Catalysis of ester aminolysis by cyclodextrins. The reaction of alkylamines with p-nitrophenyl alkanoates

Article abstract of DOI:10.1021/jo9919660

The effects of four cyclodextrins (α-CD, β-CD, hydroxypropyl-β-CD, and γ-CD) on the aminolysis of p-nitrophenyl alkanoates (acetate to heptanoate) by primary amines (n-propyl to n-octyl, isobutyl, isopentyl, cyclopentyl, cyclohexyl, benzyl) in aqueous solution have been investigated. Rate constants for amine attack on the free and CD-bound esters (k(N) and k(cN)) have ratios (k(cN)/k(N)) varying from 0.08 (retardation) to 180 (catalysis). For the kinetically equivalent process of free ester reacting with CD-bound amine (k(Nc)), the ratios k(Nc)/k(N) vary from 0.2 to 28. Either way, there is evidence of catalysis in some cases and retardation in others. Changes in reactivity parameters with structure indicate more than one mode of transition state binding to the CDs. Short esters react with short alkylamines by attack of free amine on the ester bound by its aryl group, but for longer amines, free ester reacts with CD-bound amine. Reaction of long esters with long amines, which is catalyzed by β-CD and γ-CD, involves inclusion of the alkylamino group and possibly the ester acyl group. The larger cavity of γ-CD may allow the inclusion of the ester aryl group, as well as the alkylamino group, in the transition state. Reaction between an ester bound to the CD by its acyl group and free amine appears not to be important.

Full text of DOI:10.1021/jo9919660

J . Org. Chem. 2000, 65, 6879-6889
6879
Ca ta lysis of Ester Am in olysis by Cyclod extr in s. Th e Rea ction of
Alk yla m in es w ith p-Nitr op h en yl Alk a n oa tes†
Timothy A. Gadosy, Michael J . Boyd, and Oswald S. Tee*
Department of Chemistry and Biochemistry, Concordia University, Montre´al, Que´bec, Canada H3G 1M8
Received December 28, 1999
The effects of four cyclodextrins (R-CD, â-CD, hydroxypropyl-â-CD, and γ-CD) on the aminolysis of
p-nitrophenyl alkanoates (acetate to heptanoate) by primary amines (n-propyl to n-octyl, isobutyl,
isopentyl, cyclopentyl, cyclohexyl, benzyl) in aqueous solution have been investigated. Rate constants
for amine attack on the free and CD-bound esters (kN and kcN) have ratios (kcN/kN) varying from
0.08 (retardation) to 180 (catalysis). For the kinetically equivalent process of free ester reacting
with CD-bound amine (kNc), the ratios kNc/kN vary from 0.2 to 28. Either way, there is evidence of
catalysis in some cases and retardation in others. Changes in reactivity parameters with structure
indicate more than one mode of transition state binding to the CDs. Short esters react with short
alkylamines by attack of free amine on the ester bound by its aryl group, but for longer amines,
free ester reacts with CD-bound amine. Reaction of long esters with long amines, which is catalyzed
by â-CD and γ-CD, involves inclusion of the alkylamino group and possibly the ester acyl group.
The larger cavity of γ-CD may allow the inclusion of the ester aryl group, as well as the alkylamino
group, in the transition state. Reaction between an ester bound to the CD by its acyl group and
free amine appears not to be important.
In tr od u ction
laboratory showed that the catalytic effect is less than
originally thought.9b
Cyclodextrins (CDs) form inclusion complexes with a
wide variety of species in aqueous solution.1 As a result
they have effects on the rates of organic reactions,
ranging from complete inhibition through mild retarda-
tion to significant catalysis, depending on the reaction
and the CD.1-3 These features have been observed for
basic ester cleavage in the presence of CDs, which has
been investigated by many groups,2-4 including ours.5-8
By contrast, the effect of CDs on the reactions of esters
with nucleophiles besides hydroxide ion have received
little attention. It was reported that nucleophilic attack
of R-amino acid anions on p-nitrophenyl acetate (pNPA)
can be catalyzed by CDs,9a but a later study in our
In view of the limited amount of earlier work, we have
undertaken detailed studies of the effects of cyclodextrins
on the attack of several nucleophiles on alkanoate
esters.10 Our first studies involved nucleophiles that do
not bind significantly to CDs to see how binding the
esters to CDs affects their reactivity. For the reactions
of hydroxylamine, imidazole, and the anions of 2,2,2-
trifluoroethanol (TFE) and 2-mercaptoethanol (2-ME)
with pNPA and p-nitrophenyl hexanoate (pNPH) in the
presence of CDs, it was found that the CD-bound esters
have the same or slightly reduced reactivities as the free
esters, suggesting that the carbonyl groups of esters
bound to CDs are fairly accessible to the nucleophiles in
the medium.10a However, from this initial study it was
not possible to discern the mode of binding of the ester
during attack by the nucleophile. From previous studies,7,8a
it seemed likely that pNPA would react with inclusion
of its aryl group in the CD cavity, as in 1, whereas pNPH
might well react with acyl group inclusion, as in 2.
* To whom correspondence should be addressed. Telephone: +1
(514)848-3348.FAX: +1(514)848-2868.E-mail: ostee@vax2.concordia.ca.
† Based partly on the Ph.D. thesis of T. A. Gadosy (Concordia
University, Montreal Canada, 1995).
(1) (a) Bender, M.; Komiyama, M. Cyclodextrin Chemistry; Springer-
Verlag, New York, 1978. (b) Szejtli, J . Cyclodextrins and their Inclusion
Complexes; Akademiai Kiado, Budapest, 1982. (c) Szejtli, J . Carbohyd.
Polym. 1990, 12, 375. (d) Various aspects of cyclodextrin chemistry
have recently been reviewed by multiple authors in a special issue of
Chemical Reviews: Chem. Rev. 1998, 98 (5), 1741-2076.
(2) Komiyama, M.; Bender, M. L. In The Chemistry of Enzyme
Action; Page, M. I., Ed.; Elsevier: Amsterdam, 1984; Chapter 14.
(3) Tee, O. S. Adv. Phys. Org. Chem. 1994, 29, 1.
(4) Matsui, Y.; Nishioka, T.; Fujita, T. Topics Curr. Chem. 1985,
128, 61.
(5) Tee, O. S.; Takasaki, B. K. Can. J . Chem. 1985, 63, 3540.
(6) (a) Tee, O. S., Bozzi, M.; Hoeven, J . J .; Gadosy, T. A. J . Am.
Chem. Soc. 1993, 115, 8990. (b) Tee, O. S.; Bozzi, M.; Clement, N.;
Gadosy, T. A. J . Org. Chem. 1995, 60, 3509. (c) Gadosy, T. A.; Tee, O.
S. Can. J . Chem. 1996, 74, 745.
(7) (a) Tee, O. S.; Mazza, C.; Du, X.-X. J . Org. Chem. 1990, 55, 3603.
(b) Tee, O. S.; Du, X.-X. J . Am. Chem. Soc. 1992, 114, 620. (c) Tee, O.
S.; Giorgi, J . B. J . Chem. Soc., Perkin Trans. 2 1997, 1013.
(8) (a) Gadosy, T. A.; Tee, O. S. J . Chem. Soc., Perkin Trans. 2 1994,
715; J . Chem. Soc., Perkin Trans. 2 1994, 2069. (b) Tee, O. S.; Gadosy,
T. A. J . Chem. Soc., Perkin Trans. 2 1994, 2191. (c) Tee, O. S.; Boyd,
M. J . J . Chem. Soc., Perkin Trans. 2 1995, 1237. (d) Tee, O. S.; Giorgi,
J . B. J . Am. Chem. Soc. 1995, 117, 3633.
To distinguish between the modes 1 and 2, we studied
the attack of the anions of TFE and 2-ME on a series of
(10) (a) Tee, O. S.; Gadosy, T. A. J . Chem. Soc., Perkin Trans. 2 1994,
2307. (b) Gadosy, T. A.; Tee, O. S. J . Chem. Soc., Perkin Trans. 2 1995,
71.
(9) (a) Barra, M.; de Rossi, R. H. Can. J . Chem. 1991, 69, 1124. (b)
Tee, O. S.; Gadosy, T. A.; Giorgi, J . B. Can. J . Chem. 1997, 75, 83.
10.1021/jo9919660 CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/27/2000

6880 J . Org. Chem., Vol. 65, No. 21, 2000
Gadosy et al.
