General-base catalysed hydrolysis and nucleophilic substitution of activated amides in aqueous solutions

Article abstract of DOI:10.1002/poc.607

The reactivity of 1-benzoyl-3-phenyl-1,2,4-triazole (1a) was studied in the presence of a range of weak bases in aqueous solution. A change in mechanism is observed from general-base catalysed hydrolysis to nucleophilic substitution and general-base catalysed nucleophilic substitution. A slight tendency is also observed for the more hydrophobic general bases to show higher reactivity towards 1a. Aspartame is an effective nucleophile, possibly because nucleophilic substitution is subject to intramolecular general-base catalysis. A general conclusion derived from the present results is that unexpected rate effects can only be rationalised provided that the detailed reaction mechanisms are well understood. Copyright

Full text of DOI:10.1002/poc.607

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2003; 16: 438–449
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.607
General-base catalysed hydrolysis and nucleophilic
substitution of activated amides in aqueous solutions
†
1
2
1
Niklaas J. Buurma, Michael J. Blandamer and Jan B. F. N. Engberts *
1
Physical Organic Chemistry Unit, Stratingh Institute, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands
Department of Chemistry, University of Leicester, Leicester LE1 7RH, UK
2
Received 7 October 2002; revised 30 October 2002; accepted 30 October 2002
ABSTRACT: The reactivity of 1-benzoyl-3-phenyl-1,2,4-triazole (1a) was studied in the presence of a range of weak
bases in aqueous solution. A change in mechanism is observed from general-base catalysed hydrolysis to nucleophilic
substitution and general-base catalysed nucleophilic substitution. A slight tendency is also observed for the more
hydrophobic general bases to show higher reactivity towards 1a. Aspartame is an effective nucleophile, possibly
because nucleophilic substitution is subject to intramolecular general-base catalysis. A general conclusion derived
from the present results is that unexpected rate effects can only be rationalised provided that the detailed reaction
mechanisms are well understood. Copyright  2003 John Wiley & Sons, Ltd.
KEYWORDS: hydrolysis; changes in mechanism; general-base catalysis; general-acid catalysis; nucleophilic
substitution
INTRODUCTION
lowed by loss of the (substituted) 1,2,4-triazole leaving
group (Scheme 2).
The hydrolysis reaction of the activated amide 1-benzoyl-
In this mechanism, the nucleophilic water molecule is
replaced by a stronger nucleophile. The similarity with
the water-catalysed hydrolysis goes even further, as the
nucleophilic substitution reaction can also be catalysed
by general acids and bases. Some of the possible
catalysed reaction pathways for nucleophilic substitu-
3
-phenyl-1,2,4-triazole (1a) is general-base catalysed
1
(
Scheme 1).
In highly aqueous solutions, the concentration of water
is sufficiently high for water to act (detectably) as both a
general base and a nucleophile. Hence, in the absence of
other general bases, the water-catalysed (i.e. pH-
independent) hydrolysis is the sole reaction. In the
presence of sufficiently basic cosolutes, the water-
catalysed reaction is unimportant. More basic cosolutes
are much more effective catalysts for hydrolysis than
water. Consequently, despite the relatively low molality
of added general bases, the general-base catalysed
hydrolysis pathway competes with the water-catalysed
pathway. It is stressed that even though in the water-
catalysed reaction the second water molecule in the
2–4
tion
are illustrated in Scheme 3, there are many
different pathways for reactions of nucleophiles with
amides and esters. The exact pathway depends on factors
such as leaving group ability, nucleophilicity and solvent.
Here, we shall only discuss those reactions pathways,
that, we contend, are relevant for the system under study.
If nucleophilic attack is the rate-determining step, the
reaction can be catalysed by a general base or by
hydroxide, deprotonating the nucleophile. Similarly, the
negative charge developing on the amide carbonyl can be
stabilized by general acids, resulting in general-acid
catalysis. However, if expulsion of the 3-phenyl-1,2,4-
triazole from a protonated tetrahedral intermediate is rate
determining, the reaction could be general-base cata-
lysed. Further, departure of the leaving group may be
general-acid catalysed.
In this intricate play of reactivity, reactants can assume
many different roles, resulting in a series of related but
different reaction pathways. It is therefore of paramount
importance to have a detailed understanding of these
reaction pathways if we are to fully understand the
observed rate effects induced by general bases. We
therefore studied the effect of a range of general bases on
activated complex (i.e. B = H O, Scheme 1) acts as a
2
general base, a distinction is drawn between the water-
catalysed reaction and general-base catalysed reaction.
Increased basicity of cosolutes usually leads to a
concomitant increase in nucleophilicity. This increase in
nucleophilicity provides an alternative reaction pathway:
nucleophilic attack on the carbonyl functionality, fol-
*
Correspondence to: J. B. F. N. Engberts, Physical Organic Chemistry
Unit, Stratingh Institute, University of Groningen, Nijenborgh 4, 9747
AG Groningen, The Netherlands.
E-mail: j.b.f.n.engberts@chem.rug.nl
†
Dedicated to Professor S. Kobayashi on the occasion of his retirement
from academia.
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

REACTIVITY OR ACTIVATED AMIDES IN AQUEOUS SOLUTIONS
439
Scheme 1.
Scheme 2.
Scheme 3.
the reactivity of 1a in order to identify possible competing
RESULTS AND DISCUSSION
reaction pathways. The results are used in a reinterpreta-
5
tion of the previously reported rate-accelerating effects
General-base catalysis by carboxylate ions and
water
of some a-amino acids on the hydrolysis of 1a.
We have previously shown that encounter complexes
between reactive probes and added inert cosolutes can be
stabilised by hydrophobic interactions. However, if the
The effect was determined of different general bases on
the (pseudo-)first-order rate constant of hydrolysis of 1a.
In all cases, at low molality and constant buffer ratio, the
increase in rate constants was linear with increasing
molality of added general base (Fig. 1).
6
cosolute is not inert, formation of an encounter complex
constitutes the first step for bimolecular (or higher
molecularity) reactions. Hence, if a cosolute reacts with
the reactive probe, or catalyses the reaction of the reactive
probe, more hydrophobic cosolutes are expected to show
slightly higher reactivity than hydrophilic cosolutes with
identical functional groups. Similarly, in a linear free
energy relationship, more hydrophobic reactive cosolutes
are expected to show deviations towards higher reactivity
provided that hydrophobic interactions are not of similar
importance in both processes.
In order to obtain the sole effect by the general bases,
the effect of the conjugate general acid on the observed
rate has to be calculated. In the absence of general base,
the general acids were found to decrease the rate of the
water-catalysed reaction. We attribute this decrease to
blocking by the cosolute of the reaction centre from
6–8
attack by water. We contend that a similar inhibition of
reaction occurs for the general-base catalysed hydrolysis.
Hypothetically, in the absence of inhibition by
cosolute and in the pH-range in which the reaction
without added general bases is only water-catalysed, the
rate constant is described by the equation
The molecular picture of an unreactive cosolute
blocking the reactive centre of the activated amide from
6–8
attack by water proposed by us before is expected to be
equally valid in general-base catalysed hydrolysis and
(
catalysed) nucleophilic substitution. In buffer solutions
kꢀm † ˆ kꢀm ˆ 0† ‡ m k
ꢀ1†
of general bases in which the conjugate general acid is
also present, intriguing compensating effects are possi-
ble. Increasing the hydrophobicity of the general base
will lead to an increased efficiency in general-base
catalysis (see above) and to a concomitant increase in the
rate-retarding effect of the conjugate general acid.
b
c
b b
where m is the molality of general base, k is the
b
b
(pseudo-)second-order rate constant for catalysis by the
general base and k(m = 0) is the (pseudo-)first-order
c
rate constant in the absence of cosolute. Previously, Eqn.
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

