in contrast to the more
common electrochemical methods.[9] The discovery of 1 also
highlights the use of DCC in uncovering new unexpected
structures with features that could not have been designed or
predicted.
Compared to conformers I–III, conformer IV has a larger
exposed hydrophobic surface area, which implies that con-
former IV will be the dominant species in a solvent of lower
ionic strength.[8] Indeed, in a 1H NMR experiment, the ratio of
1a/1b decreased as the acetone content was increased and no
significant amount of 1a was observed above approximately
40% (v/v) of acetone (Figure 5). In agreement with our
hypothesis, the ratio of 1a/1b increased as the ionic strength
was increased, and no significant amount of 1b was observed
in the presence of 0.5m NaNO3 (see the Supporting Informa-
tion). Although the stability of conformers I–III is dependent
Received: February 9, 2010
Revised: April 16, 2010
Published online: June 22, 2010
Angew. Chem. Int. Ed. 2010, 49, 5331 –5334
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5333

Communications
obtained from different preparative HPLC runs. The peak at
6.17 ppm in Figure 5 and 6 is an artefact that arises from the zg30
pulse program used in data acquisition.
Keywords: catenanes · host–guest systems · hydrophobic effect ·
p interactions · self-assembly
.
[7] Other p guests were also tested, and binding of 1 to 4 was found to
be specific. The use of an NDI-derived guest resulted in catenated
species in a conformation other than I and IV, whilst other
aromatic guests did not interact with 1. See the Supporting
Information for details.
[8] For examples of conformational changes of rotaxanes induced by
solvophobic effect, see: D. A. Leigh, M. ꢂ. F. Morales, E. M.
Pꢀrez, J. K. Y. Wong, C. G. Saiz, A. M. Z. Slawin, A. J. Carmi-
chael, D. M. Haddleton, A. M. Brouwer, W. J. Buma, G. W. H.
Wurpel, S. Leon, F. Zerbetto, Angew. Chem. 2005, 117, 3122 –
3127; Angew. Chem. Int. Ed. 2005, 44, 3062 – 3067.
[1] For recent examples of [2]catenanes with alternating D–A units,
see: a) D. Cao, M. Amelia, L. M. Klivansky, G. Koshkakaryan,
S. I. Khan, M. Semeraro, S. Silvi, M. Venturi, A. Credi, Y. Liu,
J. Am. Chem. Soc. 2010, 132, 1110 – 1122; b) S. Ramos, E. Alcalde,
J. F. Stoddart, A. J. P. White, D. J. Williams, L. Pꢀrez-Garcꢁa, New
J. Chem. 2009, 33, 300 – 317; c) G. Koshkakaryan, K. Parimal, J.
He, X. Zhang, Z. Abliz, A. H. Flood, Y. Liu, Chem. Eur. J. 2008,
14, 10211 – 10218; d) Y. Liu, A. Bruneau, J. He, Z. Abliz, Org.
Lett. 2008, 10, 765 – 768; e) A. Coskun, S. Saha, I. Aprahamian,
J. F. Stoddart, Org. Lett. 2008, 10, 3187 – 3190.
[9] For recent examples, see: a) Y.-L. Zhao, A. Trabolsi, J. F.
Stoddart, Chem. Commun. 2009, 4844 – 4846; b) J. M. Spruell,
W. F. Paxton, J.-C. Olsen, D. Benꢁtez, E. Tkatchouk, C. L. Stern,
A. Trabolsi, D. C. Friedman, W. A. Goddard III, J. F. Stoddart,
J. Am. Chem. Soc. 2009, 131, 11571 – 11580; c) R. Klajn, L. Fang,
A. Coskun, M. A. Olson, P. J. Wesson, J. F. Stoddart, B. A.
Grzybowski, J. Am. Chem. Soc. 2009, 131, 4233 – 4235; d) M. A.
Olson, A. B. Braunschweig, L. Fang, T. Ikeda, R. Klajn, A.
Trabolsi, P. J. Wesson, D. Benꢁtez, C. A. Nirkin, B. A. Grzybowski,
J. F. Stoddart, Angew. Chem. 2009, 121, 1824 – 1829; Angew.
Chem. Int. Ed. 2009, 48, 1792 – 1797.
[2] C. A. Hunter, J. K. M. Sanders, J. Am. Chem. Soc. 1990, 112,
5525 – 5534.
[3] a) H. Y. Au-Yeung, G. D. Pantos¸, J. K. M. Sanders, Proc. Natl.
Acad. Sci. USA 2009, 106, 10466 – 10470; b) H. Y. Au-Yeung,
G. D. Pantos¸, J. K. M. Sanders, J. Am. Chem. Soc. 2009, 131,
16030 – 16032.
[4] H. Y. Au-Yeung, P. Pengo, G. D. Pantos¸, S. Otto, J. K. M. Sanders,
Chem. Commun. 2009, 419 – 421.
[5] No significant difference was observed in the UV/Vis spectra of 1
at different temperatures.
[6] The ratio between 1a and 1b varies with the amount of polar
solvent impurities (mainly MeOH) present in the samples
5334
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5331 –5334