Ta ble 1. Dissocia tion Con sta n ts (KN) of Com p lexes
p-nitrophenyl alkanoates (C2 to C10) in the presence of
R-CD, â-CD, and hp-â-CD.10b,11 On the basis of previous
studies of reactions of alkyl derivatives in the presence
of CDs,3,7,8,12,13 and on the binding of aliphatics to
CDs,3,4,14,15 we used the sensitivity of kinetic parameters
for the CD-mediated reactions to ester chain length to
probe for acyl group binding in the transition state. It
was found that for short esters (C2 to C6) the parameters
are not sensitive to the acyl chain, but for longer esters
they increase strongly with the chain length.10b This
behavior was taken to mean (i) short esters that bind to
the CD by aryl inclusion react through aryl group
inclusion (1); (ii) short esters that bind by acyl inclusion7b,8a
have their carbonyl groups buried in the CD cavity so
that the dominant acyl-bound complexes must first
isomerize to less stable, aryl-bound complexes before
reaction occurs through 1; (iii) for longer alkanoates, the
carbonyl groups of their acyl-bound forms are exposed
enough to the medium that nucleophilic attack takes
place on the predominant acyl-bound forms (2).10b
Following studies with nonbinding nucleophiles,10 we
have turned our attention to ester cleavage by primary
amines which bind to CDs.14b Here we report extensive
studies of the reaction of p-nitrophenyl alkanoates with
simple alkylamines in the presence of R-CD, â-CD, hp-
â-CD, and γ-CD. These CDs differ in their cavity sizes11
and hence in their abilities to bind organic guests.1,3,4,10,14,15
The intent was to establish CD-mediated aminolysis and
to find out how it varies with the binding abilities of the
amine and the ester in order to probe the importance of
inclusion of the two reactants during the reaction.
Differences arising from the CD cavity sizes11 were also
of interest since a larger CD cavity might permit partial
inclusion of both reactants in the transition state. As
previously, the sensitivity of kinetic parameters to chain
length would serve as the main criterion of transition
state binding,3,7,8,10b,12,13 along with appropriate compari-
sons to initial state binding parameters.
F or m ed fr om Am in es a n d CDsa
amine
R-CD
â-CDb
hp-â-CDb
γ-CD
n-propyl
n-butyl
n-pentyl
n-hexyl
n-heptyl
n-octyl
isobutyl
isopentyl
cyclopentyl
cyclohexyl
benzyl
46.6 ( 5.9
13.3 ( 1.1
3.08 ( 0.54
1.60 ( 0.12
0.431 ( 0.128
0.134d
108
141
149 ( 17
65.6 ( 3.2
19.5 ( 2.8
9.29 ( 0.28
2.84 ( 0.37
1.12d
35.6c
42.2c
11.4c
2.62
13.7c
4.81
1.32
0.955
0.273d
0.519
16.2 ( 1.9e
5.81 ( 0.17e
13.5
36.1 ( 2.7e
16.1
1.83
5.21f
14.5 ( 1.3e
14.7 ( 0.9e
a
Values of KN in mM, determined at 25 °C, in an aqueous buffer
of pH 11.60. At this pH, the amines are ∼90% in their free base
forms (pKa ≈ 10.6), except for benzylamine (pKa ) 9.35).17 For the
more soluble amines (n-propyl, n-butyl, n-pentyl, cyclopentyl, and
cyclohexyl), the buffer was made from the amine whereas for the
others a 0.2 M phosphate buffer was used. The dissociation
constants that are presented with errors were determined in the
b
present work, as described in the Experimental Section. The
values for â-CD and hp-â-CD without errors are from a prior study
that used the displacement of a fluorescent probe,14b c Using the
competition between amines and phenolphthalein for â-CD, we
obtained very similar values of 38.4 ( 1.2 mM and 10.9 ( 0.6 mM
for n-butylamine and n-pentylamine, respectively, whereas for hp-
â-CD and the same two amines, the agreement was not quite as
d
good: 64.7 ( 2.7 mM and 17.1 ( 0.9 mM, respectively. Obtained
by extrapolation of the linear relationship between pKN and chain
length (Figure 1) because attempts to measure KN were not
successful, due to the low solubility of the amine and/or the
amine‚CD complex. e Determined by competition with phenol-
phthalein (see Experimental Section). f This value for cyclohexyl-
amine is somewhat suspect (see Table 3, footnote b, and the
Experimental Section).
Resu lts
The present studies were more demanding than previ-
ous ones10 because of the necessity to take account of the
binding of both substrates to cyclodextrins. However, we
took advantage of previous studies of ester cleavage to
provide dissociation constants for the {ester‚CD} com-
plexes,6-8 obtained under the same basic conditions (pH
11.6). These constants (KS) are given later in tables that
present kinetic results. Most of the dissociation constants
for the {amine‚CD} complexes formed by â-CD and hp-
â-CD were available from recent studies14b carried out
F igu r e 1. Chain length dependence of dissociation constants,
plotted as pKN ) -log KN, for the binding of primary n-
alkylamines to cyclodextrins, from the data from Table 1. The
symbols are: 9, R-CD; 0, â-CD; 1, hp-â-CD; 3, γ-CD.
expressly as preparation for work on aminolysis. The
remaining constants were determined by various meth-
ods described in the Experimental Section, as the need
arose. All of the dissociation constants for {amine‚CD}
complexes (KN) are collected in Table 1, and those for the
n-alkylamines are plotted in Figure 1 in the form of pKN
() -log KN) against the alkyl chain length, n.
There are two main features to note about the values
of KN. First, for a given amine the strength of binding
follows the order: R-CD > â-CD ∼ hp-â-CD > γ-CD, as
expected from the relative sizes of the CD cavities11 since
alkyl chains fit snugly in R-CD, less snugly in â-CD (and
hp-â-CD), and even less so in γ-CD.1,14,15 Second, the
strength of binding of the n-alkylamines increases mono-
tonically with the chain length (Figure 1), as found with
many derivatives.3,4,7,8,14,15 Somewhat surprisingly, the
(11) R-CD has 6 glucose units joined in a torus, â-CD has 7, and
γ-CD has 8. Thus, their cavities differ appreciably in width (∼6, ∼8,
and ∼10 Å, respectively) but not in depth (∼8 Å).1 "Hydroxypropyl-â-
cyclodextrin" (hp-â-CD) is â-CD in which most of the primary hydroxy
groups have been modified by 2-hydroxypropyl groups,1c,e so the width
of its cavity is close to that of â-CD but its depth may be altered.
(12) Bonora, G. M.; Fornasier, R.; Scrimin, P.; Tonellato, U. J . Chem.
Soc., Perkin Trans. 2 1985, 367.
(13) Tee, O. S.; Iyenger, N. R.; Takasaki, B. K. Can. J . Chem. 1993,
71, 2139.
(14) (a) Tee, O. S.; Gadosy, T. A.; Giorgi, J . B. J . Chem. Soc., Perkin
Trans. 2 1993, 1705. (b) Tee, O. S.; Gadosy, T. A.; Giorgi, J . B. Can. J .
Chem. 1996, 74, 736. (c) Tee, O. S.; Fedortchenko, A. A.; Loncke, P.
G.; Gadosy, T. A. J . Chem. Soc., Perkin Trans. 2 1996, 1243.
(15) (a) Matsui, Y.; Mochida, K. Bull. Chem. Soc. J pn. 1979, 52,
2808. (b) Satake, I.; Ikenoue, T.; Takashita, T.; Hayakawa, K.; Meda,
T. Bull. Chem. Soc. J pn. 1985, 58, 2746. (c) Satake, I.; Yoshida, S.;
Hayakawa, K.; Meda, T.; Kusumoto, Y. Bull. Chem. Soc. J pn. 1986,
59, 3991. (d) Park, J . W.; Song, H. J . J . Phys. Chem. 1989, 93, 6454.

Catalysis of Ester Aminolysis by Cyclodextrins
J . Org. Chem., Vol. 65, No. 21, 2000 6881
slopes of the plots for the four CDs are very close (0.50,
0.52, 0.49, 0.43), indicating virtually the same sensitivity
to chain length. We take these to mean that even if the
fit of an alkyl chain in a CD is not particularly snug, as
is conceivably the case with â-CD and γ-CD,15e the alkyl
chain adheres to the interior surface of the CD cavity or
that one or more of the glucose moieties of the two larger
CDs can tilt inward to provide a comparable “grip” on
the chain. In addition, a referee has suggested that the
almost constant slopes may reflect the similar extents
to which the amines must be desolvated to be included
in a CD. Regardless of its origin, there is a distinct
sensitivity of the strength of amine binding to the alkyl
chain, so that chain length dependence may be used as
a probe of alkylamine inclusion in the transition state
for aminolysis. Also, the linearity of the plots of pKN
against n (Figure 1) allowed us to estimate KN values
for n-octylamine in three cases where they were not
readily measured.16
We have carried out two distinct series of kinetics
experiments on CD-mediated aminolysis. First, we stud-
ied the cleavage of p-nitrophenyl acetate (pNPA) and
p-nitrophenyl hexanoate (pNPH) by various amines in
the presence of R-CD, â-CD, hp-â-CD, and γ-CD. Second,
we looked at the reaction of a series of p-nitrophenyl
alkanoates with n-heptylamine, in the presence of the
same CDs. As previously,10 the reaction medium was a
basic buffer of pH 11.6, in which the alkylamines (pKa ∼
10.6)17 are present largely in their free base forms.
