The Hydrolysis of the Anhydride of 2-Cyano-2-phenylpropanoic Acid Triggers the Repeated Back and Forth Motions of an Acid–Base Operated Molecular Switch

Article abstract of DOI:10.1002/chem.201904048

This work aimed to render phenomenologically autonomous the otherwise stepwise operation of a catenane-based molecular switch, which is chemically triggered by the decarboxylation of 2-cyano-2-phenylpropanoic acid (2). Given that any amount of 2 in stoich

Full text of DOI:10.1002/chem.201904048

A Journal of
Accepted Article
Title: The Hydrolysis of the Anhydride of 2-Cyano-2-phenylpropanoic
Acid Triggers the Repeated Back and Forth Motions of an Acid-
Base Operated Molecular Switch
Authors: Stefano Di Stefano, Chiara Biagini, Giorgio Capocasa, Valerio
Cataldi, Daniele Del Giudice, and Luigi Mandolini
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201904048
Link to VoR: http://dx.doi.org/10.1002/chem.201904048
Supported by

10.1002/chem.201904048
Chemistry - A European Journal
FULL PAPER
The Hydrolysis of the Anhydride of 2-Cyano-2-phenylpropanoic
Acid Triggers the Repeated Back and Forth Motions of an Acid-
Base Operated Molecular Switch
Chiara Biagini,[a],¶ Giorgio Capocasa,[a],¶ Valerio Cataldi,[a] Daniele Del Giudice, [a] Luigi Mandolini[a] and
Stefano Di Stefano*[a]
Abstract: This work has been aimed at making phenomenologically
autonomous the otherwise step-wise operation of a catenane-based
molecular switch, chemically triggered by the decarboxylation of 2-
cyano-2-phenylpropanoic acid (2). Given that any amount of 2 in
stoichiometric excess with respect to the catenane is wasted in a side
reaction, we resorted to the corresponding anhydride 5, whose slow
hydrolysis due to adventitious water in dichloromethane continuously
produces in situ the actual fuel 2. As a consequence, the machine
does not require a reloading after each cycle, but switches back and
forth as long as fuel is present.
deprotonation of 1H+ according to the previously suggested
mechanism5a summarized in Scheme 2. The switch returns to its
neutral state 1 only after complete consumption of excess fuel.
Thus, acid 2, as well as any carboxylic acid undergoing easy base
promoted decarboxylation, cannot be used in stoichiometric
excess with respect to the machine components and,
consequently, is unsuitable for fuelling an acid-base driven
molecular machine that can operate autonomously, i.e., as long
as a chemical fuel is present.3
Introduction
The wide current interest in the investigation of synthetic
prototypal molecular machines has been mainly motivated by the
challenging goal of mimicking at least in part some of the essential
features of the natural molecular machines.1 Many small-
molecule prototypes consist in interlocked systems like catenanes
or rotaxanes.2 Usually their large amplitude motions between two
states result from the sequential addition of a forward and
backward stimulus (e.g. acid-base) or irradiation with light at two
different wavelengths.3 In a few cases, the cyclic motions of a
machine are fuelled by irradiation with light at
a single
wavelength,4 or by the irreversible reaction of a single reagent5 or
reagent system.6 A remarkable example is due to Leigh and
coworkers,6a who synthetized a [2]-catenane in which one of the
two macrocycles performs autonomous and unidirectional
circumrotations around the other, under the influence of only one
stimulus. Furthermore, autonomous molecular motions in other
systems such as i) rotary motions around a covalent bond,7 ii)
acid/base guided threading/dethreading in pseudorataxane
systems,8 iii) DNA-based molecular machines,9 iv) molecular
walkers,10 have been achieved by means of chemical fuels.11
Our contributions to the field has been based on the Sauvage-
type [2]-catenane 1, whose acid-base switching from neutral state
A to protonated states B’-B”, and back again to state A is
triggered by the decarboxylation of 1 mol eq of 2-cyano-2-
phenylpropanoic acid 25a or its Cl, CH3, and OCH3 para-
derivatives,12 as shown in Scheme 1 for the reaction with the
parent acid. The distinctive feature of the system is that the base
required by back proton transfer from 1H+ is generated in situ in
the decarboxylation step B’→B”. When back proton transfer is
complete, the system is ready to perform another cycle after
addition of another mol eq of fuel.13 However, if the fuel is added
in stoichiometric excess, all excess fuel is diverted to a
mechanistic shunt where decarboxylation takes place without
Scheme 1. Back and forth motions of catenane 1 triggered by acid 2 in CH2Cl2.
Fast and quantitative proton transfer (step A→B’) forces the two phenanthroline
subunits to be intimately held together in
a tetrahedral arrangement.