4
40
N. J. BUURMA, M. J. BLANDAMER AND J. B. F. N. ENGBERTS
Equation (3) is a simpler form of Eqn. (2):
ꢀ
ꢁ
kꢀm †
c
ln
ˆ amc
ꢀ3†
kꢀm ˆ 0†
c
where m is the molality of inert cosolute c and a
c
quantifies the rate effect induced by cosolute c. Equation
(3) can be rewritten as
amc
kꢀm † ˆ kꢀm ˆ 0†e
ꢀ4†
c
c
The conjugate acid of the general-base catalyst is the
only inert cosolute. Hence, for the present case, m is the
c
Figure 1. Effect of added ethanoate/ethanoic acid (~),
butanoate/butanoic acid (*) and chloroethanoate/chloro-
ethanoic acid (*) buffers (all buffer ratios 3:1 and pH values
of 5.15 Æ 0.10, 5.20 Æ 0.10 and 3.20 Æ 0.10, respectively;
m and m are the molalities of acid and base, respectively)
molality of conjugate acid m . If Eqn. (1), describing the
a
kinetics of reaction without inhibition, is substituted into
Eqn. (4), describing the inhibiting effect of added
unreactive cosolutes, we obtain
a
b
on the hydrolysis of 1a
maa
kꢀm ; m † ˆ ‰kꢀm ˆ 0† ‡ m k Še
ꢀ5†
a
b
c
b b
(
2) has been employed to describe the effects of
where a is the (rate-retarding) effect of the acidic form of
the cosolute on both the water-catalysed and general-base
catalysed reaction. The kinetic data were fitted to Eqn.
inert cosolutes on the water-catalysed hydrolysis reac-
tions:
(
5). The results together with pK values are summarised
a
ꢀ
ꢁ
in Table 1.
Fitting data to an equation related to Eqn. (5), in which
kꢀm †
2
c
ln
ˆ
‰g À g Šm À NꢀM m ꢀ2†
cx
c6
c
1
c
2
0
kꢀm ˆ 0†
RTm
c
the acidic form is assumed to inhibit only the water-
catalysed reaction, leads to less satisfactory fits. Intro-
ducing an additional parameter to distinguish between
rate-retarding effects on the water-catalysed and on the
base-catalysed hydrolysis seems unjustifiable.
Employing the theory developed in previous work, the
effect of the formation of encounter complexes from
reactant(s) and an inert cosolute can be expressed in
terms of G(c) values (for a brief review, see Ref. 11).
These G(c) parameters were calculated using Eqn. (6) [cf.
Eqns (2) and (3)]:
where k(m ) is the (pseudo-)first-order rate constant in an
m molal aqueous solution of inert cosolute c, k(m = 0)
c
c
c
the rate constant in the absence of added cosolute, R the
gas constant and T the absolute temperature. Signifi-
cantly, [g À g ] is the difference in interaction Gibbs
cx
c
6ˆ
energies between the cosolute c and the reactants x on the
one hand and the activated complex ≠ on the other. M is
l
the molar mass of water, N is the number of water
molecules involved in the rate-determining step and f is
the practical osmotic coefficient for the aqueous solution
2
where the molality of added solute is m . For the water-
catalysed hydrolysis, N = 2. Further, the solutions are
a ˆ
Gꢀc† À NꢀMw
ꢀ6†
c
2
0
RTm
very dilute and hence, f can be taken as unity; m is the
0
À1
(
hypothetical) ideal reference state and corresponds to
where m
is 1 mol kg , N is the number of water
0
À1
1
mol kg . The term [g À g ] is denoted as G(c).
molecules incorporated in the activated complex, f is the
cx
c
6ˆ
a,c
Table 1. General-base catalysed hydrolysis of 1a at 298.2 K
Parameter
Ethanoate
Chloroethanoate
Butanoate
b
9
10
9
pKa
4.76
2.86
4.82
À4 À1
À1
kb (10
a (kg mol
s
mol kg)
16.63 (0.47)
À0.30 (0.03)
13.2 (0.2)
2.08 (0.09)
À0.11 (0.02)
13.0 (0.1)
20.61 (0.22)
À0.56 (0.01)
13.0 (0.1)
À1
)
À4 À1
k(m = 0) (10
s )
c
a
Numbers in parentheses are standard errors based on a least-squares fit of kinetic data using Eqn. (5).
b
All pK
a
values, except those for the a-amino acids and hydroxylamine, were obtained from a literature search using the Beilstein Crossfire system.
Available values were collected, assessed and averaged. Only one, representative, reference is given for every pK
a
.
c
À1
À1
À1
À2
2
Second-order and third-order rate constants are given with units s mol kg and s mol kg , respectively. These unconventional units facilitate
À1
À1
3
comparison with and introduction of G(c) values. The exceptions are rate constants for hydroxide-catalysed hydrolysis, given in s mol dm , as
hydroxide concentrations were calculated from the pH of the solutions.
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

