The Journal of Organic Chemistry
Article
This leads to an estimate for the carbon−carbon bond-breaking
step of k7 ≈ kobs × 108 = 105 s−1 at 60 °C and between 103 and
104 s−1 at 25 °C.
Scheme 4. In Situ Decomposition of Protonated Carbonic
Acid
The hydrolytic decarboxylation pathway permits a more rapid
reaction in acid solution than does the neutral dissociative
reaction mechanism because the latter requires rate-determining
formation of the minor tautomer by protonation of the conjugate
base on carbon at low acid concentrations. An important con-
sequence is that the microscopic reverse reaction in acidic
solution is carboxylation of indole. This should occur via a
Friedel−Crafts reaction of PCA, a process in which water and a
proton would be catalytic in the overall addition of CO2.
We propose, by extension, that it is also likely that a Lewis acid
could provide the necessary activation to produce a complex
of carbonic acid analogous to PCA. Interestingly, Lewis acid
promoted carboxylation reactions were demonstrated by Olah
and co-workers29 and recent studies show that regioselective
carboxylations of derivatives of both pyrrole and indole are
possible.30 Since protonated CO2 is too high in energy to exist as
a reasonable intermediate in decarboxylation (and hence, in
carboxylation), it is possible that similar Lewis acid complexes of
CO2 would be equally high in energy. On the other hand,
reactions involving Lewis acid complexes of carbonic acid would
be analogous to the more energetically feasible PCA.
Further insight into the hydrolytic decarboxylation mechanism
comes from the solvent kinetic isotope effect (SKIE) kH/kD = 1.7
at H0 = 0.98 for the decarboxylation of indole-3-carboxylic acid.
In concentrated acid solutions, the SKIE decreases to kH/kD = 1
at H0 = −2.30. This pattern of SKIE values is consistent with the
presented mechanisms. In dilute acid solutions, the mechanism
of decarboxylation is dissociative, leading to formation of CO2
from the neutral tautomer of the reactant. In this mechanism,
formation of the reactive tautomer (1*) is rate-determining via
proton transfer (where SKIE is significant). In more acidic
solutions the hydrolytic mechanism becomes dominant, with
rate-limiting carbon−carbon bond-breaking of the protonated
hydrate (2H+) (and therefore a reduced SKIE).
Once in the hydrated form, the electronic effects of the
carboxyl group on the protonation of the aromatic ring are
eliminated and replaced by that of an ortho acid (−C(OH)3) (2).
This substituent should have a small effect on the pKa for
C-protonation of the indole ring based on the effects of similar
orthoesters (σC(OMe)3,meta = −0.03 and σC(OMe)3,para = −0.04).26
This suggests that the pKa of the reactive intermediate (2) at
equilibrium would be similar to that of unsubstituted indole
(pKa = −2.4).27 Therefore, the reaction must proceed via initial
protonation of the carboxyl group to promote hydration. Once
hydrated, protonation of the aromatic ring is dramatically
facilitated, unlocking the pathway for the subsequent release of
PCA.
A simplified rate expression (eq 3) can be used to represent the
overall hydrolytic decarboxylation pathway where the conjugate
acid of indole-3-carboxylic (1H+) acid is the initial substrate.
Since the solvent isotope effect indicates that proton transfer
steps are not rate-determining, we can estimate the rate constant
for cleavage of the carbon−carbon bond (k7) in concentrated
acid solutions for indole-3-carboxylic acid (eq 4).
CONCLUSION
■
We have reported kinetic analysis for the decarboxylation of
indolecarboxylic acids over a wide range of solution acidities.
In dilute acid solutions, the rate-determining step involves
formation of the zwitterionic intermediate that is capable of
losing CO2 directly. In concentrated acid solutions, a route for
acid catalysis leads to the addition of water to the carboxyl group,
resulting in expulsion of the energetically feasible PCA. In this
hydrolytic mechanism, the rate-determining step is carbon−
carbon bond cleavage. By investigating the hydrolytic decarbox-
ylation pathway for indolecarboxylic acids, we have been able to
show that the expulsion of water from the addition intermediate
has a lower barrier than that of carbon−carbon bond cleavage to
form PCA.
EXPERIMENTAL SECTION
■
Commercial indole-2-carboxylic acid, indole-3-carboxylic acid, and
potassium chloride were used as purchased. Buffers and acid solutions
were made from reagent-grade chemicals with distilled water or
deuterium oxide.
v = kobs[1H+] = k7[2H+]
(3)
k7 = kobs[1H+]/[2H+]
Kinetics of Decarboxylation. The rates of decarboxylation of
indole-2-carboxylic acid and indole-3-carboxylic acid were measured for
reactions in hydrochloric acid. The rate of decarboxylation of indole-3-
carboxylic acid was also measured in 0.1 M buffers of chloroacetic acid,
acetic acid, and monobasic phosphate where the ionic strength (I) of all
buffered solutions was maintained at I = 1.0 M by addition of potassium
chloride. All measurements were carried out in solutions maintained at
60 °C. The reaction was monitored by the decrease in absorbance at
300 nm (indole-2-carboxylic acid) or 291 nm (indole-3-carboxylic acid)
with a UV−vis spectrometer, whose cell compartment was controlled
within 0.1 °C. Data were collected with an interfaced computer, and
the observed first-order rate constants (Table 1) were calculated from
nonlinear regression fitting to the integrated first-order rate expression.
For slow reactions, the method of initial rates was used to calculate rate
constants. For determination of solvent kinetic isotope effects, reactions
were conducted using comparable concentrations of hydrochloric acid
(in water) and deuterium chloride (in deuterium oxide).
(4)
The value of the equilibrium constant for [1H+]/[2H+] can be
estimated from a thermodynamic cycle between (1H+) and
(2H+) that follows K1, K5, and K6 as shown in Scheme 2. As noted
above, the pK1 = 0.4 for O-protonated indole-3-carboxylic acid.
An estimate of the extent of hydration of the carboxylic acid is
possible by extension of previously calculated values. The
equilibrium constant for addition of water to the carboxyl group
of methyl glycine (N-protonated) was estimated by Guthrie
and Cullimore28 to be about10−6, which is a good model for
indolecarboxylic acids based on the location of the nitrogen
substituent. Therefore, the log of the equilibrium constant
(K5) for hydration is ca. −6 while the pK6 of the C-protonated
indole derivative is approximately −2.4. Therefore, the overall equi-
librium [1H+]/[2H+] is approximately equal to 0.4−6−2.4 = −8.
6508
dx.doi.org/10.1021/jo301032f | J. Org. Chem. 2012, 77, 6505−6509