As in earlier studies,9b,10 analysis of the kinetic data
was based on consideration of four processes: hydrolysis
of the ester (S) in the medium (eq 1); ester cleavage
through an {ester‚CD} complex (eq 2); reaction of the free
ester with the nucleophile (Nuc) (eq 3); reaction of the
{ester‚CD} complex with free Nuc (eq 4).18 For these four
reactions, and the condition that [S‚CD] < [S]0 , [CD]0,
the expected variation of kobsd with [Nuc] and [CD] is
given by eq 5.
F igu r e 2. Rate constants for the reaction of p-nitrophenyl
acetate with n-heptylamine, with and without â-CD. The
symbols for different [CD]o are as follows: 9, none; 0, 2.00; 2,
5.02; 4, 7.50; 1, 10.00 mM. The curves are calculated from eq
5, using fitted constants20 obtained from multiple linear
regression analysis in terms of eq 6.
in eq 6 by dividing through by fS ) KS/(KS + [CD]). For
values of kobsd obtained at various [CD]o and [Nuc]o, eq 6
is used to analyze the results by multiple linear regres-
sion,19 from which we obtain the constant kcN for reaction
of the nucleophile with CD-bound ester (eq 4). Note that
in the present work one must take account of the binding
of the amine nucleophile to the CD (eq 7) because it
affects the values of [CD] and [Nuc] that are used in the
analysis. Details of their calculation are given in the
Experimental Section. A value of kN can be obtained from
the multiple regression, but we prefer the more accurate
value obtained from the variation of kobsd with [Nuc]o in
absence of CD, when kobsd ) ku + kN[Nuc]o.
ku
S
9
8 products
(1)
(2)
kobsd/fS ) ku + kN[Nuc] + kc[CD]/KS +
kcN[Nuc][CD]/KS (6)
kc
S + CD yK z S‚CD
\
98 products
S
KN
kN
9
CD + Nuc y z CD‚Nuc
(7)
S + Nuc
8 products
(3)
(4)
The data shown in Figure 2 are for the reaction of
pNPA with n-heptylamine in the presence of â-CD for 5
different [CD]o and various [amine]o ) [Nuc]o. Multiple
linear regression of all 33 data points in terms of eq 6
gave an excellent correlation and fitted constants20 which
were used to generate the curves in the figure. This case
was carried out in detail, as were others previously,9b to
firmly establish the validity of the treatment, especially
for an amine used at low concentrations because of its
poor solubility in water. For the remaining cases, fewer
measurements were made, as detailed later.
kcN
S‚CD + Nuc
8 products
(kuKS + kc[CD]) + (kNKS + kcN[CD])[Nuc]
(KS + [CD])
kobsd
)
(5)
Equation 5 gave good descriptions of the data in
previous studies,9b,10 and it works well for aminolysis also,
as shown by the detailed example in Figure 2. However,
because eq 5 is nonlinear and bivariate, it is not easily
used in analysis and so we convert it to the linear form
At first glance, the data in Figure 2 are puzzling,
because kobsd decreases with [amine]o, when the CD is
(16) There is a strong correlation between pKN values for the binding
of n-alkylamines to â-CD and the analogous values for n-alkyl
alcohols4,15a (r ) 0.997), with near unit slope () 0.92 ( 0.04). Moreover,
the points for isobutyl, isopentyl, and cyclohexyl are very close to the
correlation line, and the point for cyclopentyl is only 0.4 pK units away.
(17) Dean, J . A. Handbook of Organic Chemistry; McGraw-Hill: New
York, 1987.
(18) The reaction of free ester with {CD‚Nuc}, which is kinetically
equivalent to the process in eq 4, is considered explicitly in the
Discussion.
(19) Mezei, L. M. Practical Spreadsheet Statistics and Curve-fitting
for Scientists and Engineers; Prentice-Hall: Englewood Cliffs, NJ ,
1990.
(20) The regression analysis gave: ku ) 0.0986 ( 0.0047 s-1; kc
)
0.925 ( 0.003 s-1; kN ) 14.9 ( 0.9 M-1 s-1; kcN ) 244 ( 5 M-1 s-1 (r
) 0.9999). For comparison, measurements of kobsd at five different
[amine]o, in the absence of a CD, gave ku ) 0.0937 ( 0.0005 s-1 and
kN ) 16.7 ( 0.2 M-1 s-1 (r ) 0.9999).

6882 J . Org. Chem., Vol. 65, No. 21, 2000
Gadosy et al.
Ta ble 2. Secon d -Or d er Ra te Con sta n ts (kN) for th e
Rea ction of Alk yla m in es w ith p-Nitr op h en yl Aceta te
(p NP A) a n d p-Nitr op h en yl Hexa n oa te (p NP H)a
amine
[amine]o, mM
pNPA
pNPH
n-propyl
n-butyl
n-pentyl
n-hexyl
n-heptyl
n-octyl
0-250
0-100
0-25
0-10
0-5
10.6 ( 0.1
13.0 ( 0.1
13.4 ( 0.1
15.8 ( 0.3
16.7 ( 0.2
16.7b
4.97 ( 0.08
6.67 ( 0.10
6.43 ( 0.03
8.46 ( 0.05
7.05 ( 0.10
7.05b
isobutyl
isopentyl
cyclopentyl
cyclohexyl
benzyl
0-100
0-25
0-70
0-29
0-12
7.51 ( 0.14
14.1 ( 0.2
2.73 ( 0.01
1.71 ( 0.01
2.99 ( 0.05
4.19 ( 0.04
7.88 ( 0.10
1.30 ( 0.02
0.837 ( 0.016
1.21 ( 0.01
a
Conditions as given in footnote a, Table 1. Units of kN are M-1
s-1. The range of amine concentration used to find kN is given
under [amine]o. b Not easily determined because of the low solubil-
ity of the amine in water, and so the value was assumed to be the
same as for n-heptylamine.
amine, isopentylamine, cyclopentylamine, cyclohexyl-
amine, and benzylamine in the absence of CDs. For
reasons of solubility, kN for n-octylamine could not be
obtained readily, and so we take it to be the same as for
n-heptylamine, since chain length effects on kN are
minimal beyond C4.23 Table 3 presents values of kcN
obtained for pNPA and pNPH reacting with the amines
in the presence of the four CDs, while Table 4 contains
values of kN and kcN for n-heptylamine reacting with a
series of p-nitrophenyl alkanoates.
F igu r e 3. Rate constants for the reaction of p-nitrophenyl
acetate with n-pentylamine in the absence of CD (9) and in
the presence of 5.0 mM R-CD (0), 2.0 mM â-CD (b), and 5.0
mM γ-CD (O). The curves through the data for the CDs were
calculated from eq 5, using fitted constants obtained from the
analysis based on eq 6.
present. In the absence of CD (the lowest set of data in
Figure 2), kobsd increases linearly with [amine]o, as
expected,21 but with added CD more complex dependence
is observed because of the combined effects of two
equilibria (eqs 2 and 7) and four competing kinetic
processes. As [amine]o is increased, some of the free CD
becomes bound to the amine, so that [CD] decreases, and
the contribution from the ester cleavage by the CD (eq
2) is reduced. At the same time, the contributions from
the two aminolysis pathways increase but only at high
[amine]o do they dominate, as seen in Figure 2 for the
set of points where [CD]o ) 2.0 mM. As shown by the
examples in Figure 3, an upward trend in kobsd at high
[amine]o is much more evident with smaller amines
because these can be taken to higher [amine]o and they
bind to CDs more weakly.22
In general, we made two sets of measurements of kobsd
with varying [amine]o for each combination of ester,
amine, and CD (e.g., Figure 3). The first set, with no CD
present, was used to find kN accurately, and the second
set, with CD present, was required to probe for kcN. The
two sets together were used to estimate kcN from the
multiple regression analysis in terms of eq 6. If the data
did not give a satisfactory analysis, measurements were
made at another [CD]o.
Tables 3 and 4 also contain other “kinetic” parameters,
derived from kN and kcN, that will be used in the
Discussion. First, there are third-order rate constants (k3)
for reaction of the ester, CD and amine (eq 8), where k3
) kcN/KS because of the kinetic equivalence of the
processes in eqs 4 and 8. Second, from k3 we have
evaluated kNc ) k3KN for the other kinetically equivalent
process in which free ester reacts with CD-bound amine
(eq 9). Third, the tables include “dissociation constants”,
KTS, which provide a measure of the strength of binding
of the transition states of CD-mediated reactions to the
CDs.3 Such constants, which have been used extensively
by us3,6-8,9b,10 and others,24 are obtained using an ap-
proach based on transition state theory.24a For the
present purposes, KTS is defined (eq 10) as the apparent
dissociation constant of the transition state of the CD-
mediated aminolysis (TS‚CD) (eq 4, 8, or 9) into the
transition state of normal aminolysis (TS) (eq 3) and the
CD.
k3
S + CD + Nuc
98 products
(8)
(23) The variations of kN for the linear n-alkylamines in Table 2
are not considered significant. They are attributed to the fact that some
experiments were carried out in amine buffers but others employed a
phosphate buffer (see Table 1 for details), where buffer catalysis (vide
infra) possibly intrudes.
