Synthesis of interlocked 56-membered rings by dynamic self-templating

Article abstract of DOI:10.1002/anie.200903478

A love of self : Narcissistic macro cyclic rings self-assemble into highly ordered chiral [2]-catenanes displaying high component diastereoselectivity. The picture shows one such structure. One ring is shown in a space-filling representation (C green, O r

Full text of DOI:10.1002/anie.200903478

Angewandte
Chemie
DOI: 10.1002/anie.200903478
Supramolecular Chemistry
Synthesis of Interlocked 56-Membered Rings by Dynamic Self-
Templating**
Mee-Kyung Chung, Peter S. White, Stephen J. Lee, and Michel R. Gagnꢀ*
Reversible templated synthesis is a powerful tool for the
synthesis of supramolecular structures that range from solids
(e.g., MOFs) to complex molecular assemblies and new host–
guest combinations.[1] The error-correction features of rever-
sible synthesis have provided insights into host–guest ther-
modynamics while also offering
spectroscopy), and we postulate a driving force for this
assembly.
When building block d-1m[5] was combined with l,l-2ma
and trifluoroacetic acid (TFA), a mixture of cyclic oligomers
was obtained (Scheme 1); the dynamic library (DL) included
unusual structural examples in
which free energy is minimized.
One such example was the hex-
americ [2]-catenane nucleated by
acetylcholine from dipeptide
hydrazone monomers.[2] The
strong binding of acetylcholine
and the small concentration of
the catenane in its absence sug-
gested a binding site at the inter-
section of the two interlocked
macrocyclic rings. This example
represents one of numerous cases
wherein dynamic templating was
used to discover synthetic recep-
tors by dynamic combinatorial
chemistry (DCC).[3]
We report herein the sponta-
neous self-assembly of new octa-
meric
[2]-catenanes
under Scheme 1. Generation of dynamic libraries (DL) of oligomers from pairs of hydrazide/aldehyde
monomers.
dynamic assembly conditions.
The compounds are assembled
of two different peptide mono-
mers; their formation is highly diastereoselective, and an X-
ray crystal structure reveals structural features that are likely
repeated in previously discovered receptors composed of
these monomers.[4] We additionally report that the solid-state
structure is maintained in solution (as determined by NMR
cyclic dimers, trimers, tetramers, pentamers, hexamers, and
two octamers (Figure 1a). LC-MS analysis revealed that the
major cyclic octamer 3a had a mass of 2723, corresponding to
six units of 1 and two units of 2a, while the minor octamer 4a
had a mass of 2771 (five units of 1 and three units of 2a). The
formation of large oligomers is unusual, as the smaller dimers
and trimers are entropically favored. MS/MS analysis of 3a
showed the principal daughter ion to be the tetrameric
[132a1 + H]+; no other tetrameric ions were detected (Fig-
ure 1b).[6] As no heptameric, hexameric, or pentameric
fragments were detected, these data suggested an octameric
[2]-catenane wherein each tetrameric macrocycle contained
three units of 1 and a single unit of 2a. Monomers having iPr,
Bn, or CH2-indole (Trp) R groups in l,l-2mx did not generate
the corresponding catenanes.
[*] Dr. M.-K. Chung, Dr. P. S. White, Prof. Dr. M. R. Gagnꢀ
Department of Chemistry
University of North Carolina at Chapel Hill
Chapel Hill, NC 27599-3290 (USA)
Fax: (+1)919-962-2388
E-mail: mgagne@unc.edu
Dr. S. J. Lee
US Army Research Office
P.O. Box 12211, Research Triangle Park, NC 27709 (USA)
Since we employed two building blocks of different
chirality, we investigated the effect of different stereoisomers
on [2]-catenane formation. With the exception of the mirror-
image combination (l-1m and d,d-2ma), all combinations
failed to generate any octameric [2]-catenanes.[7] The high
diastereoselectivity and strong 3:1 monomer preference
[**] We thank the Defense Threat Reduction Agency (DTRA) for support
(HDTRA1-08-1-0045). We thank Dr. Marc ter Horst for helpful
discussions of NMR spectroscopy techniques, and Prof. Marcey
Waters (UNC Chapel Hill).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200903478.
Angew. Chem. Int. Ed. 2009, 48, 8683 –8686
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8683

Communications
Figure 1. a) HPLC–UV trace (289 nm) at day seven of an untemplated
DL from d-1m and l,l-2ma (1:1, 5 mm in MeCN/CHCl3 =1:3, 50 equiv
TFA); b) Collision-induced dissociation (CID) MS/MS spectrum of 3a
at a collision energy of 140 V.