Decarboxylation (step B’→B”), and subsequent slow back proton transfer from
1H+ to the counteranion, restores the resting state of the catenane (step B”→A).
An answer to the quest for a chemical stimulus that can drive the
switching of 1 for a number of cycles without reloading at each
cycle was recently published.14 The photochemical cleavage of
[a]
C. Biagini, G. Capocasa, V. Cataldi, D. Del Giudice, Prof. L.
Mandolini and Prof. S. Di Stefano
Dipartimento di Chimica, Università di Roma “La Sapienza”, and
Istituto CNR per i Sistemi Biologici (ISB-CNR), Sezione Meccanismi
di Reazione
P.le A. Moro 5, 00185 Roma, Italy.
E-mail: stefano.distefano@uniroma1.it
Supporting information for this article is given via a link at the end of
the document
1
This article is protected by copyright. All rights reserved.

10.1002/chem.201904048
Chemistry - A European Journal
FULL PAPER
16
compound 4 under controlled conditions liberates the desired
amount of 2, eqn (1). Thus, sub-stoichiometric amounts of acid 2
could be repeatedly released in situ by means of iterative
irradiations of a solution of 1 containing a large molar excess of
precursor 4.
MgCl2 failed. In all cases acid 2 was recovered unchanged.
When the same procedures were applied to 2-phenylpropanoic
acid, the corresponding anhydride was obtained in good yields.
This strongly indicates a high sensitivity of 5 to the usual aqueous
work-up conditions, most likely arising from the presence of the
strong electron withdrawing cyano substituents. Better results
were obtained by an adaptation of a general procedure consisting
in the reaction of the parent acid with a mixture of DABCO and
thionyl chloride.17 A number of repeated small-scale synthetic
runs afforded anhydride 5 in 12-15% yield, the sole contaminant
being parent acid 2, in variable amounts not exceeding 10%.
The strong tendency of anhydride 5 to undergo hydrolysis is well
illustrated by its smooth reaction with adventitious water in
dichloromethane (see SI, Figure S1). Not surprisingly, the rate of
hydrolysis was found to be strongly dependent on the solvent
batch. A standard drying procedure (see experimental) reduced
the water content of commercial samples of CD2Cl2 and CH2Cl2
to values in the order of 2 mM (0.003 % w/w),18 but a major source
of water in the reaction mixtures turned out to be anhydride 5,
whose solutions were found to contain roughly one molecule of
water per molecule of anhydride.
Under the given conditions, the operation of the chemically
triggered molecular switch was repeatable, yet the manual action
of an operator at selected time intervals was required.
Scheme 2. Reaction pathways for the switching motions of 1 triggered by acid
2 (see ref 5a). The sequence on the left-hand side from top to bottom is the
same as in Scheme 1. Excess acid opens the way to a mechanistic shunt
involving rate-limiting liberation of CO2 from the ternary adduct C. Loss of CO2
in the step C→D is slower than in step B’→B” because of stabilization of
carboxylate anion by H-bonding in the former.
Scheme 3. Chemically triggered autonomous motions of the molecular switch
1. Slow hydrolysis of excess anhydride 5 produces acid 2, which continuously
fuels repetitive motions from neutral state A to protonated states B’ - B”, and
back again to neutral state A.
In a first experiment 2.0 mM catenane 1 was mixed with 2.0 mM
anhydride 5, so that the total amount of acid 2 expected to result
from complete hydrolysis of anhydride 5 was a two-fold molar
excess with respect to 1. The reaction was monitored by 1H NMR
spectroscopy (Figure 1 and Figure 4a). The spectrum recorded at
t = ∞ revealed the complete disappearance of anhydride 5 and
the presence of the decarboxylated product 3 as the sole reaction
product, besides the catenane in its original neutral state 1.19 The
spectra recorded in the course of the reaction showed the growth
and decay of a transient species, previously shown to be ion pair
B”.5a,12a The peaks marked with a blue asterisk, with the
exception of the peak of the double bond protons at 5.52 ppm,
belong to ten of the twelve protons of the two non-equivalent
phenanthroline units.20 The relative intensities of the two signals
at 5.52 ppm and 5.44 ppm, belonging to the double bond protons
of 1H+ and 1, respectively, happen to be very nearly 1:1 in all of
the spectra taken from 184 min to 439 min. Thus, while the
concentration of anhydride 5 regularly decreases and that of the
final product 3 increases on increasing reaction time (see Figure
4a), the fraction of catenane in the protonated state 1H+ reaches
a flat maximum, which implies that the rate of formation of B”
equals the rate of its disappearance in the given time range.
