Integrating replication processes with mechanically interlocked molecular architectures

Article abstract of DOI:10.1016/j.tet.2008.05.049

A kinetic model for the integration of self-replication with the formation of a mechanically interlocked molecular architecture, namely a rotaxane, is presented. The logical steps required to convert this model into molecular structures through consideration of the design criteria highlighted by the model are discussed and executed. Ultimately, despite careful design, the rotaxane synthesised did not replicate as expected. The reasons for this failure are traced by experiment and computation to the sub-optimal association constant for the pseudorotaxane complex required to form the replicating rotaxane. Additionally, a deleterious supramolecular steric effect, operating through the proximity of the macrocyclic component of the pseudorotaxane to the transition state for the stoppering reaction is identified computationally.

Full text of DOI:10.1016/j.tet.2008.05.049

Tetrahedron 64 (2008) 8464–8475
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Integrating replication processes with mechanically interlocked molecular
architectures
*
Annick Vidonne, Douglas Philp
EaStCHEM and Centre for Biomolecular Sciences, School of Chemistry, University of St. Andrews, North Haugh, St. Andrews, Fife KY16 9ST, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A kinetic model for the integration of self-replication with the formation of a mechanically interlocked
molecular architecture, namely a rotaxane, is presented. The logical steps required to convert this model
into molecular structures through consideration of the design criteria highlighted by the model are
discussed and executed. Ultimately, despite careful design, the rotaxane synthesised did not replicate as
expected. The reasons for this failure are traced by experiment and computation to the sub-optimal
association constant for the pseudorotaxane complex required to form the replicating rotaxane. Addi-
tionally, a deleterious supramolecular steric effect, operating through the proximity of the macrocyclic
component of the pseudorotaxane to the transition state for the stoppering reaction is identified
computationally.
Received 28 March 2008
Received in revised form 6 May 2008
Accepted 9 May 2008
Available online 14 May 2008
Keywords:
Molecular recognition
Self-replication
Rotaxanes
Ó 2008 Elsevier Ltd. All rights reserved.
Kinetics
1. Introduction
synthetic self-replicators reported in this period have been based
on the kinetic model11 shown in Figure 1.
Recent advances in supramolecular synthesis have permitted
the fabrication1 of discrete, structurally intricate nanometre-scale
objects,2 molecular machinery3 and periodic structures.4 Sophisti-
cated supramolecular hierarchies5 of this sort can be constructed
through the precise arrangement of structural subunits through
algorithmic self-organisation and self-assembly. Thus, the con-
struction of structurally diverse families of novel, intelligent
materialsdpossessing programmability, addressability and desir-
able physical6 attributesdhas become feasible. Although ap-
proaches based on self-assembly are versatile, the production of the
basic subunits for any assembly process currently relies upon tra-
ditional synthetic methodologies. Drexler7 highlighted a radical
alternativedthe fabrication of replicateable networks via atom-by-
atom engineering. However, this proposal has been described8 as
implausible and unrealistic. Despite this criticism, an attractive
alternative to traditional synthetic approaches is a strategy based
on replication processes. In this approach, the building blocks for
the subsequent self-assembly processes are themselves capable of
their own synthesis through recognition-mediated processes.
In the last 15 years, several examples9 of self-replicating systems
capable of templating and catalysing their own synthesis have
appeared10 in the chemical literature. Almost all of the examples of
This minimal model of self-replication contains three distinct
reaction channels. The first is the uncatalysed bimolecular reaction
between reagents A and B to afford the template T. A key re-
quirement of the model is that A and B bear complementary rec-
ognition sites. One consequence of the presence of these
recognition sites is that A and B can associate with each other to
Bimolecular
A
B
Channel
T
T
Autocatalytic
Cycle
[A•B]
AB
Complex
Channel
Tinactive
[T•T]
[A•B•T]
Figure 1. The minimal model of self-replication. Reagents A and B can react through
three pathwaysdan uncatalysed bimolecular reaction, a recognition-mediated pseu-
dounimolecular pathway mediated by a binary complex [A$B] and a recognition-me-
diated pseudounimolecular autocatalytic cycle mediated by a ternary complex [A$B$T].
* Corresponding author.
E-mail address: dp12@st-andrews.ac.uk (D. Philp).
0040-4020/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.05.049

A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
8465
and the reaction between them is pseudointramolecular. The
product of this reaction channel forms a closed template Tinactive, in
which the recognition used to assemble the binary complex lives on
in the template, thus, although rate acceleration is achieved by this
mechanism, this template is catalytically inert. The third reaction
channel available to the system is the autocatalytic13 cycle. In this
channel, A and B bind reversibly to the open template T to form
a catalytic ternary complex [A$B$T]. In a manner similar to the
[A$B] complex, the reaction between A and B is also rendered
pseudointramolecular. Bond formation occurs between A and B to
give the product duplex [T$T], which then dissociates to return two
molecules of T to the start of the autocatalytic cycle. Thus, assuming
that the open template T presents its recognition sites in the correct
orientation, it can act as a template for its own formation, trans-
mitting molecular information through the formation of identical
copies of itself.
(a)
(b)
(c)
Heat
Figure 2. Methods of rotaxane formation. (a) Clipping: a recognition site within the
thread-like component is used to template the formation of the macrocyclic compo-
nent of the rotaxane. (b) Threading and stoppering: a pseudorotaxane complex be-
tween a linear molecule and a macrocycle is captured by covalent bond formation
between a stoppering reagent and the pseudorotaxane complex. (c) Slippage: the two
pre-formed components of the rotaxane can associate through a temporary increase in
the mean free path available through the macrocycle cavity, usually accomplished by
heating.
Our initial forays14 in the field of molecular self-replication fo-
cused on the utilisation of bifunctional precursor subunits that
incorporate mutually complementary recognition elements (es-
sential for reversible, intermolecular assembly) and reactive func-
tionalities (essential for the irreversible construction of scaffolds).
These features are combined to engineer linear self-complemen-
tary platforms capable of autocatalytic self-propagation through
the mechanism illustrated in Figure 1. The rapid developments in
the synthesis and exploitation of mechanically interlocked mole-
cules15 in chemistry, such as catenanes and rotaxanes, have been
form a binary complex, [A$B]. The formation of this complex opens
a second reaction channeldthe binary or [A$B] complex channeld
in which A and B are preorganized12 with respect to each other
Crosscatalysis
Autocatalysis
T
R
[R•R]
[R•T]
Rotaxane
Rotaxane
Thread
Rotaxane
R
S
[L•M•S•T]
[L•S•R]
[L•M•S•R]
[L•M]
M
[L•S•T]
L
S
Thread
Rotaxane
T
Thread
Thread
[R•T]
[T•T]
R
T
Figure 3. The minimal kinetic model for a replicating rotaxane based on the threading and stoppering approach. Four interlinked catalytic cycles operate emanating from the
central binding event (boxed), the reversible formation of the [L$M] pseudorotaxane complex. Above the dotted horizontal line, rotaxane synthesis is accomplished through either
crosscatalysis (top left) by thread T or autocatalysis (top right) by rotaxane R. Below the dotted horizontal line, thread synthesis is accomplished through either crosscatalysis
(bottom left) by rotaxane R or autocatalysis (bottom right) by thread T. It is assumed that reactions through binary complexes can be removed by careful design and are not shown
for clarity.

