Communications
micelle, thus forming an ion pair with 1. Therefore, as 1 is
chain of 3. Interestingly, the reaction rates appeared to
plateau with the 1C series; reactions were not accelerated
further when a longer acyl chain was used (see series 1D).
As 1C and 1D form micelles under these reaction
conditions, the large rate enhancements observed with these
two carboxylates are attributable to the micellar effect
(Scheme 1, step 2). Dehydrocondensing agents 3b–d have
alkyl chains with eight carbon atoms or more, which puts
them on the borderline in their ability to be incorporated in
micelles. Moderate accelerations in the reactions of 1B with
3c or 3d do not arise from micelle formation but rather from a
disordered aggregation owing to the hydrophobic effect. This
explanation is based on the fact that 3c, which is poorly
soluble in water, does not form micelles under the reaction
conditions. In fact, a white turbidity appeared upon addition
of 3c or 3d to initiate the reaction with 1B, whereas no
turbidity was observed with 1C and 1D, presumably as a
result of mixed micelle formation.
closely concentrated around 3, it can readily attack the
triazinyl group of 3 with concomitant liberation of 2, which
can be recycled to step 1. Step 2 is accelerated by the
preorientational effect as well as the local concentration
effect of the reactants at the micellar interface. The resulting
activated triazinyl ester 4 undergoes aminolysis with butyl-
amine (5), which is expected to be partitioned mainly in the
aqueous phase, to form the amide 6 (step 3). As this step is
[
10]
known to be much faster than hydrolysis in water, it is not
rate-limiting and thus it appeared to be unimportant to ensure
that it occurs in the micelle rather than at the aqueous phase.
We employed four sodium carboxylates 1A–D with alkyl
chains of different lengths. On the basis of the critical micelle
concentration and Kraft points of these compounds, 1A and
1
B should form a simple molecular dispersion phase in which
these electrolytes are homogeneously dissolved in a dissoci-
ated form, whereas 1C and 1D should form a micellar phase
[
13,14]
at a concentration of 15 mm (258C).
If the bimolecular
The substrate concentration dependence of the reaction
rate in the micellar system when using 3b was also examined.
The reaction rate was found to be independent of the
concentration of 1C (15, 30, and 60 mm)whereas the reaction
rate showed a linear relationship to the concentration of
butylamine 5 (5, 10, 15, and 20 mm). The results indicate that
step 3 becomes the rate-determining step in the micellar
reaction rate constants between carboxylates 1 and 3 are
assumed to be independent of their chain length, the observed
change in reaction rate can be attributed to the micellar
effect.
First, we examined the reaction of carboxylates 1A–D
[
15]
with condensing agents 3a–d to estimate the acceleration of
the rate-determining step (step 2)in the micelle. The reaction
should be close to first-order with respect to 3. The reaction
was conducted using 1 (15 mm), 5 (as hydrochloride, 20 mm),
and 3 (1.5 mm)in phosphate buffer (pH 8, 20 m m)containing
[
17]
system instead of step 2.
The rate of aminolysis of the
triazinyl ester with 5 dissolved in the aqueous phase is
independent of the length of the acyl chain. Thus, in the
micellar system, there was no significant difference between
1C and 1D, despite the difference in chain length (six carbon
atoms).
[
16]
MeOH (3%) at 258C (Table 1). The pseudo-first-order rate
[
a]
Competitive reactions between 1A and 1B with either 3a
or 3b afforded amides 6A and 6B (42:58 or 25:75, respec-
tively)after the reaction mixture was stirred for 4 h at room
temperature (Table 2). In contrast, the competitive reaction
Table 1: Relative rates for the stoichiometric reaction of 1 and 3.
Table 2: Substrate selectivity in the competitive reaction between two
carboxylates in the stoichiometric system.
Carboxylates
3
t [h]
Yield[%]
Ratio
1
1
A vs. 1B
A vs. 1B
3a
3b
3b
4
4
1
16
24
88
6A/6B 42:58
6A/6B 25:75
6A/6C 0.4:99.6
3
1A
1B
1.1
3.0
280
340
1C
56
1200
860
1D
[
a]
3
3
3
3
a
b
c
1.0
0.7
21
30
63
830
840
690
1A vs. 1C
d
1400
between 1A and 1C with 3b proceeded within 1 h in both
good yield (88%)and high selectivity (0.4:99. 6. ) These
selectivities are in good agreement with the relative reaction
rates shown in Table 1.
À3
À1
[a] Pseudo-first-order rate constant: k=1.010 min
.
constants for the reaction with respect to 3 were calculated
based on the amount of the amide 6 produced. The relative
rates were normalized to the reaction rate of 3a with 1A (rate
defined as 1). In the reactions of 3a, which has a short alkyl
chain (ethyl group), no rate acceleration was observed in the
reaction with octanoate 1B, which has an alkyl chain that is
four carbon atoms longer than that of butyrate 1A, whereas
the reaction with laurate 1C, whose alkyl chain is elongated
by an additional four carbon atoms, was accelerated by a
factor of 56. In a series of reactions with 1C, the reaction rate
increased up to 1400 times by elongation of the ester alkyl
Finally, the catalytic reaction of N,N-dimethylglycine alkyl
esters 2 and DMT-Cl, which generate condensing agent 3
in situ, was also found to be accelerated by a factor of 140 in
micelles (Table 3). The moderate acceleration in the catalytic
system relative to the stoichiometric system may be attributed
to the generation of 3 (step 1)which may become the rate-
determining step after acceleration of step 2. This catalytic
system can be considered as an acyl transferase model for
preparing lipid molecules.
The amide-forming reaction in the cyclodextrin-based
artificial enzyme was accelerated by a factor of 13 because of
7256
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7254 –7257