Unidirectional threading synthesis of isomer-free [2]rotaxanes

Article abstract of DOI:10.1002/chem.200500415

The threading of an α-cyclodextrin (α-CyD) by an unsymmetrical dumbbell generally results in two isomeric [2]rotaxanes differing in the orientation of the α-CyD. In this work, two methods have been developed for the unidirectionally threading an α-CyD to

Full text of DOI:10.1002/chem.200500415

DOI: 10.1002/chem.200500415
Unidirectional Threading Synthesis of Isomer-Free [2]Rotaxanes
Qiao-Chun Wang, Xiang Ma, Da-Hui Qu, and He Tian*[a]
Abstract: The threading of an a-cyclo-
dextrin (a-CyD) by an unsymmetrical
dumbbell generally results in two iso-
meric [2]rotaxanes differing in the ori-
entation of the a-CyD.In this work,
two methods have been developed for
the unidirectionally threading an a-
CyD to obtain isomer-free [2]ro-
taxanes.These methods use the Suzuki
coupling of a boronic acid derivative
and a halide in aqueous alkaline solu-
tion.The conformations of the two uni-
directional [2]rotaxanes-R3 and R4
were determined by 2D 1H ROESY
NMR spectra.The optical spectral
studies revealed that each of the two
[2]rotaxanes can proceed with E/Z
photoisomerization and shuttling mo-
tions of the a-CyD ring on the thread
under alternating irradiation at 330 and
275 nm, accompanied by fluorescence
intensity changes at 530 nm.The in-
duced circular dichroism (ICD) spectra
of another two analogous [2]rotaxanes
R1 and R2 were also studied.Distinc-
tive ICD signal changes resulting from
the photoisomerization with respect to
the movements of a-CyD were detect-
ed.This demonstrates that, besides the
fluorescence, ICD signal is another
way to identify the shuttling motions of
a-CyD in these [2]rotaxanes.
Keywords:
cyclodextrins
shuttles
·
·
·
molecular
photoisomerization rotaxanes
supramolecular chemistry
Introduction
major issues in these [2]rotaxanes.To improve the stability,
many CyD-based [2]rotaxanes, in which the bulky stoppers
were linked covalently to the linear part instead of coordi-
nation, were developed.[10] At the same time, many unsym-
metrical [2]rotaxanes, in which a CyD ring encircles an un-
symmetrical dumbbell, were also set up.[11] However, when
an asymmetric CyD was threaded by a nonsymmetrical
dumbbell, it would give two isomers which differ in orienta-
tion of CyD with respect to the different ends of the dumb-
bell.As a result, isomeric [2]rotaxanes were always obtained
in these systems.
It was expected that unidirectional threading of a CyD
ring by a linear subunit during the formation of a [2]ro-
taxane could be achieved in order to construct more compli-
cated rotaxanes.So far, there are few reports concerning
unidirectional threading,[12] and the reports concerning CyD-
based unidirectional [2]rotaxanes are rare.[13] Recently, we
have set up two [2]rotaxane systems for molecular shut-
tles.[14,15] The two [2]rotaxanes, R1 and R2 shown in
Figure 1, were obtained as unidirectional [2]rotaxanes.From
this point of view, we would like to widen the scope of these
work to develop new methods for controlling the orientation
of an a-CyD when synthesizing [2]rotaxanes for molecular
shuttles.Here we report the unidirectional threading an a-
CyD by the Suzuki coupling reaction between a boronic
acid and a halide in an aqueous a-CyD solution.
Cyclodextrins (CyDs) continue to be attractive wheel com-
ponents in constructing supramolecular assemblies, such as
pseudorotaxanes,[1] rotaxanes[2] and catenanes.[3] A rotaxane
is described as a molecular system in which a macrocycle
(wheel) threads a linear subunit (dumbbell) with two bulky
stoppers.Rotaxanes have attracted more and more attention
because of their challenging constructions and potential ap-
plications in areas such as nanostructured functional materi-
als,[4] molecular switches,[5] molecular logic gates,[6] molecular
wires,[7] and memory devices.[8]
The foundation for the construction of a CyD-based ro-
taxane is the interactions between the hydrophobic cavity of
the CyD and the special hydrophobic unit in the linear com-
ponent.Since the early 1980s, CyD-based [2]rotaxanes have
been prepared by encapsulation of a symmetrical linear
molecule with CyD and subsequent “locking” by coordina-
tion.[2a,9] Symmetry and the lack of stability are the two
[a] Dr. Q.-C. Wang, X. Ma, D.-H. Qu, Prof. Dr. H. Tian
Laboratory for Advanced Materials and Institute of Fine Chemicals
East China University of Science and Technology
Shanghai 200237 (P.R. China)
Fax : (+86)21-6425-2288
E-mail: tianhe@ecust.edu.cn
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
1088
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 1088 – 1096

FULL PAPER
Chem. Eur. J. 2006, 12, 1088 – 1096
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
1089

H.Tian et al.
Results and Discussion
Figure 2.The observed NOEs from the aromatic protons
Hm, Hn, Ho, Hp of the dumbbell to the internal proton H3
of the a-CyD, and from Hn, Ho, Hp, Hq to H5, as well as
from the Hq, Hr to H6, indicate that the 6-OH rim of the a-
CyD faces the isophthalic acid stopper, and that the a-CyD
ring encircles the stilbene unit.
