Solid-phase synthesis of oligo(phenylene ethynylene) rotaxanes

Article abstract of DOI:10.1002/anie.200702349

(Chemical Equation Presented) Hooks and feelers: The use of a bulky tripod-shaped stopper as a solid-phase attachment point proves essential for the controlled iterative solid-phase synthesis of an α-cyclodextrin [3]rotaxane (see picture). NMR spectroscop

Full text of DOI:10.1002/anie.200702349

Angewandte
Chemie
DOI: 10.1002/anie.200702349
Rotaxane Synthesis
Solid-Phase Synthesis of Oligo(phenyleneethynylene) Rotaxanes**
Heather W. Daniell, Eric J. F. Klotz, Barbara Odell, Tim D. W. Claridge, and
Harry L. Anderson*
Solid-phase synthesis is an attractive method for the con-
struction of complex linear oligomers, such as peptides and
oligonucleotides, because it provides complete control of the
oligomer length and sequence while avoiding the need for
purification after every growth step.[1] This strategy has been
exploited to synthesize p-conjugated oligomers for optoelec-
tronic applications, such as oligo(phenyleneethynylene)s
(OPEs),[2–4] oligothiophenes,[5,6] and oligo(phenylenetriacety-
lene)s.[7] Solid-phase synthesis is also a promising strategy for
the synthesis of monodisperse oligomeric “insulated molec-
ular wires”.[8] Recently we have developed the synthesis of
polyrotaxane-type insulated molecular wires by threading
cyclodextrin (CD) macrocycles onto a conjugated polymer
chain to achieve supramolecular control of the chemical
stability, photoluminescence, and electroluminescence effi-
Scheme 1. Functionalization of aminopropylsilylated controlled-pore
glass with the 5-iodoisophthalic acid stopper and a cleavable ester
ciency.[8,9] Herein we report the first solid-phase synthesis of
rotaxanes of this type. We have investigated the solid-phase
linker. Reagents: a) HBTU, DMAP, iPr2NEt, DMF; b) TsOH, MeOH,
CH2Cl2; c) pyBOP, iPr2NEt, DMF.[15]
synthesis of two [2]rotaxanes, 1ꢀa-CD and 2ꢀa-CD, and a
[3]rotaxane, 3ꢀ(a-CD)2. Although rotaxanes were synthe-
sized on the solid phase by Harrison and Harrison in 1967,[10]
there have been few subsequent reports of solid-phase
rotaxane synthesis,[11,12] and cyclodextrin rotaxanes[13] have
not been previously prepared on a solid phase.
Our approach to the synthesis of conjugated polyrot-
axanes uses the hydrophobic effect to drive the threading of
cyclodextrin macrocycles onto a growing polymer chain, and
thus requires aqueous conditions.[8] The challenge in trans-
lating this chemistry onto a solid support is to find a support
that is compatible with the aqueous cyclodextrin-threading
process. We chose controlled-pore glass because it has large
rigid pores which should provide a suitable environment for
aqueous host–guest chemistry, in contrast to solvent-swollen
gels, which tend to exclude high-molecular-weight solutes,
such as cyclodextrins.[14] Aminopropylsilylated controlled-
pore glass was functionalized with a 5-iodoisophthalic acid
stopper and an ester linker as shown in Scheme 1.
We first synthesized rotaxane 1ꢀa-CD using a “protected-
alkyne route” via alkyne 8 (Scheme 2). This chemistry
Scheme 2. Synthesis of rotaxane 1ꢀa-CD. Both routes gave the
rotaxane when carried out in solution (R=H), but failed on the solid
phase (R=CPG). Reagents: a) [Pd2(dba)3], PPh3, CuI, DMF, Et3N;
b) solution phase: NaOH, H2O; solid phase: Na2CO3, MeOH; c) a-
CD, Pd(OAc)2, TPPTS, CuI, H2O, Et3N (followed by LiOH in the case
of 10b); d) I2, MeCN.[15]
[*] H. W. Daniell, Dr. E. J. F. Klotz, Dr. B. Odell, Dr. T. D. W. Claridge,
Prof. H. L. Anderson
Department of Chemistry, Oxford University
Chemistry Research Laboratory
12 Mansfield Road, Oxford, OX1 3TA (UK)
Fax: (+44)1865-28-5002
E-mail: harry.anderson@chem.ox.ac.uk
Homepage: http://users.ox.ac.uk/~hlagroup/
[**] We thank the EPSRC and the EC (MRTN-CT-2006-036040) for
financial support.
proceeds smoothly in solution, but fails on the solid phase
because the terminal alkyne intermediate 6 is unstable when
bound to CPG. Another problem with this route is that Glaser
homo-coupling of alkyne 8 occurs as a minor side reaction
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 6845 –6848
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6845

