Spontaneous knotting - From oligoamide threads to trefoil knots

Article abstract of DOI:10.1002/anie.200601938

(Figure Presented) Tying up loose ends: Oligoamide thread molecules can spontaneously transform into open knots, whose terminal functional groups can subsequently be chemically linked to form closed, cyclic knots (see scheme; left: threads; middle: open k

Full text of DOI:10.1002/anie.200601938

Communications
DOI: 10.1002/anie.200601938
Molecular Knots
Spontaneous Knotting—From Oligoamide Threads to Trefoil Knots**
Jens Brüggemann, Stephan Bitter, Sonja Müller, Walter M. Müller, Ute Müller, Norbert M. Maier,
Wolfgang Lindner,* and Fritz Vögtle*
Dedicated to Professor Vincenzo Balzani on the occasion of his 70th birthday
Until now oligoamide-based molecular knots were accessible
only by intermolecular one-pot condensation of three di-
amide molecules 1 and three acid chloride molecules 2
(Scheme 1, Route A).[1–3] To explain the process of knotting
we suggested the intermediate formation of longer oligo-
amide threads 3a or 4.[2] From crystal structure analyses[2] of
amide knots like 6, as well as further experimental and
theoretical data,[4] we assumed that folding of linear thread
precursors like 3a or 4 (Scheme 1, Routes B/C) to knotted
thread parts might be preprogrammed on the basis of
favorable hydrogen-bond patterns in noncompetitive solvents
(for example, dichloromethane).[5] The intermediate forma-
tion of 3b from 5 cannot be ruled out completely, although the
reaction conditions (stoichiometry of the addition of 2a:
one equivalent in Route B, two equivalents in Route C) do
not really support this assumption. Herein, we report the first
synthesis, isolation, and characterization of threads 3a and 4,
as well as their successful conversion into the corresponding
knotanes 6 (Scheme 1).
spontaneously creates an open knot 3b (Scheme 1,
Route B).[6] To prove the existence of this intertwined
structure, we treated the isolated decaamide with various 4-
substituted pyridine dicarboxylic acid dichlorides 2a–d
(Scheme 3). In the case of the unknotted thread 3a, this
reaction should yield an achiral macromonocycle 7, whereas
the open knot 3b should lead to an isomeric (closed)
topologically chiral knot 6 with three pyridine units. This
reaction also opens up a new class of monosubstituted knots
6a–d if a substituted pyridine dicarboxylic acid dichloride 2 is
used instead (Scheme 3).
On the one hand, this strategy gives more insight into the
template mechanism of the knotting of neutral (uncharged)
molecules (without cation assistance[3]), and on the other
hand, it makes the synthesis of new trefoil knots with different
subunits possible. Such a spontaneous self-knotting process
3a!3b of low-molecular synthetic thread molecules on a
preparative scale has, to the best of our knowledge, not been
reported before.[7]
The synthesis of the elongated threads 3a and 4 as
potential precursors for the formation of amide knots is
depicted in Scheme 2. This synthesis opens up the possibility
to distinguish between Routes B and C. Therefore the
isolated threads 3a and 4 were separately treated with 2 and
with 1 and 2, respectively, and it was then examined whether
the molecular knot 6 was formed.
Indeed the reaction of the long thread 3b with pyridine
dicarboxylic acid dichloride (2a) yields the unsubstituted
knot 6a with three identical pyridine units, which we already
synthesized previously by Route A, and this product is in
accordance with our proposed mechanism for Route B. The
reaction yielded 13 mg (11%) of the pure knotane 6a.
Chromatographic enantiomer separation of the new
knotanes 6b–d and the “Bonn-knot” 6a, which was obtained
by this route for the first time, was achieved by means of
enantioselective HPLC employing chiral stationary phases
(CSPs), namely the amylose-derived Chiralpak IA[8] and the
diphenylethanediamine-based (R,R)-ULMO packings. Both
CSPs exhibited promising levels of stereodiscrimination for
the topologically chiral knotane enantiomers. However,
under optimized chromatographic conditions, the ULMO-
type CSP provided superior performance in terms of enan-
tioselectivity, efficiency, and scope of applications. The HPLC
and CD data are identical to those of samples previously
synthesized[9] by Route A. Figure 1 shows the HPLC separa-
tion and CD spectra of the new knots 6b and 6c (for a
detailed analysis of the HPLC chromatogram for the estima-
tion of purity of compound 6b, see the Supporting Informa-
tion).
Route B: Thread 3a folds by itself, then threads intra-
molecularly through the previously formed loop and thus
[*] Dipl.-Chem. J. Brüggemann,Dr. S. Bitter,Dr. S. Müller,W. M. Müller,
U. Müller,Prof. Dr. F. Vögtle
KekulØ-Institut für Organische Chemie und Biochemie
Universität Bonn
Gerhard-Domagk-Strasse 1,53121 Bonn (Germany)
Fax: (+49)228-735-662
E-mail: voegtle@uni-bonn.de
Dr. N. M. Maier,Prof. Dr. W. Lindner
Institut für Analytische Chemie
Universität Wien
Währinger Strasse 38,1090 Wien (Austria)
Fax: (+43)1-4277-9523
E-mail: wolfgang.lindner@univie.ac.at
[**] We are grateful to the Sonderforschungsbereich “Template” (SFB
624) of the Deutsche Forschungsgemeinschaft for valuable support.
We would also like to express our special thanks to Dr. Barbara
Kirchner for preliminary theoretical calculations of closed and open
knots.
Since the isolated yields of the trefoil knot 6a starting
from 3b do not exceed those obtained by condensation of
shorter threads, we assume that, depending on the choice of
conditions, a certain ratio (or a dynamic equilibrium) between
the knotted decaamide 3b and its unknotted isomer 3a exists.
Both species can react with pyridine dicarboxylic acid
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
254
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 254 –259

