Theory and example of a small-molecule motor

Article abstract of DOI:10.1002/poc.591

Conceptual principles for operation of a molecular motor stipulate that such a species must consist of an energy-consuming catalyst that exhibits within its mechanism temporal overlap of successive catalytic cycles, so as to induce irreversible intramolecular motion during operation. The exothermic hydration reaction of MeCOCH=C=NCMe₃ brought about by catalyst HO₂CCH₂OCMe(CO₂H)₂ qualifies as such a process, in the form of serial conversion between canonical carboxylic anhydride structures within the catalyst. The steady-state kinetic behavior and inactivation by methanol demonstrate the operation of an appropriate repetitive catalytic loop in this case. Copyright

Full text of DOI:10.1002/poc.591

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2003; 16: 175–182
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.591
Theory and example of a small-molecule motor
William L. Mock* and Krzysztof J. Ochwat
Department of Chemistry, University of Illinois at Chicago, Chicago, Illinois 60607-7061, USA
Received 4 April 2002; revised 24 July 2002; accepted 25 October 2002
ABSTRACT: Conceptual principles for operation of a molecular motor stipulate that such a species must consist of an
energy-consuming catalyst that exhibits within its mechanism temporal overlap of successive catalytic cycles, so as to
induce irreversible intramolecular motion during operation. The exothermic hydration reaction of Me-
epoc
COCH=C=NCMe3 brought about by catalyst HO2CCH2OCMe(CO2H)2 qualifies as such a process, in the form
of serial conversion between canonical carboxylic anhydride structures within the catalyst. The steady-state kinetic
behavior and inactivation by methanol demonstrate the operation of an appropriate repetitive catalytic loop in this
case. Copyright  2003 John Wiley & Sons, Ltd.
Additional material for this paper is available from the epoc website at http://www.wiley.com/epoc
KEYWORDS: small-molecule motor; molecular motor; catalyst; kinetics
INTRODUCTION
drives the assemblage in one of the manners stipulated.
Any nano-construct achieving this dynamic behavior (i.e.
‘motor’) is required to be a chemical catalyst, as strictly
defined. Within the course of channeling a spontaneous
energy flow, for example, in the nature of degradation of
a chemical high-potential ‘fuel,’ the catalyzing species is
to suffer according to plan a repeated systematic intra-
molecular shuttling or twisting, which is the designated
goal. The requirement for catalyzed energy expenditure
with consequent forced repetitive operation distinguishes
a true molecular motor from simpler ‘molecular
switches’ and similar devices that have been reported
previously. Initially, one need not address conversion of
that energy release into useful work. The immediate
objective (as typified here) is only to counter entropy
locally, namely, to defeat the inherent and thermodyna-
mically dictated propensity of freely interconverting
molecular conformers to proceed with an equal flux in
both directions (i.e., both clockwise and counterclock-
wise in the case of a counter-rotator) (An attempt has
been described4 to prepare a unidirectional rotator in the
form of a ‘ratchet’ device, which serially catches after
each fraction of a turn, so that ordinary thermal energy
suffices to drive the process. The device proved
unsuccessful, as indeed was obliged to be the case, since
all such spontaneous and thermoneutral microscopic
processes are necessarily reversible).
The simplest type of ‘molecular motor’ would be a single
chemical entity for which internal motion in a specific
sense can be induced to occur repetitively and auto-
matically. The designated atom movements need to be
distinguished from the ordinary thermal vibrations,
rotations and tautomerizations that organic molecules
continuously experience. Although most organic mol-
ecules in solution are constantly changing shape, methods
are only now being consciously developed for engender-
ing systematic oscillatory motion that might be captured
for performing work on such a nanoscale.1 The challenge
is to fashion a molecular mechanism that realizes this
objective. Of course, living organisms have solved the
problem. However, the biochemical resolution is for the
most part macromolecular.2 We seek a reductionist
answer: can one recognize and constructively exploit
the minimally essential features of such a molecular
machine? For example, a reciprocating motor might be a
molecule that could be chemically forced to alternate
regularly in a controlled (induced) fashion between two
related covalent structures that the molecule would
otherwise not adopt. Another type of motor might entail
counter-rotation exclusively in one dihedral sense, about
a designated covalent sigma bond connecting two
molecular halves (turrets), with that motion driven
chemically in a particular torsional direction.3 Practical
realization of these objectives requires a dissipative
system; specifically, provision of an expendable reservoir
of free energy is required, such that its controlled release
RESULTS
Design
*Correspondence to: W. L. Mock, Department of Chemistry,
University of Illinois at Chicago, Chicago, Illinois 60607-6071, USA.