Here we report that switching motions of catenane 1 could be
repeatedly triggered for a large number of cycles by continuous
release of acid 2 by the smooth reaction of a molar excess of
anhydride 5 with the tiny amount of adventitious water in the
reaction medium (Scheme 3). Since proton transfer from 2 to 1 is
instantaneous and quantitative, and since decarboxylation (step
B’→B”, Scheme 1) is complete in less than 2 min,12a the rate of
formation of the ion pair in state B” practically coincides with the
rate of formation of acid 2 and, consequently, the rate of the
overall process is determined by the rate of the back proton
transfer (step B”→A). Thus, according to Scheme 3, the
concentration of the ion pair in state B” builds up in the early
stages of the reaction, reaches a more or less extended maximum
when v1 = v2, and eventually dies out. Key to the design is that the
rate of liberation of the acid is never large enough for the machine
to be overfed with acid 2.
Results and Discussion
Early attempts at preparing anhydride 5 by dehydration of the
parent acid 2 with DCC15 or exchange with Boc2O catalyzed by
2
This article is protected by copyright. All rights reserved.

10.1002/chem.201904048
Chemistry - A European Journal
FULL PAPER
Figure 1. 1H NMR monitoring of reaction of 2.0 mM catenane 1 in the presence of 2.0 mM anhydride 5 in CD2Cl2 at 25 °C. Signals marked with blue asterisks are
related to state B”, red asterisks to catenane 1 in the neutral state A, green asterisk to decarboxylated product 3 (for a complete assignment of the signals related
to state B”, 1 and 3, see ref 5a and 12a). Reaction times are given in minutes.
Corroborating evidence came from UV-Vis monitoring of the
reaction between 1 and 5 in CH2Cl2. Ion pair B” is known to
display a strong adsorption centred at 375 nm12a whereas all of
the other species in the reaction mixture, including ion pair B’, do
not absorb, or absorb very little at the given wavelength. The red
trace in Figure 2 refers to a comparative experiment in which 2.0
mM 1 was mixed with an equimolar amount of acid 2. The rapid
growth of absorbance, corresponding to step B’→B”, is followed
by a much slower decay, corresponding to step B”→A. The blue
trace refers to the reaction of 1 with anhydride 5, carried out at the
same concentration as in the run monitored with 1H NMR
spectroscopy (Figure 1). In this case the concentration of B”
grows slowly, and after 4h reaches a plateau value that is about
half as great as the maximum value displayed by the red profile.21
In conclusion, since 2 mol eq of acid 2 were produced by compete
Figure 2. UV-vis monitoring of reaction of 2.0 mM catenane 1 in the presence
of 2.0 mM anhydride 5 (blue trace) and 2.0 mM fuel acid 2 (red trace) in CH2Cl2
at 25°C. Only ion pair B” absorbs at 375 nm. The small initial burst of
hydrolysis of 5, every molecule of catenane 1 carried out an
average number of two cycles A→(B’-B”)→A.
absorbance in the blue trace is due to the presence of tiny amounts of acid 2 in
the sample of 5. The solutions were contained in a quartz cell with length of 1.0
mm. The blue curve is a guide for the eye. The red curve is a plot of the standard
kinetic equation for two subsequent first-order processes, namely, fast
formation (B’→B”) and much slower decay (B”→A) of the ion pair in state B”
(see ref 12a). k’ = 3.6×10-2 s-1 for B” formation and k’’ = 4.8×10-4 s-1 for B” decay.
See Figure S4 and S5 for complete UV-vis monitoring.
3
This article is protected by copyright. All rights reserved.

10.1002/chem.201904048
Chemistry - A European Journal
FULL PAPER
The implicit assumption for the correct operation of the switch,
which means that the number of cycles is directly linked to the
excess of anhydride added, has been that acid produced in the
slow hydrolysis of the anhydride reacts with the fraction of
catenane in the initial state A, but not with the benzyl carbanion
in state B’’ (see Scheme 1). At first sight, the assumption is
unsound, because anion in state B’’ is clearly more basic than
neutral catenane. However, it is known22a-d that exoergonic proton
transfer reactions between O and N acids and bases are usually
diffusion controlled (k=1010 -1011M-1s-1), whereas reaction rates
involving C acids and bases are frequently much lower.23 The
assumption that the fate of acid produced by the hydrolysis of 5 is
dictated by a faster reaction with the N base, and not by the
stronger thermodynamic basicity of the carbanion, was suggested
by the peculiar behaviour of the kinetics of proton transfer from or
to carbon. A demonstration of the validity of the above assumption
is provided by the UV-Vis experiment reported in Figure 3.
