Synthetic Approaches to a Molecular Borromean Link: Two-Ring Threading with Polypyridine Templates

Article abstract of DOI:10.1002/anie.200352562

"One ring to bind them": The Borromean link, which consists of three rings connected such that no two are concatenated, is a challenging synthetic target. The synthesis of an orthogonal two-ring system that could be a precursor for such a link was prepare

Full text of DOI:10.1002/anie.200352562

Angewandte
Chemie
Communications
The Borromean link of three rings connected such that no two are
concatenated is an exceptionally challenging target and any synthetic
approach must address both representative and methodological
aspects of the problem. Methodological advances toward an orthog-
onal two-ring target are detailed by J. S. Siegel et al. on the following
pages.
Angew. Chem. Int. Ed. 2003, 42, 5701
DOI: 10.1002/anie.200352562
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5701

Communications
Macrocycles
Synthetic Approaches to a Molecular Borromean
Link: Two-Ring Threading with Polypyridine
Templates**
Jon C. Loren, Michito Yoshizawa,
Richard F. Haldimann, Anthony Linden, and
Jay S. Siegel*
The molecular Borromean link is one of the most challenging
synthetic targets generated from molecular graphs[1] of
complex topology.[2,3] Its structural complexity stems from
the interweaving of three macrocycles such that no two of the
macrocycles are concatenated nor covalently connected, yet
collectively they form an integral molecular unit. The origins
of this link in human culture go back to antiquity[4] and the
novelty of its form has been richly discussed.[5] The design of
molecular targets and retrosyntheses from topological graphs
often correlate with a specific representation (often printed in
two dimensions).[6,7] Three principal representations of the
Borromean link that inspire targets for molecular design and
synthesis are the Venn rings, chain rings, and orthogonal rings
(Figure 1).[8] The Venn rings highlight threefold symmetry
and embellish upon the triskelion[9] or related trefoil knot[10]
precursor. The chain rings betray a rack[11] or grid[12]
substructure from which functionalized end groups close the
macrocycles. The orthogonal rings beseech a single-step
template-driven synthesis[13] but readily succumb to a ring-
by-ring retrosynthesis. From this humble strategy, we begin.
To illustrate a specific strategy, we look to an orthogonal-
ring representation using metal coordination as the template
method, built up by a kinetically selective reaction sequence
(Figure 2). Two key structural intermediates along this
synthetic route are the assembly of a principal base macro-
cycle with endo- or exo-oriented metal-binding sites and the
completion of a threaded two-ring structure, in which the two
rings are orthogonal and at least one of the two rings
possesses a second set of metal-binding sites, compatible with
the threading and knitting together of the third ring.
Figure 1. Three retrosynthetic representations of the Borromean link
(Venn rings, chain rings, orthogonal rings).
relatively simple conceptual extension to a bismacrocycle of
form c (cf. Figure 2) represents a curious threaded-ring
structure of which there are few examples.[15,16] Our
Macrocycles, abstracted as form e (cf. Figure 2), have
substantial precedence in the literature;[14] however, the
[*] Prof. J. S. Siegel, Dr. A. Linden
Organic Chemistry Institute, University of Zürich
Winterthurerstrasse 190, 8057 Zürich (Switzerland)
Fax (+41)1-635-6888
E-mail: jss@oci.unizh.ch
J. C. Loren, M. Yoshizawa, Dr. R. F. Haldimann
Department of Chemistry, University of California-San Diego
La Jolla, California 92093-0358 (USA)
Fax: (+1)858-822-0386
[**] This work was supported by the USNational Science Foundation
(CHE-0213323) and the Swiss National Fond. Gail Bamber prepared
the graphical art in Figure 1. Kim Baldridge and Steve Cutchin
rendered Figure 3. Bruno Bürgi assisted with the crystal structure
determination.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Figure 2. Retrosynthesis of a Borromean link from the orthogonal-ring
representation.
5702
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200352562
Angew. Chem. Int. Ed. 2003, 42, 5702 –5705

Angewandte
Chemie
strategy focuses on those macrocycles with polypyridine
units,[15,17] which would suit a metal-template-based retrosyn-
thesis.