REACTIVITY OR ACTIVATED AMIDES IN AQUEOUS SOLUTIONS
441
practical osmotic coefficient of water and M is the molar
w
mass of water. N was set to 2, despite the fact that only
one water molecule is involved in the rate-determining
step for the general-base catalysed reaction. The error
introduced in this way is small as the term NfM is small
w
and the contribution of general-base catalysis to the
overall rate constant is generally less than 50%. The G(c)
values for ethanoate and butanoate of À 327 Æ 37
À2
and À649 Æ 13 J kg mol , respectively, correspond to
the corresponding values for the water-catalysed reac-
tion. This correspondence supports the hypothesis that
rate-retarding effects are similar for water-catalysed and
for general-base catalysed hydrolysis and is consistent
with the notion that the rate-retarding effects are (largely)
caused by blocking of the reaction centre.
Figure 2. Brùnsted plot for general-base catalysed hydrolysis
of 1a
Using the results given in Table 1, a Brønsted plot was
constructed for carboxylate general bases and water [Fig.
À2
2
, Eqn. (7)].
protonated form; G(c) = À1869 and À234 J kg mol ,
5,14
respectively.
In their deprotonated forms, they can
log k ˆ 0:29pK À 4:24
ꢀ7†
b
a
function as general bases. It was expected that especially
phenylalaninamide, being strongly rate retarding in the
protonated state and hence forming rather stable
encounter complexes, would be an effective catalyst for
hydrolysis.
The Brønsted b is 0.29 Æ 0.05, slightly lower than the
value for activated amides without the phenyl substituent
1,12,13
in the triazole ring.
Interestingly, in terms of general-base catalysis, com-
parison of ethanoate and butanoate shows that the latter is
slightly more effective. The observed difference cannot
Indeed, the reactivity of 1a in the presence of
unprotonated alaninamide and phenylalaninamide is
high. The pH dependence of the reactivity of 1a in the
presence of alaninamide and phenylalaninamide (Fig. 3)
indicates that high reactivity is indeed associated with
deprotonation of the general base.
The results for alaninamide and phenylalaninamide
were fitted to Eqn. (8), using a non-linear least-squares
procedure:
be explained on the basis of its slightly higher pK . This
a
pattern could be caused by the more hydrophobic nature
of butanoate, resulting in additional hydrophobic stabil-
isation of the encounter complexes formed between 1a
and butanoate in the initial stages of the activation
process.
kꢀm ; pH† ˆ kꢀm † ‡ k  10pHÀ14
c
c
OH
From base-catalysed hydrolysis to nucleophilic
substitution
k m
c
1
a
‡ 1
ꢀ8†
‡ 10pKaÀpH
Phenylalaninamide hydrochloride and alaninamide hy-
drochloride are both rate retarding in their fully
The first term on the right-hand side, k(m ), is the rate
c
À1
Figure 3. pH dependence of the reactivity of 1a in the presence of a-amino acid derivatives. Left: 0.11 mol kg alaninamide
À1
with pK set to 8.02 (solid line) and pK 8.58 (dotted line). Right: 0.11 mol kg phenylalaninamide, pK set to 7.45 (solid line)
a
a
and pK 7.69 (dotted line). Grey lines indicate hydroxide-catalysed hydrolysis
a
a
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

4
42
N. J. BUURMA, M. J. BLANDAMER AND J. B. F. N. ENGBERTS
a
Table 2. Reactivity of 1a in the presence of alaninamide and phenylalaninamide at 298.2 K
Parameter
Alaninamide
Phenylalaninamide
1
5
15
pKa
8.58 (lit. : 8.02)
0.51 (0.17)
7.69 (lit. : 7.45)
À1
À1
k1a (s mol kg)
0.22 (0.02)
10.3
À4 À1
k(m ) (10
kOH (10 s mol dm )
s
)
12.3
c
2
À1
À1
3
b
n.s.
7.3 (0.8)
a
Numbers in parentheses are standard errors based on a least-squares fit of kinetic data using Eqn. (8).
b
À1
À1
3
The value of 1720 (930) s mol dm obtained from the curve-fitting procedure cannot be regarded as significant considering the error margin.
constant for hydrolysis at pH 4 in the presence of mc
molal protonated cosolute c. The second term on the
right-hand side yields the rate of hydroxide-ion catalysed
methylpropan-2-ol. This maximum was attributed to a
destabilisation of the initial state of the hydroxide-ion
catalysed hydrolysis. This destabilisation of the initial
state has been shown to originate from the unfavourable
Gibbs energy of transfer of the hydroxide anion from
water to water with cosolute, more than cancelling the
corresponding favourable Gibbs energy of transfer of
1a. For neutral hydrolysis of 1a in aqueous solu-
hydrolysis with kOH the second-order rate constant for
pHÀ14
hydroxide-catalysed hydrolysis and 10
the concen-
tration of hydroxide. The third term on the right-hand
side represents the rate of reaction with the cosolute,
where k1a is the rate constant of reaction of the cosolute
19
with 1a, m is the total molality of cosolute and
tions containing 2-methylpropan-2-ol, G(c) = À392 J
c
pK
a
ÀpH À1
À2
(
1 ‡ 10
)
is the fraction of cosolute in the
kg mol . G(c) for the same reaction in the presence
14
deprotonated form. The results are summarised in Table
of phenylalaninamide hydrochloride
is À 1869 J
À2
2
.
kg mol . Compared with solutions with added 2-
methylpropan-2-ol, this pattern indicates that the stan-
dard chemical potential of 1a is lowered more in
solutions with added phenylalaninamide. Therefore,
added phenylalaninamide could lead to a rate decrease
if the stabilising effect of phenylalaninamide more than
cancels the expected destabilising effect of phenylalanin-
amide on the hydroxide anion. However, the effect of
phenylalaninamide on the standard chemical potential of
hydroxide anion is unknown. Hence, reliable estimates of
kOH cannot be made. Related to the unknown effect of the
added cosolute on the standard chemical potential of the
hydroxide anion, the effect of added cosolute on the
water self-ionization constant is unknown. Hence, using
À2
À2
The rate constants, 51 Æ 17 Â 10 and 22 Æ 2 Â 10
À1
À1
s
mol kg for alaninamide and phenylalaninamide,
respectively, are significantly higher than the values of
À3
À3 À1
À1
1
2.2 Â 10 and 8.33 Â 10
s
mol kg predicted on
the basis of the Brønsted plot for general-base catalysed
hydrolysis and their literature pK s of 8.02 and 7.45,
a
15
respectively [the pK values obtained from fitting to
a
Eqn. (8) lead to only marginally higher predicted values].
We attribute this marked difference in reactivity of 1a for
these general bases to a change in reactivity from
general-base catalysed hydrolysis to nucleophilic sub-
stitution. The UV/Vis properties of amides, the expected
products of nucleophilic substitution reactions, are only
slightly different from those of the products of hydro-
lysis. Hence the spectral changes upon reaction (followed
at a single wavelength) give no information about the
actual reaction occurring, other than that the amide
functionality is reacting. However, comparison with the
À
[OH ] based on the observed pH should be done with
caution.
The values for pK obtained using Eqn. (8) as given in
a
Table 2 are both higher than literature values. Together
with the observed pattern in the deviation between the
experimental and calculated values (Fig. 3, right-hand
side), this trend indicates that Eqn. (8) underestimates the
rate constants at higher pH. As nucleophilic substitution
can be general-base catalysed, the increasing molalities
of unprotonated amine not only increase nucleophile
molalities, but also increase catalyst molalities. In
addition, at lower pH, the rate constants might be
underestimated as a result of inhibition of reaction by
the protonated amine. In order to test the hypothesis of
general-base catalysed nucleophilic substitution, the
effect of both ethanoate/ethanoic acid and butanoate/
butanoic acid buffers on the rate of reaction of 1a with
phenylalaninamide was determined (Fig. 4).
16,17
literature
reveals that the difference spectrum upon
reaction is in accord with formation of an amide instead
of a carboxylate. The structurally related activated amide
1
-acetyl-1,2,4-triazole has also been shown to undergo
general-base catalysed hydrolysis and nucleophilic sub-
18
stitution by a variety of nucleophiles.
The rate constant for hydroxide-catalysed hydrolysis
of 1a in the presence of phenylalaninamide is consider-
À1
À1
3
ably lower than the rate constant of 1130 s mol dm
for hydroxide-catalysed hydrolysis of 1a without co-
13
solute. Previously, the effect of 2-methylpropan-2-ol on
the hydroxide-ion catalysed hydrolysis of 1a has been
12
studied. At low mole fractions of added 2-methylpro-
pan-2-ol, the second-order rate constants (kOH) showed a
maximum when plotted against the mole fraction of 2-
According to the plot on the left-hand side of Fig. 4, the
rate of reaction between 1a and alaninamide increases
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