catenane 1. The reaction progress was monitored by 1H NMR
spectroscopy (Figure 4). Sets of complete spectra as a function
of time are reported in the SI, Figure S2 and Figure S3. Significant
portions are shown in Figure 4b and in Figure 4c. The first
observation is that in both cases the amount of adventitious water
turned out to be large enough for the anhydride to undergo
complete hydrolysis. Thus, the average number of cycles
performed by every molecule of 1 was four and eight in the runs
in which the initial concentration of 5 was 4.0 mM and 8.0 mM,
respectively. Furthermore, also in these cases the relative
intensities of the double bond signals at 5.52 ppm and 5.44 ppm
show that the fraction of 1H+ (state B”) reaches a maximum value
and remains constant, or very nearly so, for a long time. This
fraction is 0.70 in the time range from 110 min to 260 min when
[5]o = 4.0 mM and ≈ 0.9 in the time range from 28 min to 183 min
when [5]o = 8.0 mM. It appears therefore that in the above
experiments the rate of acid 2 production is never high enough to
overfeed the machine with fuel, and it seems very likely that the
rate of fuel production in the run in which [5]o = 8.0 mM closely
approaches the upper limit beyond which the excess fuel is
entirely wasted.24
Comparison of the run in which [5]o = 2.0 mM with those in which
[5]o = 4.0 mM and 8.0 mM shows that the time required by the
concentration of the ion pair in state B” to reach its maximum
value is shorter the larger the initial concentration of anhydride 5
(and water). For example, when [5]o = 8.0 mM this time is roughly
one order of magnitude shorter than when [5]o = 2.0 mM. This
indicates, not surprisingly, that the rate of hydrolysis significantly
increases on increasing reactant concentrations.
Although the rate equation for the hydrolysis reaction is unknown,
reliable estimates of the rates v1 of acid production at the quasi-
stationary state in the different runs could be obtained
remembering that v1 = v2 (Scheme 3) when the concentration of
the ion pair in state B” reaches its maximum value. Since the
decay of B” is a first-order process5a,12a (k’’ = 4.8×10-4 s-1, see
caption to Figure 2), and since [B”]max could be estimated from
NMR spectra, v1 values of 4.8×10-7 M s-1, 6.7×10-7 M s-1, and
8.6×10-7 M s-1 were calculated from eqn (2) for the runs in which
[5]o = 2.0 mM, 4.0 mM and 8.0 mM in the given order.25
Figure 3. UV-Vis monitoring of the reaction of 0.30 mM catenane 1 with the
stoichiometric amount of acid 2 in CH2Cl2 at 25°C. Addition of the acid was
carried out in two steps. At t=0, half mol eq of acid was added. The arrow
indicates the time at which the second half of acid was added. The solution was
contained in a quartz cell of 10 mm optical path. See Figure S6 for complete
UV-vis monitoring.
v1 = k’’[B”]max
(2)
It appears therefore that rates of acid production at the
quasi-stationary state show but a modest dependence on
reactant concentrations, in contrast with the marked dependence
exhibited by the times required to reach the quasi-stationary state.
That this oddity is more apparent than real can be understood
remembering that hydrolysis of 5 is subject to catalysis19 (either
base or nucleophilic) by the phenanthroline units of 1 and
considering that the fraction of catenane 1 in its catalytically active
neutral form is lower the larger the initial concentration of
anhydride 5. This fraction is 1/2 when [5]o = 2.0 mM, but drops to
1/10 when [5]o = 8.0 mM. In other words, the system deactivates
its own catalyst in the phase in which the quasi-stationary state is
approached. This has a beneficial action on the operation of the
system because the possibility of overfeeding the molecular
switch with fuel acid is attenuated.
In this experiment, the reaction of 0.30 mM catenane 1 with the
stoichiometric amount of acid 2 was carried out in two steps.
Addition of half of the acid (0.5 mol eq) at t=0 was followed by a
rapid growth of absorbance at 375 nm, until a plateau value was
reached, which indicates that the reaction B’→B’’ was over. At
this point the system was a 1:1 mixture of neutral catenane in
state A and ion pair in state B’’. Addition of the second half of the
acid caused a further increase in absorbance, until a maximum
value was reached which was consistent with a virtual complete
transformation of catenane 1 and added acid to the ion pair in
state B’’. This is tantamount to saying that in the second step
protonation of the neutral catenane is the preferred pathway, or
that protonation of the carbanion, which would not cause any
increase in the concentration of B’’, can be definitely ruled out.
Additional runs were carried out in which the switching of
catenane 1 (2.0 mM) was triggered by 4.0 mM and 8.0 mM
precursor 5, whose complete hydrolysis would produce a 4-fold
and 8-fold molar excess, respectively, of acid 2 relative to
4
This article is protected by copyright. All rights reserved.

10.1002/chem.201904048
Chemistry - A European Journal
FULL PAPER
Figure 4. Selected portions of 1H NMR spectra taken during reactions of 2.0 mM 1 with (a) 2 mM 5, (b) 4 mM 5, and (c) 8 mM 5 in CD2Cl2 at 25 °C. Colour code as
in Figure 1. For the sake of graphical convenience. Traces between 6.5 ppm and 4.5 ppm have been zoomed in (5x scale approx.) Reaction times are given in
minutes.