Figure 2. a) The X-ray structure of 3e displaying two identical inter-
locked tetramers. One ring is shown as a space-filling representation,
while the symmetry-related ring is shown in a stick representation.
b) View from the bottom of (a) along the C2 axis. Two possible
enchained diastereomers are illustrated in (1) and (2). Octamer 3e
adopts the diastereomeric form (1). c) A single tetrameric ring of 3e
displaying four b-turns (b-1 to b-4).
suggested a tightly ordered structure that was exceptionally
sensitive to component stereochemistry. We also examined
l,l-2m monomers with variable aryl substituents to study
the effect of electronic and steric changes on [2]-catenane
preference. Electron-withdrawing or -donating groups in the
para position and the 2-naphthyl substituent did not signifi-
cantly affect the [2]-catenane equilibrium concentration.[8]
However, when a 1-naphthyl group was employed (l,l-
2me), larger quantities of 3e (29%) and 4e (19%) were
generated. MS/MS analysis of 4e showed two tetrameric
daughter ions ([132e1 + H]+ and [122e2 + H]+), three trimeric
ions, and only two dimeric daughter ions. The fact that [2e2 +
H]+ was not an observed dimeric daughter ion implied that
the [2]-catenane contained one [1·1·1·2e] ring and one
symmetric tetramer with alternating 1 and 2e units, that is,
[1·2e·1·2e].
CH–p[8,11] and hydrogen-bonding interactions are well
known to stabilize host–guest structures.[12] As an aryl or
naphthyl group in 2x was required to form the catenanes, it
can reasonably be inferred that the intercalation of the
naphthyl group between a proline and an AIB residue serves
a stabilization function. Metrical parameters suggests good
geometries to accommodate CH–p interactions between the
naphthyl group and the proline residue (3.9–4.1 ꢀ) and the
naphthyl group and one AIB methyl group (ca. 3.5 ꢀ).[13] The
AIB methyl group preferentially sits over the naphthyl B ring
(Figure 1a), while the proline CH units interact with both.
Edge-to-face p-stacking interactions are also observed in the
intertwined central loop of the structure. The crystal structure
additionally revealed that of the eight NH groups, six
In CHCl3, this same reaction proved to be more selective
for 3e, and an even higher selectivity was observed with a
biased DL (d-1m/l,l-2me = 3:1). Under these conditions, 3e
could be isolated in 64% yield after purification by flash
column chromatography.
=
hydrogen-bond to an amide C O unit, four of which help to
rigidify a b turn and two of which are intermacrocyclic; the
remaining two hydrogen atoms point towards solvent (Fig-
ure 4a). Finally, like the Sanders hexameric [2]-catenane
receptor, 3e is observed as a single enchained diastereomer
(Figure 2b, (1)).
Unlike the Sanders hexameric [2]-catenane receptor,
which gives broad line spectra in the absence of acetylcholine,
3e provides sharp spectra with well-dispersed signals;
[D5]pyridine as solvent gave especially high quality spectra
(Supporting Information, SI Figure 13). Particularly diagnos-
tic were upfield-shifted resonances for the single AIB methyl
group and proline CH units participating in the CH–p
interactions detected in the X-ray analysis. These close
contacts were also apparent in the ROESY spectrum
(Figure 3). One AIB methyl group had cross peaks to the
naphthyl B protons, while the upfield proline hydrogen atoms
are close to both A and B hydrogens (Figure 3). Close contact
Fortunately, 3e could be obtained as single crystals, and
high quality X-ray diffraction data were obtained.[9] As
expected, the octameric [2]-catenane consisted of two iden-
tical interlocked tetramers related by a C2 axis penetrating a
tightly packed central loop, which also relates two floppier
loops (Figure 2a). Each tetrameric ring is saddle-shaped with
three b turns created by the d-Pro-AIB unit (AIB = 2-amino-
isobutyric acid) and a fourth formed by the l-Pro-l-arylGly
moiety; each b turn is connected by a rigid, extended 1,4-
hydrazone/benzamide linkage (Figure 2c). The unique 2e
unit is located in a dense lower core with the naphthyl ring
sandwiched between the b-2 and b-4 turns of the other
macrocycle. A F,Y analysis indicates three type II’ b turns
(b-1, b-2, b-3) for the 1 unit and a type VIII turn for the 2e
unit (b-4).[10]
8684 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 8683 –8686

Angewandte
Chemie
consistent with the solid-state structure being maintained in
solution.