Table 2 contains the values of kN found for pNPA and
pNPH reacting with six linear n-alkylamines, isobutyl-
(24) (a) The approach was pioneered by Kurz in the context of
catalysis by acids and bases: Kurz, J . L. J . Am. Chem. Soc. 1963, 85,
987. Acc. Chem. Res. 1972, 5, 1. Some other examples of its use are in
(b) Wolfenden, R. Acc. Chem. Res. 1972, 5, 10. (c) Leinhard, G. E.
Science (Washington, D. C.) 1973, 180, 149. (d) Kraut, J . Science
(Washington, D. C.) 1988, 242, 533. (e) Wolfenden, R.; Kati, W. M.
Acc. Chem. Res. 1991, 24, 209. (f) Pregel, M. J .; Dunn, E. J .; Buncel,
E. J . Am. Chem. Soc. 1991, 113, 3545. (g) Palmer, D. R. J .; Buncel, E.;
Thatcher, G. R. J . J . Org. Chem. 1994, 59, 5286. (h) Kirby, A. J . Angew.
Chem., Int. Ed. Engl. 1996, 35, 707. (i) Davies, D. M.; Foggo, S. J .;
(21) Rate expressions for ester aminolysis often contain a term in
[amine]2, but generally only for nonaqueous solutions and when
[amine] is high. In none of our experiments did we see any second
order dependence on amine. Likewise, for the aminolysis of phenyl
acetate by n-alkylamines in water, no term in [amine]2 was detected
by Bruice and co-workers: Bruice, T. C.; Donzel, A.; Huffman, R. W.;
Butler, A. R. J . Am. Chem. Soc. 1967, 89, 2106.
(22) At first sight, more clear-cut aminolysis data could be obtained
by carrying out experiments at a lower pH (10.6, say), where the pH-
dependent background processes (eqs 1 and 2) are 10 times slower.
However, at such a pH, binding of both the amine and its protonated
form to the CD would complicate a full analysis.
Paradis, P. M. J . Chem. Soc., Perkin Trans.
2 1998, 1597. (j)
Pirrinccioiglu, N.; Williams, A. J . Chem. Soc., Perkin Trans. 2 1998,
37.

Catalysis of Ester Aminolysis by Cyclodextrins
J . Org. Chem., Vol. 65, No. 21, 2000 6883
Ta ble 3. Con sta n ts for th e Rea ction of Alk yla m in es w ith
p-Nitr op h en yl Aceta te (p NP A) a n d p-Nitr op h en yl
Hexa n oa te (p NP H) in th e P r esen ce of Cyclod extr in sa
Ta ble 4. Con sta n ts for th e Rea ction of n -Hep tyla m in e
w ith p-Nitr op h en yl Alk a n oa tes in th e P r esen ce of
Cyclod extr in sa
kcN
,
k3,
kNc
,
KTS
,
KS,
mM
kcN
,
k3,
kNc
,
KTS,
amine
M-1 s-1
kcN/kN M-2 s-1 M-1 s-1 kNc/kN mM
ester
M-1 s-1
kcN/kN M-2 s-1 M-1 s-1 kNc/kN mM
(a) pNPA + R-CD (KS ) 10.5 mM)
(a) R-Cyclodextrin
10.5 126 ( 5 7.5 12000
n-propyl
n-butyl
n-pentyl
n-hexyl
n-heptyl
n-octyl
13.7 ( 0.8
16.7 ( 1.2
12.2 ( 1.0
19.2 ( 3.3
126 ( 4
1.3
1.3
0.91
1.2
7.5
1300
1600
61
21
5.7
1.6
0.27 12
0.19 8.6
0.31 1.4
0.35 0.38
8.1
8.2
acetate
propanoate 7.4 51.9 ( 3
5.2
3.0
1.8
1.5
1.9
0.31 1.4
0.25 1.7
0.24 1.8
0.20 2.2
0.27 1.6
4.3
7000
4200
3600
4400
1200
1800
12000
44000
3.6
butanoate 5.0 21.0 ( 8.3
2.8
1.6
1.8
2.9
5.2
5.8
pentanoate 3.4 12.1 ( 2.5
hexanoate
2.9 12.8 ( 2.0
457 ( 18
27
(b) â-Cyclodextrin
15 31000
8.8 20000
(b) pNPA + â-CD (KS ) 7.92 mM)
acetate
propionate
butanoate
7.9 244 ( 5
29
1.8
1.6
2.9
3.3
5.4
0.54
n-propyl
n-butyl
n-pentyl
n-hexyl
n-heptyl
n-octyl
14.3 ( 0.8
17.2 ( 2.0
24.7 ( 0.5
181 ( 2
1.4
1.3
1.8
11
15
76
1.3
4.3
3.6
11
4.3
1800 195
18
5.9
6.0
4.3
0.69
0.54
0.11
6.0
1.9
2.2
5.2 105 ( 3
19
22
26
38
0.59
0.33
0.29
0.18
2200
3100
23000
31000
160000
1300
77
36
60
29
44
20
44
17
6.0
2.7
3.8
1.8
2.6
2.7
3.1
6.1
2.6
7.9
2.7 60.7 ( 8.3
8.2 22000
6.9 27000
9.1 40000
pentanoate 2.0 54.0 ( 2.5
hexanoate
1.6 63.9 ( 2.0
244 ( 5
(c) Hydroxypropyl-â-cyclodextrin
1265 ( 4
9.94 ( 2.0
60.1 ( 2.9
acetate
8.2 59.1 ( 6.4
3.5
1.3
1.1
1.8
1.7
7200
3200
3100
7200
7600
9.5
0.57 2.3
0.35 3.8
0.55 2.4
isobutyl
isopentyl
propanoate 5.1 16.0 ( 2.8
butanoate 2.7 8.36 ( 0.57
4.1
4.1
9.7
7600
1200
cyclopentyl 9.78 ( 0.67
cyclohexyl 19.2 ( 2.5
pentanoate 1.9 14.0 ( 0.9
1.3
1.4
1.1
0.93
2400
1600
4.4
24
0.71
1.8
hexanoate
1.6 12.1 ( 1.0
10
benzyl
12.9 ( 0.5
(d) γ-Cyclodextrin
(c) pNPA + hp-â-CD (KS ) 8.20 mM)
acetate
36 736 ( 24
44
63
36
36
25
28
20000
58
67
50
53
46
47
3.5
5.5
6.8
6.9
6.6
6.7
0.82
0.51
0.42
0.41
0.43
0.42
n-propyl
n-butyl
n-pentyl
n-hexyl
n-heptyl
n-octyl
5.13 ( 0.78
8.55 ( 1.08
6.67 ( 1.14
57.2 ( 5.1
59.1 ( 6.4
185 ( 4
0.48
0.66
0.50
3.6
3.5
11
1.8
630
1040
810
7000
7200
23000
590
88
44
11
34
9.5
12
9.5
8.3 17
3.4 13
0.83 17
propanoate 32 749 ( 8
butanoate 15 264 ( 2
pentanoate 15 282 ( 2
hexanoate 11 179 ( 5
heptanoate 12 200 ( 1
23000
18000
19000
16000
17000
2.1
2.3
0.57 2.3
0.70 0.74
cyclopentyl 4.82 ( 0.30
3.5
4.6
a
At 25 °C, in a 0.2 M phosphate buffer of pH 11.60. The values
of KS are taken from earlier studies.7,8 For reaction of the esters
with n-heptylamine in the absence of CDs the rate constants (kN,
in M-1 s-1) are: acetate, 16.7 ( 0.2; propanoate, 12.0 ( 0.1;
butanoate, 7.41 ( 0.04; pentanoate, 7.78 ( 0.06, hexanoate, 7.05
( 0.10; heptanoate, 7.05 (assumed to be the same as the hex-
anoate).