E-mail: wlmock@uic.edu
As a demonstration of feasibility, a simple oscillatory
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 175–182

176
W. L. MOCK AND K. J. OCHWAT
Scheme 1
motor that spontaneously operates in solution has been
assembled and examined. Analysis of it serves to
illustrate some of the principles and techniques entailed
in successful operation of a molecular machine of the
type sought. The ‘fuel’ is an acylketenimine (Me-
1 → 2 is rate limiting, the prevalent species during the
steady state consists of the carboxy-activated cyclic
anhydride (upper-right structure labelled 1 within the
loop). The catalytic cycle proceeds in partly aqueous
medium with a spontaneous hydrolysis of the carboxy–
anhydride linkage (step 1 → 2). This ring-opening
reaction takes place more readily than hydrolysis of the
activated enol ester moiety of 1, as suggested by previous
mechanistic investigations of acylketenimines5 and
structurally related carbodiimides6 and by prototype rate
measurements. The released activated-acyl dicarboxylate
(lower-right structure 2 within the loop) then rapidly
recyclizes to a six-membered ring anhydride (3) with
release of ketocarboxamide catalysis product (step
2 → 3). This transformation can be estimated to be very
fast. The loop is completed with consumption of
acylketenimine ‘fuel’ by reconstitution of the original
activated anhydride through an amino enol ester cyclic
anhydride (via step 3 → 4) and subsequent acyl group
migration (which is known to be rapid)7 giving back the
original anhydride (step 4 → 1). Successful motor
function requires that this latter carboxyl reactivation
(specifically, bimolecular step 3 → 4) be faster than the
anhydride hydrolysis (step 1 → 2), but slower than the
preceding cyclization (step 2 → 3). Under such circum-
stance, the catalytic cycle will proceed as indicated,
without intervention of free tricarboxylic acid, and
without activation of carboxyl groups except in the third
step. Evidence for motor operation is exhibited in
subsequent Figures.
COCH=C=NCMe3), provided in excess, which in
aqueous medium undergoes exothermic hydration to the
corresponding carboxamide (MeCOCH2CONHCMe3),
through a stepwise process that may be conveniently
monitored spectrophotometrically. That conversion is
catalyzed by a tricarboxylate ‘motor,’ HO2CCH2OCMe-
(CO2H)2, which oscillates between two anhydride forms
as a consequence of this energy expenditure. The key to
successful operation lies in the establishment of appro-
priate relative rates for individual steps within a catalytic
loop, as presented in Scheme 1. The working cycle is
indicated by the structures lettered in bold face (1–4), as
will be fully examined. The lighter-faced species
branching off from this core process represent initiation
(upper-left of the diagram), and also potential side
reactions (dashed arrows) that must be considered, but
which in actuality are not of sufficient consequence to
vitiate motor function. Evidence for successful operation
is provided subsequently. The chemical mechanism will
be inspected first.
The overall catalyzed reaction produced by the loop is
AcCH
=
C
=
NtBu ‡ H2O → AcCH2CONHtBu. For satis-
factory motor operation, the cardinal kinetic relationship
within Scheme 1 is k2, k4 >k3[AcCH=C=
NtBu]
>k1[H2O], for which adequate reaction conditions may
in fact be found, as subsequently described. Under these
appropriate circumstances, and specifically because step
But what justification can be provided for dubbing this
catalyst as a reciprocating motor? The net consequence of
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 175–182

SMALL-MOLECULE MOTOR
177
one traversal of the afore-described loop is to exchange
the geminal carboxyls of the catalyst within otherwise
equivalent cyclic anhydrides (1a → 1b). A second
traversal of the cycle returns the molecule to its original
state (1b → 1a). Energy released in hydration of acyl-
ketenimine is being directed so as to shuttle the
—OCH2CO2— moiety back and forth between otherwise
equivalent anhydride linkages, in an epimerization
reaction. As summarized in 1a and 1b, serial consump-
tion of two acylketenimines shifts the bold-faced
carbonyl and then returns it to the starting position. So
long as fuel is being consumed, this interchange process
continues; when the reserve of acylketenimine becomes
exhausted, it ceases. The motor function consists of this
repetitive and energetically forced interconversion of
transient molecular structures.