Crystals of 1 were grown from acetone/diethyl ether as the
PF6 salt, and the crystal structure was elucidated in space
¯
group P1 with molecules residing on Ci special positions
The macrocycle 6 can be synthesized in a multistep
procedure from the previously reported terpyridine 2[18]
(Scheme 1). Elaboration of 2 into a related terpyridine 4
with alkyne-terminated side chains, proceeds by standard
demethylation and realkylation methods. Terpyridine 4 forms
an intermediate homoleptic copper(ii) complex[19] with four
dangling arms, terminating in alkynes. These arms are too
short for the ends to form a [2]catanane, as seen in related
work,[20] but are optimal to form the so-called “figure-eight”
complex.[21] The combined one-pot operation of forming 5
from 4 and copper ions followed by addition of excess copper
ion, under modified Eglinton conditions in ethanol without
pyridine, leads directly to the 66-membered macrocycle 6 in
over 90% yield. Hydrogenation of the diynes in 6 affords
macrocycle 7, which bears two endo-oriented terpyridines and
saturated bridging arms. Preformation of the ruthenium
complex 10 by standard methods[22] provides the reagent
necessary to complex two 4,4’’-diarylterpyridyl–Ru units
(LRu) to macrocycle 7, thus forming 11. These LRu units
serve as exo-oriented precursors of ring 2 threaded within
ring 1. The ends of the 4,4’’-aryl units are juxtaposed with each
other and macrocyclization to form the threaded two-ring
architecture 1 is effected by Williamson ether synthesis with
the known biselectrophile 6,6’-bisbromomethyl-2,2’-bipyri-
dine.[23] The bridging bipyridine units represent the endo-
oriented portions of a tetrahedral or trigonal bipyramidal
metal-binding site through which the third ring would
ultimately be threaded, oriented, and cyclized.
(Figure 3).[24] The two rings are clearly threaded within one
another in an orthogonal orientation. The largest dimension
of the complex is roughly 29 Š from the hydrogen atoms in
the saturated arm of ring 1 to their symmetry-related
partners. The Ru–Ru span is 16.4 Š; on the outer ring
(ring 1) the span from the para-hydrogen atom of the central
ring of the terpyridine to its symmetry partner is approx-
imately 27 Š. The dimensions of ring 2 are slightly smaller,
with a long diagonal of 26.1 Š measured between manisyl aryl
hydrogen atoms. The conformation of ring 1 is not planar and
bows into a chair form. The two endo-oriented terpyridines of
ring 1 sit in parallel planes, offset by about 5 Š. One might
anticipate considerable uncertainty about the exact confor-
mation of the saturated arms and this is supported by residual
disorder, which required a two-conformation model for these
arms in the crystal structure analysis. The conformation of
ring 2 is also readily likened to a macrocyclic chair. The exo-
oriented terpyridines sit in planes parallel to one another but
shifted by roughly 6.5 Š. The bridging bipyridines, adopt the
expected anti conformation in the absence of a metal. The
idealized symmetry of ring 2 would be C2h, with the twofold
axis bisecting each of the 2,2’ bonds of each bipyridine, and
the horizontal mirror plane bisecting the exo-oriented terpyr-
idines. The structure shows that there is ample space for the
arms of the third ring to be threaded inside of ring 2 and
outside of ring 1. The dimensions of this third ring will again
need to be over 20 Š to span even the smallest region of
ring 1.
Scheme 1. Reaction conditions: a) molten py·HCl, 1858C, 4 h, 97%; b) Cs2CO3, TsOCH2CH2OCH2CCH (2.2 equiv), DMF, 808C, 10 h, 69%;
c) Cu(OAc)2·H2O (0.5 equiv), ethanol, room temperature, 1 h; d) 1) Cu(OAc)2·H2O (10.0 equiv), ethanol, high dilution, reflux, 72 h; 2) KCN (aq),
CH2Cl2, 91%; e) H2 80 psi, Pd/C (10%), ethanol/CH2Cl2 (1:1), 6 h, 91%; f) 10 (2.0 equiv), CH2Cl2/EtOH/ethylene glycol (2:2:1), reflux, 12 h, 65%;
g) 6,6’-bisbromomethyl-2,2’-bipyridine (2.0 equiv), Cs2CO3, acetonitrile, reflux, 72 h, 49%.
Angew. Chem. Int. Ed. 2003, 42, 5702 –5705
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5703

Communications
Table 1: UV/Vis, fluorescence, and electrochemical reduction potentials
for 1 and 12.
UV/Vis[a]
Fluorescence (l)[a]
Excitation Emission
labs
loge
E1/2 [V][b]
1
201
273
312
480
202
271
313
480
5.6
5.0
5.1
4.4
5.0
4.8
5.9
4.1
312
313
407
419
0.99
12
0.95
[a] In acetonitrile. [b] 0.1m tetrabutylammonium hexafluorophosphate in
acetonitrile vs Ag/AgNO3 in acetonitrile.