cyclohexyl
b
(d) pNPA + γ-CD (KS ) 35.6 mM)
n-propyl
n-butyl
n-pentyl
n-hexyl
n-heptyl
n-octyl
33.9 ( 1.8
24.9 ( 2.1
51.2 ( 2.3
263 ( 7
736 ( 24
2960 ( 20
3.2
1.9
3.8
17
44
950 142
13
11
700
1400
46
28
69
58
93
3.5 19
2.1
4.3
3.5
5.6
9.3
2.1
0.81
0.20
7400
21000
83000
180
kNc
(e) pNPH + R-CD (KS ) 2.90 mM)
S + CD‚Nuc
KTS ) [TS][CD]/[TS‚CD] ) kNKS/kcN ) kN/k3 )
kNKN/kNc (10)
8 products
(9)
n-propyl
n-butyl
n-pentyl
n-hexyl
n-heptyl
n-octyl
4.13 ( 0.26
5.47 ( 0.38
5.02 ( 0.46
4.45 ( 0.99
12.8 ( 0.5
114 ( 26
0.83
0.82
0.78
0.53
1.8
1400
1900
1700
1500
4400
39000
66
25
5.3
2.5
1.9
5.3
13
3.8
3.5
3.5
0.83 3.7
0.29 5.5
0.27 1.6
0.75 0.18
16.2
Discu ssion
(f) pNPH + â-CD (KS ) 1.60 mM)
n-propyl
n-butyl
n-pentyl
n-hexyl
n-heptyl
n-octyl
1.58 ( 0.19
3.84 ( 0.22
6.89 ( 0.12
43.5 ( 2.1
63.9 ( 2.0
270 ( 17
0.32
0.58
1.1
5.1
9.1
990 110
21
13
7.6
8.4
5.4
6.6
4.2
6.3
5.0
2.8
1.5
2400
85
49
71
38
47
17
49
25
Over view . The aminolysis of aryl esters25 takes place
most simply by formation of a zwitterionic intermediate
T( which undergoes fast deprotonation at high pH,
followed by the loss of aryloxide ion from its anion, T-
(Scheme 1). In some instances T- is formed directly from
the reactants by general base assistance from hydroxide
ion, a second molecule of amine, or the buffer base.26
Regardless of the finer details of the mechanism, the rate-
limiting transition state should resemble T( or T- and,
by inference from Hammond’s Postulate,27 factors that
can stabilize these intermediates are expected to lower
the overall activation barrier.
4300
27000
40000
170000
1100
8500
1900
4900
2300
0.31
0.18
0.042
3.9
0.93
0.69
0.17
0.52
38
isobutyl
isopentyl
1.72 ( 0.25
13.6 ( 0.4
0.41
1.7
2.3
9.3
3.1
cyclopentyl 3.01 ( 0.06
cyclohexyl 7.78 ( 0.17
20
9.0 11
34
benzyl
3.71 ( 0.05
28
(g) pNPH + hp-â-CD (KS ) 1.59 mM)
n-propyl
n-butyl
n-pentyl
n-hexyl
n-heptyl
n-octyl
0.662 ( 0.336
0.512 ( 0.194
2.24 ( 0.11
9.32 ( 0.45
12.1 ( 1.0
0.13
0.077
0.35
1.1
1.7
3.3
420
320
59
14
19
28
10
12
2.0 21
12
1400
5900
7600
14500
450
3.0
3.3
1.4
1.1
5.5
4.6
1.4
0.93
0.49
2.9
23.1 ( 2.6
7.5
7.2
cyclopentyl 0.709 ( 0.039
0.55
(25) (a) Background references on ester aminolysis can be found in:
Adalsteinsson, H.; Bruice, T. C.; J . Am. Chem. Soc. 1998, 120, 3440,
and Koh, H. J .; Shin, C. H.; Lee, H. W.; Lee, I. J . Chem. Soc., Perkin
Trans. 2 1998, 1329. (b) DeTar, D. F.; Delahunty, C. J . Am. Chem.
Soc. 1983, 105, 2734. (c) Henge, A. C.; Hess, R. A. J . Am. Chem. Soc.
1994, 116, 11256.
(26) Because of general base catalysis, pseudo-first-order rate
constants for the aminolysis of esters may contain terms in [amine]2,
[amine][OH-], and [amine][buffer], depending on the reactants and
the solvent.25 However, as remarked above,21 our plots of kobsd against
[amine]o in the absence of a CD were all strictly linear.
cyclohexyl
b
(h) pNPH + γ-CD (KS ) 10.7 mM)
n-propyl
n-butyl
n-pentyl
n-hexyl
n-heptyl
n-octyl
6.63 ( 0.8
4.85 ( 2.0
12.3 ( 0.5
22.7 ( 0.7
179 ( 5
1.3
0.73
1.9
2.7
25
98
620
450
1200
2100
17000
64000
92
30
22
20
48
72
19
4.5 15
3.5
2.3
6.7
8.0
5.6
4.0
0.42
0.11
688 ( 4
10
a
See footnote a, Table 1. The values of KS are taken from earlier
work.7,8 For the reactions of pNPA and pNPH with cyclohexy-
b
(27) (a) Hammond, G. S. J . Am. Chem. Soc. 1955, 77, 334. (b)
Maskill, H. The Physical Basis of Organic Chemistry; Oxford University
Press: Oxford, 1985; pp 261-263.
lamine in the presence of hp-â-CD the data analysis was unac-
ceptable (see Experimental Section).

6884 J . Org. Chem., Vol. 65, No. 21, 2000
Gadosy et al.
Sch em e 1
Reactions proceeding through the CD-bound termo-
lecular transition states 3 to 7 are kinetically equivalent
and distinction between them is possible only by analyz-
3,7,8
ing how kinetic parameters (e.g., kcN/kN, k3, and KTS
)
In a CD-mediated aminolysis, stabilization of the
intermediates T( and T-, and the transition states
adjacent to them, may occur by binding of the CD to one
of three loci: the alkyl portion (R) of the alkylamino
group, the alkyl group (R′) of the proacyl group, and the
aryl group (Ar) of the nucleofuge (Scheme 1). A preference
for binding one of these three groups will depend on their
nature, but in the absence of any unusual interactions it
should follow established trends. If both R and R′ are
quite small, the CD is expected to bind preferentially to
the aryl group (Ar ) p-nitrophenyl). For a long ester, and
short amine, the preference may be for inclusion of R′ of
the proacyl group. However, for reaction by a long amine
with a short ester the transition state may be bound by
the alkylamino group, R. Therefore, at the outset we
envisage three main possibilities for aryl ester aminolysis
mediated by a cyclodextrin: (i) reaction of the aryl-bound
ester with amine external to the CD cavity (3); (ii)
reaction of the acyl-bound ester with the amine outside
the CD cavity (4); (iii) reaction between CD-bound amine
with the free ester (5).
vary with the structures of the ester, the amine, and the
CD. If the reaction occurs with aryl group inclusion, 3,
the parameters should be insensitive to R-NH2 and R′-
CO, while for reaction through 4 they should vary with
the ester acyl group R′CO, and for transition state 5 they
should vary with the alkylamino group, RNH2. Likewise,
for the transition state 6, the parameters should show
sensitivity to the alkyl groups of both the amine and the
ester, but if it involves 7 then there should be sensitivity
to Ar and RNH2, but not to R′CO.
Even a cursory viewing of Tables 3 and 4 will note that
the various kinetic parameters for CD-mediated ami-
nolysis vary substantially with the amine, the ester and
the CD. For instance, the third-order rate constants k3
(eq 8) vary from 590 to 160000 M-2 s-1 for pNPA, and
from 320 to 170000 M-2 s-1 for pNPH (Table 3). More-
over, there are specific trends in the parameters as the
chain lengths of R and R′ are increased, and there are
differences between the effects of the four CDs.
Rea ctivity Ra tios (kcN/kN a n d kNc/kN). The ratio kcN
/
kN is a measure of the relative reactivities of the CD-
bound and free forms of the ester toward the nucleophile,
assuming the CD-mediated reaction is that in eq 4,
whereas kNc/kN measures the relative reactivities of the
CD-bound and free amine toward unbound ester for
reaction as in eq 9. Note also, that by virtue of the form
of KTS in eq 10, the logarithms of the reactivity ratios
reflect differences between transition state binding and
substrate (ester or amine) binding, as shown in eq 11.
log(kcN/kN) ) pKTS - pKS; log(kNc/kN) ) pKTS - pKN
(11)
For the reaction of nonbinding nucleophiles with
p-nitrophenyl alkanoate esters values of kcN/kN are small
and close to one (0.14 to 1.7),10 and for the anions of TFE
and 2-ME there are biphasic variations in these ratios
(and other parameters) with ester chain length that
indicate a switch from reaction mode 1 for the short
esters to 2 for the longer esters.10b The present study
shows that for the aminolysis of the same esters in the
presence of CDs the corresponding ratios vary much more
widely, from 0.08 to 180 (Table 3).
If the alkylamine and the alkanoate ester both have
fairly long alkyl groups then a competition between
binding of R-NH2 and R′-CO is set up, and it is not clear
what the preferred mode of transition state binding might
be (4 or 5). Moreover, with γ-CD, and maybe even with
â-CD, there is the possibility of partial inclusion of both
of the alkyl groups, R and R′ (6). Likewise, if the CD
cavity is large enough to permit it, there might be some
inclusion of the aryl group of the ester and the alkyl
group of the amine (7).