We now consider the unidirectionality of the two [2]ro-
taxane.For the synthesis of R3, a-CyD was first stirred with
5 to form inclusion complexes, which might contain two iso-
mers, that is, [2]pseudorotaxanes 5-CD-a and 5-CD-b as
shown in Figure 3.This is confirmed by the 1H NMR spec-
trum.After mixing 5 with a-CyD in D2O for 50 min (Fig-
ure 3b), the presence of two isomeric [2]pseudorotaxanes
can be clearly observed.However, as time elapses, the
isomer 5-CD-b disappears almost completely after 36 h (Fig-
ure 3c).These facts indicate that the threading rates of stil-
Synthesis and co-conformational identification of R3: The
synthesis of the shuttle R3 is similar to that of R2 (shown in
Figure 1), which was obtained from the combination of com-
pound 5 with a-CyD and the subsequent coupling with com-
pound 6.[15] Thus, after 5 was stirred with a-CyD in H2O for
50 h, it was coupled with 6 under the catalysis of Pd(OAc)2,
and the molar ratios of 5/a-CyD/6/Pd(OAc)2 were
1:1.2:1:0.15. R3 was obtained with a yield of 6.4% by chro-
matography.
It has been confirmed that only one isomer was present in
R2, where the 2,3-rim of the a-CyD faces the ANS stopper,
as shown in Figure 1.[15] The same fact was also found in R3,
whose co-conformation was confirmed by 2D 1H ROESY
NMR (500 MHz) spectrum in D2O at 298 K, as shown in
Figure 2.Selected 2D 1H ROESY NMR (500 MHz) spectrum of R3 in D2O recorded at 298 K.
Figure 3. 1H NMR spectra of a) 5 and the mixture of 5 and a-CyD at b) 50 min, and c) 36 h after the addition of a-CyD to 5.
1090
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 1088 – 1096

Rotaxanes
FULL PAPER
bene unit through the narrow 6-rim and through the wide
2,3-rim are not much different, but the dethreading rate of
the former [2]pseudorotaxane is much faster than that of
the latter, giving mostly the thermodynamically stable
isomer, where the 2,3-rim of the a-CyD faces the ANS stop-
per.Finally, the resulting [2]pseudorotaxane 5-CD-a under-
went Suzuki coupling with the boronic acid 6 which was
added to the solution subsequently, and as a result, the
[2]rotaxane R3 formed, where the 2,3-rim of the a-CyD
faces the ANS stopper, as shown in Figure 1.
From this point of view, it is not surprising that the [2]ro-
taxane R2 was obtained as the sole isomer.Analogous to
R3, it was synthesized from the reaction of a short-chain
phenylboronic acid derivative with a long-chain phenylha-
lide one, which can also form an inclusion compound with
a-CyD.The pure [2]pseudorotaxane was also obtained after
mixing 3 with a-CyD for 48 h in H2O.2D 1H ROESY NMR
spectrum of the pseudo-rotaxane revealed that the direction
of the CyD coincides with its analogue 3-CD-a, in which the
2,3-rim of the CyD faces the ANS unit.[16] The subsequent
coupling with 4 would no doubt result in the formation of
the co-conformational-pure [2]rotaxane R2, as shown in
Figure 1.
which account for the co-conformation of R1.However, the
2D 1H ROESY NMR revealed that no [2]peudorotaxane
formed at the beginning of the mixing of 1 with a-CyD.
Unlike the thermodynamic threading of R2 and R3, it
seems that the boronic acid 1 formed a diol complex with a-
CyD on the 2,3-rim, and the unidirectional [2]rotaxane R1 is
more like a result of a kinetic threading.[16] The synthesis of
R4 is quite similar to that of R1, the difference between
them is only that the isophthalic acid stopper in 1 is changed
to NS unit in 7 (Figure 1).It has been confirmed that R4
also have unique co-conformation, where the 2,3-rim of the
CyD faces the NS stopper.This co-conformation was con-
1
firmed by the 2D H ROESY NMR (500 MHz) spectrum of
R4 in D2O, as shown in Figure 4.The two methylene pro-
tons Ha and Hj were separated into two singlets at d 5.25
and 5.37, compared with one overlapping singlet at d 5.26 in
the dumbbell PR4.[16] The downfield shift of proton Ha in
R4 was believed to be due to threading through the a-CyD
ring.Furthermore, the NOEs shown in Figure 4 observed
from the aromatic protons Hc, Hd, He of the dumbbell to
the internal proton H5 of the a-CyD, and from Hb, Hc, Ha
(no Hj) to H3, as well as from the He, Hf to H6, indicate
that the 2,3-OH rim of the a-CyD is close to the NS stopper,
and that the a-CyD ring sits over the stilbene unit.
Synthesis and co-conformational identification of R4: The
[2]rotaxane R1 was synthesized from the Pd-catalyzed
Suzuki coupling of 1 with 2 in an alkaline aqueous a-CyD
solution, and it was confirmed that only one isomer, where
the 2,3-rim of the a-CyD faces the isophthalic acid stopper,
was obtained in the synthesis of R1, as shown in Figure 1.[14]
The other possible isomer where the 2,3-rim of the a-CyD
faces the ANS stopper was not present.