The synthesis of the two-ring intermediate en route to a
Borromean link in the orthogonal-ring representation marks
a milestone. The challenge of this target lead to a highly
efficient strategy for the synthesis of macrocycle 7, which in
turn made 100 mg quantities of 1 readily accessible. The
elucidation of the structure of 1 provides evidence that the
molecule adopts a structure close to the idealized scheme and,
more importantly, gives insight into the stereochemistry and
structural dimensions necessary to assemble the last ring of
the Borromean link. Beyond the mere technical details, these
studies illustrate an important aspect of the pursuit of
chemistry: the human conception of molecular structure.[31]
Received: August 4, 2003 [Z52562]
Published Online: November 20, 2003
Keywords: macrocycles · nitrogen heterocycles · ruthenium ·
.
supramolecular chemistry · template synthesis
Figure 3. Space-filling model of the crystal structure of 1: colored to
emphasize connectivity; aspect to show chair form of ring 1 (bottom)
[1] D. M. Walba, Tetrahedron 1985, 41, 3161.
[2] For a review of the field of chemical topology, see: a) J. P.
Sauvage, C. Dietrich-Buchecker, Molecular Catenanes, Rotax-
anes and Knots: A Journey Through the World of Molecular
Topology, Wiley-VCH, Weinheim, 1999, p. 368; b) G. Schill,
Catenanes, Rotaxanes, and Knots, Vol. 22, Academic, New York,
1971, p. 192 (Organic Chemistry, A Series of Monographs);
c) J. M. Lehn, Supramolecular Chemistry: Concepts and Per-
spectives, VCH, Weinheim, 1995, p. 262.
The photophysical and electrochemical properties of 1,
relative to its parent mono ruthenium heteroleptic complex
12, indicate that the ruthenium centers behave independently
of one another (Table 1). Excitation near 312 nm results in
[3] For pioneering work in the field, see: a) E. Wasserman, J. Am.
Chem. Soc. 1960, 82, 4433 – 4434; b) G. Schill, A. Lüttringhaus,
Angew. Chem. 1964, 76, 567; Angew. Chem. Int. Ed. Engl. 1964,
3, 546; c) A. Lüttringhaus, F. Cramer, H. Prinzbach, F. M.
Henglein, Justus Liebigs Ann. Chem. 1958, 613, 158; d) A.
Lüttringhaus, F. Cramer, H. Prinzbach, Angew. Chem. 1957, 69,
137.
[4] P. R. Cromwell, E. Beltrami, M. Rampichini, Mathematical
Intelligencer 1998, 20, 53 – 62; see also: http://www.liv.ac.uk/
~ spmr02/rings/.
[5] C. Liang, K. Mislow, J. Math. Chem. 1994, 16, 27 – 35.
[6] In contrast, the one existing molecular effigy of a Borromean
link comes from the creative use of molecular biological
methods for connecting DNA analogous to the well-known
Holiday-junction construction. This form, albeit topologically
equivalent to the Borromean link, is not visually equivalent to a
simple Borromean-link representation. This lack of direct visual
correlation between the actual structure and the ideal represen-
what appears to be a ligand-centered emission at 407 and
419 nm for 1 and 12, respectively. These values coincide with
the emission wavelength range of the free ligands.[18] Excita-
tion at 480 nm results in no detectable emission for either.
Cyclic voltametry (CV) shows only quasi-reversible behavior,
and electrodeposition occurs on the electrodes after multiple
scans for both 1 and 12.
5704
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 5702 –5705

Angewandte
Chemie
tation is also characteristic of Whiteside's novel micromachined
model of a Borromean link.
1997, 109, 1365 – 1367; Angew. Chem. Int. Ed. Engl. 1997, 36,
1308 – 1310.
[21] a) N. Belfrekh, C. Dietrich-Buckecker, J.-P. Sauvage, Inorg.
Chem. 2000, 39, 5169 – 5172; b) A. G. Fallis, M. A. Heuft, Angew.
Chem. 2002, 114, 4702 – 4705; Angew. Chem. Int. Ed. 2002, 41,
4520 – 4523.
[22] P. B. Sullivan, J. M. Calvert, T. J. Meyer, Inorg. Chem. 1980, 19,
1404 – 1407.