A final experiment, in which 2.0 mM 1 was reacted with 4.0 mM 5
in CD2Cl2 saturated with water,26 illustrates well how an
exceedingly fast anhydride hydrolysis causes a malfunction in the
molecular switch (Figure 5). The complete absence of signals of
neutral catenane 1 in the first spectrum at time t = 3 min is
consistent with the expected faster production of acid 2 in a
medium containing a relatively large amount of water. The major
component of the reaction mix is the ternary adduct C5a (Scheme
2), whose distinctive signals persist over a large fraction of the
reaction time. The total lack of signals (including the highly
structured pattern of intense signals in the high field region) in the
1H NMR spectra of the switching experiments (Figure 1, and
Figures S2 and S3) provides a definite confirmation that the rate
of production of acid 2 never exceeded the threshold value above
which the way to the wrong sequence B’→C→D→B’ in Scheme
2 is opened.
.
5
This article is protected by copyright. All rights reserved.

10.1002/chem.201904048
Chemistry - A European Journal
FULL PAPER
Figure 5. 1H NMR monitoring (25°C, 300 MHz) of the reaction between 2.0 mM 1 and 4.0 mM 5 in CD2Cl2 saturated with H2O. Colour code as in Figure 1, orange
asterisks are related to state C in Scheme 2 (see ref 5a). Reaction times are given in minutes.
Experimental Section
Conclusions
Instruments and Methods
In this work we have shown that the switching motions of an acid-
base operated molecular switch could be continuously triggered
by the slow hydrolysis of the anhydride of acid 2, the latter being
1H NMR spectra were recorded on a 300 MHz spectrometer. The spectra
were internally referenced to the residual proton signal of the solvent at
5.30 ppm. Spectrophotometric measurements were carried out on a diode
the actual fuel. The rate of hydrolysis was never high enough to
array spectrophotometer equipped with a thermostatted cell compartment.
overfeed the machine with acid 2, thus avoiding any waste of acid
Mass spectrometric measurements were carried out by an ESI-TOF
spectrometer.
in a side decarboxylation pathway in which motions of the
molecular switch are not involved.
Thus, our catenane-based molecular switch performs a number
of cycles27 until fuel exhaustion without requiring additional
actions during its operation, such as supply of material or radiant
Materials
energy. While it is clear that the slow and continuous release of
acid
2 does not transform switch 1 into an intrinsically
All reagents and solvents were purchased at the highest commercial
quality and were used without further purification, unless otherwise stated.
CH2Cl2 and CD2Cl2 were dried over molecular sieves (3 Å) for 24 h and
stored in a desiccator loaded with activated silica gel Rubin. Before use
the molecular sieves were activated at 300 °C for 24 h. All the glassware
was flame-dried and subsequently left to reach room temperature under
an inert gas flow. All the experiments were carried out in thermostatted
UV-vis cells and NMR tubes. Catenane 1 (mixture of EE, EZ and ZZ
geometrical isomers) was available from a previous investigation.28 Acid 2
was available from a previous investigation.5a
autonomous machine such as the one described by Leigh and co-
workers,6a it represents an effective trick for making an acid-base
operated molecular machine operationally autonomous. It is felt
that anhydride 5 could be used as a pre-fuel for the operation of
other acid-base operated molecular machines.
6
This article is protected by copyright. All rights reserved.

10.1002/chem.201904048
Chemistry - A European Journal
FULL PAPER
2-Cyano-2-phenylpropanoic anhydride (5). Anhydride
5
was
Keywords: Molecular switch • Chemical fuel • Autonomous
motions • Catenane • Anhydride hydrolysis
synthesized adapting the general procedure described by Kazemi and
Kiasat.17 In an accurately dried Schlenck tube, 56 mg (0.5 mmol) of 1,4-
diazabicyclo[2.2.2]octane (DABCO) were dissolved in 1.5 mL of dry
CH2Cl2. The solution was cooled in an ice bath, then 36 µl of thionyl
chloride (0.50 mmol) were carefully added. The bright yellow mixture was
stirred for 5 min, then 87.5 mg (0.50 mmol) of 2-cyano-2-phenylpropanoic
acid were added under an inert atmosphere. The mixture immediately
turned white, then the ice bath was removed and the reaction was
vigorously stirred for 30 min. The mixture was quenched with 4 mL of a
10% solution of NaHCO3. The organic phase was quickly diluted with 2 mL
of CH2Cl2, separated, washed with brine, dried over Na2SO4 and the
solvent was evaporated under reduced pressure. Each run of this
procedure afforded about 20-25 mg of the desired product (yield 12-15%),
which, being pure except for minor amounts (<10%) of parent acid, was
used without further purification. Samples of anhydride 5 were freshly
prepared before use in all of the experiments carried out in the presence
of the catenane (preparation of the anhydride in the morning, reaction with
catenane in the afternoon).