Angewandte
Chemie
The disubstituted long thread 3c was
obtained by reaction of 1 with the
methoxy-substituted pyridine dicarbox-
ylic acid dichloride 2b (see the Support-
ing Information). Subsequent cycliza-
tion with acid chloride 2a afforded the
disubstituted knotane 6e (Scheme 4),
also in accordance with our conclusions
above. HPLC separation and the corre-
sponding CD spectra are similar to those
shown in Figure 1.
Furthermore we achieved the syn-
thesis of knotane 6d by cyclization of the
long thread 3b with the methoxy-sub-
stituted isophthalic acid dichloride 2d.
This knotane, owing to the ring-closure
reaction with the methoxyisophthalic
unit, consists of four isophthalic units
and just two pyridine units, whereas all
other previously synthesized amide
knots contained three isophthalic units
and three pyridine units (Scheme 3).
Thus a new class of trefoil knots[11] that
was previously not accessible is opened
up (for HPLC separation and CD spec-
trum, see the Supporting Information).
Route C: The proposed mechanism
for Route B above (perhaps also via the
monoacylic derivative of 3b) is not the
only possible synthetic pathway. Kno-
tanes are also obtained by reaction of
the shorter thread 4 with diamine 1,
possibly by formation of the supramolec-
ular complex 5 and sequential cycliza-
tion with two pyridine dicarboxylic acid
dichloride molecules 2 (Scheme 1). Here
we assume that the hexaamide diamine 4
arranges itself into a cisoid conformation
as a result of hydrogen bonding between
the pyridine nitrogen atoms and the
amide hydrogen atoms and subsequently
forms a helical loop. This loop should be
additionally stabilized at the points of
intersection of the molecule by further
hydrogen bonding. The loop could then
act as a host (template) for diamine 1, so
that the organic complex 5 could be
formed after the threading process, and 5
could then act as a “bimolecular tem-
plate”. Cyclization with successively
added pyridine dicarboxylic acid dichlor-
ide 2a affords the knotane 6a (4% yield;
Scheme 5). Hence the final cyclization
step might even proceed via the open
knot 3b as well.
Scheme 1. Synthetic pathways and mechanistic alternatives for the formation of amide knots
from 1 and 2a or from the elongated thread molecules 3a and 4. Route A: one-pot synthesis
from building blocks 1 and 2a. Route B: intramolecular self-knotting of thread 3a. Route C:
possible intermolecular host–guest interaction (“bimolecular template” 5) between two thread
molecules (4 and 1).
The racemate of 6a was separated
dichloride 2a to yield either the knot or the unknotted
monomacrocycle 7 as well as polycondensation products, thus
resulting in a decrease of the yield of the trefoil knot.
into enantiomers, and HPLC and CD data proved to be
identical to earlier samples. Route C therefore proved to be
an alternative reaction mechanism, and the knotanes 6e (5%
Angew. Chem. Int. Ed. 2007, 46, 254 –259
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
255