Communications
during Sonogashira cross-coupling. Thus we turned to the
to provide access to longer monodisperse insulated molecular
wires.
alternative “masked-iodide route” using triazene 9 and aryl
iodide 10 (Scheme 2). In the solution phase, this route gives
clean samples of the rotaxane 1ꢀa-CD (> 90% yield by
HPLC; 11% yield of isolated product).[16] The instability
problem of terminal alkyne 6b is avoided, but we were still
unable to synthesize the rotaxane on the solid phase. In the
absence of a-CD, support-bound aryl iodide 10b reacts
cleanly with alkyne 6a under aqueous Sonogashira coupling
conditions identical to those used to prepare 1ꢀa-CD in
solution to give the dumbbell 1 after cleavage from the
support with lithium hydroxide. However when this reaction
is carried out on the solid phase in the presence of a-CD, no
rotaxane is formed, and dumbbell 1 is still the main product.
After testing a variety of reaction conditions for solid-
phase synthesis of 1ꢀa-CD without success, we conjectured
that the environment on the surface of the support might be
too crowded to allow threading of the cyclodextrin. To
alleviate this congestion, we tried attaching a larger stopper to
the support in place of iodoisophthalic acid 5a to keep the
OPE chains further apart. Use of the bulky tricarboxylic acid
support-bound stopper 11 enabled the [2]rotaxane 2ꢀa-CD
to be synthesized on the solid support and allowed us to
extend this chemistry by another iteration to achieve the
solid-phase synthesis of the [3]rotaxane 3ꢀ(a-CD)2
(Scheme 3). Rotaxanes 2ꢀa-CD and 3ꢀ(a-CD)2 were also
synthesized by the same routes using solution-phase chemis-
try, giving identical products (2ꢀa-CD: 97% conversion by
HPLC, 12% yield of isolated product; 3ꢀ(a-CD)2: 63%
conversion by HPLC, 2.4% yield of isolated product). The
efficiency of rotaxane formation on the solid phase is similar
to that in solution (HPLC analysis of solid-phase cleavage
products shows that 2ꢀa-CD and 3ꢀ(a-CD)2 are formed with
83% and 85% conversion, respectively). The successful
construction of [3]rotaxane 3ꢀ(a-CD)2 by solid-phase syn-
thesis gives us confidence that this strategy could be extended
The presence of two different stoppers in rotaxanes 2ꢀa-
CD and 3ꢀ(a-CD)2 means that these compounds could exist
as several stereoisomers with different threading orientations.
Both rotaxanes are formed predominantly as one isomer, in
solution and on the solid phase, with the wide 2,3-rims of the
cyclodextrins pointing towards the large tetraarylmethane
stopper. In NMR spectroscopic investigations of [2]rotaxane
2ꢀa-CD, NOEs were observed from the CD resonance H3 at
the wide rim to aromatic protons b, c, d, e, and f, while proton
H5 correlates with protons d, e, f, and g (see Figure 1 and
Scheme 3 for atom-labeling scheme). Interestingly, protons
H6 and H6’ at the narrow rim correlate strongly with protons
e and g, but only weakly with proton f, implying that the a-CD
is in fast exchange between two locations over the two alkyne
units, spending little time on the central phenylene unit.
The two anomeric resonances (H1A and H1B of CDA and
CDB, respectively, Figure 1) of 3ꢀ(a-CD)2 are well resolved,
making it possible to assign most of the resonances of each
cyclodextrin via spin diffusion. The NOESY spectrum of
3ꢀ(a-CD)2 (Figure 1a) shows a remarkably selective set of
correlations: b–H3A; c–H3A; d–H3A, H5A; e–H5A, H6A, H6’A;
g–H3B; h–H3B, H5B; and i–H5B, H6B, H6’B. This simple pattern
of interactions and the absence of any significant NOE from
either cyclodextrin to proton f implies that the two cyclo-
dextrins have well defined locations over the outer two alkyne
units, as seen in the energy-minimized structure (Fig-
ure 1b).[17] At first sight it seems surprising that the two
cyclodextrin units in the [3]rotaxane 3ꢀ(a-CD)2 are localized
on the end alkyne units without visiting the central alkyne,
whereas in the [2]rotaxane 2ꢀa-CD the cyclodextrin hops
between both alkynes. The explanation is simply that the
length of an OPE repeat unit (6.9 Š) is less than the length of
a cyclodextrin (8.2 Š),[8] so that if the two a-CDs were to sit
over adjacent alkynes, one of them would be pushed over a
Scheme 3. Solution-phase and solid-phase synthesis of rotaxanes 2ꢀa-CD and 3ꢀ(a-CD)2. Reagents and conditions: a) Solution phase: MeOH,
H2SO4; solid phase: CPG-OH, pyBOP, iPr2NEt; b) 5, [Pd2(dba)3], PPh3, CuI, Et3N; c) I2; d) Solution phase: ester hydrolysis with LiOH, then 6a, a-
CD, Pd(OAc)2, TPPTS, CuI, Na2CO3, H2O; solid phase: 6a, a-CD, Pd(OAc)2, TPPTS, CuI, Na2CO3, H2O, then LiOH.[15]
6846 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 6845 –6848