ment (solvent aqueous dioxane, 25°C, pH 5.5, concen-
tration of AcCH
NtBu initially 2 M). It certifies true
=C=
catalytic behavior, as monitored by consumption of
acylketenimine ‘fuel.’ The ratio of substrate to catalyst
employed indicates an average of 11 turnovers taking
place in the example provided, although a greater number
is attainable with a diminished amount of catalyst, which
remains active at the finish. The reaction velocity has
been observed to depend directly upon catalyst concen-
tration but, as demonstrated by the plot, it is of nearly
zero order with respect to substrate acylketenimine
through 90% of completion, which is an expected kinetic
consequence of catalytic step 1 → 2 being rate-limiting
(i.e. slow hydrolysis of cyclic anhydride intermediate, not
directly involving a substrate molecule). In order to
estimate the pertinent rate constants, the reaction
progress curve shown has been fitted to an appropriate
kinetic expression for Scheme 1 by the method of least
squares, in inverse mode, as described in the Supple-
mentary Material. The initial, nearly linear portion of the
curve corresponds to a rate constant of 0.024 Æ 0.01 sÀ1
,
referring to the substrate-loaded catalyst 1 (k1). That
value satisfactorily matches an independent rate constant
for spontaneous cyclic anhydride hydrolysis, as deter-
mined stoichiometrically for separately prepared analo-
gue 5 under the reaction conditions, kw = 0.066 Æ 0.002
Evidence that the motor operates as described
An indication of successful performance comes from
kinetic examination of substrate conversion, which
provides energy for the process. Because the acylkete-
nimine reactant absorbs light in the ultraviolet region
more strongly than does the hydration-product amide, the
water addition may be conveniently followed spectro-
photometrically. Figure 1 displays a reaction progress
curve for an individual acylketenimine hydration experi-
s
À1, a value which does not depend on pH. These
observations confirm that hydrolytic ring opening is the
slow step during catalysis. Towards the terminus of the
trace presented in Fig. 1 a serious curvature is noted, as
the reactant becomes exhausted and the ultraviolet
absorption of product levels off at its final value.
Depletion of substrate has led to a change in rate-limiting
step, attributable to the bimolecular reaction of Ac-
CH
=
C
=
NtBu ‡ 3 → 4 becoming slower than solvolytic
step 1 → 2, owing to dilution. The curve-fitting process
alluded to previously yields a rate constant of
À1
0.46 Æ 0.02 M
sÀ1 for the final decay of substrate
(k3). This value adequately matches an independently
measured rate constant of 0.125 Æ 0.002 M sÀ1 for the
1
bimolecular reaction of non-catalytic model imidocar-
boxylic acid 6 with the substrate acylketenimine under
reaction conditions approximating that of the catalytic
reaction (aqueous dioxane, pH 5.5). The initial, zero-
order portion of the curve presented in Fig. 1, comprising
90% of the reaction, is not perfectly straight. This is a
consequence of a competing spontaneous hydration of
acylketenimine, not involving the catalytic cycle, as may
be detected in a blank reaction. The fitted rate constant
for this latter process is 0.00017 Æ 0.00003 sÀ1, a value
similar to that measured for the control reaction in
absence of tricarboxylate catalyst, which may also
include other minor competing reactions of substrate as
well. For Fig. 1, by calculation 96% of the substrate
hydration occurs through the agency of the catalyst, with
the remaining 4% of the acylketenimine being hydrated
in a competing spontaneous H2O addition, as depicted by
Figure 1. Hydration consumption of substrate N-tert-
butylacetylketenimine catalyzed by HO2CCH2OCMe-
(CO2H)2, as measured spectrophotometrically. Initial sub-
strate concentration, 2.0 M; catalyst concentration, 0.18 M;
solvent, aqueous dioxane; apparent pH, 5.5; temperature,
25°C
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 175–182

178
W. L. MOCK AND K. J. OCHWAT
the dashed line at the top of the plot. Intramolecular steps
2 → 3 (k2) and 4 → 1 (k4) should be more rapid than
either of the detected catalytic steps, in order to yield the
observed kinetic behavior. Previous studies in an
analogous case have indicated that step 4 → 1 is too
rapid to measure, even at À60°C7 (the same catalytic
behavior would follow were step k4 rate limiting).
Likewise, step 2 → 3 must be fairly fast,8 because if that
were not so then all three carboxyl groups of the catalyst
would become converted to enol esters, with loss of
repetitive action as in the following demonstration. We
conclude that the kinetic behavior illustrated in Fig. 1 is
satisfactorily accommodated by the main loop of Scheme
1.
the catalytic reaction short of completion (although the
competing spontaneous hydration of ketenimine con-
tinues). Calculations based on the efficiency of inactiva-
tion suggest that methanol reacts with the cyclic
anhydride approximately eightfold more rapidly than
does water on a molar basis. Such evidence, in
conjunction with prototype rate constants previously
described for the individual steps of the catalytic loop (k1
and k3), substantiates the designated mode of operation of
the motor in alcohol-free solution.