8466
A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
is the only one that can be combined with the synthesis of a linear
template through replication. Replication requires the formation of
covalent bond at some point along the length of a rigid, linear
molecule. Of the three possible modes of rotaxane synthesis
available, only threading and stoppering has the same requirement
for covalent bond formation. In the other two cases, the linear
component is already intact at the time of rotaxane formation.
Recognising this requirement allows us to envisage the expanded
replication scheme depicted in Figure 3. The template T (which
forms the thread-like component of the rotaxane) has been ex-
tended compared to the template T shown in Figure 1, and now
incorporates an additional binding site (blue in Fig. 3).
The introduction of this additional binding site opens up a much
more complex network of catalytic pathways. This kinetic scheme
for the formation of replicating rotaxane actually contains four
interlinked catalytic cyclesdthe autocatalytic cycle in the lower
right of the kinetic scheme corresponds to the autocatalytic cycle
present in Figure 1. All of these four cycles emanate from the central
binding event (Fig. 3), the reversible formation of the [L$M] com-
plex, with pseudorotaxane geometry, through the association of
macrocycle M with the linear component L. Above the dotted
horizontal line, the [L$M] complex can react with the stoppering
reagent S to form rotaxane R (this process is analogous to the bi-
molecular channel in Fig. 1). Rotaxane R can then enter an auto-
catalytic cycle (top right, Fig. 3)dcollecting the [L$M] complex and
stoppering reagent S and catalysing the formation of a second
molecule of rotaxane R. Alternatively, rotaxane R can enter
a crosscatalytic cycle (lower left, Fig. 3)dcollecting the unbound
linear component L and stoppering reagent S and catalysing the
formation of a molecule of thread T. Conversely, below the dotted
horizontal line, unbound linear component L can react with the
stoppering reagent S to form thread T (once again, this process is
analogous to the bimolecular channel in Fig. 1). Thread T can then
enter a crosscatalytic cycle (top left, Fig. 3)dcollecting the [L$M]
complex and stoppering reagent S and catalysing the formation of
a molecule of rotaxane R. Alternatively, thread T can enter an au-
tocatalytic cycle (lower right, Fig. 3)dcollecting the unbound linear
component L and stoppering reagent S and catalysing the forma-
tion of a second molecule of thread T.
It is clear from this analysis that the position of the equilibrium
between the linear component L and macrocycle M and the
corresponding [L$M] complex is central to the success of this en-
terprise. Replication processes operate best at low concentrationsd
where the relative importance of the bimolecular reaction channel,
in a kinetic sense, is minimised. This objective clearly requires the
association constant Ka for the [L$M] complex to be high. Typically,
our successful replicating systems17 operate in a concentration
range from 10 to 25 mM. Using this knowledge, we can assess the
required target Ka for the additional binding site that will be used in
the L component (Fig. 3) to associate macrocycle M.
In order to make this assessment, we employed (Fig. 4a) a sim-
ple kinetic scheme in which rotaxane synthesis is viewed as two
parallel kinetic pathways operating either side of an equilibrium.
Assuming that L and [L$M] have the same reactivity, we can assess
the ratio of rotaxane to thread component in the reaction mixture
across a range of Ka values for the [L$M] complex through kinetic
simulation. Taking k1¼k2¼1.67ꢀ10ꢁ3 Mꢁ1 sꢁ1 and initial concen-
trations of L, S and M of 20 mM, Figure 4b illustrates how the
[rotaxane]/[thread] composition in the reaction mixture after 8 h
varies with Ka. We would view an acceptable level of selectivity for
rotaxane as being at least 10:1, therefore we had to target a Ka for
the [L$M] complex of >5000 Mꢁ1. Accordingly, we identified pyr-
idone as potential binding site, and designed and synthesised (see
Supplementary data) macrocycle 1 as a host to complex guests
(Fig. 5) incorporating this recognition motif. During the course of
this work, Chiu and co-workers independently reported18 the
Figure 4. (a) Simple kinetic model for rotaxane formation. Stoppering reagent S can
react with either unbound linear component L (k1) or the complex between L and
macrocycle M (k2). The [L$M] complex has a stability defined by Ka. (b) Ratio of ro-
taxane R/thread T as a function of Ka assuming k1¼k2¼1.67ꢀ10ꢁ3 Mꢁ1 sꢁ1
.
contemporaneous with the emergence of synthetic replicating
systems. This upsurge in interest16 in mechanically interlocked
molecules has been made possible by the exploitation of recogni-
tion-mediated processes for their synthesis. However, despite the
obvious advantages, the combination of the recognition-mediated
synthesis of a mechanically linked architecture and the amplifica-
tion of this structure by replicationdsimilar in spirit to the repli-
catable networks described by Drexlerdhas remained an
unexplored area. Here, we report our first attempts at integrating
the synthesis of a [2]rotaxane within a self-replicating framework.
2. Results and discussion
Conceptually, the integration of the synthesis of a [2]rotaxane
within a self-replicating system requires a number of design
choices to be made. Firstly, the method of rotaxane synthesis
(Fig. 2) must be chosen.
Consideration of how the processes highlighted in Figures 1 and
2 might be integrated leads rapidly to the realisation that the
threading and stoppering approach to rotaxane synthesis (Fig. 2b)

A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
8467
1
Figure 5. Partial H NMR spectra (400.1 MHz, 25 ꢂC, CDCl3, 25 mM) of (a) 4-pyridone 2, (b) an equimolar mixture of macrocycle 1 and 4-pyridone 2, (c) macrocycle 1 and (d) an
equimolar mixture of macrocycle 1 and extended pyridone 3. Dashed lines are shown to connect resonances for specific protons in the bound and unbound states. The 1H NMR
spectrum of 3 is not shown as a result of the limited solubility of this compound in CDCl3.
synthesis of this macrocycle and its incorporation in rotaxanes
through its complexation of ureas and carbamates.
phenylene protons (He and Hf, ꢁ0.31 and ꢁ0.35 ppm, respectively)
and that of the 4-pyridone (ꢁ1.03 and ꢁ0.98 ppm) are character-
istic of protons residing in the shielding zone of an aromatic ring.
Additionally, the signals of the macrocycle methylene protons are
shifted upfield (Hd and Hg, ꢁ0.05 and ꢁ0.29 ppm, respectively).
These chemical shift changes suggest that the pyridone guest is
located within the cavity of the macrocycle. From an 1H NMR ti-
tration experiment, in which successive amounts of pyridone 2
were added to a solution of macrocycle 1, we were able to estimate
The 1H NMR spectrum of an equimolar mixture of macrocycle 1
and 4-pyridone in CDCl3 (Fig. 5b) shows significant changes in the
chemical shifts of the resonances for the complex relative to those
of the free species. The resonance arising from the macrocycle NH
protons (Hc) is shifted downfield (þ1.90 ppm) as a result of the
hydrogen bonds between these protons and the carbonyl group of
4-pyridone. The upfield shifts of the resonances for the macrocycle
Figure 6. Partial 1H NMR spectra (400.1 MHz, 25 ꢂC, CDCl3, 25 mM) of (a) benzanilide, (b) an equimolar mixture of macrocycle 1 and benzanilide and (c) macrocycle 1. Dashed lines
are shown to connect resonances for specific protons in the bound and unbound states.