REACTIVITY OR ACTIVATED AMIDES IN AQUEOUS SOLUTIONS
443
Figure 4. Reactivity of 1a with alaninamide and phenylalaninamide in the presence of ethanoate/ethanoic acid and butanoate/
butanoic acid buffers. Left: the contributions of general-base catalysed hydrolysis and nucleophilic substitution to the observed
rates as a function of total molality of alaninamide hydrochloride and alaninamide (m1a ‡ m ) at total concentrations of
1b
À1
ethanoate/ethanoic acid (3:1) buffers (m2b ‡ m ) of (*) 0.27, (*) 0.44, (~) 0.56 and (!) 0.75 mol kg . Right: Second-order
2a
rate constant k1b for nucleophilic substitution by phenylalaninamide as a function of total molality (m ‡ m ) of (~) ethanoate/
2a
2b
ethanoic acid and (!) butanoate/butanoic acid molality
with increasing buffer concentration. The contribution of
general-base catalysed hydrolysis and (general-base
catalysed) nucleophilic substitution is calculated from
the observed rate constant k(m ,m ,m ,m ) and the
based on the available data, a value for knuc for
alaninamide cannot be determined.
Using hydroxylamine, an even higher reactivity than
that observed for alaninamide and phenylalaninamide
1a
1b 2a 2b
calculated rate constant of the water-catalysed reaction in
the presence of only the rate retarding protonated
cosolutes k(m ,m ). The values obtained in this way
were not corrected for rate-retarding effects. A similar
pattern is found for the reaction of phenylalaninamide
with 1a. This pattern indicates that general-base cata-
lysed nucleophilic substitution is indeed occurring, which
was found, despite its lower pK (Fig. 5).
a
À1
À1
The observed rate constant of 53.1 Æ 7.4 s mol kg
À1
À1
(104.9 Æ 14.8 s mol kg based on only deprotonated
1a
2a
À1
À1
hydroxylamine, k
= 0.69 Æ 0.03 s mol kg)
HONH .HCl
2
is exceptionally high. Error margins are based on
À1
inclusion of the apparent outlier at 5.2 mmol kg and
the notion that a negative intercept is physically
unrealistic. The negative intercept could be caused by
two factors. First, a small change in protonation (the pH
decreases slightly) upon dilution of the hydroxylamine
buffer stock solution. Second, a second-order (in total
hydroxylamine molality) term corresponding to general-
base or general-acid catalysed nucleophilic substitution.
Both effects would lead to an overestimate of the second-
explains the deviating pK values from the curve fits.
a
From a plot of the corresponding second-order rate
constants as a function of total buffer molality m ‡ m
2
a
2b
(
Fig. 4, right-hand side), the rate constant for uncatalysed
nucleophilic substitution knuc can be determined. For
phenylalaninamide,
value of 0.153 Æ 0.007
s mol kg is obtained (cf. Table 2). Unfortunately,
a
À1
À1
Figure 5. Reactivity of 1a in the presence of hydroxylamine and hydroxylamine hydrochloride. Concentrations are total
concentrations. Left: hydroxylamine and hydroxylamine hydrochloride in 1:1 ratio, pH = 5.91 Æ 0.06. Dotted lines indicate error
margins (see text). Right: hydroxylamine hydrochloride, pH = 3.76 Æ 0.03
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