[7] a) N. C. Seeman, Angew. Chem. 1998, 110, 3408 – 3428; Angew.
Chem. Int. Ed. 1998, 37, 3220 – 3238; b) C. Mao, W. Sun, N. C.
Seeman, Nature 1997, 386, 137 – 138; c) H. Wu, S. Brittain, J.
Anderson, B. Grzybowski, S. Whitesides, G. M. Whitesides, J.
Am. Chem. Soc. 2000, 122, 12691 – 12699.
[8] These representations exemplify but do not limit the number of
ways this topology can be visually appreciated or structurally
realized; however it is through the connections between these
representations and the designed molecular targets that we
enrich our conceptualization of molecular structure and envision
retrosyntheses. For an example of how this has been interpreted
in crystal engineering, see: L. Carlucci, G. Ciani, D. M.
Proserpio, Cryst. Eng. Comm. 2003, 5, 269 – 279.
[23] U. S. Schubert, J. L. Kersten, A. E. Pemp, C. D. Eisenbach, G. R.
Newkome, Eur. J. Org. Chem. 1998, 2573 – 2581.
[24] Crystal-structure analysis of 1: crystals from acetone/Et2O;
¯
C204H232F24N16O24P4Ru2, Mr = 4074.0; triclinic, space group P1,
a = 14.8913(6), b = 15.6077(7), c = 24.010(1) Š, a = 98.044(2),
b = 90.089(2), g = 90.936(2)8, V= 5524.6(4) Š3, Z = 1, 1 =
1.224 gcmÀ3
,
T= À 1138C, crystal dimensions: 0.20 ” 0.25 ”
0.30 mm, Nonius Kappa CCD area detector diffractometer,
graphite-monochromated MoKa radiation, l = 0.71073 Š, m =
0.249 mmÀ1, 2qmax = 558, 62117 total reflections, 18576 symme-
try-independent reflections, Rint = 0.072, 11249 reflections with
I > 2s(I). Data reduction performed with HKL Denzo and
Scalepack;[25] absorption correction applied.[26] The structure was
solved by using the Patterson and Fourier expansion routines of
DIRDIF94.[27] The SQUEEZE routine[28] of the program
PLATON[29] was employed to omit the contribution to the
reflection data from the highly disordered solvent molecules,
estimated to be six acetone molecules per asymmetric unit. This
significantly improved the refinement results. The six CH2
[9] C. R. Woods, M. Benaglia, S. Toyota, K. Hardcastle, J. S. Siegel,
Angew. Chem. 2001, 113, 771 – 773; Angew. Chem. Int. Ed. 2001,
40, 749 – 751.
[10] a) O. Safarowsky, M. Nieger, R. Fröhlich, F. Vögtle, Angew.
Chem. 2000, 112, 1699 – 1701; Angew. Chem. Int. Ed. 2000, 39,
1616 – 1618; b) G. Rapenne, C. Dietrich-Buchecker, J.-P. Sauv-
age, J. Am. Chem. Soc. 1999, 121, 994 – 1001; c) G. Schill, R.
Henschel, J. Boeckmann, Justus Liebigs Ann. Chem. 1974, 709 –
733; d) H. Adams, E. Ashworth, G. A. Breault, C. A. Hunter,
P. C. Mayers, Nature 2001, 411, 763.
[11] M. Benaglia, F. Ponzini, C. R. Woods, J. S. Siegel, Org. Lett. 2001,
3, 967 – 969.
À
groups of the hexamethylene chain and one PF6 anion are
[12] a) G. S. Hanan, D. Volkmer, U. S. Schubert, J. M. Lehn, G.
Baum, D. Fenske, Angew. Chem. 1997, 109, 1929 – 1931; Angew.
Chem. Int. Ed. Engl. 1997, 36, 1842 – 1844; b) S. Toyota, C. R.
Woods, M. Benaglia, R. Haldimann, K. Warnmark, K. Hardcas-
tle, J. S. Siegel, Angew. Chem. 2001, 113, 773 – 776; Angew.
Chem. Int. Ed. 2001, 40, 751 – 754.
[13] For a review on template-directed approaches, see: a) Breault,
C. A. Hunter, P. C. Mayers, Tetrahedron 1999, 55, 5265 – 5293;
b) S. Anderson, H. L. Anderson, J. K. M. Sanders, Acc. Chem.
Res. 1993, 26, 469 – 475; c) T. J. Hubin, D. H. Busch, Coord.