8468
A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
and the ether oxygen atoms. These changes are consistent with the
pseudorotaxane structure depicted in Figure 5, where the guest is
threaded through the cavity of the host. From a series of 1H NMR
dilution experiments, involving a 1:1 mixture of macrocycle 1 and
dialkynylpyridone 3, we were able to estimate the association
constant for the [1$3] complex as 1500ꢃ300 Mꢁ1 in CDCl3 at 25 ꢂC.
It is worth noting at this point that the complex between 1 and 3 is
in slow exchange on the chemical shift time scale at low concen-
trations, suggesting that some of the binding between 1 and 3 may
be constrictive19 in nature.
Although these results suggest that macrocycle 1 is a suitable
host for pyridones and that the complex between the macrocycle
and the extended guest 3 has the correct pseudorotaxane geome-
try, this recognition system has two fatal flaws. Firstly, the solubility
of extended 2,6-substituted pyridone systems, such as 3, is, at best,
limited in the non-polar solvents necessary for efficient molecular
recognition. Secondly, all of our synthetic efforts to incorporate this
binding site into the linear component L of our replicating rotaxane
design proved unsuccessful.
O
5
N
N
H
O
N
O
O
O
O
N
N
O
4
N
O
N
O
N
H
H
H
O
CHCl3
1
O
N
O
N
O
N
H
H
Therefore, the choice of binding site, which binds the macro-
cycle in building block L, had to be reassessed. Amides offered an
attractive alternative binding sitedtheir synthetic chemistry is
much better developed than that of 2,6-substituted pyridone sys-
tems and they appeared to be adequate guests for macrocycle 1. We
established that diarylamides20 are capable of associating with
macrocycle 1 with a pseudorotaxane-type geometry by performing
an additional binding experiment. The chemical shift changes ob-
served in the 1H NMR spectrum of an equimolar mixture of mac-
rocycle 1 and benzanilide in CDCl3 (Fig. 6b) are comparable to those
observed for 4-pyridone and demonstrate that amides are suitable
guests for macrocycle 1. The strength of the complex formed be-
tween macrocycle 1 and benzanilide was estimated by measuring
the association constant Ka by an 1H NMR titration experiment and
was found to be 160ꢃ20 Mꢁ1 in CDCl3 at 25 ꢂC. This value falls well
short of our ideal value set out previously. However, given the
synthetic tractability of amides when compared to 4-pyridones, we
were prepared to accommodate their sub-optimal binding con-
stants with macrocycle 1 in our revised designs.
In addition to the binding site, the design of our self-replicating
rotaxane requires two further features: two reactive sites (bright
red cartoons in Fig. 3) and an additional pair of complementary
recognition sites (green cartoons in Fig. 3). We chose to use the 1,3-
dipolar cycloaddition between a nitrone and a maleimide as the
covalent bond forming reaction between the L and S components,
and the association through two hydrogen bonds between an
amidopyridine and a carboxylic acid as the complementary rec-
ognition sites. These choices were guided by our previous success17
in exploiting these structural features in diastereoselective recog-
nition-mediated reactions21 and in the implementation of mini-
mal22 and reciprocal23 self-replicating systems. In order to ensure
that a simple rotaxane (i.e., one that is not capable of self-replica-
tion) can be formed using the cycloaddition reaction between
a nitrone and a maleimide, amide-containing maleimide 4 (Fig. 7),
which does not contain a recognition site on the stopper, macro-
cycle 1 and nitrone 5, were synthesised (see Supplementary data).
A mixture of maleimide 4 and macrocycle 1 was pre-equilibrated in
CDCl3 at ꢁ5 ꢂC to allow the formation of a pseudorotaxane. The
addition of nitrone 5 gave, after 3 days, rotaxane 6 in 45% yield, as
a 2:1 mixture of its trans and cis diastereoisomers, and thread 7 in
42% yield, as a 5:1 mixture of its trans and cis diastereoisomers. In
addition, 13% of unreacted maleimide 4 was recovered from the
reaction mixture. The 1H NMR spectrum of the cis diastereoisomer
of rotaxane 6 (Fig. 8) reveals the features expected of a [2]rotaxane.
Of particular note are the resonances arising from the macrocycle
NH protons and the adjacent methylene groups (Fig. 8). In the free
macrocycle, the resonance arising from the NH protons appears as
O
O
H
H
O
N
N
N
H
O
O
O
O
H
N
6
45%
N
O
O
N
H
H
O
N
N
N
H
O
O
H
7
42%
N
Figure 7. Design and synthesis of model rotaxane 6 and the corresponding thread 7
from macrocycle 1, maleimide 4 and nitrone 5. Both 6 and 7 are formed as mixtures of
their trans and cis diastereoisomers.
the association constant for the [1$2] complex as 500ꢃ100 Mꢁ1 in
CDCl3 at 25 ꢂC. Given the association between macrocycle 1 and 4-
pyridone, we were able to exploit 4-pyridone as a template during
the synthesis of macrocycle 1 itself: the yield of the cyclisation step
increased from 29% to 39% using 1 equiv of 4-pyridone as a tem-
plate (see Supplementary data).
We also prepared (see Supplementary data) an extended guest
molecule 3 containing a 4-pyridone moiety and assessed its asso-
ciation with macrocycle 1. The 1H NMR spectrum of an equimolar
mixture of macrocycle 1 and 3 in CDCl3 (Fig. 5d) displays dramatic
chemical shift changes, the origins of which are similar to those
observed in the case of the complexation of 4-pyridone by 1; the
resonance arising from Hc is shifted downfield (þ2.20 ppm) and
both of He and Hf (ꢁ0.29 ppm) along with those of Hd and Hg (ꢁ0.04
and ꢁ0.32 ppm, respectively) are shifted upfield. In addition, the
separation of the originally overlapping signals for the protons of
Hh and Hi suggests hydrogen bonding between the pyridone NH

A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
8469
1
Figure 8. H NMR spectrum (400.1 MHz, 25 ꢂC, CDCl3) of rotaxane cis-6 together with proton assignments. Resonances arising from the macrocyclic component are labelled with
small letters and those arising from the thread component are labelled with capital letters. The resonances arising from the c and d protons on the macrocycle are shown as inset
partial spectra. Note that the sample also contains around 8% of trans-6. Resonances arising from trans-6 are particularly evident in the pyridine CH3 and aryl C(CH3)3 regions of the
spectrum.
a triplet at
d 8.10 and that arising from the adjacent CH2 as a doublet
at 4.62. Upon rotaxane formation, the two methylene protons are
d
rendered magnetically inequivalent as a result of the end-to-end
asymmetry imposed by the thread. Further, the mirror plane
present in the free macrocycle, which passes through the macro-
cycle perpendicular to the pyridine ring, is absent in the rotaxane.
This loss is a result of the fact that the isoxazolidine ring formed in
the cycloaddition reaction does not possess a matching element of
symmetry. Therefore, the two NH protons and the four methylene
protons are now all magnetically inequivalent and this leads to the
observed complex signal patterns for the resonances arising from
these protons (inset, Fig. 8). In addition, these resonances also show
marked chemical shift changes as a result of the dramatic change in
their local environments upon the interlocking of the thread and
the macrocycle. The macrocycle amide NH proton resonances (c
O
O
N
N
O
N
H
H
O
O
H
H
O
N
N
N
N
H
O
O
O
9
O
H
O
O
H
N
protons, Fig. 8) shift downfield from
rocycle methylene proton resonances (d protons, Fig. 8) are now
located at 4.81 and 4.32. There is also a strong upfield shift of the
amide NH resonance of the macrocycle binding site (G proton,
Fig. 8) in the rotaxane ( 7.89) compared to the free thread ( 8.55).
d 8.10 to d 9.06 and the mac-
d
d
O
d
d
N
N
H
O
N
This observation is once again consistent with the macrocycle being
bound in the expected location in the rotaxane.
With these results in hand, we replaced one of the tert-butyl
groups in maleimide 4 with a carboxylic acid recognition site
affording the potential self-replicating rotaxane 9 (Fig. 9). In this
design, the amidopyridine was incorporated into the stopper of
compound 5 (building block S, Fig. 3), which also bears the nitrone
reactive site. Compound 8 (building block L, Fig. 3) was designed by
combining the maleimide, the amide-binding site and the carbox-
ylic acid on the stopper. Both ends of nitrone 5, the 4,6-dime-
thylpyridine ring and the tert-butyl phenyl ring are large enough to
prevent the macrocycle forming24 a pseudorotaxane with this
component. Maleimide 8 was synthesised (see Supplementary
data), but its extremely poor solubility in non-polar solvents forced
us once again to reconsider the design of the system.
5
O
O
N
N
O
N
O
H
H
O
O
N
1
N
H
O
8
O
O
HO
From our previous experience in designing recognition-medi-
ated reactions, we envisaged that a solution to the solubility issues
Figure 9. Retrosynthetic analysis of the design of replicating rotaxane 9 based on the
model rotaxane 6.

8470
A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
Figure 10. (a) Final design of a potential replicating rotaxane 13. Nitrone 11 can react with maleimide 10 to afford the thread 12. Alternatively, nitrone 11 can react with the
pseudorotaxane complex [1$10] to afford the rotaxane 13. The trans diastereoisomers of both thread 12 and rotaxane 13 are expected to be replicating. (b) Stick representation of the
calculated structure (OPLS2005/GB/SA CHCl3 solvation) of the product duplex of trans-13. Carbon atoms are coloured green in the linear components and grey in the macrocyclic
components of the [2]rotaxanes, oxygen atoms are coloured red, nitrogen atoms are coloured blue and hydrogen atoms are coloured white. Most hydrogen atoms have been
removed for clarity. Hydrogen bonds are shown as dashed lines.
was to reverse the location of the recognition sites on the L and S
components of the rotaxane design. Accordingly, we designed a new
system where the amidopyridine recognition site is situated on the
stopper of malemide 10 (building block L) (Fig. 10a) and the car-
boxylic acid recognition site on the stopper of nitrone 11 (building
block S) (Fig.10a). Once again, the substitution pattern on nitrone 11
was designed to prevent pseudorotaxane formation between the
nitrone and macrocycle 1. Overall, these changes leave us with
a design in which the trans diastereoisomer25 of both the thread and
the corresponding rotaxane is predicted computationally (Fig. 10b)
to provide the correct geometry to support replication. Gratifyingly,
all of the components of the potential replicator were sufficiently
soluble in CDCl3 to permit the study of their behaviour.
Reaction of nitrone 10 with maleimide 11 in CDCl3 at 25 ꢂC from
starting concentrations of the two reagents of 20 mM afforded the
thread 12 as a 7.9:1 trans/cis mixture of its diastereoisomers. After
8 h, the conversion was 49% (Fig. 11). When this reaction was re-
peated under the same conditions in the presence of 1 equiv
(20 mM) of macrocycle 1, the conversion after 8 h dropped (Fig. 11)
to only 33%. More worryingly, the ratio of rotaxane 13 to thread 12
was 0.45:1, i.e., there was more thread present than rotaxane. The
diastereoisomeric ratio of the thread in this reaction was now only
6:1 in favour of the trans stereoisomer. Interestingly, the ratio of
diastereoisomers observed for rotaxane 13 was 1.4:1 in favour of
the cis stereoisomer. We imagined that, given the sub-optimal
binding constant26 (w200 Mꢁ1) for the association of macrocycle 1
with maleimide 10, the poor performance of the system reflected
the fact that insufficient pseudorotaxane [1$10] was present in so-
lution. In order to address this problem, the reaction was repeated,
again under the same conditions, but now in the presence of 3 equiv
(60 mM) of macrocycle 1dthe conversion after 8 h (Fig.11) dropped
to only 26%. Gratifyingly, the ratio of rotaxane 13 to thread 12 was
now 2.6:1, i.e., there was now more rotaxane present than thread.
The diastereoisomeric ratio of the thread 12 in this reaction was still
only 6:1 in favour of the trans stereoisomer and, in this case, the
ratio of diastereoisomers observed for rotaxane 13 was 1.1:1 in fa-
vour of the cis stereoisomer. Although the addition of macrocycle
operated in the manner expecteddincreasing the rotaxane/thread
ratiodit also reduced the reactivity of the system markedly. Taken
together, the results of the kinetic experiments were disappointing.
The finally selected design targeted a trans selective replicator.
Although the formation of the thread appears to be somewhat
selective27 for the trans diastereoisomer as expected, this selectivity
is lost when macrocycle 1 is added to the reaction mixture. Indeed,
the formation of rotaxane 13 is slightly selective for the cis
diastereoisomer. Additionally, the overall conversion decreases
rapidly when the macrocycle concentration is increased implying
that the pseudorotaxane complex [1$10] and other, higher order,
complexes incorporating [1$10], for some reason, are somewhat
unreactive.