4
44
N. J. BUURMA, M. J. BLANDAMER AND J. B. F. N. ENGBERTS
Figure 6. Observed rates of reaction for 1a. Left: in the presence of (*) n-propylamine and (*) n-pentylamine. Buffer ratios are
1
:9 (base:acid), pH = 9.54 Æ 0.07for n-propylamine and 9.53 Æ 0.03 for n-pentylamine. Right: in the presence of benzylamine.
Buffer ratio is 1:1 (base:acid) and pH = 9.34 Æ 0.04
2
0
order rate constant of reaction. Based on the pK of 5.96
tion of general base (Fig. 6). We attribute this pattern to
general-base catalysed nucleophilic substitution, but
a
À4
4
and Eqn. (7), the expected rate constant is 40.1 Â 10
À1
À1
s
mol
kg for general-base catalysis. This large
general-acid catalysis (by the conjugate acids) cannot be
3,24,25
discrepancy between the predicted value for general-
base catalysed hydrolysis and the observed value again
strongly suggests that the reaction pathway that is
followed is not general-base catalysed hydrolysis. The
high reactivity of hydroxylamine is often accounted for in
excluded.
The observed rate constants were fitted to the equation
2
c
kꢀm † ˆ k ‡ k2ndm ‡ k3rdm
ꢀ9†
c
pH
c
21
terms of the a-effect, therefore the results are indicative
for nucleophilic attack on 1a. Hydroxylamine has two
nucleophilic centres (of different reactivity), further
enhancing reactivity. Nucleophilic attack by hydroxyl-
amine will occur on the amide functionality of 1a,
eventually leading to N-hydroxybenzamide. Surprisingly,
the hydroxyl moiety is most nucleophilic in hydroxyl-
amine in the case when p-nitrophenyl acetate is the
substrate undergoing nucleophilic attack, resulting in an
initial excess of the product from nucleophilic attack by
the hydroxyl moiety. However, this initial unstable
where kpH is the (pseudo-)first-order rate constant for
reaction in the absence of added general base at the
experimental pH, k2nd is the second-order rate constant
based on total buffer concentration for nucleophilic
substitution (and a minor fraction general-base catalysed
hydrolysis) by cosolute c, m is the molality of cosolute c
c
and k3rd is the third-order rate constant for the general-
base catalysed nucleophilic substitution. Assuming that
general-base catalysed hydrolysis makes a negligible
contribution to the observed rate, knuc can be calculated
(Table 3) from k2nd
.
22,23
product can react further to N-hydroxybenzamide.
The zwitterionic form of hydroxylamine is not present in
According to Table 3, n-pentylamine provides the most
effective non-catalysed nucleophilic substitution. How-
ever, the increase in reactivity appears to be too large to
20,23
detectable amounts,
practically excluding the O-
deprotonated hydroxylamine as the nucleophile.
be caused solely by the difference in pK and different
a
Using n-propyl- and n-pentylamine as well as
benzylamine, again high reactivities were found but the
observed rate constants were not linear with concentra-
steric effects. This increase in reactivity points towards
more favourable interactions between n-pentylamine and
1a compared with those between n-propylamine and 1a.
a,b
Table 3. Nucleophilic substitution of 1a in aqueous solution at 298.2 K in the presence of amine/ammonium HCl buffers
Parameter
n-Propylamine
n-Pentylamine
Benzylamine
2
6
27
28
pKa
10.66
10.64
9.36
À1
À1
k2nd (s kg mol
)
6.9 (1.8)
69 (18)
5.4 (1.0)
14.5 (4.6)
145 (46)
4.6 (2.4)
5.5 (0.8)
11.1 (1.6)
10.7 (0.6)
À1
À1 c
knuc (s kg mol
k3rd (10 s kg mol
)
2
À1
2
À2
)
a
The numbers in parentheses are standard errors based on a least-squares fit of the kinetic data using Eqn. (9).
b
c
29
k
nuc = 10k2nd, the rate effect of the protonated amine in this concentration range is expected to be negligible.
Solvent kinetic isotope effect of 1.6 was found for the reaction. However, many factors influence this isotope effect, including the shift in pK
O to D O as the solvent.
a
of the amine
upon changing from H
2
2
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

REACTIVITY OR ACTIVATED AMIDES IN AQUEOUS SOLUTIONS
445
The second-order terms (in amine), however, are
comparable (although care has to be taken in interpreting
the value for n-pentylamine) suggesting that self-
association of n-pentylamine is comparable to self-
association of n-propylamine. Previously, for nucleophi-
lic substitution of sufficiently hydrophobic p-nitrophenyl
esters by hydrophobic amines, the second-order (in
nucleophile) process was found to be more effective
30,31
because the reactants tend to cluster.
The reactivity of benzylamine in comparison with that
of the alkylamines is surprisingly high taking into
account its lower pK . Given the present data set, the
a
second-order term (in amine) cannot be compared with
second-order terms for the alkylamines. The buffer ratio
is different and the influence of basicity (or acidity) on
general-base (or general-acid) catalysis of nucleophilic
substitution is unknown, because Brønsted plots for
general-base catalysis and general-acid catalysis of
Figure 7. Plots of k and k
(indicated by k ) vs pK for
1a
b
nuc
a
1
a. (~) General-base catalysis by: water (1), chloroethanoate
(
2), ethanoate (3) and butanoate (4). (*) Nucleophilic
substitution by amines, viz. phenylalaninamide (6), alanin-
amide (7), benzylamine (9), propylamine (10) and pentyl-
amine (11). (!) Nucleophilic substitution by other com-
pounds, viz. hydroxylamine (5), aspartame (8) and hydroxide
(12). The line through the data for amines was drawn to
guide the eye
18
nucleophilic substitution have not been determined.
An experiment, conducted under the conditions of
the kinetic runs, was performed on a milligram (of
substrate) scale. Benzylamine was used as nucleophile
1
and 1b instead of 1a was used as substrate. Both H NMR
13
and C NMR spectra of the product correspond to
both groups. From the pK of the carboxylic acid
a
32–34
39,40
literature spectra of N-benzylbenzamide,
corrobor-
functionality
of 3.2, we conclude that in the molality
À1
ating the view that nucleophilic substitution did indeed
range up to 0.044 mol kg of aspartame, the contribu-
tion of the carboxylate group to the observed rate of
reaction is negligible. The pH–rate profile for reaction of
1a with aspartame at three concentrations of aspartame is
given in Fig. 8.
take place. Reaction between 1b and phenylamine in
35
benzene was shown to yield N-phenylbenzamide.
Moreover, 1b has been used as a benzoylation reagent
36
in dry cyanomethane, both consistent with the possi-
bility of nucleophilic substitution of 1b and the related
From the non-linear least-squares fits to the observed
rate profiles, the second-order rate constant for nucleo-
philic substitution on 1a by aspartame is 7.8 Æ 0.2
1
a.
An overall Brønsted plot of rate constants for general-
À1
À1
base catalysed hydrolysis k and uncatalysed nucleophilic
s
mol kg. As can be seen from Fig. 7, this rate
b
substitution knuc of 1a (both indicated by k ) shows
constant for nucleophilic substitution is clearly much
higher than expected. We attribute this increase to two
factors. First, aspartame is a rather hydrophobic mol-
ecule, which could form hydrophobically stabilised
encounter complexes (cf. the effect of benzylamine).
Second, the nucleophilic substitution by aspartame can
be intra-molecularly general-base catalysed by the
carboxylate functionality, thereby greatly enhancing the
nucleophilicity of aspartame (Scheme 5). It is interesting
to view aspartame as a cyclised, non-nucleophilic general
base catalyst for hydrolysis. The molecule consists of two
a-amino acids, and has a hydrophobic ‘binding site’ and a
‘catalytic center’.
1a
considerable scattering (Fig. 7). Note that in all LFERs,
only the linear terms describing the bimolecular (un-
catalysed) nucleophilic substitution reaction have been
used.
For general bases with comparable nucleophilic
groups, a correlation is obtained between observed rate
constant and pK . However, the correlation does not
a
extend over different groups of bases, in agreement with
37,38
extensive literature data.
Remarkably, the observed
rate constants for aromatic amines acting as a nucleophile
seem to be consistently higher than those for non-
aromatic amines with the same pK would have been, i.e.
a
the data points for aromatic amines lie above and to the
left of a line through the non-aromatic amines.
An interesting case is presented by aspartame (Asp-
PheOMe) (Scheme 4). Aspartame offers a number of
functional groups, the carboxylic acid and the amine
being the most important for the present study. The
carboxylate group is expected to be a general-base
catalyst in the hydrolysis reaction of 1a. The amine group
will act as nucleophile, the relative importance of both
reactions being dependent on the degree of protonation of
Scheme 4.
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