It should be noted that the CyD-free dumbbells PR3 and
PR4, resulting from the direct coupling of the boronic acids
and the halides, were also isolated from the reaction mix-
tures of R3 and R4, respectively (see Supporting Informa-
tion).[16]
Detection of the shuttling motion
One might reasonably think that, as for 3 and 5, the inter-
action of 1 with a-CyD would result in the formation of a
thermodynamically more stable [2]pseudorotaxane isomer,
ICD spectra: The circular dichroism spectra have been
widely used to interpret the absolute conformation of chiral
compounds.[17] When an achiral guest chromophore is locat-
Figure 4.Selected 2D 1H ROESY NMR (500 MHz) spectrum of R4 in D2O recorded at 298 K.
Chem. Eur. J. 2006, 12, 1088 – 1096
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
1091

H.Tian et al.
ed in a chiral host, the guest becomes optically active and
induced circular dichroism (ICD) signals originates.Harata,
Kajtµr, and Shimizu et al.have developed a rule for the in-
clusion complexes of cone-shaped chiral hosts such as cyclo-
dextrins.According to the rule, a positive ICD signal arises
when the electric transition dipole moment is aligned paral-
lel to the axis of the chiral host, but the signal is negative
and half as strong when it is oriented in a perpendicular
manner (Figure 5).[18–21] Recently, a similar rule has also
been developed by Kodaka for ICD effects outside the
chiral host cavity, where the sign of the ICD signal is oppo-
site to that arising inside the host cavity (Figure 5).[22]
The prolonged irradiation of aqueous R1-alk solution
(10 mol solid Na2CO3 were added to the R1 and the iso-
phthalic acid unit was changed to the anionic form) at
335 nm would result in the photoisomerization of the E to Z
form and the shuttling of the a-CyD ring from the stilbene
unit to the biphenyl unit, which was previously confirmed
by 2D ROESY NMR spectrum.[14] In this work, the ICD
spectra with respect to the photoisomerization of aqueous
R1-alk (2.0”10ꢀ5 m) solution are shown in Figure 6.
Before the irradiation, the ICD spectrum shows a weak
negative Cotton effect (De=ꢀ0.21 Lmolꢀ1 cmꢀ1) at 431 nm
(Figure 6A, curve a).This is because the ANS unit is far
from the a-CyD and the a-CyD affects little the naphthali-
mide group.Aqueous R1-alk solution is then irradiated at
335 nm for 100 min to reach the photostationary state.The
E!Z photoisomerization leads to the decrease in the ICD
signal at around 434 nm (De=ꢀ1.64 Lmolꢀ1 cmꢀ1) and the
presence of an isobestic points at 288 nm as shown in Fig-
ure 6A (curve d).The result coincides with the Kodakaꢁs
rule as the ANS unit in the Z form is closed to the a-CyD
ring due to the shuttling of the a-CyD ring.This co-confor-
Figure 6.A) ICD spectra of R1 (2.0”10ꢀ5 m) aqueous solution (a) and
after the irradiation at 335 nm for 75 min (c), 100 min (d).The spectral
change can be shifted back by the irradiation at 280 nm (40 min, curve
b).B) ICD spectra of R2 (2.0”10ꢀ5 m) aqueous solution (a) and after the
irradiation at 360 nm for 15 min (c), 25 min (d).The spectral change can
be shifted back by the irradiation at 430 nm (10 min, curve b).
allel to the Harataꢁs rule, the strong positive ICD signal at
around 300 nm (De=1.06 Lmolꢀ1 cmꢀ1) before irradiation
should be attributed to the stilbene p system in the cavity of
mation results in a more negative ICD signal (Figures 5 and
max
abs
6) at around 430 nm, which is the l of the ANS unit.Par-
Figure 5.Harataꢁs rule and the Kodakaꢁs rule, and the ICD signals of E/Z R1 and E/Z R2 according to the rules.The arrows on the left represent the
electric transition dipole moments of the guests.
1092
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 1088 – 1096

Rotaxanes
FULL PAPER
the a-CyD ring.However, the signal blue shifts to 284 nm
after the photoisomerization for the sake that the p system
becomes weaken in the Z form.Nevertheless, the intensity
of the ICD signal change little because the p system still lo-
cates in the cavity of the a-CyD ring, as shown in Figure 6.
The ICD spectra for R2 are more complex because all of
the NS unit, ANS unit and the azo group generate ICD sig-
nals.According to the Kodakaꢁs rule, in the E form of R2,
because the ANS unit is closed to the a-CyD ring, a nega-
It can be seen that the process of the a-CyD shuttling
from the stilbene unit to the biphenyl unit (E!Z photoiso-
merization) in R1-alk is accompanied by an obvious in-
crease (6.8 times) in the ICD value at around 440 nm. The
shifted ICD spectra can be shifted reversibly back when the
a-CyD shuttles back (Z!E photoisomerization).These
facts reveal that ICD spectrum can identify distinctively the
shuttling motions of CyD ring in a molecular shuttle.The
different ICD signals before and after the shuttling can also
represent the “0” and “1” states of these CyD-based [2]ro-
taxanes.Furthermore, the ICD spectrum provides possibili-
ties to identify the movements of these CyD-based [2]ro-
taxanes in solid state, in which the ICD signals would
become more intensive.[23]
tive Cotton effect should appear at around 430 nm, which is
max
abs
the l
of the ANS unit; and the NS unit is far away from
the a-CyD ring, thus a weak negative ICD signal would be
induced at around 340 nm.Simultaneously, the azo group is
included in the cavity of the a-CyD, and according to the
Harataꢁs rule, a characteristic strong positive Cotton effect
would have to appear at around 360 nm.As a consequence,
the aqueous solution of R2 (2.0”10ꢀ5 m) shows a positive
Cotton effect (De=1.31 Lmolꢀ1 cmꢀ1) at 355 nm, as well as a
negative Cotton effect (De=ꢀ0.78 Lmolꢀ1 cmꢀ1) at 442 nm
(Figure 6B, curve a).