Included on the periphery of Scheme 1 are a number of
side reactions (lighter-faced type) that called for
consideration when implementing the motor. By inspec-
tion it may be seen that each branch directed away from
the catalytic loop, designated by a dashed arrow, would in
some fashion vitiate the motor function. Since the
catalyst does repetitiously operate successfully, most
such side reactions may be excluded as kinetically
insignificant. Others can be discounted by rate measure-
ments for model reactions, as previously indicated. The
one side reaction which we are at present unable to
exclude is racemization of 3 via a bicyclic hemi-ortho-
anhydride (lower-left branch-off from cycle in Scheme
1). Latent diversions of reaction intermediates are of
consequence, for they potentially could be fatal to the
catalytic process. By way of illustration, initial attempts
to employ a carbodiimide in place of an acylketenimine
as ‘fuel’ for the motor apparently succumbed to catalyst
destruction by intramolecular N-acylation of an isourea
intermediate by the cyclic anhydride, in the manner
typified by the potential diversion of 4 in the upper-center
of Scheme 1. Hence, recourse was made to N-tert-
butylacetylketenimine, which escapes that fate (4 → 1
faster). The necessity for avoiding such faults is a major
impediment as regards the design of molecular motors,
since even minor side reactions can deplete a repetitively
operating catalyst. As another example of design
complication, control of pH presented a serious problem
encountered in developing this system. The bimolecular
activation step 3 → 4 is in actuality catalyzed by acid
Methanol deliberately incorporated into the reaction
mixture in minor amount serves as an inactivator. By
replacing H2O in step 1 → 2, it converts the catalyst into
an inoperative ester 7, a species independently isolated
(R = H) and characterized as an ineffectual catalyst.
Reaction progress curves for two such kinetic runs are
presented in Fig. 2. Inclusion of 0.0045 mole fraction of
CH3OH (vs available H2O) progressively retards cata-
lysis. Raising the mole fraction to 0.023 effectively halts
‡
(H ), as is also the competing spontaneous hydration of
Figure 2. Demonstration of catalyst poisoning in hydration
of acylketenimine as brought about by HO2CCH2OCMe-
(CO2H)2, carried out in absence and in presence of methanol
(mole fractions relative to H2O as indicated). Inset: partition-
ing of substrate consumption in the instance of CH3OH mole
fraction 0.023; approximate division into that consumed by
tricarboxylate catalyst (exhausted after half of the substrate
has been converted, owing to formation of methyl ester 7),
and into the portion of acylketenimine hydrated sponta-
neously (upper slow exponential decay, continuing una-
bated)
acylketenimine. Buffering of the kinetic reaction solu-
tions was necessary, but many common buffers of an
appropriate pKa were found to react by nucleophilic
addition to the acylketenimine (present in excess), with
obstruction of kinetic analysis. The problem was solved
by preparation of a new buffer, 1-trifluoroethyl-2,2,6,6-
tetramethyl-4-piperidinol bisulfate (8), pKa 4.54 Æ 0.03
in aqueous solution, the nitrogen of which is sufficiently
sterically hindered as not to react with the substrate, and
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 175–182

SMALL-MOLECULE MOTOR
179
within which the solubility-conferring bisulfate anion is
non-nucleophilic. By employment of this buffer, a fixed
pH for the acylketenimine-consuming reaction could be
selected that yielded satisfactory relative rates for
individual steps within the catalytic loop of Scheme 1
step of the immediately preceding cycle. This last
requirement is the unique and critically defining feature
of a nano-scale motor. Molecular chemical processes are
microscopically reversible, and no accessible counterpart
to physical momentum exists that can be exploited to
sustain motion between successive catalytic turnovers
(i.e. no ‘memory’ effect, as with a macroscopic
flywheel). The ‘relaxed state’ of any simple, mono-
phasic catalyst immediately after completion of its
energy-expenditure cycle generally will allow undoing
of work accomplished during the cycle (e.g. random
molecular motions thermally reversing the change
accomplished). It is essential to void this ‘dead time’
by having the succeeding cycle commence within a poly-
phasic motor prior to completion of the previous cycle.
That was achieved as efficiently as possible within the
foregoing example, by arranging relative rate constants:
‡
(i.e. with step k3[H ] rendered sufficiently faster than k1,
which latter conversion has no acid dependence).
DISCUSSION
Strategy for molecular motor design
k2, k4 > k3[AcCH=C=
NtBu] > k1[H2O]. In that
The potential significance of this catalytic investigation is
that it allows us to articulate formally a theory of
molecular motors, one that the previous minimal case
exemplifies. We identify three essential characteristics
exhibited by such a motor, which may serve as a guide for
the purposes of the invention of suchlike devices: the first
attribute is consumption of energy in the form of a
chemical (or other) fuel, so as to drive the process. The
second attribute is catalysis of that energy dissipation
involving two or more discrete chemical steps with
structurally distinct intermediates, such that a repetitive
molecular movement within the catalyst may be
identified (i.e. motor function corresponds to serial
conversion between intermediates). The third essential
attribute is poly-phasic convolution yielding overlap of
successive catalytic cycles, such that the first step of an
ensuing cycle commences prior to completion of the final
manner, the possibility of an undesired re-formation of
the selfsame cyclic anhydride (motor reversal) was
specifically avoided, by acylketenimine activation of
the next carboxyl prior to hydrolysis of an existing cyclic
anhydride, with that latter step only subsequently
finishing release of latent energy from the previously
consumed acylketenimine. Maintenance of a continuous
chemical dis-equilibrium within the catalyst-motor
provides a thermodynamic bias, debarring mechanical
relapse at any time during motor operation. The
microscopic reversibility that generally characterizes
molecular transformations is so counteracted, thereby
yielding the processive vectorial change that we equate
with motor function. To help visualize this, release of
Gibbs free energy (chemical potential) as a function of
time during the catalytic process in the case of the
reciprocating motor is portrayed graphically in Fig. 3.