A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
8471
initially suspected that the low reactivity of the bound maleimide
could originate from the mode of binding of the macrocycle. If the
NH of the macrocycle formed hydrogen bonds with the carbonyl of
the maleimide (Fig. 13a) instead of the carbonyl of the amide
(Fig. 13b), the attack of the nitrone would be prevented through the
steric obstruction of the
p-face of the maleimide double bond.
However, analysis of the complex formed between macrocycle 1 and
model maleimide 14 using variable temperature 1H NMR spectros-
copy in CDCl3, together with a ROESY experiment28 performed at
ꢁ30 ꢂC, proved unambiguously that the macrocycle binds the am-
ide-binding site only (see Supplementary data), and, therefore, the
pseudorotaxane has the expected geometry shown in Figure 13b.
In order to investigate the reduced reactivity of the bound
maleimide further, we turned to electronic structure calculations.
Given the large size of the system, we were forced to use semi-
empirical methods for these calculations. We wished to use a semi-
empirical method that was capable of locating a high quality
transition state structure for the dipolar cycloaddition. Hence, we
located the transition state for the dipolar cycloaddition between
N-phenylmaleimide and diphenylnitrone using density functional
methods (B3LYP/6-31Gþ(d,p)). We then screened the available
semi-empirical methodsdAM1,29 PM3,30 RM131 and PM632
d
against this transition state structure. The RM1 method gave the
closest match to the DFT transition state structure and we therefore
chose to proceed with the calculations involving the macrocycle
using this method. We calculated the transition states accessed by
the dipolar cycloaddition reaction between a model maleimide 15
and a model nitrone 16 using RM1 in the presence and in the ab-
sence of macrocycle 1. The transition state structures located are
shown in Figure 14. It is immediately apparent from the calculated
structures that the association of macrocycle 1 with the amide-
binding site exerts a remote steric effect33 on the transition state.
The tert-butyl group of the nitrone is located close to the bottom
diethylene glycol loop of the macrocycle (Fig. 14, region X)din fact,
in region X (Fig. 14), part of the diethylene glycol loop and the tert-
butyl group are essentially in van der Waals contact. This in-
teraction is undoubtedly somewhat destabilising34 to the transition
state in the reaction between [1$10] and 11 when compared to the
reaction between 10 and 11 in isolation. This observation also helps
to explain the change observed in the diastereoselectivity of the
cycloaddition between thread 12 (6:1 trans/cis) and rotaxane 13
(1:1.4 trans/cis). In rotaxane 13, the cis diastereoisomer is slightly
favoured, probably reflecting the fact that the tert-butyl group on
nitrone 11 interacts more unfavourably with the macrocycle in the
trans transition state compared with either the carboxymethylene
or the propyloxy groups of 11 in the cis transition state.
3. Conclusions
Figure 11. (a) Concentration of products (thread 12¼green columns, rotaxane 13¼blue
columns) as a function of added macrocycle 1 for the kinetic experiments performed
with macrocycle 1. (b) Total conversion ([12]þ[13] expressed as a percentage, blue
points) and the ratio of rotaxane 13/thread 12 (red points) as a function of the amount
of added macrocycle 1.
We have presented a kinetic model for the integration of self-
replication with the formation of a mechanically interlocked mo-
lecular architecture, namely a rotaxane. Ultimately, despite careful
design, the system synthesised did not function as expected. This
failure is a consequence of deficiencies in two important design
criteriadone of which we recognised and the other was un-
anticipated. As a result of solubility and synthetic difficulty, we
were forced to compromise with respect to the macrocycle binding
In order to understand the results of these kinetic experiments,
we returned to the simplified kinetic model (Fig. 4) for rotaxane
formation presented earlier. We used this model to predict (Fig. 12)
the behaviour of the system with different loadings of macrocycle 1
present. We used a value for k1, as before, of 1.67ꢀ10ꢁ3 Mꢁ1 sꢁ1 and
set the Ka for the pseudorotaxane complex as 200 Mꢁ1. The rate
constant for rotaxane formation was then fitted to obtain the best
match with the selected metrics shown in Figure 12. Gratifyingly,
we were able to reproduce the experimental data for the selected
sitedsubstituting
a ) with an amide
pyridone (Kaw1500 Mꢁ1
(Kaw150 Mꢁ1). This compromise is clearly sub-optimal, since the
association constant for the pseudorotaxane complex ([L$M], Fig. 3)
must be as high as possible. This requirement is essential since the
formation of this complex shuts down all kinetic pathways that
involve free L and, hence, lead to the thread T. The depletion of free
L therefore concentrates the kinetic flux in the autocatalytic cycle
involving the rotaxane R (top right, Fig. 3). Secondly, the reactive
site on L must be placed sufficiently far away from the binding site
metrics using a value of k2¼4.61ꢀ10ꢁ4 Mꢁ1 sꢁ1
.
The analysis presented above suggests that maleimide 10 be-
comes around four times less reactive within the [1$10] complex. We

8472
A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
Figure 12. Comparison of experimental (blue) and calculated (green) metrics for the kinetic experiments performed using macrocycle 1, maleimide 10 and nitrone 11. The chosen
metrics are (a) total conversion to cycloadduct expressed as a percentage, (b) ratio of rotaxane 13/thread 12, (c) concentration of thread 12 after 8 h and (d) concentration of
rotaxane 13 after 8 h. Calculations used the kinetic scheme shown in Figure 4 with values of Ka¼200 Mꢁ1 and k1¼1.67ꢀ10ꢁ3 Mꢁ1 sꢁ1. The best fit of all four metrics at the three
concentrations of added macrocycle was obtained for k2¼4.61ꢀ10ꢁ4 Mꢁ1 sꢁ1
.
to prevent the introduction of a supramolecular steric effect
through the proximity of the macrocyclic component M. We are
currently addressing these issues through improved designs using
aggressive optimisation of the computational approaches described
here. In addition, there are a number of other potential methods of
integrating replication processes with mechanically interlocked
molecular architectures and these are also currently under in-
vestigation in our laboratory.
presence of sodium/benzophenone under a N2 atmosphere and
was collected by distillation. Acetonitrile (CH3CN) and dichloro-
methane (CH2Cl2) were dried by heating under reflux over calcium
hydride and distilled under N2.
Thin-layer chromatography (TLC) was performed on aluminium
plates coated with Merck Kieselgel 60 F254. Developed plates were
air-dried and scrutinised under an UV lamp (366 nm), and where
necessary, stained with ninhydrin or potassium permanganate to
aid visualisation. Column chromatography was performed using
Kieselgel 60 (0.040–0.063 mm mesh, Merck 9385) or MP Silica
(silica gel, 0.032–0.063 mesh).
4. Experimental
Melting points were determined using an Electrothermal 9200
melting point apparatus and are reported uncorrected. Micro-
analyses (C, H, N) were carried out at the University of St. Andrews.
1H Nuclear Magnetic Resonance (NMR) spectra were recorded
4.1. General procedures
Chemicals and solvents were purchased from standard com-
mercial suppliers and were used as-received unless otherwise
stated. Tetrahydrofuran (THF) was dried by heating to reflux in the
on
a Bruker Avance II 400 (400.1 MHz), a Bruker Avance