4
46
N. J. BUURMA, M. J. BLANDAMER AND J. B. F. N. ENGBERTS
46
links between the two scales, but a unifying scale for
nucleophilicity does not exist.
We compare rate constants for uncatalysed nucleo-
philic substitution and general-base catalysed hydrolysis
of 1a with the rate constants for the same reactions
of p-nitrophenyl acetate (p-NPA). There are ample
experimental data on nucleophilic substitution on
2,23,47–50
p-NPA
and the available data for p-NPA form
a reliable basis for linear free energy relationships
51
(
LFERs). Most importantly, however, the change in
reaction pathway from general-base catalysed hydrolysis
to nucleophilic substitution of 1a is mirrored by p-NPA.
Both are carbonyl compounds showing approximately
the same reactions with added general bases. Also, the
anions of both 3-phenyl-1,2,4-triazole and 4-nitrophenol
Figure 8. pH±rate pro®le for 1a in the presence of different
total molalities m ‡ m of aspartame: (*) 0.007; (*) 0.028;
a
b
À1
(
!) 0.044 mol kg . The stability of aspartame under these
are reasonably good leaving groups based on the pK
conditions is reasonable in comparison with the reaction
under study (see Ref. 40). The pH of aspartame solutions was
adjusted (from a pH in the range 4±7) immediately before
performing the measurement
a
53
52
values of the parent compounds of 9.58 and 7.15,
respectively.
When log k is plotted as a function of log k for
1
a
pÀNPA
different bases/nucleophiles, a linear correlation is found
Fig. 9) with a slope of 0.91. The correlation includes data
(
points for hydroxylamine (an a-effect nucleophile) and
hydroxide anion (a charged nucleophile).
General-base catalysed hydrolysis and uncata-
lysed nucleophilic substitution on 1a; a linear free
energy relationship
Interestingly, the LFER spans regions of different
reactivity; general-base catalysed hydrolysis for both 1a
5
4
and p-NPA by water (1) and ethanoate (3) and
nucleophilic substitution by hydroxylamine (5), benzyl-
amine (9), propylamine (10), pentylamine (11) and
hydroxide (12). This pattern is in line with the argument
Despite the fact that comparing basicities of nucleophiles
with kinetic nucleophilic reactivities towards carbon
compounds is not a proper rate-equilibrium comparison
41
42
55
(
Parker and Hine and Weimar developed the concept
formulated by Fersht and Jencks that ‘such a correlation
of ‘carbon basicity’) reactivity of general bases in
nucleophilic substitution is roughly correlated with their
basicity. However, there are many factors influencing
shows only that the two compounds being compared have
similar transition states for each individual reaction under
consideration; if there is a change in the nature of the
transition state with changing nucleophile … the
correlation shows that this change takes place in a
similar manner for both compounds.’
We conclude that the enhanced reactivity compared
with what would be expected on the basis of general-base
catalysis alone is indeed caused by a change in mech-
43
reactivity other than basicity, the most important being
21
the elusive a-effect and steric factors. Often, nucleo-
philicity varies with basicity within a series of
compounds, which can be attributed to differences in
solvation of different nucleophilic groups and a differ-
ence in hardness/softness of the nucleophile. In view of
the failure of basicity as an indicator of nucleophilicity,
alternative scales of nucleophilicity have been developed,
4
4
most importantly Richie’s
N (for nucleophilic addi-
‡
2
45
tion on sp carbon) and the Swain and Scott n parameter
for nucleophilic substitution on sp carbon). There are
3
(
Figure 9. Linear free energy relationship between nucleo-
philic substitution on 1a and nucleophilic substitution on p-
NPA by different bases/nucleophiles. Numbering as in Fig. 7
Scheme 5.
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

REACTIVITY OR ACTIVATED AMIDES IN AQUEOUS SOLUTIONS
447
anism to nucleophilic substitution. The linear free energy
relationship also allows reliable estimates to be made
of the rates of uncatalysed nucleophilic substitution
on 1a.
cosolute molality m . Finally, the observed kinetic
c
solvent isotope effect of 2.49 for the hydrolysis of 1a in
À1
0.5 mol kg glycine at pH 4 is very similar to the 2.69
observed for the reaction in water at pH 4 without
cosolute. Unfortunately, however, this value will be
influenced by the increase in pK of both the a-amino acid
carboxylate and the a-amino acid amine functionality
a
General-base catalysis, nucleophilic substitution
and inhibition
upon changing the solvent from H O to D O. The
2
2
5
7
increase in pK for the glycine carboxylate group of
a
As described above, a range of possible interactions
between cosolute and hydrolytic probe sometimes result
in rather complex reactivity patterns. For example,
phenylalaninamide and alaninamide show rate-decreas-
ing effects in their protonated forms, but are strong
nucleophiles in their deprotonated forms. Similarly, the
effect of the carboxylate buffers is a composite effect of
the rate retardation by the protonated form and catalysis
by the carboxylate.
Previously, rate effects induced by a-amino acids
have been studied at a pH of 4.0, at which the a-amino
acids are present mainly in their zwitterionic forms. Rate-
enhancing effects were found for a range of a-amino
acids and even for some of their derivatives. Rate-
enhancing effects were not found to correlate with the
0.39 results in a larger fraction of the a-amino acid
carboxylate group becoming neutralised. Together with
the expected kinetic solvent isotope effect, this leads to a
decrease in rate constant. Consequently, the observed rate
of the nucleophilic substitution reaction will be de-
creased, even though no proton transfer takes place in the
uncatalysed nucleophilic substitution reaction. The
5
7
increase in pK of the a-amino acid amine functionality
a
of 0.63 results in fewer free amine groups being available
for reaction, also leading to a decrease in observed rate
constant. Hence the similarity in kinetic solvent isotope
effects could be merely coincidental, which interfered
with the interpretation of the result obtained for a-amino
acids.
56
We contend that for the least hydrophobic a-amino
acids, rate-retarding effects are negligible and only rate
enhancements are observed caused by general-base
pK of the carboxylic acid moiety. Furthermore, the
a
kinetic solvent isotope effect did not change significantly
and no linearity of the slope of kobsd versus molality of
cosolute was observed. Hence, the kinetic effects were
5
6
catalysed hydrolysis and nucleophilic substitution.
The effect of general-base catalysis of hydrolysis can
‘
not governed by general-base catalysis of the a-amino
be calculated using Eqn. (7), pK values of the a-amino
acids and pH (Table 4).
If the residual observed rate enhancements are
attributed to uncatalysed nucleophilic substitution by
free amine, excellent correlation is again observed (Fig.
10) with data for p-NPA.
Hence, even though the fraction of unprotonated amine
functionalities is of the order of ppm, the unprotonated
amine functionality of a-amino acids is strongly
nucleophilic, rendering nucleophilic substitution kineti-
cally detectable at a pH as low as 4.0 for hydrophilic a-
amino acids. More hydrophobic a-amino acids, however,
show rate retardation as the main effect at a pH of 4.0.
However, the observed rate retardations will be a
combined effect of inhibition by the hydrophobic a-
amino acid, general-base catalysis and nucleophilic
substitution.
a
58
acid carboxylate group, but involve medium effects
56
instead.’ The kinetic analyses presented here, however,
indicate that apart from general-base catalysis by the a-
amino acid carboxylate group, nucleophilic substitution
by the a-amino acid amine group is also a possible
reaction pathway. Consequently, the observed rate effects
are rather difficult to interpret, as a linear Brønsted plot is
not to be expected. In addition, given the observed G(c)
values and the data in Fig. 4.1 of Ref. 56, it is difficult to
draw conclusions about the linearity of the observed
rate effects with molality of cosolute. This is especially
the case since possible general-base or general-acid
catalysis will produce deviations from linearity in plots
of k(m ) versus molality. The deviation is towards
higher k(m ) at higher molalities, which can accidentally
produce good linear plots of ln[k(m )/k(m = 0)] against
c
c
c
c
a
Table 4. Reactivity of 1a in aqueous solution at 298.2 K in the presence of a-amino acids
b
À4 À1
c
À4 À1
2
À1 À1
a-Amino acid
pKa
G(c)
k(m = m ) (10
s
)
m k (10
b b
s
)
knuc (10 kg mol
s )
c
0
Glycine
Alanine
Valine
2.35, 9.78
2.34, 9.69
2.32, 9.62
2.36, 9.60
875 (21)
558 (16)
467 (9)
24.5 (0.5)
19.0 (0.3)
17.6 (0.2)
18.4 (0.4)
2.8 (1.3)
5.8 (1.5)
2.0 (0.8)
1.2 (0.6)
1.4 (0.6)
2.7 (1.2)
2.7 (1.2)
2.8 (1.3)
Leucine
518 (21)
a
Numbers in parentheses are standard errors.
b
c
In the calculation of the errors in G(c), m
b
k
b
a a
and knuc, the errors in the pK values were set to 0.05 and 0.1 for the first and second pK , respectively.
Calculated using Eqn. (7).
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