Absorption spectra: The UV/Vis absorption spectra of [2]ro-
taxanes R3 and R4, as well as the corresponding spectra of
dumbbells PR3 and PR4 in H2O are shown in Supporting
Information.[16] The absorption maxima at around 427 nm
for R3 (curve a in Figure S4A)[16] and PR3 attributes to
ANS unit, and the maxima at around 330 nm ascribes the
stilbene unit.For R4 (curve a in Figure S4B)[16] and the cor-
responding dumbbell PR4, there is an obvious increase in
absorption at the band around 330 nm with the NS unit.It
can be observed that the two absorption spectra of R3 and
PR3, as well as the two of R4 and PR4, are very similar,
which means that the encircling of a-CyD on the dumbbells
does not influence the absorption spectra.
It has been found that the irradiation on a stilbene based
[2]rotaxane would induce E!Z photoisomerization, which
results in absorption spectral changes.[2f,14] For [2]rotaxanes
R3 and R4, the same characteristic spectral changes in aque-
ous solution (9.1”10ꢀ6 m) were also found, as also shown in
Figure S4A.[16] The irradiation on the R3 solution at 330 nm
for 40 min results in photoisomerization to give a E/Z mix-
ture, characterized by a rise of the absorption at around
275 nm (DA = 0.016) and a decrease in absorption at
330 nm (DA = 0.013), as well as the presence of two isobes-
tic points at 294 and 379 nm.The maximum absorption of
ANS unit at around 427 nm changes little, as this stopper is
separated from the residues by the methylene group.Similar
results were obtained for R4 solution after irradiation at
330 nm for 50 min, inducing an increase in absorption at
around 275 nm (DA=0.020) and decrease in absorption at
330 nm (DA=0.017), as well as the appearance of two iso-
bestic points at 296 and 361 nm (as illustrated in Figure
S4B).[16]
It has also been confirmed that the irradiation of aqueous
R2 solution at 360 nm would induce the E!Z photoisome-
rization, and the a-CyD ring would move from the azo unit
to the biphenyl unit.[15] As a consequence, the azo group
would be located outside the a-CyD ring and the ANS unit
would become far away from the CyD ring.(Figure S3 in
the Supporting Information shows the absorption changes of
R2 upon the E!Z isomerization.[16]) Besides the absorption
decrease at 360 nm, an additional absorption at around
430 nm would appear during the course of E!Z photoiso-
merization (DA = 0.02). As a result, an ICD signal would
originate at 430 nm.According to Kodakaꢁs rule, the ICD
signals are negative for both the ANS unit and the Z form
of the azo unit (Figure 5).Although the increase of the ab-
sorption at 430 nm is relatively small compared with the ab-
sorption of the ANS unit (A=0.13), it is likely that the azo
unit is very closed to the a-CyD rim, which results in a
more strong negative ICD, and that the ANS unit is far
away from the a-CyD ring, which results in a small less neg-
ative ICD.As a result, after 25 min irradiation at 360 nm,
the aqueous R2 solution (2.0”10ꢀ5 m) gives a more negative
ICD signal (De=ꢀ1.11 Lmolꢀ1 cmꢀ1) at 451 nm (Figure 6B,
curve d) compared with the original one.Meanwhile, a de-
crease in De (D(De)=ꢀ0.37 Lmolꢀ1 cmꢀ1) at 355 nm is also
observed after the irradiation.This spectral change should
be due to the more negative ICD signal at around 340 nm of
NS unit when its position is close to the a-CyD ring
(Figure 5).
Compared with R3 and R4 solutions, the corresponding
aqueous solution of dumbbell PR3 and PR4 also undergo
E!Z photoisomerization after the irradiation at 330 nm
(Figures S5 and S6 in Supporting Information),[16] but reach-
es to the photostationary state within shorter time (21 and
29 min for PR3 and PR4, respectively).There is a relative
larger increase in the absorption at around 275 nm (DA=
0.027 and DA=0.045 for PR3 and PR4, respectively) and a
relative large decrease at 330 nm (DA=0.025 and DA=
0.042 for PR3 and PR4, respectively).These results were
It should be noted that the photoisomerized solutions of
both R1-alk and R2 can be induced reversible Z!E photo-
isomerization by irradiation at 280 and 430 nm, respective-
ly.[14,15] These processes can also be addressed by the ICD
spectra, as shown in Figure 6.The ICD spectra of the two
isomerized solutions of R1-alk and R2 after the irradiation
(at 280 nm for R1-alk and 430 nm for R2) can be shifted
back to their photostationary states (curve b in Figure 6A
and B), which are almost the same as the original E ones.
Chem. Eur. J. 2006, 12, 1088 – 1096
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
1093

H.Tian et al.
also found in R1[14] indicating that the dumbbells undergo
easier E!Z photoisomerization than the corresponding
[2]rotaxanes do.This again demonstrates the fact that E!Z
isomerization becomes more difficult in the presence of the
a-CyD ring.[2f]
Fluorescence spectra: Figure 7 shows the emission spectra of
R4 (curve a) and model compound NS (1,8-naphthalimide-
3-sulfonic acid sodium salt, curve b) in aqueous solution
Figure 7.The fluorescence emission spectra of R4 (curve a) and the
model compound NS (1,8-naphthalimide-3-sulfonic acid sodium salt,
curve b, half reduced) in aqueous solution (both 9.1”10ꢀ6 m) excited at
330 nm.The shoulder around 395 nm in the spectrum of R4 is due to a
very small contribution from the emission of NS unit.