A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
8473
O
O
N
N
H
14
O
(a)
(b)
Figure 13. Stick representations of the calculated structures (OPLS2005/GB/SA CHCl3
solvation) of the pseudorotaxane complexes formed between macrocycle and
1
maleimide 14. (a) Macrocycle 1 amide NH groups hydrogen bond to a maleimide ring
C]O oxygen and the maleimide amide NH hydrogen bonds to an ether oxygen atom
on macrocycle 1. (b) Macrocycle 1 amide NH groups hydrogen bond to maleimide
amide C]O oxygen atom. In this geometry, the maleimide amide NH does not interact
with the ether oxygens on macrocycle 1. Carbon atoms are coloured green, oxygen
atoms are coloured red, nitrogen atoms are coloured blue and hydrogen atoms are
coloured white. Most hydrogen atoms have been removed for clarity. Hydrogen bonds
are shown as dashed lines.
(300.1 MHz), a Bruker Avance (500.0 MHz) or a Varian UNITYplus
500 (500.1 MHz) spectrometer using deuterated solvent as the lock
and the residual solvent as the internal reference in all cases. 13C
NMR spectra were recorded on a Bruker Avance 300 (75.5 MHz) or
a Bruker Avance II 400 (100.6 MHz) spectrometer using the PEN-
DANTor DEPTQ pulse sequences. All spectra were recorded at 298 K
unless otherwise stated. All coupling constants (J) are quoted to the
nearest 0.1 Hz. In the assignment of 1H NMR spectra the symbols br,
s, d, t, q and m denote broad, singlet, doublet, triplet, quartet and
multiplet, respectively.
Electron impact mass spectrometry (EIMS) and high-resolution
mass spectrometry (HRMS) were carried out on a Micromass GCT
orthogonal acceleration time of flight mass spectrometer. Chemical
Ionisation Mass Spectrometry (CIMS) was carried out on a Micro-
mass GCT orthogonal acceleration time of flight mass spectrometer.
Electrospray mass spectrometry (ESMS) and high-resolution mass
spectrometry (HRMS) were carried out on a Micromass LCT or-
thogonal time of flight mass spectrometer.
Figure 14. Stick representations of the calculated (RM1 semi-empirical method)
transition state structures leading to the trans diastereoisomer for (a) the reaction
between model maleimide 15 and model nitrone 16 and (b) the reaction between
model nitrone 16 and the complex between model maleimide 15 and macrocycle 1. In
region X, the tert-butyl group of 16 is essentially in van der Waals contact with the
diethylene glycol loop in macrocycle 1, as illustrated in the space filling model (c).
Carbon atoms are coloured green, oxygen atoms are coloured red, nitrogen atoms are
coloured blue and hydrogen atoms are coloured white. Most hydrogen atoms have
been removed for clarity. Hydrogen bonds are shown as dashed lines and forming
bonds as dotted lines.
a concentration of 600 mM. The extent of reaction for kinetic ex-
periments was determined using the deconvolution of the appro-
priate product resonances using the tools available in Topspin
(Version 2.0 pl 3, Bruker Biospin, Germany, 2006).
4.2. General procedure for kinetic experiments
Masses of reagents were measured using a Sartorius BP 211D
balance (ꢃ0.01 mg) and reagent solutions prepared by dissolving
compounds in the appropriate volume of CDCl3. Subsequent ex-
perimental samples, suitable for kinetic experiments, were pre-
pared by using Hamilton gas-tight syringes to transfer solution to
either a Wilmad 507PP or a 528PP NMR tube, which was then fitted
with a polyethylene pressure cap to minimise solvent evaporation.
Reaction mixtures were monitored systematically by 500 MHz 1H
NMR spectroscopy over 18 h. Concentrations were calibrated using
an internal standard ((Me3Si)4Si) added to the reaction mixtures at
4.3. Computational methods
All molecular mechanics calculations were performed on
a Linux workstation using the OPLS2005 forcefield together with
the GB/SA solvation model for chloroform as implemented in
Macromodel (Version 9.5, Schro¨dinger Inc., 2007). Electronic
structure calculations were carried out using GAMESS-US35 or
MOPAC200736 running on a Linux cluster. The 64-bit Linux ver-
sion dated 24 Mar 2007 (Revision 3) was used in all density
functional calculations. The transition state for the reaction