Communications
Figure 1. HPLC separation (tert-butyl alcohol/n-heptane 50:50; flow
rate 0.5 mLminꢀ1; T=658C; black curve) and CD-spectra (red curve)
of 6b (a) and 6c (b),synthesized through Route B using (R,R)-ULMO
material as the chiral stationary phase (developed by W. Lindner
et al.[10]).
Scheme 3. Reaction of the isolated decaamide 3b with 2a–d and
formation of the unsubstituted “Bonn-knot” 6a and the new monosub-
stituted knots 6b–d (Route B).
substitution patterns that are not accessible by Route A.
However, even the use of linear precursors does not
necessarily result in high yields because of the dynamic
nature of the process.
Scheme 2. Synthesis of the isolated threads 3a and 4.
yield) and 6 f (2% yield) were also synthesized by this route,
through reaction with acid chlorides 2b and 2c, respectively
(Scheme 5). Therefore Route B as well as Route C both open
up new perspectives for the synthesis of new trefoil knots with
Open knots: The “interception” of extended knotted
threads would be interesting, not only with regard to the
formation of cyclic knotanes, but also regarding the synthesis
of open, stoppered amide knots that do not consist of metal
256
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 254 –259

Angewandte
Chemie
Scheme 5. Alternative synthesis of knotanes 6a, 6e,and 6 f by reaction
of the short thread 4 with diamine 1 and acid chlorides 2a–c,possibly
via the supramolecular complex 5 (Route C).
obtained a pure compound with a mass corresponding to
either the stoppered open knot 9a or the stoppered thread 9b
(29% yield; for MALDI-TOF spectrum, see the Supporting
Information). This synthetic pathway corresponds to Route B
(Scheme 1).
A compound of the mass corresponding to 9a/9b is not
only available by Route B. We also obtained this substance
from a one-pot synthesis similar to Route A (Scheme 1) with
yields of around 0.5%. In this case, the stopper component 8
was added slowly after a reaction time of one hour, so that it
could intercept (that is, terminate) the intermediate threads
3b and 3a. Since 9a is chiral, whereas 9b is achiral, this was
the starting point for further investigations. However,
attempts to resolve the enantiomers of this compound on
the (R,R)-ULMO CSP[10], employing the chromatographic
conditions optimized for cyclic knots, failed (for the HPLC
chromatogram, see the Supporting Information). This implies
that there might be an equilibrium at room temperature
between stoppered thread 9b and the open knot 9a as a result
of the insufficient steric demand of the stopper (as deduced
from newer theoretical calculations and modelling[12]). The
barrier of this equilibrium should be less than 30 kcalmolꢀ1 at
Scheme 4. Synthesis of the dimethoxy-substituted long thread 3c and
subsequent cyclization with 2a to yield knotane 6e (Route B).
complexes.[3] Interception of the decaamide diamine 3b with a
monoacid chloride with significant steric demand seemed
feasible because of the expandable molecular scaffold. After
purification of the decaamide diamine 3a (or 3b) and a
subsequent separate reaction with stopper 8 (Scheme 6), we
Angew. Chem. Int. Ed. 2007, 46, 254 –259
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
257