Figure 3. Diagram showing cascade of free energy in course of catalysis by reciprocating motor
Copyright  2003 John Wiley & Sons, Ltd. J. Phys. Org. Chem. 2003; 16: 175–182

180
W. L. MOCK AND K. J. OCHWAT
Energy dissipation from each molecule of ‘fuel’ is
represented by downward-sloping arrows connecting
successive intermediates. With regard to a specific
reaction sequence designated by integer n, energy
dissipation commences (in the form of initiating step
3 → 4) before that from sequence n À1 has finished (by
ultimate step 1 → 2), and this pattern persists for
succeeding cycles (n ‡ 1, …). In the diagram, vertical
divisions correspond to the transitory arrest point within
the individual sequence detailed in Scheme 1 (created by
the loop’s rate-limiting step), with intermediates also
summarized below the time-axis of the free-energy
diagram.
We believe that any true molecular motor must in
some manner exhibit this last catalytic feature (third
attribute), as was recognized previously by Jencks in a
biochemical context.9 The challenge in molecular motor
design is to incorporate this abstract precept into
nanostructures having a desired function. Of overriding
importance is the principle that a stipulated molecular
motion should arise from a continuous discharge of
chemical potential, in consequence of a poly-phasic
overlap of energy-consumptive catalytic cycles. Concern
for the management of energy flow is as critical as
molecular structure assembly, in constructing a nano-
scale motor.
pressure. The residue was dissolved in chloroform, which
was washed successively with 10% hydrochloric acid,
saturated sodium bicarbonate and water. The solution
was then dried over anhydrous sodium sulfate. Fractional
distillation afforded 18.1 g (65.5%) of triethyl 2-oxa-
1,3,3-butanetricarboxylate as a colorless oil, b.p. 110–
120°C (0.2 mm Hg); 1H NMR (200 MHz, CDCl3), ꢀ 4.35
(2 H, s), 4.24 (4 H, q, J = 7 Hz), 4.20 (2 H, q, J = 7 Hz),
1.68 (3 H, s), 1.28 (6 H, t, J = 7 Hz), 1.27 (3 H, t,
J = 7 Hz). A solution of 18 g (65 mmol) of this triester in
60 ml of methanol was stirred at 0°C and slowly treated
with 100 ml of 10% sodium hydroxide in methanol. The
mixture was allowed to come to room temperature, and
was stirred overnight. The obtained white precipitate was
filtered, washed with methanol and dried under reduced
pressure. The white powder was dissolved in a minimum
amount of water and was acidified with sulfuric acid. The
solution was then extracted continuously with diethyl
ether for 24 h. The extract was concentrated under
reduced pressure, giving 10.9 g (87%) of 2-carboxy-
methoxy-2-methylmalonic acid as a colorless oil which
eventually crystallized, m.p. 129°C; 1H NMR (500 MHz,
CD3COCD3), ꢀ 4.38 (2 H, s), 1.66 (3 H, s); 13C NMR
(125 MHz, CD3COCD3), ꢀ 171.11, 170.07, 82.03, 64.33,
‡
21.04. Anal. Calcd for C6H9O7 [M ‡ H] : m/z 193.0348.
Found: 193.0356 (LSIMS, glycerol). The substance could
not be dried for elemental analysis without decarboxyla-
tion.
Substrate N-tert-butylacetylketenimine was obtained
from commercially available 2-tert-butyl-5-methylisoxa-
zolium perchlorate by treatment with alkali.5 Catalyst
surrogates 5 and 6 (models) and catalytically inert ester 7
and novel buffer 8 were obtained as follows.
CONCLUSION
A theory of molecular motors has been formulated and a
‘bare-bones’ specimen has been constructed and oper-
ated. The present motor does not actually accomplish any
work, other than the overcoming of entropy in generating
unidirectional conformational changes within the cataly-
tic molecule. The plainness of the example provided
ought not to obscure the conceptual advance. We regard
it as a design virtue that the intrinsic objective can be
achieved with such a compact molecule.