Figure 8.A) The fluorescence emission spectra of R3 (9.1”10ꢀ6 m) aque-
ous solution (a), and after UV light (330 nm) irradiation for 40 min (b),
the spectra change can be shifted back by the irradiation at 275 nm for
23 min (c); B) The fluorescence emission spectra of R4 (9.1”10ꢀ6 m)
aqueous solution (a), and after UV light (330 nm) irradiation for 50 min
(b), the spectra change can be shitted back by the irradiation at 275 nm
for 28 min (c); all of the spectra were recorded with the excitation at
427 nm.
(both 9.1”10ꢀ6 m) with excitation at 330 nm, the maximum
absorption wavelength of the NS unit.Compared with the
strong emission maximum at 395 nm of compound NS, the
fluorescent intensity of NS unit in R4 at around 395 nm is
1
reduced to = of that of compound NS.At the same time, a
50
distinct emission maximum at 530 nm, which is the charac-
teristic emission maximum of the ANS unit, is obtained.
The quenching of the NS unit should be attributed to the
singlet–singlet energy transfer from the NS unit to the ANS
unit;[4a] the first excited singlet state of which lies at a lower
energy state.Energy transfer in R4 was confirmed by the
fluorescence excitation spectra of R4 for ANS emission
measured at 530 nm (Figure S7).[16] The spectrum at around
330 nm region is parallel to the absorption of R4 in the
region.
As energy transfer from the NS unit to the ANS unit
quenches the fluorescence of the NS unit, only the emission
of the ANS unit was taken into account in R4 and PR4.
Compound R4 has an emission maximum at 530 nm in
aqueous solution (9.1”10ꢀ6 m) with the excitation at 427 nm,
which is the maximum absorption wavelength of the ANS
unit, as shown in Figure 8A.However, the fluorescent inten-
sity of R4 (216 a.u.) is higher than that of PR4 (175 a.u.)[16]
by about 23%.This is consistent with an 86% increase in
the intensity of R3 (151 a.u., curve a in Figure 8B) over that
of PR3 (81 a.u.)[16] at the same fluorescence band, and also a
decrease in fluorescent intensity of dumbbell PR1-alk over
that of R1-alk as observed in our previous work.[14] These
facts should be ascribed to the rigidity of the a-CyD ring to
the encircled stilbene unit, in which the vibration and the
rotation of the bonds are hindered.It might also be consid-
ered that the stoppers increase fluorescence when the CD
approaches due to displacement of solvent and due to
shielding from molecular oxygen quenching.As a result,
non-radiative excited energy loss from the excited singlet
state of these [2]rotaxane is reduced.
It is known that the photoisomerization of encircled stil-
bene or azobenzene unit would result in the shuttling
motion of the a-CyD ring.[2f,14,15] Moreover, shifting the a-
CyD macrocycle away from a luminescent stopper would
cause a decrease of the fluorescence intensity of the stopper.
Shuttling of the a-CyD ring to another stopper would cause
an increase in the intensity of the emission.[14,15] The same
were found in R3 and R4, as shown in Figure 8.On the one
hand, once the shuttling of a-CyD ring to the biphenyl unit
performed, the macrocycle becomes closer to the ANS unit
in R4.Such movement of a-CyD results in an increasing by
about 36% in the intensity of the fluorescence band at
lmax =530 nm.On the other hand, as a-CyD ring moves
away from the ANS unit in R3, a decreasing by about 29%
in the intensity at the same fluorescence band is observed.
1094
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 1088 – 1096

Rotaxanes
FULL PAPER
measured at 298 K in aqueous solution.The UV light irradiation at
330 nm on these solutions was carried out to induce E/Z photoisomeriza-
tion.After prolonged irradiation for 40 min on R3 (50 min for R4,
21 min for PR3 and 29 min for PR4) solution, no additional change in
the absorption was observed, which meant that the solution reached the
photostationary state.The fluorescence emission spectra were recorded
without delay.The wavelength of the irradiating UV light was then
changed to 275 nm in order to induce Z/E photoisomerization.After the
prolonged irradiation of the solution for 23 min (28 min for R4, 12 min
for PR3 and 16 min for PR4), the spectral change in absorption and
emission was recorded.
The fact that there is little change in the fluorescent intensi-
ty of the corresponding dumbbell PR3 and PR4 after the
photoisomerization,[16] again confirms that only the move-
ment of the a-CyD ring is responsible for the obvious spec-
tral changes.
Reversibility: After the photoisomerization of R3 and R4
from the E to Z form, the equilibrium can be reversed by ir-
radiation at 275 nm.This process is accompanied by de-
crease in the absorption at around 275 nm and increase in
absorption at 330 nm (curve c in Figure S4A and S4B in the
Supporting Information[16]), as along with an increase in the
fluorescence intensity at lmax =530 nm for R3 (curve c in
Figure 8A) and a decrease for R4 (curve c in Figure 8B).
Owing to the reversibility of the photoisomerization process,
the photoinduced shuttling motion of the a-CyD ring can be
carried out repeatedly with reversible fluorescent signal
changes.