Communications
Scheme 6. Synthesis of the open stoppered amide knot 9a,or the stoppered thread 9b,by reaction of the decaamide diamine 3b with two stopper
molecules 8 in a manner analogous to Route B.
room temperature.[13] For further clarification of this aspect,
the synthesis of even bulkier stopper components is necessary,
but such components are not yet known.
For the spectroscopic data of 3b, 3c, and 4, see the Supporting
Information.
Knotanes 6a–e by Route B: Decaamide 3b or 3c was dissolved in
dichloromethane (30 mL), and triethylamine (0.5 mL) was added. A
The spontaneous reversible self-knotting of oligoamide
solution of the appropriate acid chloride 2a–d in dichloromethane
threads, which is demonstrated here for the first time, opens
up new perspectives. New knotane architectures may be
obtained by, for example, changing the spacer between amide
groups and by further extension of the thread, thus converting
the molecules into cyclic and open knots with extended loops
(expanded knotanes) as well as open-chain knots with longer
threads. By further extension of the thread part, it might be
possible for knots to be successively tied together like a string
of pearls. Polymers of this kind should be very elastic, since
open knots can be tightened and loosened like shoelaces.
Finally, our results regarding self-threading, self-knotting, and
self-templating can be expected to stimulate the development
of new intertwined topologies. Just as the information needed
for the knotting is already contained in the constitution and
the amide sequence of 3a, other cyclic and open entangle-
ments and knottings might be achievable by the skillful design
of starting materials.
(10 mL) was slowly added over a period of 1 h. The reaction mixture
was stirred at room temperature for 48 h, the solvent was removed
under reduced pressure, and the residue was purified by column
chromatography (dichloromethane/ethyl acetate 4:1). All melting
points are greater than 2208C.
Knotanes 6a, 6e, and 6 f by Route C: Hexaamide 4 and diamine 1
were dissolved in dichloromethane (100 mL) and stirred at room
temperature for 48 h. After the addition of triethylamine (0.5 mL), a
solution of the appropriate acid chloride 2a–c in dichloromethane
(30 mL) was added slowly over 2 h. The reaction mixture was stirred
for an additional 12 h at room temperature, the solvent was removed
under reduced pressure, and the residue was purified by column
chromatography (dichloromethane/ethyl acetate 4:1). All melting
points are greater than 2208C.
For further experimental details, spectroscopic data of 6a, 6e, and
6 f, as well as the synthesis of 9a/9b, please refer to the Supporting
Information.
Received: May 16, 2006
Published online: October 10, 2006
Keywords: enantiomer separation · molecular knots ·
nanostructures · supramolecular chemistry · template synthesis
.
Experimental Section
Synthesis of the short (4) and the long (3b/3c) threads: Compound 1
(3.00 g, 3.87 mmol) was partially dissolved in chloroform (5 mL), and
dichloromethane (95 mL) was added. After addition of triethylamine
(2 mL), a solution of 2a (0.14 g, 0.69 mmol) or 2b (0.16 g, 0.70 mmol)
in dichloromethane (75 mL) was added slowly over a period of 2 h at
room temperature. The reaction mixture was stirred overnight, the
solvent was removed under reduced pressure, and the residue was
purified two times by column chromatography (dichloromethane/
ethyl acetate 3:1 und 2:1) to afford 4 (518 mg, 0.31 mmol, 16%) or the
methoxy-substituted analogue 4-OCH3 (375 mg, 0.22 mmol, 17%) as
well as 3b (151 mg, 0.06 mmol, 5%) or 3c (130 mg, 0.05 mmol, 4%).
All melting points are greater than 2208C.
[1] O. Safarowsky, M. Nieger, R. Fröhlich, F. Vögtle, Angew. Chem.
2000, 112, 1699 – 1701; Angew. Chem. Int. Ed. 2000, 39, 1616 –
1618.
[2] Review: F. Vögtle, O. Lukin, Angew. Chem. 2005, 117, 2 – 23;
Angew. Chem. Int. Ed. 2005, 44, 2 – 23.
[3] Reviews on molecular knots of different types: Molecular
Catenanes, Rotaxanes and Knots (Eds.: J.-P. Sauvage, C. Die-
trich-Buchecker), Wiley-VCH, Weinheim, 1999; R. F. Carina, C.
Dietrich-Buchecker, J.-P. Sauvage, J. Am. Chem. Soc. 1996, 118,
258
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 254 –259