3-Methoxycarbonyl-3-methyl-[1,4]-dioxane-2,6-dione
(5). To a stirred and cooled (0°C) solution of 4.0 g
(20.8 mmol) of 2-carboxymethoxy-2-methylmalonic acid
in 50 ml of methylene chloride, 4.4 g (21 mmol) of
trifluoroacetic anhydride were added dropwise. The
mixture was allowed to come to room temperature and
was stirred for 30 min. The solution was then concen-
trated. While under reduced pressure, a white solid
precipitated out. The rest of the solvent and trifluoro-
acetic acid were carefully decanted and the precipitate
was washed twice with 1:1 methylene chloride–hexane
solution. The solid was then dried under reduced
pressure. We obtained 3.37 g (93%) of 3-carboxy-3-
methyl-[1,4]-dioxane-2,6-dione (3) as a white solid, m.p.
145°C; 1H NMR (500 MHz, CD3COCD3), ꢀ 4.85 (1 H, d,
J = 18 Hz), 4.78 (1 H, d, J = 18 Hz), 1.80 (3 H, s); 13C
NMR (125 MHz, CD3COCD3), ꢀ 167.60, 164.02, 163.64,
78.72, 62.63, 21.59; IR, ꢁ (cmÀ1) 1827, 1778, 1720. A
stirred and cooled (0°C) solution of 1.58 g (9.1 mmol) of
this material in 30 ml of diethyl ether was treated with
15 ml (10.7 mmol) of a 3% diethyl ether solution of
diazomethane. After the addition the mixture was
immediately subjected to solvent stripping and the
EXPERIMENTAL
Materials. The catalyst structure was readily secured
synthetically: HO2CCH2OCMe(CO2H)2 obtained from
EtO2CCH2OH and BrCMe(CO2Et)2 plus base, followed
by saponification:10
2-Carboxymethoxy-2-methylmalonic acid. A solution
of 25.3 g (0.1 mol) of diethyl 2-bromo-2-methylmalonate
and 10.9 g (0.105 mol) of ethyl glycolate in 100 ml of dry
tetrahydrofuran was stirred at À10°C. To that mixture
2.28 g (0.105 mol) of sodium hydride were added
portionwise. After the addition, the stirred mixture was
gradually brought to room temperature over a period of
1 h, followed by 1 h of reflux. Salts were separated by
filtration and the solvent was removed under reduced
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 175–182

SMALL-MOLECULE MOTOR
181
residue was then distilled, yielding 1.64 g (96%) of 3-
methoxycarbonyl-3-methyl-[1,4]-dioxane-2,6-dione (5)
as a colorless oil, b.p. 100–105°C (0.2 mm Hg); 1H
NMR (400 MHz, CD3COCD3), ꢀ 4.79 (1 H, d,
J = 18 Hz), 4.73 (1 H, d, J = 18 Hz), 3.85 (3 H, s), 1.75
(3 H, s); 13C NMR (100 MHz, CD3COCD3), ꢀ 167.36,
163.92, 163.42, 78.75, 62.57, 53.78, 21.96. This
substance allowed an independent estimate of the first-
order hydrolysis rate for 1.
symmetry present in 7 confirms that the less-hindered
carbonyl of the cyclic anhydride moiety in 3 was the site
of nucleophilic addition. The substance was an ineffec-
tual catalyst. Anal. Calcd for C7H10O7: C, 40.78; H, 4.89.
Found: C, 40.76; H, 4.98%.
1-(2,2,2-Tri¯uoroethyl)-2,2,6,6-tetramethyl-4-piperi-
dinol bisulfate (8, buffer). A solution of 15.1 g
(50 mmol) of perfluoro-n-butanesulfonyl fluoride in
20 ml of methylene chloride was stirred and chilled
to À 78°C while treated with a solution of 5.1 g
(50.5 mmol) of triethylamine and 5 g (50 mmol) of
2,2,2-trifluoroethanol in 20 ml of methylene chloride.
After addition the mixture was allowed to warm to room
temperature, and was washed with 1 M hydrochloric acid,
1 M sodium hydroxide and water. The solution was dried
over anhydrous sodium sulfate, and was concentrated
under reduced pressure. Distillation of the residue gave
17.0 g (89%) of 2,2,2-trifluoroethyl perfluoro-n-butane-
sulfonate as a colorless oil, b.p. 141°C; 1H NMR
(400 MHz, CDCl3), ꢀ 4.73 (q, J = 7 Hz). A mixture of
15.3 g (40 mmol) of this material and 6.3 g (40 mmol) of
2,2,6,6-tetramethyl-4-piperidinol was stirred at 130°C
for 2 h. The resulting material was purified by silica
chromatography with 1:1 ethyl acetate–hexane as eluent.
We obtained 4.4 g (46%) of 1-(2,2,2-trifluoroethyl)-
2,2,6,6-tetramethyl-4-piperidinol as white crystals, m.p.