In the downfield portion of the H NMR spectra, four
1
&
^
*
hydrazone NH resonances ( , , 2 ) were observed from d =
10.7–11.3 ppm (Figure 4 and Supporting Information, SI
Figure 13). Their assignments were deduced from cross
peaks in the ROESY spectrum (Supporting Information, SI
Figure 20). The remaining amide NH groups could be differ-
entiated from imine CH resonances by the lack of correlations
to the carbon atoms in the HSQC spectrum (Supporting
~
Information, SI Figures 18 and 19). The amide NH group ( )
in the b-4 turn was uniquely coupled to the adjacent benzylic
CH group (COSY; Supporting Information, SI Figures 14 and
~
15). The amide NH group ( ) in the b-2 turn was also deduced
by cross-peak analysis of the ROESY spectrum (Supporting
Information, SI Figures 21 to 23).
The downfield hydrazone resonances each nucleate their
respective b turns by hydrogen bonding, while the amide NH
groups in b-1 and b-3 apparently orient towards solvent.
Although no direct NMR spectroscopic evidence for hydro-
gen bonding was apparent for the amide NH groups in b-2 and
~ ~
b-4 ( , ), the X-ray structure showed these NHs to partic-
ipate in intermacrocyclic hydrogen bonding.
H/D exchange reactions were informative. In [D4]MeOH,
two of four hydrazone NH groups exchanged quickly (in less
Figure 3. A portion of the 2D 599.8 MHz ROESY spectrum of 3e (F1
d=0.5–1.85 ppm; F2 d=5.75–8.25 ppm; [D5]pyridine, 208C).
^
than 2 h), while a third ( ) was gone after four days; a fourth
*
NH group ( ) was untouched (Figure 4). The two amide NH
groups in the b-1 and b-3 turns were also found to exchange.
When this experiment was repeated in [D5]pyridine/D2O, all
NH groups exchanged except for the two amide NH groups
~
(
~
and ) in b-2 and b-4 (d = 8.98 and 8.51 ppm, respectively).
In the X-ray crystal structure, it is these two turns which
sandwich the naphthyl group.
The exchange experiments are thus consistent with a
densely packed, relatively inaccessible region that is com-
posed of the b-2 and b-4 turns. Additional floppier elements
are also apparent, with b-1 and b-3 being more prone to
solvation and H/D exchange. Cooperative hydrogen bonding
and CH–p interactions thus serve to rigidify and stabilize a
core of the structure that is efficiently nucleated by the arene
ring.
The structural results described herein articulate a unique
solution to the minimization of free energy in a mixed
monomer system. The three-dimensional assembly interlocks
two 56-membered macrocycles and is perfectly diastereose-
lective with respect to the enchainment stereoisomer. It has a
significant preference for a 3:1 ratio of macrocyclic ring
components, and it is highly selective for a specific combina-
tion of diastereomeric monomers.[14] X-ray structural inves-
tigations reveal a dense self-complementary structure that
efficiently utilizes functionality arranged by the Pro-AIB
templated b turns to grasp the complementary structure.
Liberal use of intra- and intermacrocyclic hydrogen bonds,
p stacking, and CH–p interactions conspire to stabilize an
entropically disfavored structure that persists in solution. This
analysis serves as an excellent structural model for cases in
which these subunits assemble into efficient hosts for the
binding of guests.
Figure 4. a) Illustration of the 3e structure; dotted lines refer to
hydrogen bonds noted in the X-ray structure. b) A portion (d=7.8–
11.4 ppm) of the 599.8 MHz H NMR spectra of 3e ([D4]MeOH,
1
208C), recorded at different times from mixing.
between the upfield AIB methyl group and Ha was consistent
À
with a C C distance of 3.0 ꢀ in the X-ray structure. In short,
the observed through-space interactions are completely
Angew. Chem. Int. Ed. 2009, 48, 8683 –8686
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8685

Communications
[5] The m in the compound numbering refers to monomer, which is
slightly different from the repeating unit in the cyclic com-
pounds.
[6] LC-MS analyses also showed that there were no larger fragments
than tetramers. See the Supporting Information (SI Figure 2).
[7] The combination of d-1m/l,d-2ma, d-1m/d,l-2ma, d-1m/d,d-
2ma, l-1m/l,d-2ma, l-1m/l,l-2ma, or l-1m/l,d-2ma failed to
generate the octamer. See details in the Supporting Information,
SI Figure 3.