8474
A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
between N-phenylmaleimide and diphenylnitrone was located by
generation of an initial guess using the linear synchronous transit
(LST) method and then refinement at the HF/6-31G(d) level of
theory within GAMESS. This model transition state was then used
to construct an initial guess for the transition state leading to the
appropriate cycloadduct. This guess was refined at the B3LYP/
6-31Gþ(d,p) level of theory to a transition state structure pos-
sessing single imaginary vibration that corresponded to the re-
action coordinate. This transition state was then used for
comparison purposes for those calculated using the semi-empir-
ical methods AM1, PM3, RM1 and PM6. All semi-empirical elec-
tronic structure calculations were carried out using MOPAC2007.
Version 8.032L was used in all calculations. In all cases, Eigen
vector following was used to locate a transition state structure
possessing single imaginary vibration that corresponded to the
reaction coordinate.
Acknowledgements
The financial support provided by the University of St. Andrews
and EaStCHEM (graduate studentship to A.V.) is gratefully ac-
knowledged. We thank Tomas Lebl and Melanja Smith for their
assistance with NMR and Caroline Horsburgh for mass
spectrometry.
Supplementary data
Supplementary data associated with this article can be found in
the online version, at doi:10.1016/j.tet.2008.05.049.
References and notes
1. Ruben, M.; Rojo, J.; Romero-Salguero, F. J.; Uppadine, L. H.; Lehn, J.-M. Angew.
Chem., Int. Ed. 2004, 43, 3644–3662.
2. Lukin, O.; Vo¨gtle, F. Angew. Chem., Int. Ed. 2005, 44, 1456–1477.
3. (a) Sauvage, J.-P. Acc. Chem. Res. 1998, 31, 611–619; (b) Balzani, V.; Credi, A.;
Raymo, F. M.; Stoddart, J. F. Angew. Chem., Int. Ed. 2000, 39, 3348–3391; (c)
Schalley, C. A.; Beizai, K.; Vo¨gtle, F. Acc. Chem. Res. 2001, 34, 465–476; (d)
Feringa, B. L. Molecular Switches; Wiley-VCH: Weinheim, 2001; (e) Dietrich-
Buchecker, C.; Jimenez-Molero, M. C.; Sartor, V.; Sauvage, J.-P. Pure Appl. Chem.
2003, 75, 1383–1393; (f) Flood, A. H.; Ramirez, R. J. A.; Deng, W.-Q.; Muller, R. P.;
Goddard, W. A., III; Stoddart, J. F. Aust. J. Chem. 2004, 57, 301–322; (g) Tian, H.;
Wang, Q.-C. Chem. Soc. Rev. 2006, 35, 361–374; (h) Balzani, V.; Credi, A.; Silvi, S.;
Venturi, M. Chem. Soc. Rev. 2006, 35, 1135–1149; (i) Credi, A. J. Phys.: Condens.
Matter 2006, 18, S1779–S1795; (j) Kay, E. R.; Leigh, D. A.; Zerbetto, F. Angew.
Chem., Int. Ed. 2007, 46, 72–191; (k) Green, J. E.; Choi, J. W.; Boukai, A.;
Bunimovich, Y.; Johnston-Halperin, E.; Delonno, E.; Luo, Y.; Sheriff, B. A.; Xu, K.;
Shin, Y. S.; Tseng, H.-R.; Stoddart, J. F.; Heath, J. R. Nature 2007, 445, 414–417.
4. Chworos, A.; Severcan, I.; Koyfman, A. Y.; Weinkam, P.; Oroudjev, E.; Hansma,
H. G.; Jaeger, L. Science 2004, 306, 2068–2072.
5. (a) Carbone, A.; Seeman, N. C. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 12577–12582;
(b) Seeman, N. C.; Belcher, A. M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 6451–6455;
(c) Seeman, N. C. Nature 2003, 421, 427–431; (d) Seeman, N. C. Chem. Biol. 2003,
10, 1151–1159; (e) Liao, S.; Seeman, N. C. Science 2004, 306, 2072–2074; (f) Yan,
H. Science 2004, 306, 2048–2049; (g) Samori, B.; Zuccheri, G. Angew. Chem., Int.
Ed. 2005, 44, 1166–1181; (h) Gothelf, K. V.; Brown, R. S. Chem.dEur. J. 2005, 11,
1062–1069.
4.4. Synthetic procedures
Preparative methods and characterisation data for compounds
1, 3, 4, 5, 8, 10 and 11 and the intermediates leading to them are
provided in Supplementary data.
4.4.1. Rotaxane 6
A mixture of macrocycle 1 (110 mg, 0.23 mmol), maleimide 4
(140 mg, 0.23 mmol) and chloroform (4.6 mL) was stirred at ꢁ5 ꢂC
for 1 h. Nitrone 5 (93 mg, 0.23 mmol) was then added to the re-
action mixture at ꢁ5 ꢂC. The reaction mixture was stirred at room
temperature for 3 days. The solvent was evaporated under reduced
pressure. The different products of the reaction were isolated by
column chromatography (SiO2, gradient from 9:1 to 3:7 cyclohex-
ane/ethyl acetate) that afforded thread 7 (97 mg, 42%, trans/cis 5:1)
as a pale brown solid and rotaxane 6 (153 mg, 45%, trans/cis 2:1) as
a yellow solid. Representative data for rotaxane (cis-6). 1H NMR
6. (a) Lloyd Carroll, R.; Gorman, C. B. Angew. Chem., Int. Ed. 2002, 41, 4378–4400;
(b) Becerril, H. A.; Stoltenberg, R. M.; Wheeler, D. R.; Davis, R. C.; Harb, J. N.;
Wooley, A. T. J. Am. Chem. Soc. 2005, 127, 2828–2829; (c) Niemeyer, C. M. Science
2002, 297, 62–63.
7. Drexler, K. E. Engines of Creation: The Coming Era of Nanotechnology; Anchor:
Doubleday, NY, 1986.
(400.1 MHz, CDCl3):
d¼9.08–9.04 (dt, J¼7.5, 3.8 Hz, 2H), 8.81 (br s,
1H), 8.40–8.37 (m, 2H), 8.05 (d, J¼8.4 Hz, 2H), 8.02 (s, 1H), 7.89 (s,
1H), 7.77 (t, J¼7.8 Hz, 1H), 7.65–7.63 (m, 1Hþ2H), 7.56 (d, J¼8.9 Hz,
2H), 7.53 (d, J¼1.4 Hz, 2H), 7.45–7.40 (m, 8H), 7.29 (d, J¼8.8 Hz, 2H),
7.08 (d, J¼8.8 Hz, 2H), 6.98 (d, J¼9.0 Hz, 2H), 6.74–6.72 (m, 1Hþ4H),
6.61 (d, J¼7.9 Hz, 4H), 5.31 (d, J¼7.8 Hz, 1H), 4.98 (d, J¼8.9 Hz, 1H),
4.81 (dt, J¼15.0, 7.6 Hz, 2H), 4.36–4.28 (m, 2H), 4.17–4.13 (m,
1Hþ4H), 3.94–3.90 (m, 2H), 3.76–3.72 (m, 2H), 3.54–3.51 (m, 4H),
2.38 (s, 3H), 2.33 (s, 3H), 1.36 (s, 18H), 1.27 (s, 9H); 13C NMR
8. (a) Smalley, R. E. Sci. Am. 2001, 285, 76–77; (b) Whitesides, G. M. Sci. Am. 2001,
285, 78–83.
9. For reviews, see: (a) Orgel, L. E. Nature 1992, 358, 203–209; (b) Famulok, M.;
Nowick, J. S.; Rebek, J., Jr. Acta Chem. Scand. 1992, 46, 315–324; (c) Wintner,
E. A.; Conn, M. M.; Rebek, J., Jr. Acc. Chem. Res. 1994, 27, 198–203; (d) Wintner,
E. A.; Rebek, J., Jr. Acta Chem. Scand. 1996, 50, 469–485; (e) Bag, B. G.; von
Kiedrowski, G. Pure Appl. Chem. 1996, 68, 2145–2152; (f) Lee, D. H.; Severin,
K.; Ghadiri, M. R. Curr. Opin. Chem. Biol. 1997, 1, 491–496; (g) Robertson, A.;
Sinclair, A. J.; Philp, D. Chem. Soc. Rev. 2000, 29, 141–152; (h) Li, X.; Chmie-
lewski, J. Org. Biomol. Chem. 2003, 1, 901–904; (i) Paul, N.; Joyce, G. F. Curr.
Opin. Chem. Biol. 2004, 8, 634–639; (j) Patzke, V.; von Kiedrowski, G. Arkivoc
2007, 293–310.
10. Some recent examples include: (a) Kindermann, M.; Stahl, I.; Reimold, M.;
Pankau, W. M.; von Kiedrowski, G. Angew. Chem., Int. Ed. 2005, 44, 6750–6755;
(b) Isaac, R.; Chmielewski, J. J. Am. Chem. Soc. 2002, 124, 6808–6809; (c) Li, X.;
Chmielewski, J. J. Am. Chem. Soc. 2003, 125, 11820–11821; (d) Matsumura, S.;
Takahashi, T.; Ueno, A.; Mihara, H. Chem.dEur. J. 2003, 9, 4829–4837; (e) Lee,
D. H.; Granja, J. R.; Martinez, J. A.; Severin, K.; Ghadiri, M. R. Nature 1996, 382,
525–528.
(100.6 MHz, CDCl3):
d¼173.73,171.47,164.95,163.86,163.29,156.32,
152.18, 150.74, 149.24, 148.62, 144.04, 139.06, 138.96, 138.69, 137.22,
137.18, 136.59, 135.66, 134.81, 134.08, 131.52, 130.67, 128.83, 128.25,
128.11, 127.96, 126.30, 125.91, 125.70, 125.57, 125.43, 123.17, 120.83,
120.08,119.79,119.54,111.88, 90.54, 87.38, 76.82, 73.53, 71.13, 70.82,
68.89, 54.76, 43.53, 35.00, 34.50, 31.39, 31.32, 23.74, 21.47; MS (ES):
m/z (%)¼1482 (100) [M]þ, 1483 (95) [MþH]þ; HRMS (MALDI): m/z
calcd for [MþH]þ C93H93N8O10 1481.7015, found 1481.7056. Rep-
resentative data for thread (trans-7). 1H NMR (400.1 MHz, CDCl3):
d
¼8.59 (s, 1H), 8.55 (s, 1H), 8.02 (s, 1H), 7.91–7.89 (m, 2Hþ2H), 7.76
11. For comprehensive kinetic analyses of minimal systems, see: (a) von
Kiedrowski, G. Bioorganic Chemistry Frontiers; Springer: Berlin, Heidelberg,
1993; Vol. 3, 113–146; (b) Reinhoudt, D. N.; Rudkevich, D. M.; de Jong, F. J. Am.
Chem. Soc. 1996, 118, 6880–6889.
12. (a) Cram, D. J. Angew. Chem., Int. Ed. Engl. 1988, 27, 1009–1020; (b) Constable,
E. C. Comprehensive Supramolecular Chemistry; Atwood, J. L., Davis, J. E. D.,
(t, J¼1.5 Hz, 1H), 7.64 (d, J¼8.4 Hz, 2H), 7.60 (d, J¼8.9 Hz, 2H), 7.43
(d, J¼8.6 Hz, 4H), 7.37 (d, J¼8.6 Hz, 4H), 7.25 (d, J¼8.9 Hz, 2H), 7.08
(d, J¼8.9 Hz, 2H), 6.76 (s, 1H), 6.45 (d, J¼8.9 Hz, 2H), 5.82 (s, 1H),
5.04 (d, J¼7.5 Hz, 1H), 3.97 (d, J¼7.9 Hz, 1H), 2.41 (s, 3H), 2.33 (s,
3H), 1.32 (s, 18H), 1.27 (s, 9H); 13C NMR (100.6 MHz, CDCl3):
Macnicol, D. D., Vogtle, F., Eds.; Elsevier: New York, NY, 1996; p 218; (c)
Rowan, S. J.; Hamilton, D. G.; Brady, P. A.; Sanders, J. K. M. J. Am. Chem. Soc.
1997, 119, 2578–2579; (d) Rowan, S. J.; Sanders, J. K. M. J. Org. Chem. 1998, 63,
1536–1546.
¨
d
¼174.25, 172.84, 165.14, 164.66, 156.50, 152.27, 150.72, 150.44,
146.39, 146.20, 143.03, 138.74, 137.22, 135.33, 134.11, 131.58, 129.68,
127.92, 127.07, 126.99, 126.77, 126.47, 125.58, 124.64, 120.91, 120.61,
119.54, 114.12, 111.82, 91.54, 87.16, 77.48, 69.22, 57.33, 34.96, 34.31,
31.54, 31.25, 23.87, 21.43; MS (ES): m/z (%)¼402 (100), 1007 (30)
[MþH]þ; HRMS (ES): m/z calcd for [MþH]þ C66H64N5O5 1006.4907,
found 1006.4901.
13. See also: Soai, K.; Shibata, T.; Sato, I. Acc. Chem. Res. 2000, 33, 382–390 and
references cited therein Blackmond, D. G. Asymm. Synth. 2007, 181–185.
14. (a) Allen, V. C.; Philp, D.; Spencer, N. Org. Lett. 2001, 3, 777–780; (b) Quayle,
J. M.; Slawin, A. M. Z.; Philp, D. Tetrahedron Lett. 2002, 43, 7229–7233.
15. For reviews, see: (a) Dietrich-Buchecker, C. O.; Sauvage, J.-P. Chem. Rev. 1987,
87, 795–810; (b) Sauvage, J.-P. Acc. Chem. Res. 1990, 23, 319–327; (c) Amabilino,