2-Carboxy-2,4-dimethylmorpholine-3,5-dione (6). To
50 ml of a 1 M solution of methylamine in dioxane at 0°C
were added with stirring 20 ml of a 1 M solution of 3-
carboxy-3-methyl-[1,4]-dioxane-2,6-dione (3) in diox-
ane. After the addition was complete, external cooling
was removed and mixture was stirred for 15 min at room
temperature. Next, excess methylamine and solvent were
removed under reduced pressure. The oil obtained was
dissolved in 50 ml of diethyl ether and washed 3Â with
20 ml portions of 1 M hydrochloric acid. The solution was
then dried over anhydrous sodium sulfate and concen-
trated under reduced pressure. The resulting crude oil was
purified by silica chromatography with ethyl acetate as
eluent. We obtained 2.34 g (57%) of 2-(N-methylamino-
carbonylmethoxy)-2-methylmalonic acid as a colorless
1
oil; H NMR (400 MHz, CD3COCD3), ꢀ 4.16 (2 H, s),
2.76 (3 H, s), 1.65 (3 H, s). A solution of 1.02 g (5 mmol)
of this material in 20 ml of diethyl ether was stirred at
room temperature while it was treated with 1.07 g
(5.1 mmol) of trifluoroacetic anhydride. The mixture
became warm, and was stirred at room temperature for
20 min. Then, solvent was removed under reduced
pressure and the concentrate obtained was purified by
silica chromatography with ethyl acetate as eluent. We
obtained 0.58 g (62%) of 2-carboxy-2,4-dimethylmor-
pholine-3,5-dione (6) as a colorless oil: 1H NMR
(500 MHz, CDCl3), ꢀ 4.64 (1 H, d, J = 17.5 Hz), 4.54 (1
H, d, J = 17.5 Hz), 3.23 (3 H, s), 1.80 (3 H, s); 13C NMR
(125 MHz, CDCl3), ꢀ 171.27, 169.43, 168.92, 65.20,
26.49, 22.65. This substance allowed an independent
estimate of the rate of activation for 3 (for bimolecular
1
128°C; H NMR (400 MHz, CDCl3), ꢀ 3.99 (1 H, m),
3.13 (2 H, q, J = 9 Hz), 1.86 (2 H, d, J = 12 Hz), 1.46 (2
H, t, J = 11.5 Hz), 1.15 (6 H, s), 1.06 (6 H, s); 13C NMR
(100 MHz, CDCl3), ꢀ 63.89, 57.15, 49.92, 46.82, 34.51,
22.79. A solution of 2.4 g (0.01 mol) of this material and
1.8 g (0.0113 mol) of pyridine–sulfur trioxide complex
was refluxed in 25 ml of toluene for 15 h. The crude
product precipitated from the chilled reaction mixture
and was filtered. Recrystallization from water gave 2.9 g
(91%) of 1-(2,2,2-trifluoroethyl)-2,2,6,6-tetramethyl-4-
1
piperidinol bisulfate as white crystals, m.p. 222°C; H
NMR (500 MHz, D2O), ꢀ 4.74 (1 H, m), 4.10 (2 H, q,
J = 9 Hz), 2.34 (2 H, d, J = 12 Hz), 1.94 (2 H, t,
J = 13 Hz), 1.43 (12 H, s); 13C NMR (125 MHz, D2O),
ꢀ 69.79, 69.15, 46.05, 42.73, 39.82, 28.88, 21.30. Anal.
Calcd for C11H20F3NO4S: C, 41.37; H, 6.31; N, 4.39.
Found: C, 41.23; H, 6.32; N, 4.36%.
reaction with AcCH=C=
NtBu). It was converted into a
dicyclohexylamine salt for elemental analysis, white
crystals, m.p. 147°C. Anal. Calcd for C19H32N2O5: C,
61.93; H, 8.75; N, 7.60. Found: C, 61.82; H, 8.89; N,
7.60%.
Measurements. Estimates of pH for kinetic runs are
glass-electrode meter readings in buffered aqueous
dioxane (1:1, v/v), referenced to calibration standards
in water alone. Catalytic kinetic runs as represented by
Figs 1 and 2 were monitored spectrophotometrically at a
suitable wavelength employing a 1 mm cell, with
absorption measurements serially recorded at 6 s inter-
vals subsequent to mixing of reactants. In most instances
recrystallized 3 was the form in which the catalyst was
introduced to initiate reaction. Solvent was maintained at
50% aqueous in runs that were 2 M in substrate, by
reducing appropriately the amount of dioxane co-solvent.