Synthetic routes to R3 and R4: The synthetic schemes for the two [2]ro-
taxanes are shown in Scheme S1 and S2.[16] Boronic acid 7, which was ob-
tained from the reaction of 1,8-naphthalimide-3-sulfonic acid sodium salt
with a benzyl bromide A3, was coupled with the halide 2 in the a-CyD
aqueous solution catalyzed by Pd(OAc)2 resulted in the formation of R4.
Similarly, R3 was prepared from the coupling of the halide 5 with the
boronic acid 6.
Compound R3: A mixture of 5 (0.35 g, 0.47 mmol), a-CyD (0.55 g,
0.56 mmol) in distilled water (15 mL) was stirred at 608C for 48 h, then
Na2CO3 (0.7 g, 6.6 mmol), boronic acid 6 (0.15 g, 0.71 mmol) and Pd-
(OAc)2 (15 mg) were added to the suspension, the resulting mixture was
stirred under Ar at 808C for 24 h, acidified with acetic acid, evaporated
to dryness, the resulting dark solid was purified by column chromatogra-
phy to give pure R3 (52 mg, 6.4%) as yellow powder. No obvious m.p.
was found, which might due to the decomposing of the a-cyclodextrin.
1H NMR (500 MHz, D2O, 25 8C, TMS): d=8.98 (s, 1H), 8.82 (s, 1H), 8.76
Conclusion
In summary, an unidirectional [2]pseudorotaxane can be
thermodynamically prepared from interaction of a naphtha-
limide–stilbene compound with a-CyD, and an unidirection-
al [2]rotaxane can be obtained from the consequent Suzuki
coupling of the [2]pseudorotaxane with a boronic acid.The
[2]rotaxane R3 was synthesized by this thermodynamic
means.On the other hand, the Suzuki coupling of a “long
chain” boronic acid and a halide stopper in aqueous a-CyD
solution also results in the formation of an unidirectional
[2]rotaxane R4.Unlike the thermodynamic threading, the
threading of a-CyD in R4 is likely a kinetic one, in which
the “long” boronic acid might act as an orientation auxiliary
group.The alternating irradiation with two different fre-
quencies of UV light on the two [2]rotaxanes in aqueous
solutions results in reversible E/Z photoisomerization and
the shuttling motions of a-CyD, accompanied by obvious
changes in the fluorescent intensity at 530 nm.The ICD
spectra study of two previous prepared [2]rotaxanes, R1 and
R2, reveals that the reversible shuttling motions of the a-
CyD ring can also induce distinctive ICD signal changes,
which can in term identify the movements of the a-CyD re-
sulting from the photoisomerization.
3
(s, 1H), 8.42 (s, 1H), 8.28 (s, 2H), 7.83 (d, J(H,H)=8.7 Hz, 2H), 7.78 (d,
3J(H,H)=8.7 Hz, 2H), 7.60 (d, 3J(H,H)=8.6 Hz, 2H), 7.46 (d, 3J(H,H)=
8.6 Hz, 2H), 7.14 (d, 3J(H,H)=16.5 Hz, 1H), 6.95 (d, 3J(H,H)=16.5 Hz,
1H), 5.38 (s, 2H), 4.95 (s, 6H), 3.94 (dd, 3J(H,H)=9.0, 9.2 Hz, 6H), 3.83
3
3
(d, J(H,H)=8.7 Hz, 6H), 3.73 (dd, J(H,H)=10.2, 10.6 Hz, 6H), 3.68 (d,
3J(H,H)=9.0 Hz, 6H), 3.54 (dd, J(H,H)=9.2, 10.6 Hz, 6H), 3.48 ppm (d,
3
3J(H,H)=10.2 Hz,
6H);
elemental
analysis
calcd
(%)
for
C71H82N2Na2O42S2·11.5H2O (1952.7): C 43.66, H 5.42, N 1.43; found: C
43.47, H 5.31, N 1.57.
Compound R4: A mixture of boronic acid 7 (0.5 g, 0.93 mmol), a-CyD
(1.2 g, 1.2 mmol) and Na2CO3 (0.2 g, 1.9 mmol) in distilled water (25 mL)
was stirred at 608C for 50 h, then Na2CO3 (0.35 g, 3.3 mmol), 2 (0.6 g,
0.95 mmol) and Pd(OAc)2 (15 mg) were added to the suspension, the re-
sulting mixture was stirred under Ar at 808C for 20 h, acidified with
acetic acid, evaporated to dryness, the resulting dark solid was purified
by column chromatography to give pure R4 (88 mg, 4.9%) as yellow
powder.No obvious mp. .was found, which might due to the decompos-
1
ing of the a-cyclodextrin. H NMR (500 MHz, D2O, 25 8C, TMS): d=8.93
3
(s, 1H), ~8.70 (m, 4H), 8.39 (s, 1H), 8.30 (s, 1H), 7.78 (dd, J(H,H)=8.2,
7.8 Hz, 1H), ~7.66 (m, 6H), 7.45 (m, 6H), 7.10 (d, 3J(H,H)=16.2 Hz,
1H), 6.93 (d, 3J(H,H)=16.2 Hz, 1H), 5.37 (s, 2H), 5.25 (s, 2H), 4.94 (d,
3
3J(H,H)=2.4 Hz, 6H), 4.13 (t, J(H,H)=9.4 Hz, 6H), 3.69 (m, 12H), 3.63
(m, 12H), 3.57 (d, 3J(H,H)=10.7 Hz, 6H), 3.53 (dd, 6H), 3.47 (dd,
3J(H,H)=2.4, 9.6 Hz, 6H); elemental analysis calcd (%) for
C82H88N3Na3O43S3·13H2O (2203.0): C 44.71, H 5.18, N 1.91; found: C
44.47, H 5.34, N 1.77.