Experimental Section
Generation of DLs for LC-MS analyses: 5 mm DLs were prepared on
a 1 mL scale. d-1m (5 mmol) and l,l-2mx (5 mmol) were separately
dissolved in 1 mL of a mixture of MeCN and CHCl3 (1:3). 0.5 mL of
the d-1m solution was mixed with 0.5 mL of the l,l-2mx solution and
TFA (50 equiv, 250 mmol, 19 mL). The resultant solution was allowed
to sit for several days until a steady state was reached (as determined
by LC-MS analyses).
Generation and isolation of 3e: Three 5 mm biased DLs (d-1m/
l,l-2me = 3:1) were prepared on a 20 mL scale. Each DL was
prepared by dissolving the mixture of d-1m (29.4 mg, 75.0 mmol) and
l,l-2me (12.3 mg, 25.0 mmol) in CHCl3 (20.0 mL). DLs were allowed
to sit for four days, and then Et3N (50 equiv, 5000 mmol, 0.7 mL) was
added to each DL. To remove the precipitates, the solutions were
filtered and all three were combined. The volatiles were removed in
vacuo, and CHCl3 (60 mL) was added. This solution was washed with
water (3 ꢁ 60 mL) and then dried over anhydrous MgSO4. More
CHCl3 (60 mL) was added, and the mixture was heated to 508C (to
alleviate the poor solubility of 3e in CHCl3) and then filtered to
remove MgSO4. Volatile components were removed in vacuo, and
pure 3e was obtained in 64.0% (67.8 mg) yield after flash column
chromatography on silica gel (8% MeOH/CHCl3). Recrystallization
of 3e from either pyridine by slow evaporation at RT or pyridine/
water at low temperature yielded crystals suitable for X-ray analysis.
[8] K. M. Thomas, D. Naduthambi, N. J. Zondlo, J. Am. Chem. Soc.
2006, 128, 2216 – 2217.
[9] Crystal data for 3e: C88.5H121.5N18.5O25.5 = C76H82N16O12·2.5
(C5H5N)·13.5(H2O), Mr = 1852.54, crystal dimensions 0.28 ꢁ
0.37 ꢁ 0.44 mm, orthorhombic, space group P21212, cell param-
eters a = 26.5088(8), b = 18.1541(5), c = 20.0246(5) ꢀ, V=
9636.7(5) ꢀ, T= 100(2) K, Z = 4, 1calcd = 1.277 MgcmÀ3
, m =
0.790 mmÀ1, radiation CuKa, l = 1.54178 ꢀ. 129815 reflections
were collected to a max q angle of 70.488 (0.82 ꢀ resolution), of
which 18203 were independent (average redundancy 7.132,
completeness = 99.3%, Rint = 0.0346, Rsig = 0.0208) and 16792
(92.25%) were greater than 2s(F2). Data were corrected for
absorption effects using the multiscan method (SADABS). Min/
max apparent transmission = 0.886, min transmission coeffi-
cient = 0.7419, max transmission coefficient = 0.8247. The struc-
ture was solved by direct methods (SHELXL-97) and refined by
full-matrix least-squares methods on F2 with 1214 parameters.
R1 = 0.0946 (I > 2s(I)), wR2 = 0.2736 for 16792 data and R1 =
0.0994, wR2 = 0.2786 for all data, GOF = 1.082; max/min
residual density 1.043/À0.509 eꢀ3 with an RMS deviation
0.096 eꢀ3. CCDC 736965 contains the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Received: June 26, 2009
Published online: October 8, 2009
Keywords: molecular recognition · self-assembly ·
.
supramolecular chemistry · template synthesis
[1] For several representative reviews of this broad topic, see:
a) P. C. Haussmann, J. F. Stoddart, Chem. Rec. 2009, 9, 219 – 226;
b) K. Ariga, J. P. Hill, Y. Wakayama, Phys. Status Solidi A 2008,
205, 1249 – 1257; c) J.-M. Lehn, Chem. Soc. Rev. 2007, 36, 151 –
160; d) J. R. Nitschke, Acc. Chem. Res. 2007, 40, 103 – 112; e) S. J.
Rowan, S. J. Cantrill, G. R. L. Cousins, J. K. M. Sanders, J. F.
Stoddart, Angew. Chem. 2002, 114, 938 – 993; Angew. Chem. Int.
Ed. 2002, 41, 898 – 952.
[2] R. T. S. Lam, A. Belenguer, S. L. Roberts, C. Naumann, T.
Jarrosson, S. Otto, J. K. M. Sanders, Science 2005, 308, 667 – 669.
Other examples include: a) K. R. West, R. F. Ludlow, P. T.
Corbett, P. Besenius, F. M. Mansfeld, P. A. G. Cormack, D. C.