A. Vidonne, D. Philp / Tetrahedron 64 (2008) 8464–8475
8475
D. B.; Stoddart, J. F. Chem. Rev. 1995, 95, 2725–2828; (d) Vo¨gtle, F.; Du¨nnwald,
T.; Schmidt, T. Acc. Chem. Res. 1996, 29, 451–460; (e) Breault, G. A.; Hunter, C.
A.; Mayers, P. C. Tetrahedron 1999, 55, 5265–5293; (f) Fujita, M. Acc. Chem. Res.
1999, 32, 53–61; (g) Hubin, T. J.; Busch, D. H. Coord. Chem. Rev. 2000, 200–202,
5–52; (h) Schalley, C. A.; Weilandt, T.; Bru¨ ggemann, J.; Vo¨gtle, F. Top. Curr.
Chem. 2004, 248, 141–200; (i) Arico´ , F.; Badjic, J. D.; Cantrill, S. J.; Flood, A. H.;
Leung, K. C.-F.; Liu, Y.; Stoddart, J. F. Top. Curr. Chem. 2005, 249, 203–259; (j)
Vickers, M. S.; Beer, P. D. Chem. Soc. Rev. 2007, 36, 211–225; (k) Loeb, S. J. Chem.
Soc. Rev. 2007, 36, 226–235; (l) Kay, E. R.; Leigh, D. A. Pure Appl. Chem. 2008, 80,
17–29.
Safarowsky, O.; Nieger, M.; Vo¨gtle, F. Liebigs Ann. 1997, 2269–2273; (ee) Ha¨ndel,
M.; Plevoets, M.; Gestermann, S.; Vo¨gtle, F. Angew. Chem., Int. Ed. 1997, 36, 1199–
1201; (ff) Leigh, D. A.; Murphy, A.; Smart, J. P.; Slawin, A. M. Z. Angew. Chem., Int. Ed.
1997, 36, 728–732; (gg) Lane, A. S.; Leigh, D. A.; Murphy, A. J. Am. Chem. Soc. 1997,
119, 11092–11093; (hh) Johnston, A. G.; Leigh, D. A.; Murphy, A.; Smart, J. P.;
Deegan, M. D. J. Am. Chem. Soc. 1996, 118, 10662–10663; (ii) Vo¨gtle, F.; Ja¨ger, R.;
Ha¨ndel, M.; Ottens-Hildebrandt, S.; Schmidt, W. Synthesis 1996, 353–356; (jj)
Vo¨gtle, F.; Ahuis, F.; Baumann, S.; Sessler, J. L. Liebigs Ann. 1996, 921–926; (kk)
Vo¨gtle, F.; Ha¨ndel, M.; Meier, S.; Ottens-Hildebrandt, S.; Ott, F.; Schmidt, T. Liebigs
Ann. 1995, 739–743.
16. Until the advent of synthetic strategies that exploited molecular recognition for
the construction of mechanically interlocked architectures, the study of such
systems was confined to more theoretical aspects of such assembly. See: Walba,
D. M. Tetrahedron 1985, 41, 3161–3212.
17. Kassianidis, E.; Philp, D. Angew. Chem., Int. Ed. 2006, 45, 6334–6348.
18. (a) Huang, Y.-L.; Hung, W.-C.; Lai, C.-C.; Liu, Y.-H.; Peng, S.-M.; Chiu, S.-H. Angew.
Chem., Int. Ed. 2007, 46, 6629–6633. The syntheses of similar macrocycles have
been reported: (b) Furusho, Y.; Matsuyama, T.; Takata, T.; Moriuchi, T.; Hirao, T.
Tetrahedron Lett. 2004, 45, 9593–9597; (c) Leigh, D. A.; Thomson, A. R. Org. Lett.
2006, 8, 5377–5379.
21. (a) Philp, D.; Robertson, A. Chem. Commun. 1998, 879–880; (b) Bennes, R.; Philp,
D.; Spencer, N.; Kariuki, B. M.; Harris, K. D. M. Org. Lett. 1999, 1, 1087–1090; (c)
Robertson, A.; Philp, D.; Spencer, N. Tetrahedron 1999, 55, 11365–11384; (d)
Bennes, R.; Babiloni, M. S.; Hayes, W.; Philp, D. Tetrahedron Lett. 2001, 42, 2377–
2380; (e) Kassianidis, E.; Pearson, R. J.; Philp, D. Chem.dEur. J. 2006, 12, 8798–
8812; (f) Bennes, R.; Philp, D. Org. Lett. 2006, 8, 3651–3654.
22. (a) Kassianidis, E.; Pearson, R. J.; Philp, D. Org. Lett. 2005, 7, 3833–3836; (b)
Pearson, R. J.; Kassianidis, E.; Philp, D. Tetrahedron Lett. 2004, 45, 4777–4780; (c)
Pearson, R. J.; Kassianidis, E.; Slawin, A. M. Z.; Philp, D. Chem.dEur. J. 2006, 12,
6829–6840.
19. The [1$3] complex has a larger association constant than the [1$2] complex
despite the fact that both 2 and 3 having the same binding site. The slow ex-
change on the chemical shift time scale exhibited by the [1$3] complex sug-
gests that the on and off rates for this complex are significantly lower than
those for the [1$2] complex. This change in the on and off rates probably arises
from a steric barrier imposed by a close complementarity between the mean
free path through the cavity of macrocycle 1 and the phenylethynyl sub-
stituents present in 3, rather than a greatly altered intrinsic binding of the
pyridone ring to the macrocycle. The small size of macrocycle 1 is evident in
other situations in this work (cf. complex [1$10]). For a discussion of slippage
and constrictive binding, see: Fyfe, M. C. T.; Raymo, F. M.; Stoddart, J. F. Stim-
ulating Concepts in Chemistry; Vo¨gtle, F., Stoddart, J. F., Shibasaki, M., Eds.; VCH:
Weinheim, 2000; pp 211–220.
23. Kassianidis, E.; Philp, D. Chem. Commun. 2006, 4072–4074.
24. During the course of this work, we also discovered that nitrones can bind
macrocycle 1. Therefore, we designed nitrone
5 in order to prevent
pseudorotaxane formation between macrocycle 1 and the nitrone.
25. The stereochemical labels trans and cis refer to the relative stereochemistry
in the cycloadduct of the proton derived from the nitrone with respect to the
ring junction protons (derived from the maleimide). In the trans di-
astereoisomer, the proton derived from the nitrone is on the opposite face of
the fused ring system as the ring junction protons derived from the maleimide.
In the cis diastereoisomer, the proton derived from the nitrone is on the same
face of the fused ring system as the ring junction protons derived from the
maleimide.
26. The association constant for the [1$10] complex was estimated at 230 Mꢁ1 from
the 400 MHz 1H NMR spectrum recorded at room temperature in CDCl3. Under
these conditions, exchange between bound and unbound species is slow on the
chemical shift time scale and, therefore, a crude estimate of the association
constant can be obtained simply by integrating resonances arising from the
two species.
20. Several groups have exploited the use of amides for the construction of ro-
taxanes over the past 15 years. For some examples, see: (a) Alvarez-Perez, M.;
Goldup, S. M.; Leigh, D. A.; Slawin, A. M. Z. J. Am. Chem. Soc. 2008, 130, 1836–
1838; (b) Fioravanti, G.; Haraszkiewicz, N.; Kay, E. R.; Mendoza, S. M.; Bruno, C.;
Marcaccio, M.; Wiering, P. G.; Paolucci, F.; Rudolf, P.; Brouwer, A. M.; Leigh, D. A.