[1]
(a) M. A. Watson, S. L. Cockroft, Chem. Soc. Rev., 2016, 45, 6118-6129;
(b) S. Kassem, T. van Leeuwen, A. S. Lubbe, M. R. Wilson, B. L. Feringa,
D. A. Leigh, Chem. Soc. Rev., 2017, 46, 2592-2621; (c) M. Baroncini, L.
Casimiro, C. de Vet, J. Groppi, S. Silvi, A. Credi, ChemistryOpen, 2018,
7, 169–179.
[2]
[3]
[4]
(a) J.-P. Sauvage Angew. Chem. Int. Ed., 2017, 56, 11080-11093; (b) J.
F. Stoddart Angew. Chem. Int. Ed., 2017, 56, 11094-11125.
S. Erbaş-Çakmak, D. A. Leigh, C. T. McTernan, A. L. Nussbaumer,
Chem. Rev., 2015, 115, 10081-10206.
(a) A. M. Brouwer, C. Frochot, F. G. Gatti, D. A. Leigh, L. Mottier, F.
Paolucci, S. Roffia, G. W. H. W, Science, 2001, 291, 2124–2128. (b) G.
W. H. Wurpel, A. M. Brouwer, I. H. M. van Stokkum, A. Farran, and D. A.
Leigh, J. Am. Chem. Soc., 2001, 123, 11327-11328. (c) W. Abraham, L.
Grubert, U. W. Grummt, K. Back, Chem. Eur. J., 2004, 10, 3562-3568.
(d) V. Balzani, M. Clemente-Leon, A. Credi, B. Ferrer, M. Venturi, A. H.
Flood, J. F. Stoddart, Proc. Natl. Acad. Sci. 2006, 103, 1178−1183.
(a) J. A. Berrocal, C. Biagini, L. Mandolini, S. Di Stefano, Angew. Chem.
Int. Ed., 2016, 55, 6997-7001; (b) S. Erbas-Cakmak, S. D. P. Fielden, U.
Karaca, D. A. Leigh, C. T. McTernan, D. J. Tetlow, M. R. Wilson, Science,
2017, 358, 340–343; (c) A. Ghosh, I. Paul, M. Adlung, C. Wickleder, M.
Schmittel, Org. Lett., 2018, 20, 1046−1049.
[5]
[6]
1H-NMR (300 MHz, 25°C, CD2Cl2). Anhydride 5 was obtained as a 1:1
mixture of meso and d,l compounds. δ (ppm): 7.39 (m, 10×2 = 20 H, Ar-
H), 1.95 (s, 6 H,-CH3 of one of the two diasteroisomers), 1.92 (s, 6 H,-CH3
of the other diasteroisomer). 13C-NMR (75 MHz, 25°C, CD2Cl2). δ (ppm):
161.21, 133.10, 133.04, 129.42, 125.75, 125.70, 117.42, 117.33, 49.14,
49.09, 23.17, 23.08.
(a) M. R. Wilson, J. Solà, A. Carlone, S. M. Goldup, N. Lebrasseur, D. A.
Leigh, Nature, 2016, 534, 235-240; (b) Q. Shi, C.-F. Chen, Chem. Sci.,
2019, 10, 2529–2533.
[7]
[8]
W. L. Mock, K. J. Ochvat, J. Phys. Org. Chem., 2003, 16, 175-182.
O. Buyukcakir, F. T. Yasar, O. A. Bozdemir, B. Icli, E. U. Akkaya, Org.
Lett., 2013, 1012-1015.
Reaction between 2.0 mM catenane and 4.0 mM anhydride in CD2Cl2
saturated with H2O
[9]
(a) Y. Chen, M. Wang, C. Mao, Angew. Chem. Int. Ed., 2004, 43, 3554 –
3557. (b) A. J. Turberfield, J. C. Mitchell, B. Yurke, A. P. Mills, Jr., M. I.
Blakey, F. C. Simmel, Pys. Chem. Lett. 2003, 90, 118102-118102. (c) W.
U. Dittmer, F. C. Simmel, Nano Lett., 2004, 4, 689-691.
1.16 mg of 1 were dissolved in 480 µL of CD2Cl2. To this solution, 120 µL
of a 0.02M stock solution of 5 were added, immediately followed by 5µl of
H2O (this amount of water is enough to saturate the CD2Cl2 solution26).
After mixing, the reaction was monitored by 1H NMR spectroscopy.