Sherrington, J. M. Goodman, M. C. A. Stuart, S. Otto, J. Am.
Chem. Soc. 2008, 130, 10834 – 10835; b) H. Y. Au-Yeung, G. D.
Pantos¸, J. K. M. Sanders, Proc. Natl. Acad. Sci. USA 2009, 106,
10466 – 10470.
[3] For reviews, see: a) S. Ladame, Org. Biomol. Chem. 2008, 6,
219 – 226; b) R. F. Ludlow, S. Otto, Chem. Soc. Rev. 2008, 37,
101 – 108; c) J.-M. Lehn, Chem. Soc. Rev. 2007, 36, 151 – 160;
d) P. T. Corbett, J. Leclaire, L. Vial, K. R. West, J.-L. Wietor,
J. K. M. Sanders, S. Otto, Chem. Rev. 2006, 106, 3652 – 3711.
[4] a) H. Schiltz, M.-K. Chung, S. J. Lee, M. R. Gagnꢂ, Org. Biomol.
Chem. 2008, 6, 3597 – 3600; b) M.-K. Chung, C. M. Hebling, J. W.
Jorgenson, K. Severin, S. J. Lee, M. R. Gagnꢂ, J. Am. Chem. Soc.
2008, 130, 11819 – 11827; c) J. Liu, K. R. West, C. R. Bondy,
J. K. M. Sanders, Org. Biomol. Chem. 2007, 5, 778 – 786; d) S. M.
Voshell, S. J. Lee, M. R. Gagnꢂ, J. Am. Chem. Soc. 2006, 128,
12422 – 12423; e) S. L. Roberts, R. L. E. Furlan, S. Otto, J. K. M.
Sanders, Org. Biomol. Chem. 2003, 1, 1625 – 1633; f) R. L. E.
Furlan, Y.-F. Ng, S. Otto, J. K. M. Sanders, J. Am. Chem. Soc.
2001, 123, 8876 – 8877; g) G. R. L. Cousins, S.-A. Poulsen,
J. K. M. Sanders, Chem. Commun. 1999, 1575 – 1576.
[10] Proline at the i + 1 position has long been known to promote
b-turn formation. It has also been reported that d-proline in this
position favors type I’ and II’ turns, while l-proline at the same
position prefers type I and II as the common turns and type III as
the most common nonclassic b turn: a) J. R. Lai, B. R. Huck, B.
Weisblum, S. H. Gellman, Biochemistry 2001, 41, 12835 – 12842;
b) C. Das, G. A. Naganagowda, I. L. Karle, P. Balaram, Biopo-
lymers 2001, 58, 335 – 346; c) T. S. Haque, S. H. Gellman, J. Am.
Chem. Soc. 1997, 119, 2303 – 2304; d) H. Santa, M. Ylisirniꢃ, T.
Hassinen, R. Laatikainen, M. Perꢄkylꢄ, Protein Eng. 2002, 15,
651 – 657. For the dihedral angles, see the Supporting Informa-
tion, SI Figure 9.
[11] E. A. Meyer, R. K. Castellano, F. Diederich, Angew. Chem.
2003, 115, 1244 – 1287; Angew. Chem. Int. Ed. 2003, 42, 1210 –
1250.
[12] a) H.-J. Schneider, Angew. Chem. 2009, 121, 3982 – 4036; Angew.
Chem. Int. Ed. 2009, 48, 3924 – 3977; b) D. Pal, P. Chakrabarti,
J. Mol. Biol. 1999, 294, 271 – 288; c) L. Biedermannova, K. E.
Riley, K. Berka, P. Hobza, J. Vondrasek, Phys. Chem. Chem.
Phys. 2008, 10, 6350 – 6359.
[13] The distances were measured between the carbon atom of an
AIB methyl group and three imaginary centroids of the naphthyl
unit or between a centroid of the proline unit and the naphthyl
centroids (see the Supporting Information SI Figure 10). They
all were regarded as CH–p interactions on the basis of the
distance cutoff (d(X···M) < 4.3 ꢀ). See a) T. Steiner, G. Koellner,
J. Mol. Biol. 2001, 305, 535 – 557; b) M. Brandl, M. S. Weiss, A.
Jabs, J. Sꢅhnel, R. Hilgenfeld, J. Mol. Biol. 2001, 307, 357 – 377.
[14] The exact reason for favoring the d-Pro-AIB unit at the other
three positions is not precisely known, although we note that the
l-Pro-l-naphthylGly unit is partially tolerated in b-2 position of
4e.
8686 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 8683 –8686