Angewandte
Chemie
Sonogashira coupling but could be combined with Suzuki
coupling or Heck coupling to generate most varieties of p-
conjugated oligomer. This methodology should be applicable
to the construction of diverse insulated molecular wires, with
precisely defined sequences, lengths, and end groups.
Received: May 29, 2007
Published online: August 2, 2007
Keywords: cyclodextrins · molecular wires · rotaxanes ·
.
solid-phase synthesis · supramolecular chemistry
[1] P. Seneci, Solid-Phase Synthesis and Combinatorial Technolo-
gies, Wiley, New York, 2000.
[2] J. K. Young, J. C. Nelson, J. S. Moore, J. Am. Chem. Soc. 1994,
116, 10841 – 10842; E. L. Elliot, C. R. Ray, S. Kraft, J. R. Atkins,
J. S. Moore, J. Org. Chem. 2006, 71, 5282 – 5290.
[3] L. Jones, J. S. Schumm, J. M. Tour, J. Org. Chem. 1997, 62, 1388 –
1410.
[4] S. Anderson, Chem. Eur. J. 2001, 7, 4706 – 4714.
[5] P. R. L. Malenfant, J. M. J. FrØchet, Chem. Commun. 1998,
2657 – 2658.
[6] C. A. Briehn, P. Bäuerle, Chem. Commun. 2002, 1015 – 1023.
[7] N. F. Utesch, F. Diederich, C. Boudon, J.-P. Gisselbrecht, M.
Gross, Helv. Chim. Acta 2004, 87, 698 – 718.
[8] M. J. Frampton, H. L. Anderson, Angew. Chem. 2007, 119, 1046 –
1083; Angew. Chem. Int. Ed. 2007, 46, 1028 – 1064.
[9] F. Cacialli, J. S. Wilson, J. J. Michels, C. Daniel, C. Silva, R. H.
Friend, N. Severin, P. Samori, J. P. Rabe, M. J. OꢀConnell, P. N.
Taylor, H. L. Anderson, Nat. Mater. 2002, 1, 160 – 164.
[10] I. T. Harrison, S. Harrison, J. Am. Chem. Soc. 1967, 89, 5723 –
5724.
[11] J. A. Bravo, D. Orain, M. Bradley, Chem. Commun. 2002, 194 –
195.
[12] K. D. Johnstone, N. Bampos, J. K. M. Sanders, M. J. Gunter,
Chem. Commun. 2003, 1396 – 1397; Y.-F. Ng, J.-C. Meillon, T.
Ryan, A. P. Dominey, A. P. Davis, J. K. M. Sanders, Angew.
Chem. 2001, 113, 1807 – 1810; Angew. Chem. Int. Ed. 2001, 40,
1757 – 1760.
[13] G. Wenz, B.-H. Han, A. Müller, Chem. Rev. 2006, 106, 782 – 817.
[14] The synthesis of 1ꢀa-CD and 2ꢀa-CD was also attempted on
hydrophilic swollen gels, but rotaxane formation on these
supports is less efficient than on controlled-pore glass. For
example, we synthesized 2ꢀa-CD using the route shown in
Scheme 3 with oligo(oxyethylene glycol) ether functionalized
polystyrene in place of CPG (phenol ether linked PEG400 on
polystyrene cross-linked with 2 mol% divinylbenzene; poly-
(oxyethylene glycol) polymer (POP) resin, as described in M. C.
McCairn, S. R. Tonge, A. J. Sutherland, J. Org. Chem. 2002, 67,
4847 – 4855). The product was a 50:1 mixture of 2 and 2ꢀa-CD
(HPLC analysis after cleavage of the ester link with LiOH).
Thus it appears that a-CD is excluded from the gel-phase
support; however, it is known that large molecules, such as
enzymes, enter some hydrophilic gels; see: A. Basso, B. A.
Maltman, S. L. Flitsch, G. Margetts, I. Brazendale, C. Ebert, P.
Linda, S. Verdelli, L. Gardossi, Tetrahedron 2005, 61, 971 – 976.
[15] Abbreviations: DMAP: N,N-dimethyl-4-pyridinamine; HBTU:
O-(1H-benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium hexa-
fluorophospate; pyBOP: benzotriazol-1-yloxytris(pyrrolidino)-
phosphonium hexafluorophospate; TPPTS: tris(3-sulfonatophe-
nyl)phosphane trisodium salt; dba: trans,trans-dibenzylidene-
acetone; Ts: para-toluenesulfonyl.
Figure 1. a) Partial NOESY spectrum of 3ꢀ(a-CD)2 ([D6]DMSO, 298 K,
500 MHz). Cross-peaks to protons on the two cyclodextrin units (A
and B) are shown in red and blue, respectively; resonances a–i are
assigned to the aromatic protons as shown in Scheme 3. b) Calculated
structure of 3ꢀ(a-CD)2 showing the van der Waals surface of 3 and
the solvent-accessible surface of the a-CDs (1.4-ꢀ probe radius).[17]
bulky phenylene residue.[18] This is the first cyclodextrin
[3]rotaxane to be formed selectively as the head-to-tail
isomer;[19] previous examples have been synthesized as tail-
to-tail isomers (with both 6-rims at the center),[20,21] or as
mixtures of stereoisomers.[22,23] This work represents the first
solid-phase synthesis of a cyclodextrin rotaxane, the first
solid-phase synthesis of any [3]rotaxane, and the first
stereoselective synthesis of a cyclodextrin [3]rotaxane with a
head-to-tail stereochemistry (with the 2,3-rims of both CDs
pointing in the same direction). Use of the large tripodal
anchor 11 as a grafting point during solid-phase synthesis is
critical for avoiding crowding on the solid support, and will
enable the final insulated molecular wires to be anchored to
TiO2 semiconductor nanoparticles.[24] The general synthetic
strategy used to construct [3]rotaxane 3ꢀ(a-CD)2 by repeated
coupling with masked aryl iodide monomers is not limited to
[16] A similar rotaxane to 1ꢀa-CD was reported parallel to this
investigation: J. S. Park, J. N. Wilson, K. I. Hardcastle, U. H. F.
Bunz, M. Srinivasarao, J. Am. Chem. Soc. 2006, 128, 7714 – 7715.
Angew. Chem. Int. Ed. 2007, 46, 6845 –6848
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6847