[10] (a) W. B. Sherman, N. C. Seeman, Nano Lett., 2004, 4, 1203-1207. (b)
P. Yin, H. Yan, X. G. Daniell, A. J. Turberfield, J. H. Reif, Angew. Chem.
Int. Ed., 2004, 43, 4906-4911. (c) J. Bath, S. J. Green, A. J. Turberfield,
Angew. Chem. Int. Ed., 2005, 44, 4358-4361. (d) Y. Tian, Y. He, Y. Chen,
P. Yin, C. Mao, Angew. Chem. Int. Ed. 2005, 44, 4355 –4358.
[11] Chemical fuels have been also used to drive chiral information, see: (a)
A. Carlone, S. M. Goldup, N. Lebrasseur, D. A. Leigh, A. Wilson, J. Am.
Chem. Soc., 2012, 134, 8321-8232. (b) S. Corra, C. de Vet, J. Groppi, M.
La Rosa, S. Silvi, M. Baroncini, and A. Credi, J. Am. Chem. Soc. 2019,
141, 9129-9133, or to drive the dissipative self-assembly of transient
species and nanostructures, and catalysis. See for example: (c) H.
Fanlo-Virgs, A.-Ne. R. Alba, S. Hamieh, M. Colomb-Delsuc, S. Otto,
Angew. Chem. Int. Ed., 2014, 53, 11346 –11350. (d) F. della Sala, S.
Neri, S. Maiti, J. L-Y Chen, L. J Prins, Curr. Opin. Biotechnol., 2017, 46,
27–33. (e) C. Biagini, S. D. P. Fielden, D. A. Leigh, F. Schaufelberger, S.
Di Stefano, D. Thomas, Angew. Chem. Int. Ed., DOI:
10.1002/anie.201905250.
Conflicts of interest
There are no conflicts to declare.
Authors Contribution¶
CB and GC contributed equally to this work
[12] (a) C. Biagini, S. Albano, R. Caruso, L. Mandolini, J. A. Berrocal, S. Di
Stefano, Chem Sci., 2018, 9, 181–188. (b) P. Franchi, C. Poderi, E.
Mezzina, C. Biagini, S. Di Stefano, M. Lucarini, J. Org. Chem. DOI:
10.1021/acs.joc.9b01164.
[13] Along the manuscript the term “fuel” is used as a synonymous of
“stimulus” although machine 2 is definitely a switch and not a motor (see
ref 3). However, since acids undergoing decarboxylation at convenient
rate have been demonstrated to be able to trigger the motions of both
molecular switches (see ref 5a, 5c, 12a and 12b) and motors (see ref 5b)
and since the fuel system is the main object of the present article, we feel
that “fuel” and “stimulus” are appropriately interchangeable.
Acknowledgements
Progetto di Ateneo 2018 protocol number RG1181641DCAAC4E
and Progetto Avvio alla Ricerca 2018 protocol number
AR11816426964D18 are acknowledged. The authors thank Dr.
Alessandro Latini for technical assistance.
[14] C. Biagini, F. Di Pietri, L. Mandolini, O. Lanzalunga, S. Di Stefano, Chem.
Eur. J., 2018, 24, 10122-10127.
[15] Z. Selinger, Y. Lapido, J. Lipid Res., 1966, 7, 174-175.
[16] G. Bartoli, M. Bosco, A. Carlone, R. Dalpozzo, E. Marcantoni, P.
Melchiorre, L. Sambri, Synthesis, 2007, 22, 3489-3496.
7
This article is protected by copyright. All rights reserved.

10.1002/chem.201904048
Chemistry - A European Journal
FULL PAPER
[17] F. Kazemi, A. R. Kiasat, Phosphorus, Sulfur, and Silicon, and the Related
Elements, 2003, 178, 2287–2291.
[23] As a classical example, the rate constant of the reaction between H3O+
and (CH3CHNO2)ˉ is as low as 16 M-1s-1 whereas the corresponding
reaction of borate anion, a species with very similar basicity, is diffusion-
controlled (see ref 22b, p 324).
[18] Estimates of water content was based on a comparison of the intensity
of the more or less sharp singlet of free water at 1.55 ppm with the
intensity of the signals of other species of known concentration. In the
case of reaction mixtures the comparison was based on the first
spectrum recorded after addition of anhydride 5.
[24] An additional run was carried out in the presence of 2.0 mM 1 and 12.0
mM 5. After a first phase in which the behavior of the system was
superimposable to that of the experiment carried out with 8 mM 5, a
number of intense 1H NMR signals appeared in the spectrum related to
unidentified species probably due to the insurgence of undesired side
reactions.