The cleavage of pNPA by n-alkylamines exhibits kcN
/
kN ratios that are small, not far from one, and nearly
constant for short amines (3, 4, 5, or 6 carbons), after
which they increase dramatically (Figure 4), up to 180
for catalysis by γ-CD (Table 3). Similar trends are
observed for the aminolysis of pNPH, although the ratios
are generally smaller (Table 3). This largely biphasic
behavior, which is seen for all four CDs (Figure 4), is
evidence that short alkylamines react with the CD-bound

Catalysis of Ester Aminolysis by Cyclodextrins
J . Org. Chem., Vol. 65, No. 21, 2000 6885
meaning that aminolysis of the esters by n-heptylamine
is definitely catalyzed by γ-CD.
The ratios kNc/kN for â-CD and hp-â-CD are intermedi-
ate between those for R-CD and γ-CD (Table 4), as
expected if the differences are related to cavity size.
Conceivably, R-CD holds the alkylamino group rigidly so
that the aminolysis is impaired (kNc/kN < 1), but the much
wider cavity of γ-CD11 is large enough that it can partially
accommodate the aryl group of the ester moiety, as well
as the alkylamino group (7), or at least interact favorably
with it. In this regard, it is important to note that by
themselves p-nitrophenyl alkanoates bind to γ-CD by aryl
group inclusion,8b and not by acyl group inclusion, which
is the case for the longer (>C4) esters binding to the other
CDs.7,8
Values of kNc/kN for reaction of the esters with n-
heptylamine in the presence of â-CD and hp-â-CD are
not constant, and they increase about 3-fold from the
acetate to the hexanoate (Table 4). These increases
conceivably denote partial inclusion of the longer ester
chains, as well as of the amine chain (6). Comparing hp-
â-CD to â-CD, both reactivity ratios are smaller and
closer to one for the alkylated CD derivative, perhaps
because its hydroxypropyl groups on the primary side to
the cavity1c form an intrusive floor28 that reduces its
ability to bind to the transition state for aminolysis,
whatever the moiety included in the cavity.
F igu r e 4. Dependence of reactivity ratios kcN/kN (Table 3) on
amine chain length for the CD-mediated reaction of pNPA with
n-alkylamines. The symbols are for the CDs as follows: 9,
R-CD; ], â-CD; 1, γ-CD.
ester (3) but longer amines are bound to the CD and they
react with free ester (5). Alternatively, if the CD-
mediated reaction entails reaction of free ester with CD-
bound amine (eq 9) the relevant ratios, kNc/kN, show
exactly opposite dependencies on amine chain length as
they decrease for the short amines and they are es-
sentially constant for the longer amines. This behavior
is also consistent with a change in the mode of reaction
from 3 to 5, as the amine chain becomes longer. Note
that in most cases where reaction is mediated by â-CD
and γ-CD both reactivity ratios are greater than one,
which means that the aminolysis is truly catalyzed in
those cases, regardless of the precise mechanism.
Th ir d -Or d er Rea ctivity (k3). A more unbiased view
of CD-mediated aminolysis is obtained by looking at the
rate constants for the third-order process in eq 8. These
constants (k3) are a measure of the Gibbs energy of the
transition state composed of {amine + ester + CD},
relative to the separate reactants in the initial state,
without any assumptions being made about the reaction
mechanism. Variations of k3 with the structures of the
reactants may provide insight into the structure(s) of the
termolecular transition states.3
For the reaction of various amines with pNPA and
pNPH, mediated by the four CDs, k3 varies over a 270-
fold range for pNPA and over a 530-fold range for pNPH
(Table 3 and Figure 5). Less than a factor of 10 of these
large ranges are due to differences in amine/ester reac-
tivity (see kN values, Table 2), and so they must be due
to variations in the transition state stabilization afforded
by the different CDs.
The CD-mediated reaction of n-heptylamine with p-
nitrophenyl alkanoates shows ratios of kcN/kN that vary
only marginally with acyl chain length but substantially
with the CD (Table 4). This insensitivity to ester chain
length is evidence against reaction by acyl group binding
(4), but it does not preclude some participation of the aryl
group and it is consistent with alkylamino group binding
(5), as suggested for longer amines, in the previous
paragraph. Most noteworthy, the ratios kcN/kN for any
given ester increase in the order: γ-CD > â-CD > R-CD
g hp-â-CD, with those for γ-CD being 10 to 30 times
greater than those for hp-â-CD. Also, the ratios for γ-CD
are all much greater than one (25 to 63), implying
substantial catalysis if the reaction takes place between
bound ester and free amine (eq 4, and 3).
As seen in Figure 5, the variations of k3 on amine chain
length are biphasic, being comparatively invariant for
short amines and increasing steeply for the longer ones.
With R-CD the point of inflection for both esters is at C6
while for the wider γ-CD it is at C4. The picture is less
consistent for â-CD and hp-â-CD, with the breakpoint
being at C4 (pNPH/hp-â-CD), or C5 (pNPA/â-CD and
pNPA/hp-â-CD), or barely discernible (pNPH/â-CD). Nev-
ertheless, Figure 5, like Figure 4, provides strong evi-
dence of a general change in the mode of transition state
binding from aryl group inclusion (3) for short amines to
alkylamino group inclusion for long amines (5), and the
possibility of reaction with alkylamino group inclusion
and some acyl group or aryl group binding (6 or 7).
For the reaction of n-heptylamine with p-nitrophenyl
alkanoates in the presence of R-CD, the ratios kNc/kN fall
in the narrow range 0.20-0.31 (Table 4) and are insensi-
tive to the ester chain length. With γ-CD, values of kNc
/
kN are 10 to 30 times larger and remarkably constant
for the C4 to C7 esters (6.8, 6.9, 6.6, 6.7). These results
for R-CD and γ-CD are compatible with transition state
binding of the alkylamino group (5), and not of the acyl
group (4), but there must be some additional interaction
to account for the 10- to 30-fold greater reactivity
observed with γ-CD. Also with respect to the results for
γ-CD in Table 4, note that both kcN/kN and kNc/kN are
greater than one (25-60 and 3.5-6.9, respectively),
(28) The cleavage of nitrophenyl alkanoate esters by hp-â-CD is less
efficient than that by â-CD.8a Other functionalities on the primary side
of â-CD that provide an intrusive floor to the cavity generally lead to
slower ester cleavage. For examples, see: Breslow, R.; Czarniecki, M.
F.; Emert, J .; Hamaguchi, H. J . Am. Chem. Soc. 1980, 102, 762. Fujita,
K.; Shinoda, A.; Imoto, T. J . Am. Chem. Soc. 1980, 102, 1161.

6886 J . Org. Chem., Vol. 65, No. 21, 2000
Gadosy et al.
a combination of alkylamino inclusion and aryl inclusion
(7), bearing in mind that p-nitrophenyl alkanoates bind
to γ-CD by aryl group inclusion.8b
Tr a n sition Sta te Sta biliza tion (pKTS). In our view,3
the best parameter for probing transition state binding
is KTS, the apparent dissociation constant of the CD-
containing transition state (eq 10), and pKTS ) -log KTS
.
Since pKTS is a measure of the stabilization of a transition
state by the CD, variations of it with structure can
provide information about transition state binding.3,24 For
the present results, the variations of pKTS with the amine,
the ester, and the CD are shown in Figures 6, 7, and 8.
The dependence of pKTS on the binding ability of the
amine (pKN) for aminolysis of pNPA and pNPH in the
presence of R-CD is distinctly biphasic (Figure 6a). For
short amines (C3 to C6), pKTS is insensitive to pKN, as
expected for aryl inclusion in the transition state (3) but
note that pKTS is decidedly larger for pNPH than for
pNPA, implying that there is some additional stabiliza-
tion due to the longer acyl chain of pNPH. For amines
beyond C6, the values of pKTS rise steeply with pKN, with
slopes near 1 (1.3 and 1.4) that are perfectly in accord
with the reaction taking place between CD-bound amine
and free ester (5). Furthermore, for the long amines the
values of pKTS for pNPA and pNPH are essentially the
same, also consistent with a transition state that has the
ester acyl chain outside the R-CD cavity, as in 5.