Angewandte
Chemie
9110 – 9116; C. Dietrich-Buchecker, J.-P. Sauvage, B. X. Colas-
son, Top. Curr. Chem. 2005, 249, 261 – 283; L.-E. Perret-Aebi, A.
von Zelewsky, C. Dietrich-Buchecker, J.-P. Sauvage, Angew.
Chem. 2004, 116, 4582 – 4585; Angew. Chem. Int. Ed. 2004, 43,
4482 – 4485; D. B. Amabilino, J. F. Stoddart, Chem. Rev. 1995, 95,
2725 – 2828; C. A. Hunter, P. C. Mayers, Nature 2001, 411, 763.
[4] C. A. Schalley, W. Reckien, S. Peyerimhoff, B. Baytekin, F.
Vögtle, Chem. Eur. J. 2004, 10, 4777 – 4789.
[5] For peptide folding, compare: E. C. B. Johnson, T. Durek,
S. B. H. Kent, Angew. Chem. 2006, 118, 3361 – 3365; Angew.
Chem. Int. Ed. 2006, 45, 3283 – 3287.
[6] Mathematically, knots are only properly defined in their cyclic
state (closed loop): P. Virnau, Y. Kantor, M. Kardar, J. Am.
Chem. Soc. 2005, 127, 15102 – 15106.
[9] F. Vögtle, A. Hünten, E. Vogel, S. Buschbeck, O. Safarowsky, J.
Recker, A.-H. Parham, M. Knott, W. M. Müller, U. Müller, Y.
Okamoto, T. Kubota, W. Lindner, E. Francotte, S. Grimme,
Angew. Chem. 2001, 113, 2534 – 2537; Angew. Chem. Int. Ed.
2001, 40, 2468 – 2471.
[10] G. Uray, W. Lindner, Chromatographia 1990, 30(5 – 6), 323 – 327;
N. M. Maier, G. Uray, J. Chromatogr. A 1996, 732(2), 215 – 230.
[11] With other arene units such as, for instance, thiophene- or furan-
2,5-dicarboxylic acid dichlorides, we obtained compounds with a
mass of the corresponding hexamers; however, owing to their
(achiral) HPLC behavior, we assume that they do not exist in a
knotted structure. We attribute this to the widened binding
angles of the five-membered arenes: M. Knott, Diploma thesis,
Universitꢀt Bonn, 2000; A. Böhmer, Ph.D. thesis, Universitꢀt
Bonn, 2006.
[7] a) For nucleic acid knots, see also: N. C. Seeman, J. Am. Chem.
Soc. 1992, 114, 9652 – 9655; S. M. Du, B. D. Stollar, N. C.
Seeman, J. Am. Chem. Soc. 1995, 117, 1194 – 1200; b) on the
basis of calculations, polymerization reactions should yield
different knot molecules under certain conditions.[6]
[12] We thank Dr. B. Kirchner, Dr. W. Reckien, and M. Eggers of the
Institut für Physikalische und Theoretische Chemie der Rheini-
schen Friedrich-Wilhelms-Universitꢀt Bonn for preliminary
classical and Car–Parrinello simulations.
[8] Solvent
n-hexane/dichloromethane
.
50:50,
flow
rate
[13] Variable-temperature 1H NMR experiments between ꢀ50 and
558C showed no substantial changes.
0.3 mLminꢀ1
Angew. Chem. Int. Ed. 2007, 46, 254 –259
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
259