Acknowledgements
Experimental Section
This work is financially supported by NSFC/China (90401026 and
20273020), partially by Education Committee of Shanghai and Scientific
Committee of Shanghai.We thank Professor Yu Liu (NanKai University/
China) for his relevant advice on the synthesis.The detail work on this
direction originated from the discussion with Professor Harry L.Ander-
son (Oxford University/UK).
1
Instruments: H NMR spectra were measured on a Bruker AM 500 spec-
trometer.UV/Vis spectra were done on a Varian Cary 100 spectropho-
tometer (1 cm quartz cell used).Fluorescent spectra were recorded on a
Varian Cary Eclipse Fluorescence Spectrophotometer (1 cm quartz cell
used).The photoirradiation was carried on a CHF-XM 500-W high-pres-
sure mercury lamp with suitable filters in a sealed Ar-saturated 1 cm
quartz cell.ICD spectra were recorded on a Jasco J-720 spectropolarime-
ter in a 1 cm quartz cell.
[1] a) A.Harada, M.J.Kamachi, Macromolecules 1990, 23, 2821–2823;
UV-visible absorption and fluorescence emission spectral measurements:
The absorption and emission spectra of the samples (all 9.1”10ꢀ6 m) were
b) Y.Liu, L.Li, H.Y.Zhang, Y.L.Zhao, X.Wu,
2002, 35, 9934–9938; c) Y.Liu, Y.L.Zhao, H.Y.Zhang, H.B.Song,
Macromolecules
Chem. Eur. J. 2006, 12, 1088 – 1096
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
1095

H.Tian et al.
Angew. Chem. 2003, 115, 3382–3385; Angew. Chem. Int. Ed. 2003,
42, 3260–3263; d) H.Fujita, T.Ooya, N.Yui, Macromolecules 1999,
32, 2534–2541.
[10] a) A.Harada, J.Li, M.Kamachi, Chem. Commun. 1997, 1413–1414;
b) Y.Kawaguchi, A.Harada, Org. Lett. 2000, 2, 1353–1356; c) C.A.
Stainier, M.J. OꢁConnell, W. Clegg, H.L. Anderson,
Chem.
[2] a) H.Ogino, J. Am. Chem. Soc. 1981, 103, 1303–1304; b) S.Ander-
Commun. 2001, 493–494; d) H.Onigi, B.Carrozzini, G.L.Cascara-
son, T.D.W. Claridge, H.L. Anderson,
1367–1370; Angew. Chem. Int. Ed. Engl. 1997, 36, 1310–1313; c) H.
Murakami, A.Kawabuchi, K.Kotoo, M.Kunitake, N.Nakashima, J.
Angew. Chem. 1997, 109,
no, C.J.Easton, A.J.Edwards, S.F.Lincoln, A.D.Rae,
J. 2003, 9, 5971–5977.
Chem. Eur.
[11] a) R.Isnin, A.E.Kaifer,
J. Am. Chem. Soc. 1991, 113, 8188–8190;
Am. Chem. Soc. 1997, 119, 7605–7606; d) Y.Kawaguchi, A.Harada,
J. Am. Chem. Soc. 2000, 122, 3797–3798; e) H.Onagi, C.J.Easton,
S.F. Lincoln, Org. Lett. 2001, 3, 1041–1044; f) C.A. Stainier, S.J.
b) J.E.H. Buston, F. Marken, H.L. Anderson,
Chem. Commun.
2001, 1046–1047; c) J.W.Park, H.J.Song,
Org. Lett. 2004, 6, 4869–
4872.
Alderman, T.D.W. Claridge, H.L. Anderson,
Angew. Chem. 2002,
[12] a) A.Arduini, F.Calzavacca, A.Pochini, A.Secchi, Chem. Eur. J.
114, 1847–1850; Angew. Chem. Int. Ed. 2002, 41, 1769–1772; g) Y.
2003, 9, 793–799; b) A.Credi, S.Dumas, S.Silvi, M.Venturi, A.Ar-
duini, A.Pochini, A.Secchi, J. Org. Chem. 2004, 69, 5881–5887;
c) an unidirectional threading was reported during our submission:
Liu, H.Wang, Y.Chen, C.F.Ke, M.Liu,
J. Am. Chem. Soc. 2005,
127, 657–666.
[3] a) D.Armspach, P.R.Aston, C.P.Moore, N.Spencer, J.F.Stoddart,
TJ..Wear, DJ..Williams, Angew. Chem. 1993, 105, 901–905;
A.Arduini, F.Ciesa, M.Fragassi, A.Pochini, A.Secchi,
Chem. 2005, 117, 282–285; Angew. Chem. Int. Ed. 2005, 44, 278–
281.
Angew.
Angew. Chem. Int. Ed. Engl. 1993, 32, 854–858; b) D.Armspach,
P.R. Aston, R. Ballardini, A. Balzani, C.P. Godi, C.P. Moore, L.
[13] Anderson and co-workers had obtained a unidirectional [2]rotaxane
based on hexakis(2,3,6-tri-O-methyl)-a-CD.MR. .Craig, MG. .
Prodi, N.Spencer, J.F.Stoddart, T.J.Wear, D.J.Williams,
Chem.