Figures 6b and 6c show plots of pKTS vs pKN for the
aminolysis of pNPA and pNPH in the presence of â-CD
and hp-â-CD. With hp-â-CD the dependence is also
essentially biphasic, in keeping with a switch from mode
3 to mode 5, the breakpoints being at C4 for pNPA and
at C3 for pNPH. Likewise, for pNPA reacting in the
presence of â-CD, there is a change in behavior at C3 or
C4, but for pNPH there is no clear inflection as pKTS rises
consistently with pKN (Figure 6b). The values of pKTS for
pNPA and pNPH reacting with n-propylamine are es-
sentially equal, as expected if both react with aryl group
inclusion (3). For amines longer than n-propyl, pKTS for
pNPH increases with pKN linearly (r ) 0.992) and with
a slope of 0.80, implying that the ester reacts with
inclusion of the alkylamino group, as in 5 (or 6). However,
since for both â-CD and hp-â-CD, the pKTS values are
larger for pNPH than for pNPA there appears to be
additional stabilization due to the longer acyl chain of
pNPH, possibly due to partial acyl inclusion, as in 6.
F igu r e 5. Chain length dependence of the third-order rate
constants (k3, Table 3) for the reaction of ester (pNPA or
pNPH) + n-alkylamine + CD. (a) Ester ) pNPA; (b) ester )
pNPH. In both cases the symbols for the CDs as follows: 9,
R-CD; 0, â-CD; 1, hp-â-CD; 3, γ-CD.
By contrast, values of k3 for the reaction of p-nitro-
phenyl alkanoates with n-heptylamine show little de-
pendence on the ester chain (Table 4). Broadly speaking,
for a given ester the reactivity varies as: â-CD > γ-CD
> R-CD > hp-â-CD, which reflects the relative amounts
of transition state stabilization3 afforded by the CDs (see
below). Therefore, there must be stabilizing factors with
â-CD and γ-CD which are reduced or absent in the case
of R-CD and hp-â-CD. For R-CD, k3 decreases from the
C2 to C4 esters and then remains constant, parallel to kN
for reaction in the absence of the CD (Table 2), and
consistent with alkylamino group inclusion and no acyl
group inclusion in the transition state (i.e., 5). With â-CD
and hp-â-CD, k3 values decrease somewhat and then
increase, suggesting at least two factors are involved and
that there is switch between their dominance. Conceiv-
ably, for the long esters there is partial acyl chain
inclusion, as well as of the alkylamino group (6). Most
remarkably, the six values of k3 for catalysis by γ-CD are
almost constant (17000 to 23000 M-2 s-1), consistent with
a transition state involving alkylamino inclusion (5) or
As shown in Figure 6d, for catalysis by γ-CD, there is
little difference between the values of pKTS for the
aminolysis of pNPA and pNPH, suggesting that the
disposition of the two esters in their transition states are
virtually identical. Both esters exhibit a similar biphasic
variation of pKTS with pKN, with little change for the C3
and C4 amines and then a dramatic increase. For the two
shortest amines, the reaction probably occurs between
aryl-bound ester and free amine (3), as for the three other
CDs (see above). From the C4 amine onward, pKTS vs pKN
has slopes of 1.1 and 1.2, which are entirely consistent
with alkylamino group binding in the transition state,
as in 5 or 7 (discussed below).
To probe further the transition state of CD-mediated
aminolysis, we consider how pKTS varies with the chain
lengths of the esters reacting with n-heptylamine in the
presence of R-CD, â-CD, and γ-CD (Figure 7). In the case
of R-CD, the values are low and they hardly vary with
the ester, as expected if there is transition state binding

Catalysis of Ester Aminolysis by Cyclodextrins
J . Org. Chem., Vol. 65, No. 21, 2000 6887
F igu r e 6. Correlation of transition state stabilization (pKTS) with the strength of amine binding (pKN) for the CD-mediated
reaction of pNPA (9) and pNPH (0) with n-alkylamines: (a) R-CD; (b) â-CD; (c) hp-â-CD; (d) γ-CD.
0.2-0.3 (Table 4), implying that n-heptylamine bound to
R-CD is 3-5 times less reactive than the free amine.
For reaction in the presence of γ-CD, pKTS values show
little variation with ester chain length, and they are
much greater than those for R-CD (Figure 7) because the
reaction is definitely catalyzed by γ-CD: both kcN/kN and
k
Nc/kN > 1 (Table 4). This behavior is consistent with
transition state binding of the n-heptylamino group (5),
but the lower barrier with γ-CD is not simply due to
binding of this group because n-heptylamine binds more
weakly to γ-CD than to R-CD (Table 1). Conceivably,
there is partial inclusion of the p-nitrophenyl group of
the ester (7) or some favorable interaction between it and
the rim of the γ-CD cavity.
In contrast to those for R-CD and γ-CD, pKTS values
for â-CD increase with ester chain length (Figure 7), as
do those for hp-â-CD. If pKTS is plotted against pKS for
ester binding, the overall slopes are ∼0.7 for both CDs,
implying appreciable sensitivity to ester structure. Fur-
thermore, from the C3 to C6 esters the slopes of pKTS
against pKS are essentially 1 (0.95 for â-CD; 1.27 for hp-
â-CD), indicating that the transition state binding is as
F igu r e 7. Dependence of transition state stabilization (pKTS
)
on ester chain length for the CD-mediated reaction of n-
heptylamine with p-nitrophenyl alkanoates (acetate to hep-
tanoate). The symbols are: 0, R-CD; 9, â-CD; 3, γ-CD.
of the n-heptylamino group but not of the acyl group of
the ester (5). For this reaction pathway (eq 9), kNc/kN
)

6888 J . Org. Chem., Vol. 65, No. 21, 2000
Gadosy et al.
for aminolysis of the esters by all 11 amines in the
presence of â-CD. The two sets of data correlate quite
well, but in almost all cases pKTS for pNPH is larger, and
by a fairly constant amount, implying that there is an
additional stabilizing interaction in the transition state
due to the longer hexanoyl acyl chain of pNPH. Thus,
Figure 8 provides additional support for the idea that CD-
catalyzed reaction of longer esters and larger amines
takes place with some inclusion of the ester acyl chain
(6).
Con clu sion s
Detailed kinetic studies have shown that the aminoly-
sis of p-nitrophenyl alkanoate esters by primary alkyl-
amines is mediated by cyclodextrins and in several cases
it is catalyzed by them. From the variations of kinetic
parameters with structure it is proposed that at least two
and possibly four different modes of transition state
binding are operative: structures 3 and 5, and possibly
6 and 7. The principal conclusions are as follows:(i) the
reaction of short alkanoate esters with short amines
(<C6) occurs by the attack of free amine on ester bound
to the CD by its aryloxy group (3); (ii) for longer amines,
the reaction is between free ester and CD-bound amine
(5); (iii) the reaction of long esters with long amines is
catalyzed by â-CD (and to a lesser extent by hp-â-CD)
and it may involve by partial inclusion of the alkyl
moieties of both reactants (6); (iv) γ-CD catalyzes ami-
nolysis and its larger cavity of appears to allow the
partial inclusion of the aryloxy group of the ester as well
as the alkyl group of the amine in the transition state
(7); (v) for none of the ester/amine/CD systems studied
did reaction between acyl-bound ester and free amine (4)
appear to be important; (vi) for any given ester, the
catalysis is largest for long amines and the larger CDs.
Finally, we point out that in a companion study of the
aminolysis of naphthyl acetates we recently found ap-
preciable catalysis by â-CD.29
F igu r e 8. Comparison of transition state stabilization (pKTS
)
by â-CD for the reactions of pNPH and pNPA with alkyl-
amines: linear amines, 9; branched and cyclic amines, 0. For
almost every amine the transition state stabilization for pNPH
is greater than that for pNPA.
sensitive as substrate binding to the structure of the
esters, as well as being very sensitive to the amine
structure (Table 3 and Figure 6b). Thus, it seems justifi-
able to suggest that with â-CD and hp-â-CD the ami-
nolysis transition state is stabilized by some acyl group
binding as well as by alkylamino group binding (6).
Br a n ch ed a n d Cyclic Am in es. Up to this point we
have said nothing about the nonlinear amines: isobutyl-,
isopentyl-, cyclopentyl-, cyclohexyl-, and benzylamine,
aside from citing their KN values (Table 1). As antici-
pated, these nucleophiles are less reactive than n-
alkylamines toward pNPA and pNPH for steric reasons25b
(kN, Table 2), except for isopentylamine where the
branching is far enough from the nitrogen that it reacts
as fast as n-pentylamine. Also, as expected, the branched
and cyclic amines react about twice as rapidly with pNPA
as with pNPH.
The effects of â-CD on the aminolysis of pNPA and
pNPH by the nonlinear amines are comparable to those
for the n-alkylamines, as judged by the ratios kcN/kN and
kNc/kN (Table 3), which are greater than one in almost
all cases, meaning that the reactions are catalyzed.
Correspondingly, the nonlinear and linear amines show
similar transition state binding (pKTS) in relation to
amine binding (pKN), but there are minor differences. For
reaction with pNPA, the two branched amines (and
cyclohexylamine) have very similar parameters to the
linear amines, but cyclopentylamine and benzylamine
bind more strongly in the transition state. For reaction
with pNPH, all three cyclic amines show superior transi-
tion state binding whereas the two branched amines bind
less well in the transition state, compared to the linear
amines. These differences presumably reflect the bulki-
ness of the amine nucleophile and how it influences the
stability of the transition state structure.