Chem. Rev. 2000, 200, 5 – 52; d) F. Diederich, P. J. Stang,
Templated Organic Synthesis, Wiley-VCH, Weinheim, 2000,
p. 410.
disordered and two conformations were modelled for each. The
structure was refined on F2 by full-matrix least-squares methods
using SHELXL97.[30] The H atoms were placed in geometrically
calculated positions and were allowed to ride on their parent
atom. The refinement of 1136 parameters using 18576 reflec-
tions and 739 restraints gave R(F) (I > 2s(I) data) = 0.103,
Rw(F2) (all data) = 0.306, goodness of fit on F2 = 1.040, D1max
=
2.27 eŠÀ3. CCDC-216666 contains the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge via www.ccdc.cam.ac.uk/conts/retrieving.html (or from
the Cambridge Crystallographic Data Centre, 12, Union Road,
Cambridge CB21EZ, UK; fax: (+ 44)1223-336-033; or deposit@
ccdc.cam.ac.uk).
[25] Z. Otwinowski, W. Minor, Methods Enzymol. 1997, 276, 307 –
326.
[14] a) D. Parker in Macrocycle Synthesis: A Practical Approach
(Ed.: C. J. Moody, L. M. Harwood), Oxford University, Oxford,
1996; b) E. C. Constable, Coordination Chemistry of Macro-
cyclic Compounds, Oxford University, Oxford, 1999.
[15] M. Schmittel, A. Ganz, D. Fenske, Org. Lett. 2002, 4, 2289 – 2292.
[16] S.-H. Chiu, A. R. Pease, J. F. Stoddart, A. J. P. White, D. J.
Williams, Angew. Chem. 2002, 114, 280 – 284; Angew. Chem. Int.
Ed. 2002, 41, 270 – 274.
[17] a) M. Schmittel, A. Ganz, Synlett 1997, 710 – 712; b) U. Leh-
mann, A. D. Schlüter, Eur. J. Org. Chem., 2000, 3483 – 3487;
c) O. Henze, D. Lentz, A. D. Schlüter, Chem. Eur. J. 2000, 6,
2362 – 2367; d) G. R. Newkome, H.-W. Lee, J. Am. Chem. Soc.
1983, 105, 5956 – 5957; e) T. W. Bell, F. Albert, J. Am. Chem. Soc.
1986, 108, 8109 – 8111.
[26] R. H. Blessing, Acta Crystallogr. Sect. A 1995, 51, 33.
[27] P. T. Beurskens, G. Admiraal, G. Beurskens, W. P. Bosman, R.
de Gelder, R. Israel, J. M. M. Smits, DIRDIF94: The DIRDIF
program system, Technical Report of the Crystallography
Laboratory, University of Nijmegen, The Netherlands, 1994.
[28] P. van der Sluis, A. L. Spek, Acta Crystallogr. Sect. A 1990, 46,
194 – 201.
[29] A. L. Spek, PLATON, Program for the Analysis of Molecular
Geometry, University of Utrecht, The Netherlands, 2003.
[30] G. M. Sheldrick, SHELXL97, Program for the Refinement of
Crystal Structures, University of Göttingen, Germany, 1997.
[31] Note added in Proof: An example of a related topological
structure can be found in: O. V. Dolomanov, A. J. Blake, N. R.
Champness, M. Schröder, C. Wilson, Chem. Commun. 2003,
682 – 683.
[18] J. C. Loren, J. S. Siegel, Angew. Chem. 2001, 113, 776 – 779;
Angew. Chem. Int. Ed. 2001, 40, 754 – 757.
[19] U. S. Schubert, C. Eschbaumer, P. Andres, H. Hofmeier, C. H.
Weidl, E. Herdtweck, E. Dulkeith, A. Morteani, N. E. Hecker, J.
Feldmann, Synth. Met. 2001, 121, 1249 – 1252.
[20] a) By optimizing the arm length, the mixture of catenane and
“figure-eight” products can be controlled (J. Loren, P. Gantzel,
J. S. Siegel, unpublished results); b) the length and flexibility of
the arms has been noted as a crucial factor in a related RCM-
based synthesis of catenanes; see: a) M. Weck, B. Mohr, J. P.
Sauvage, R. H. Grubbs, J. Org. Chem. 1999, 64, 5463 – 5471; b) B.
Mohr, M. Weck, J. P. Sauvage, R. H. Grubbs, Angew. Chem.
Angew. Chem. Int. Ed. 2003, 42, 5702 –5705
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5705