Communications
[17] a) The geometry of 3ꢀ(a-CD)2 in water was calculated using
[20] M. R. Craig, T. D. W. Claridge, M. G. Hutchings, H. L. Ander-
son, Chem. Commun. 1999, 1537 – 1538.
[21] D.-H. Qu, Q.-C. Wang, X. Ma, H. Tian, Chem. Eur. J. 2005, 11,
5929 – 5937.
[22] A. G. Cheetham, T. D. W. Claridge, H. L. Anderson, Org.
Biomol. Chem. 2007, 5, 457 – 462.
[23] K. Sakamoto, Y. Takashima, H. Yamaguchi, A. Harada, J. Org.
Chem. 2007, 72, 459 – 465.
MacroModel with the Merck Molecular Force Field; see: T. A.
Halgren, J. Comput. Chem. 1999, 20, 720 – 729; T. A. Halgren, J.
Comput. Chem. 1999, 20, 730 – 748. b) Cyclodextrins are nor-
mally represented with the 6-rim narrower than the 2,3-rim, as in
Scheme 3, but the relative diameters of the rims can change with
conformation. In the energy-minimized structure of 3ꢀ(a-CD)2
(Figure 1b), CDB adopts a conformation in which the 6-rim is
wider than the 2,3-rim.
[24] E. Galoppini, W. Guo, J. Am. Chem. Soc. 2001, 123, 4342 – 4343;
Q. Wei, E. Galoppini, Tetrahedron 2004, 60, 8497 – 8508.
[18] A. Müller, G. Wenz, Chem. Eur. J. 2007, 13, 2218 – 2223.
[19] H. Yamaguchi, T. Oshikiri, A. Harada, J. Phys. Condens. Matter
2006, 18, S1809 – S1816.
6848 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 6845 –6848