2-(Methoxycarbonylmethoxy)-2-methylmalonic acid
(7). A solution of 0.174 g (1 mmol) of 3-carboxy-3-
methyl-[1,4]-dioxane-2,6-dione (3) in 10 ml of methanol
was stirred at room temperature for 3 min. Removal of
excess methanol under reduced pressure resulted in
0.204 g (99%) of 2-(methoxycarbonylmethoxy)-2-
methylmalonic acid (7) as a colorless oil; 1H NMR
(400 MHz, CDCl3), ꢀ 10.20 (2 H, s), 4.31 (2 H, s), 3.73 (3
H, s), 1.64 (3 H, s); 13C NMR (100 MHz, CDCl3), ꢀ
171.61, 171.40, 81.68, 64.11, 52.47, 20.64. The spectral
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 175–182

182
W. L. MOCK AND K. J. OCHWAT
Acc. Chem. Res. 2001; 34: 514–522; (c) van Delden RA, Koumura
N, Harada N, Feringa BL. Proc. Natl. Acad. Sci. USA 2002; 99:
4945–4949; (d) Feringa BL, Koumura N, Van Delden RA, Ter
Wiel MKJ. Appl. Phys. 2002; 75: 301–308; (e) Wang H, Oster G.
Appl. Phys. 2002; 75: 315–323; (f) Sestelo JP, Kelly TR. Appl.
Phys. 2002; 75: 337–343.
2. (a) Vale RD, Milligan RA. Science 2000; 288: 88–95; (b)
Mahadevan L, Matsudaira P. Science 2000; 288: 95–99.
3. (a) Kelly TR, De Silva H, Silva RA. Nature (London) 1999; 401:
150–152; (b) Koumura N, Zijlstra RWJ, van Delden RA, Harada
N, Feringa BL. Nature (London) 1999; 401: 152–155.
4. (a) Kelly TR, Tellitu I, Sestelo JP. Angew. Chem., Int. Ed. Engl.
1997; 36: 1866–1868; (b) Kelly TR, Sestelo JP, Tellitu I. J. Org.
Chem. 1998; 63: 3655–3665.
5. (a) Woodward RB, Woodman DJ. J. Am. Chem. Soc. 1966; 88:
3169–3170; (b) Woodward RB, Woodman DJ. J. Am. Chem. Soc.
1968; 90: 1371–1372; (c) Woodman DJ, Murphy ZL. J. Org.
Chem. 1969; 34: 3451–3456; (d) Woodman DJ, Davidson AI. J.
Org. Chem. 1970; 35: 83–87; (e) Woodman DJ, Davidson AI. J.
Org. Chem. 1973; 38: 4289–4295.
Hydrolysis of control substance 5 was followed similarly
by observing an exponential decay in the spectral
absorption of the anhydride (265 nm) following addition
to aqueous dioxane, yielding an independent estimate of
the catalytically limiting rate constant. The velocity did
not depend on pH. The characteristic rate constant of
carboxyl activation was separately estimated by employ-
ing control substance 6 to combine with substrate
acylketenimine under pseudo-first-order conditions (6
in excess) by the same technique. The addition reaction
was found to be acid catalyzed for pH values that provide
an ionized carboxyl of model 6 (measured pKa
2.65 Æ 0.05 in aqueous dioxane by pH dependence of
13C NMR); the inferred mechanism entails addition of the
carboxylate anion to an transiently protonated acyl-
ketenimine (pKa <2 for conjugate acid of substrate).6
Throughout, tolerances listed are standard errors from
least-squares analysis.
6. Ibrahim IT, Williams A. J. Chem. Soc., Perkin Trans. 2 1982;
1459–1466.
7. (a) Niederhauser A, Neuenschwander M. Helv. Chim. Acta 1973;
56: 1318–1330; (b) Niederhauser A, Neuenschwander M. Helv.
Chim. Acta 1973; 56: 1331–1344; (c) Gais H-J, Lied T. Angew.
Chem., Int. Ed. Engl. 1978; 17: 267–268; (d) Gais H-J. Angew.
Chem., Int. Ed. Engl. 1978; 17: 597–598.
Supplementary material
8. (a) Gaetjens E, Morawetz H. J. Am. Chem. Soc. 1960; 82: 5328–
5335; (b) Bruice TC, Bradbury WC. J. Am. Chem. Soc. 1968; 90:
3808–3812.
9. (a) Jencks WP. Annu. Rev. Biochem. 1997; 66: 1–18; (b) Jencks
WP. Biosci. Rep. 1995; 15: 283–287; (c) Jencks WP. Protein Sci.
1994; 3: 2459–2464; (d) Urry DW. Angew. Chem., Int. Ed. Engl.
1993; 32: 819–841.
10. Vijverberg, CAM, Peters JA, Bove´e WMMJ, Vroon H, Kieboom
APG, van Bekkum H. Recl. Trav. Chim. Pays-Bas 1983; 102: 255–
263.
Derivation of rate equations employed in fitting kinetic
data is available at the epoc website at http://www.
wiley.com/epoc.
REFERENCES
1. (a) Feringa BL. Acc. Chem. Res. 2001; 34: 504–513; (b) Kelly TR.
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 175–182