4
48
N. J. BUURMA, M. J. BLANDAMER AND J. B. F. N. ENGBERTS
Buffer solutions containing n-propylamine and n-pentyl-
amine were prepared by addition of the appropriate
amount of 1.000 M aqueous HCl within 2 min prior to
monitoring the reaction of 1a in order to prevent
evaporation of the volatile amines from the solutions.
Water was distilled twice in an all-quartz distillation unit.
All reactions were monitored at 273 nm (or the lowest
possible wavelength above 273 nm if a given cosolute
had absorption bands at that wavelength) and at
2
5.0 Æ 0.1°C. Amide 1a was injected as 5–7 ml of a
stock solution containing 1a in cyanomethane into about
.8 ml of an aqueous solution of cosolute in a concentra-
2
tion range in which the reaction could be followed in a
stoppered 1.000 cm quartz cuvette. The resulting con-
centrations were about 10 mol dm or less. The pH of
all solutions was checked at the end of each kinetic
experiment using either a Ross semi-micro combination
pH electrode or a Sentron ISFET pH probe and was found
Figure 10. (*) Data for nucleophilic substitution on 1a by a-
amino acids, glycine (a), alanine (b), valine (c) and leucine (d),
compared with (X) the LFER as described in Fig. 9. Error
margins in log k1a are 0.6, 0.9, 1.1 and 1.0 for glycine,
alanine, valine and leucine, respectively
À5
À3
to correspond well (not more than 0.2 pK units below)
a
CONCLUSIONS
with the predicted pH from the pK and the buffer ratio.
a
1
NMR spectra were recorded on Varian Gemini 200 ( H:
1
In aqueous solutions containing general bases, activated
amide 1a is subject to water-catalysed hydrolysis,
general-base catalysed hydrolysis with a Brønsted b of
200 MHz) and VXR 300 ( H: 300 MHz) spectrometers.
0
.29 and nucleophilic substitution with a Brønsted b of
1.5 for amine nucleophiles. In certain cases, nucleo-
Acknowledgement
ꢁ
philic substitution is general-base and/or general-acid
catalysed. Reactivities of more hydrophobic general
bases seem to be consistently higher than reactivities of
hydrophilic general bases supporting an explanation
based on the formation of hydrophobically stabilised
encounter complexes. In future studies, rate effects of
cosolutes, in particular rate-enhancing effects, should be
scrutinised for unexpected catalytic effects or changes in
mechanism. In the present study we have shown that
small fractions of compounds present in, e.g., a
deprotonated state, albeit in ppm, can induce large rate
effects.
Dr Matt Fielden is thanked for helpful mechanistic
discussions.
REFERENCES
1. Karzijn W, Engberts JBFN. Tetrahedron Lett. 1978; 1787–1790.
2
. Johnson SL. Adv. Phys. Org. Chem. 1967; 5: 237–330.
3
. Bruice TC, Donzel A, Huffman RW, Butler AR. J. Am. Chem. Soc.
1
967; 89: 2106–2121.
4. Bruice TC, Hegarty AF, Felton SM, Donzel A, Kundu NG. J. Am.
Chem. Soc. 1970; 92: 1370–1378.
5
6
7
8
9
. Streefland L, Blandamer MJ, Engberts JBFN. J. Am. Chem. Soc.
996; 118: 9539–9544.
. Buurma NJ, Pastorello L, Blandamer MJ, Engberts JBFN. J. Am.
Chem. Soc. 2001; 123: 11848–11853.
. Rispens T, Lensink MF, Berendsen HJC, Engberts JBFN.
submitted for publication.
. Lensink MF, Mavri J, Berendsen HJC. J. Comput. Chem. 1999; 20:
1
EXPERIMENTAL
8
86–895.
Materials. All buffers were made from commercially
available acids, alkylamines or benzylamine (from Acros
or Aldrich) using aqueous NaOH or aqueous HCl of
known concentration (Titrisol). Aspartame was kindly
provided by Professor Dr H. E. Schoemaker (DSM/
University of Amsterdam). 1-Benzoyl-3-phenyl-1,2,4-
triazole (1a) and 1-benzoyl-1,2,4-triazole were synthe-
. Headley AD, Starnes SD, Wilson LY, Famini GR. J. Org. Chem.
994; 59: 8040–8046.
10. Niazi MSK. J. Chem. Eng. Data 1993; 38: 527–530.
1. Engberts JBFN, Blandamer MJ. J. Phys. Org. Chem. 1998, 11:
41–846.
2. Karzijn W, Engberts JBFN. Recl. Trav. Chim. Pays-Bas 1983; 102:
513–515.
1
1
1
1
8
3. Karzijn W. PhD Thesis, Rijksuniversiteit Groningen, 1979.
14. Streefland L, Blandamer MJ, Engberts JBFN. J. Phys. Chem. 1995;
13,59,60
sised according to literature procedures.
9
9: 5769–5771.
1
5. Chambers RW, Carpenter FH. J. Am. Chem. Soc. 1955; 77: 1522–
1
526.
Kinetic Experiments. Aqueous solutions were prepared
by weight immediately before use. Buffers were prepared
by partially (to the desired buffer ratio) neutralising the
1
1
6. Ley H, Specker H. Chem. Ber. 1939; 72: 192–202.
7. Moser CM, Kohlenberg AI. J. Chem. Soc. 1951; 804–809.
18. Fox JP, Jencks WP. J. Am. Chem. Soc. 1974; 96: 1436–1449.
9. Blokzijl W, Engberts JBFN, Blandamer MJ. J. Am. Chem. Soc.
1990; 112: 1197–1201.
0. Kallies B, Mitzner R. J. Phys. Chem. B 1997; 101: 2959–2967.
21. Edwards JO, Pearson RG. J. Am. Chem. Soc. 1961; 84: 16–24.
1
corresponding acid by adding the appropriate amount of
À1
1
.000 mol l aqueous NaOH by volume or by weight.
2
Buffer ratios were routinely accurate to within 1%.
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449