[19] Consumption of 2.0 mM anhydride 5 in the presence of 2.0 mM catenane
1 (t1/2 = 8 h) turned out to be faster than the background reaction (t1/2
=
29 h, compare Figure S1 with Figure 1). A similar result was obtained in
the presence of 4.0 mM neocuproine (t1/2 = 3 h), suggesting that the
phenanthroline moiety acts as a general base or nucleophilic catalyst.
[20] The 1H NMR spectrum of B” differs markedly from that of B’ (see ref 12),
where there is no trace of the upfield shifted signals of the phenanthroline
hydrogens of 1H+. Whereas B’ is an ion pair in strict sense, in that the
two partners are held together only by electrostatic attraction, an
important, additional contribution to the driving force for association of
the partners in B” was suggested (see ref 5a) to come from an electron-
donor-acceptor (EDA) interaction between the electron rich benzyl
carbanion and the electron poor phenanthroline core of 1H+. As a
consequence, the protons of one phenanthroline unit are exposed to the
shielding effect of the ring current of the aromatic counterion.
[25] In
the
case
of
[5]o = 2.0 mM,
[B”]max = 0.5×[1]o = 1.0 mM,
v1 = 4.8×10-7 M s-1;
in
the
case of [5]o = 4.0 mM,
[B”]max = 0.7×[1]o = 1.4 mM, v1 = 6.7×10-7 M s-1; eventually in the case of
[5]o = 8.0 mM, [B”]max = 0.9×[1]o = 1.8 mM, v1 = 8.6×10-7 M s-1.
[26] The solubility of water in dichloromethane is about 0.13 M (0.17 %, w/w;
see IUPAC-NIST Solubility Database, NIST Standard Reference
Database 106) which means that the amount of water in this experiment
roughly exceeds that in the previous experiments by an order of
magnitude.
[27] It can be speculated that if catenane 1 was exposed to the action of a
substoichiometric amount of any stable acid, then the same motions
triggered by acid 1 would also occur due to spontaneous proton
exchange between catenane molecules. As a matter of fact the latter
exchange reaction, if any, is not under the control of the experimenter: in
principle it would go on for an infinite time, while because of the presence
of anhydride 5 the back and forth motions of switch 1 are triggered and
continue as long as the former is present. Furthermore, the proton
exchange between catenane 1 molecules, which is slow on 1H NMR time
scale (see Figure S7) when trifluoroacetic acid is used as
substoichiometric protonating agent, is expected to be much slower
when the counterion is the carbanion in B” that strongly interacts with
the phenanthroline core of 1H+ (see ref 20).
[21] Slow solvent evaporation from the cell became appreciable after a few
hours. This prevented reliable measurements to be carried out at
prolonged reaction times. The slow evaporation of the volatile
dichloromethane solvent clearly shows that the small Teflon stopper of
our 1.0 mm spectrophotometric cell is not a perfect seal. Since we found
no appreciable water absorption by a dichloromethane sample in a NMR
tube after 24h, and since the timescale of the spectrophotometric
experiment (Figure 2) compares well with the timescale of the 1H NMR
experiment carried out at the same initial concentrations (Figure 1), we
argue that the amount of atmospheric water possibly taken in by the
reaction mixture in the UV-Vis cell does not affect the reaction rate to a
significant extent.
[28] (a) J. A. Berrocal, M. M. L. Nieuwenhuizen, L. Mandolini, E. W. Meijer, S.
Di Stefano, Org. Biomol. Chem. 2014, 12, 6167–6174; (b) J. A. Berrocal,
L. M. Pitet, M. M. L. Nieuwenhuizen, L. Mandolini, E. W. Meijer, S. Di
Stefano, Macromolecules 2015, 48, 1358–1363.
[22] (a) D. J. Cram in Fundamentals of Carbanion Chemistry, Academic
Press, New York and London, 1965, ch. 1. (b) L. P. Hammett in Physical
Organic Chemistry, Mc Graw-Hill Book Edition, 1970, ch. 10. (c) J. R.
Jones in The Ionization of Carbon Acids, Academic Press Inc (London),
Ltd., London, 1973, ch. 2. (d) E. F. Caldin in The Mechanisms of Fast
Reactions in Solution, IOS Press, Amsterdam, 2001, ch. 8.
…
8
This article is protected by copyright. All rights reserved.

10.1002/chem.201904048
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents
FULL PAPER
A chemically triggered
Chiara Biagini, Giorgio Capocasa,
Valerio Cataldi, Daniele Del
Giudice, Luigi Mandolini and
Stefano Di Stefano*
molecular machine performs
multiple cycles of back and
forth motions, without need of
additional supply of material or
radiant energy during its
operation.
Page No. – Page No.
The Hydrolysis of the
Anhydride of 2-Cyano-2-
phenylpropanoic Acid Triggers
the Repeated Back and Forth
Motions of an Acid-Base
Operated Molecular Switch
9
This article is protected by copyright. All rights reserved.