Eur. J. 1995, 1, 33–35.
Hutchings, T.D.W. Claridge, H.L. Anderson,
Angew. Chem. 2001,
[4] a) V.Balzani, M.Venturi, A.Credi,
Molecular Devices and Ma-
113, 1105–1108; Angew. Chem. Int. Ed. 2001, 40, 1071–1074.
chines, VCH, Weinheim, Germany, 2003; b) M.Born, H.Ritter,
Adv. Mater. 1996, 8, 149–151; c) D.Nelpal, S.Samal, K.E.Geckel-
er, Macromolecules 2003, 36, 3800–3802.
[14] Q.C. Wang D.H. Qu, J. Ren, K. Chen, H. Tian,
Angew. Chem.
2004, 116, 2715–2719; Angew. Chem. Int. Ed. 2004, 43, 2661–2665.
[15] D.H.Qu, Q.C.Wang, J.Ren, H.Tian,
2088.
Org. Lett. 2004, 6, 2085–
[5] a) M.Bauer, W.M.Müller, U.Müller, K.Rissanen, F.Vçgtle,
bigs Ann. 1995, 649–656; b) B.L. Feringa, Molecular Switches,
VCH, Weinheim, Germany, 2001; c) M.Asakawa, M.Higuchi, Mat-
tersteig, G.Nakamura, T.A.R.Pease, F.M.Raymo, T.Shimizu, J.F.
Stoddart, Adv. Mater. 2000, 12, 1099–1102; d) G.Bottari, DA. .
Leigh, E.Pꢂrez, J. Am. Chem. Soc. 2003, 125, 13360–13361; e) B.L.
Feringa, Acc. Chem. Res. 2001, 34, 504–513.
Lie-
[16] Supporting Information: Detailed synthetic procedures and
1H NMR data, possible mechanism of threading synthesis as well as
the absorption and emission spectra of PR3 and PR4.
[17] C.Pierre, ORD and CD in Chemistry and Biochemistry, Academic
Press, New York, 1972.
[18] K.Harata, H.Uedaira, Bull. Chem. Soc. Jpn. 1975, 48, 375–378.
[6] a) CP..Collier, EW. .Wong, M.Belohradsky, FM. .Raymo, JF. .
[19] M.Kajtµr, C.Horvath-Toro, E.Kuthi, J.Szejtli,
Acta Chim. Acad.
Stoddart, P.J.Kuekes, R.S.Williams, J.R Heath,
Science 1999, 285,
Sci. Hung. 1982, 110, 327–355.
391–394; b) C.P.Collier, M.Belohradsky, F.M.Raymo, J.F.Stod-
dart, J.R.Heath, J. Am. Chem. Soc. 2000, 122, 5831–5840; c) D.A.
Leigh, Mꢃ. F. .Morales, EM. .Pꢂrez, JK. Y. .Wong, CG. .Saiz,
[20] H.Shimizu, A.Kaito, M.Hatano,
2678–2684.
Bull. Chem. Soc. Jpn. 1979, 52,
[21] X.Zhang, W.M.Nau,
Angew. Chem. 2000, 112, 555–557; Angew.
A.M.Z. Slawin, A.J. Carmichael, D.M. Haddleton, A.M. Brouw-
Chem. Int. Ed. 2000, 39, 544–547.
er, W.J. Buma, G.W.H. Wurpel, S. Leꢄn, F. Zerbetto,
Chem. 2005, 117, 3122–3127; Angew. Chem. Int. Ed. 2005, 44, 3062–
3067.
Angew.
[22] a) M.Kodaka, J. Phys. Chem. 1991, 95, 2110–2112; b) M.Kodaka, J.
Am. Chem. Soc. 1993, 115, 3702–3705; c) M.Kodaka, J. Phys.
Chem. A 1998, 102, 8101–8103.
[7] a) P.N. Taylor, M.J. OꢁConnell, L.A. McNeill, M.J. Hall, R.T.
Aplin, H.L. Anderson, Angew. Chem. 2000, 112, 3598–3602;
Angew. Chem. Int. Ed. 2000, 39, 3456–3460; b) F.Cacialli, JS. .
Wilson, J.J.Michels, C.Daniel, C.Silva, R.H.Friend, N.Severin,
P.J.Samorꢅ, P.Rabe, M.J.OꢁConnell, P.N.Taylor, H.L.Anderson,
Nat. Mater. 2002, 1, 160–164; c) P.N.Taylor, A.J.Hagan, H.L.An-
derson, Org. Biomol. Chem. 2003, 1, 3851–3856.
[23] The intensity of the ICD signals is relatively weak in solutions be-
cause the concentration of these rotaxanes is low.It should be noted
that the E form of R1-alk or R2 is not completely transferred to the
Z form (for R1-alk, the ratio of E/Z at the photostationary is about
63/37, and for R2 is about 59:41),[14,15] and this fact should also be
responsible for the relative small change in the intensity of ICD
signal.
[8] I.Willner, V.Pardo-Yissar, E.Katz, KT. .Ranjit,
J. Electroanal.
Chem. 2001, 497, 172–177.
[9] a) T.V.S.Rao, D.S.Lawrence,
J. Am. Chem. Soc. 1990, 112, 3614–
Received: April 14, 2005
Revised: July 12, 2005
3615; b) R.S.Wylie, D.H.Macartney, J. Am. Chem. Soc. 1992, 114,
3136–3138.
Published online: October 24, 2005
1096
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 1088 – 1096