REACTIVITY OR ACTIVATED AMIDES IN AQUEOUS SOLUTIONS
449
2
2. Jencks WP. J. Am. Chem. Soc. 1958; 80: 4581–4584 and 4585–
588.
40. Skwierczynski RD, Connors KA. Pharm. Res. 1993; 10: 1174–
1180.
4
2
2
2
2
3. Jencks WP, Carriuolo J. J. Am. Chem. Soc. 1960; 82: 1778–1786.
4. Jencks WP, Carriuolo J. J. Am. Chem. Soc. 1960; 82: 675–681.
5. Bruice TC, Mayahi MF. J. Am. Chem. Soc. 1960; 82: 3067–3071.
6. Lin B, Islam N, Friedman S, Yagi H, Jerina DM, Whalen DL. J.
Am. Chem. Soc. 1998; 120: 4327–4333.
41. Parker AJ. Proc. Chem. Soc. 1961; 371.
42. Hine J, Weimar Jr RD. J. Am. Chem. Soc. 1965; 87: 3387–3396.
43. Bunnett JF. Annu. Rev. Phys. Chem. 1963; 14: 271–290.
44. Ritchie CD. Can. J. Chem. 1986; 64: 2239–2250.
45. Swain CG, Scott CB. J. Am. Chem. Soc. 1953; 75: 141–147.
46. Bunting JW, Mason JM, Heo CKM. J. Chem. Soc., Perkin Trans. 2
1994; 2291–2300.
2
2
2
7. Hoerr CW, McCorkle MR, Ralston AW. J. Am. Chem. Soc. 1943;
6
5: 328–329.
8. Arrowsmith CH, Guo HX, Kresge AJ. J. Am. Chem. Soc. 1994;
16: 8890–8894.
9. Hol P, Streefland L, Blandamer MJ, Engberts JBFN. J. Chem. Soc.,
Perkin Trans. 2 1997; 485–488.
47. Um IH, Lee GJ, Yoon HW, Kwon DS. Tetrahedron Lett. 1992; 33:
2023–2026.
1
48. Acher F, Wakselman M. J. Org. Chem. 1984; 49: 4133–4138.
49. Jencks WP, Gilchrist M. J. Am. Chem. Soc. 1968; 90: 2622–2637.
50. Tee OS, Gadosy TA, Giorgi JB. Can. J. Chem. 1997; 75: 83–91.
51. Gregory MJ, Bruice TC. J. Am. Chem. Soc. 1967; 89: 2121–2125.
52. Kr o¨ ger C-F, Freiberg W. Z. Chem. 1965; 5: 381–382.
53. Mock WL, Morsch LA. Tetrahedron 2001; 57: 2957–2964.
54. Butler AR, Gold V. J. Chem. Soc. 1962; 1334–1339.
55. Fersht AR, Jencks WP. J. Am. Chem. Soc. 1970; 92: 5442–5452.
56. Streefland L. PhD Thesis, University of Groningen, 2000;
available online at http://www.ub.rug.nl/eldoc/dis/science/e.
streefland/.
57. Li NC, Tang P, Mathur R. J. Phys. Chem. 1961; 65: 1074–1076.
58. Fasman GD (ed). Handbook of Biochemistry and Molecular
Biology, Proteins (3rd edn). CRC Press: Cleveland, OH, 1976.
59. Staab HA, L u¨ king M, D u¨ rr FH. Chem. Ber. 1962; 95: 1275–1283.
60. Mooij HJ, Engberts JBFN, Charton M. Recl. Trav. Chim. Pays-Bas
1988; 107: 185–189.
30. Oakenfull D. J. Chem. Soc., Perkin Trans. 2 1973; 1006–1012.
31. Oakenfull D. J. Chem. Soc., Chem. Commun. 1970; 1655–1656.
32. Keck GE, Wager TT, McHardy SF. Tetrahedron 1999; 55: 11755–
1
1772.
3
3
3. Darbeau RW, White EH. J. Org. Chem. 1997; 62: 8091–8094.
4. Knowles HS, Parsons AF, Pettifer RM, Rickling S. Tetrahedron
2
000; 56: 979–988.
3
5. Purygin PP, Pankov SV. Russ. J. Org. Chem. 1996; 32: 871–873.
6. Humpf HU, Berova N, Nakanishi K, Jarstfer MB, Poulter CD. J.
Org. Chem. 1995; 60: 3539–3542.
3
3
7. Heo CKM, Bunting JW. J. Chem. Soc., Perkin Trans. 2 1994;
2
279–2290.
3
8. Bruice TC, Lapinski R. J. Am. Chem. Soc. 1958; 80: 2265–2267.
9. Maheswaran MM, Divakar S. Indian J. Chem., Sect. A 1991; 30:
3
3
0–34.
Copyright  2003 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2003; 16: 438–449