J. Am. Chem. Soc. 2008, 130, 2593–2601; (c) Serreli, V.; Lee, C.-F.; Kay, E. R.;
Leigh, D. A. Nature 2007, 445, 523–527; (d) Berna, J.; Brouwer, A. M.; Fazio,
S. M.; Haraszkiewicz, N.; Leigh, D. A.; Lennon, C. M. Chem. Commun. 2007,
1910–1912; (e) Chatterjee, M. N.; Kay, E. R.; Leigh, D. A. J. Am. Chem. Soc.
2006, 128, 4058–4073; (f) See Ref. 18c. (g) Kishan, M. R.; Parham, A.;
Schelhase, F.; Yoneva, A.; Silva, G.; Chen, X.; Okamoto, Y.; Vo¨gtle, F. Angew.
Chem., Int. Ed. 2006, 45, 7296–7299; (h) Aucagne, V.; Leigh, D. A.; Lock, J. S.;
Thomson, A. R. J. Am. Chem. Soc. 2006, 128, 1784–1785; (i) Berna, J.; Leigh,
D. A.; Lubomska, M.; Mendoza, S. M.; Perez, E. M.; Rudolf, P.; Teobaldi, G.;
Zerbetto, F. Nat. Mater. 2005, 4, 704–710; (j) Leigh, D. A.; Venturini, A.;
Wilson, A. J.; Wong, J. K. Y.; Zerbetto, F. Chem.dEur. J. 2004, 10, 4960–4969;
(k) Keaveney, C. M.; Leigh, D. A. Angew. Chem., Int. Ed. 2004, 43, 1222–1224;
(l) Kidd, T. J.; Loontjens, T. J. A.; Leigh, D. A.; Wong, J. K. Y. Angew. Chem., Int.
Ed. 2003, 42, 3379–3383; (m) Lukin, O.; Kubota, T.; Okamoto, Y.; Schelhase,
F.; Yoneva, A.; Muller, W. M.; Muller, U.; Vo¨gtle, F. Angew. Chem., Int. Ed.
2003, 42, 4542–4545; (n) Gatti, F. G.; Lent, S.; Wong, J. K. Y.; Bottari, G.;
Altieri, A.; Morales, M. A. F.; Teat, S. J.; Frochot, C.; Leigh, D. A.; Brouwer, A. M.;
Zerbetto, F. Proc. Natl. Acad. Sci. U.S.A. 2003,100,10–14; (o) Hannam, J. S.;Kidd, T. J.;
Leigh, D. A.; Wilson, A. J. Org. Lett. 2003, 5, 1907–1910; (p) Watanabe, N.; Kihara,
N.; Furusho, Y.; Takata, T.; Araki, Y.; Ito, O. Angew. Chem., Int. Ed. 2003, 42, 681–
683; (q) Bottari, G.; Dehez, F.; Leigh, D. A.; Nash, P. J.; Perez, E. M.; Wong, J. K. Y.;
Zerbetto, F. Angew. Chem., Int. Ed. 2003, 42, 5886–5889; (r) Da Ros, T.; Guldi, D. M.;
Morales, A. F.; Leigh, D. A.; Prato, M.; Turco, R. Org. Lett. 2003, 5, 689–691; (s)
Alteri, A.; Gatti, F. G.; Kay, E. R.; Leigh, D. A.; Martel, D.; Paolucci, F.; Slawin, A. M. Z.;
Wong, J. K. Y. J. Am. Chem. Soc. 2003,125, 8644–8654; (t) Clegg, W.; Gimenez-Saiz,
C.; Gatti, F. G.; Leigh, D. A.; Nepogodiev, S. A.; Slawin, A. M. Z.; Teat, S. J.; Wong, J. K.
Y. J. Am. Chem. Soc. 2001, 123, 5983–5989; (u) Watanabe, N.; Furusho, Y.; Kihara,
N.; Takata, T.; Kinbara, K.; Saigo, K. Bull. Chem. Soc. Jpn. 2001, 74, 149–155; (v)
Furusho, Y.; Shoji, J.; Watanabe, N.; Kihara, N.; Adachi, T.; Takata, T. Bull. Chem. Soc.
Jpn. 2001, 74, 139–147; (w) Hunter, C. A.; Low, C. M. R.; Packer, M. J.; Spey, S. E.;
Vinter, J. G.; Vysotsky, M. O.; Zonta, C. Angew. Chem., Int. Ed. 2001, 40, 2678–2682;
(x) Reuter, C.; Mohry, A.; Sobanski, A.; Vo¨gtle, F. Chem.dEur. J. 2000, 6,1674–1682;
(y) Reuter, C.; Vo¨gtle, F. Org. Lett. 2000, 2, 593–595; (z) Leigh, D. A.; Murphy, A.;
Slawin, A. M. Z.; Teat, S. J. J. Am. Chem. Soc. 1999, 121, 4124–4129; (aa) Seel, C.;
Parham, A. H.; Safarowsky, O.; Hu¨bner, G. M.; Vo¨gtle, F. J. Org. Chem. 1999, 64,
7236–7242; (bb) Heim, C.; Affeld, A.; Nieger, M.; Vo¨gtle, F. Helv. Chim. Acta 1999,
82, 746–759; (cc) Schmidt, T.; Schmieder, R.; Mu¨ller, W. M.; Kiupel, B.; Vo¨gtle, F.
Eur. J. Org. Chem. 1998, 2003–2007; (dd) Ja¨ger, R.; Baumann, S.; Fischer, M.;
27. The dipolar cycloaddition between a nitrone and an N-aryl maleimide is rather
insensitive to electronic effects and usually gives a trans/cis ratio between 2.7:1
and 3.3:1.
28. The 500 MHz 1H NMR spectrum in CDCl3 of this pseudorotaxane shows broad
resonances at 25 ꢂC as a result of intermediate exchange on the chemical shift
time scale. Therefore, the ROESY experiment was performed at ꢁ30 ꢂC (slow
exchange limit).
29. Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem. Soc. 1985,
107, 3902–3909.
30. Stewart, J. J. P. J. Comput. Chem. 1989, 10, 209–220.
31. Rocha, G. B.; Freire, R. O.; Simas, A. M.; Stewart, J. J. P. J. Comput. Chem. 2006, 27,
1101–1111.
32. Stewart, J. J. P. J. Mol. Model. 2007, 13, 1173–1213.
33. Altered reactivity of functional groups incorporated within mechanically in-
terlocked structures is a well-established phenomenon. For some examples,
see: (a) Leigh, D. A.; Pe´rez, E. M. Chem. Commun. 2004, 2262–2263; (b) Zehnder,
D. W.; Smithrud, D. B. Org. Lett. 2001, 3, 2485–2487; Oku, T.; Furusho, Y.; Takata,
T. Org. Lett. 2003, 5, 4923–4925; Craig, M. R.; Hutchings, M. G.; Claridge, T. D.
W.; Anderson, H. L. Angew. Chem., Int. Ed. 2001, 40, 1071–1074; Parham, A. H.;
Windisch, B.; Vo¨gtle, F. Eur. J. Org. Chem. 1999, 1233–1238; Leigh, D. A.; Murphy,
A.; Smart, J. P.; Slawin, A. M. Z. Angew. Chem., Int. Ed. 1997, 36, 728–732;
Asakawa, M.; Brown, C. L.; Menzer, S.; Raymo, F. M.; Stoddart, J. F.; Williams, D.
J. J. Am. Chem. Soc. 1997, 119, 2614–2627.
34. Calculation of the energies of activation using thermochemical data derived
from the RM1 calculations estimates that the activation barrier for the reaction
of the [1$15] complex is around 6 kcal molꢁ1 higher than that for the reaction of
15 in the absence of macrocycle. Whilst the absolute value of this difference is
unlikely to be correct at this low level of theory, it does support, at least
qualitatively, the conclusion that binding of the macrocycle introduces a su-
pramolecular steric effect into the system.
35. (a) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.;
Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S.; Windus, T. L.;
Dupuis, M.; Montgomery, J. A. J. Comput. Chem. 1993, 14, 1347–1363; (b)
Gordon, M. S.; Schmidt, M. W. Theory and Applications of Computational
Chemistry: The First Forty Years; Dykstra, C. E., Frenking, G., Kim, K. S., Scuseria,
G. E., Eds.; Elsevier: Amsterdam, 2005; p 1167.
36. Stewart, J. J. P. MOPAC2007; Stewart Computational Chemistry: Colorado
Springs, CO, USA, 2007; http://openmopac.net.