Exp er im en ta l Section
Ma ter ia ls. The amines, 1-anilino-8-naphthalenesulfonic
acid (ANS), R-CD, â-CD, phenolphthalein, and most other
chemicals were purchased as the best grades available from
the Aldrich Chemical Co., whereas γ-CD and "hydroxypropyl-
â-cyclodextrin" (hp-â-CD) were obtained from Wacker Chemie
(Munich, Germany). As in earlier studies,8,14 the hp-â-CD had
an average molecular weight of 1500, corresponding to the
alkylation of about six of the seven primary hydroxy groups
of â-CD by 2-hydroxypropyl groups,1c,28 and this formulation
was confirmed by electrospray mass spectroscopy. Most of the
esters were obtained from the Sigma Chemical Company
except for p-nitrophenyl heptanoate which had been synthe-
sized previously.8a Aqueous buffers were made up from the
best grades of the buffer reagents available and with water,
doubly distilled from glass.
Kin etic Exp er im en ts. These were carried out much the
same way as in previous work on ester cleavage.8,10 Equal
volumes of the amine dissolved in a basic aqueous buffer were
mixed (1:1) with an aqueous solution of the ester (10-100 µM,
depending on solubility) and reacted in a stopped-flow spec-
trophotometer. The concentrations of all species after mixing
were half of those initially present in the individual syringes.
For the experiments in the presence of a CD, one syringe
contained the amine in buffer, while the other syringe con-
tained an aqueous solution of the ester and the CD (which aids
A more important point is that the nonlinear amines,
as well as the n-alkylamines, show stronger transition
state binding for reaction with pNPH than with pNPA.
This point is emphasized by Figure 8 which shows the
values of pKTS for pNPH plotted against those for pNPA,
(29) Tee, O. S.; Boyd, M. J . Can. J . Chem. 1999, 77, 950.

Catalysis of Ester Aminolysis by Cyclodextrins
J . Org. Chem., Vol. 65, No. 21, 2000 6889
in solubilization of the esters). The solutions containing amine
were all set to pH 11.60 in a buffer. For the less soluble amines
(n-hexyl, n-heptyl, n-octyl, isobutyl, isopentyl, and benzyl) the
buffer was 0.4 M phosphate; for the remaining amines the
buffer was made using the amine itself. The amine concentra-
tions used were varied with the amine solubility, with five to
seven different [amine]o ranging from 0 to 4.00 mM (for
n-octylamine) to 0-250 mM (for n-propylamine).
The kinetics of cleavage of the nitrophenyl esters were
followed by monitoring the first order appearance of p-
nitrophenoxide ion at 405 nm, using an Applied Photophysics
SX17MV Stopped-flow Spectrophotometer, with the temper-
ature of the cell kept at 25.0 ( 0.1 °C. Normally, 5 to 10 good
absorbance traces were collected and averaged before estima-
tion of kobsd by nonlinear least-squares fitting of an exponential
growth curve to the data (400 points). With the least soluble
esters, present at very low concentration, more traces were
averaged if necessary.
Da ta An a lysis. Values of kobsd obtained at different levels
of [CD]o and [amine]o were analyzed in terms of eq 6 (Nuc )
amine) by multiple linear regression analysis of the depen-
dence of kobsd/fS on [CD], [amine], and [CD][amine] using Lotus
123 spreadsheet software.19 From this analysis the rate
constant kcN was obtained from the fitted coefficient for the
term in [CD][amine], which equals kcN/KS () k3). Although a
value of kN can be obtained from the same analysis, it was
more accurately determined from measurements made in the
absence of CD when kobsd ) ku + kN[amine]o.
Analysis in terms of eq 6 requires values of [CD] and [amine]
corrected for the formation of the {CD‚amine} complex (eq 7).
To find these concentrations, we make use of the definition of
KN and the mass balance for the species in eq 7 and solve the
quadratic in [CD] obtained by the expansion of: KN ) [CD]-
[Nuc]/[CD‚Nuc] ) [CD]([Nuc]o - [CD]o + [CD])/([CD]o - [CD]).
The required solution of the quadratic is [CD] ) { -b + (b2 +
4KN[CD]o)1/2}/2, where b ) (KN - [CD]o + [Nuc]o). By difference,
[Nuc] ) [Nuc]o - [CD‚Nuc] ) [Nuc]o - ([CD]o - [CD]).
Bin d in g Stu d ies. The data analysis required the dissocia-
tion constants of the CD complexes of the esters and the
amines. Values of KS were taken from earlier studies,7,8 and
most of the KN values for the amines binding to â-CD and hp-
â-CD were determined previously using displacement of a
fluorescent probe.14b Additional values were obtained using a
spectral displacement method30 based on the complexation of
the dye phenolphthalein (PP).31
the spectral changes32 caused by these amines gave values of
K
using the fluorescent probe method, but for hp-â-CD the
agreement was not quite as good (see Table 1, footnote c).
Application of the method to the binding of isobutylamine,
isopentylamine, and benzylamine to â-CD and hp-â-CD gave
the values of KN presented in Table 1.
N in excellent agreement with those found earlier14b for â-CD,
The binding of amines to R-CD was studied using p-
nitrophenoxide ion (pNPO-) as a UV-visible spectral probe.
From analysis of the increases in absorbance at five wave-
lengths (415, 417, 420, 425, 430 nm) with [R-CD] (0-10 mM,
10 concentrations), the average Kd for the complex of pNPO-
was determined to be 0.732 ( 0.010 mM, in 0.20 M aqueous
phosphate buffer of pH 11.6. This value is just outside the
range of most of the literature values, which range from 0.28
to 0.63 mM, but which were obtained at lower pH, in weak
buffers, and by various means.30 Using Kd ) 0.732 mM for
pNPO-, and the changes in absorbance in the range 415-420
nm brought about by added amines, values of KN for the
amines and R-CD were estimated (Table 1).32
Attempts to probe the binding of amines to γ-CD using the
displacement of phenolphthalein were unsuccessful (cf. ref
31e), and so we developed a method based on induced circular
dichroism (icd).33 Measurement of the icd signal of the probe
molecule ANS at 276 nm as a function of [γ-CD] led to a Kd of
7.04 ( 0.28 mM, in very good agreement with the value of
7.81 ( 0.17 mM found by fluorescence measurements.14b Using
Kd ) 7.04 mM, and the reductions in the icd spectra caused
by added amines, values of KN for amines binding to γ-CD were
estimated (Table 1). Again, analysis of the data was virtually
the same as that used for the displacement of absorbance (see
above) and fluorescence probes.14b
It should be noted that if a KN value was to be in serious
error, for whatever reason, then the data analysis based on
eq 6 would not work well. This situation only seemed to arise
for the reactions of pNPA and pNPH by cyclohexylamine in
the presence of hp-â-CD where KN ) 5.21 mM (determined
twice)14b gave poor data analyses. Better analyses could be
obtained with a KN value of 1.8 mM, close to that for â-CD.
Ack n ow led gm en t. We thank the Natural Sciences
and Engineering and Research Council of Canada for
an operating grant, a post-graduate scholarship (to
T.A.G.) and an equipment grant toward the purchase
of the stopped-flow spectrophotometer. We are also
grateful to Dr. J oanne Turnbull for the use of the
circular dichroism spectrometer and to Mr. Craig Fen-
wick for running electrospray mass spectra. We also
thank Prof. R. A. McClelland (University of Toronto) for
helpful, critical comments.
From the decreases in absorbance of 50 µM PP at three
wavelengths (542, 552, and 562 nm), brought about by the
addition of the CD (0-5.00 mM, nine concentrations), Kd
values for the complexation of PP were determined to be 0.112
( 0.002 mM for â-CD, 0.117 ( 0.002 mM for hp-â-CD, and
1.24 ( 0.04 mM for γ-CD, in the 0.20 M aqueous phosphate
buffer of pH 11.6, at 25° C. The values found for â-CD and
γ-CD are about twice those found earlier at lower pH (10.0-
10.5) in weak carbonate buffers31 which is not unreasonable
given the different media.
J O9919660
The Kd values for PP binding were used in competition
experiments to estimate KN values for amine binding. To test
the methodology, we studied the displacement of PP from â-CD
and hp-â-CD by n-butylamine and n-pentylamine. Analysis of
(32) The data analysis for the spectral displacement method is
essentially identical to that developed for the displacement of a
fluorescent probe.14b
(33) The chromophores of an achiral guest bound to a chiral host,
such as a CD, can exhibit induced circular dichroism12,34 (icd) and
measurement of icd spectral changes as a function of [host] have been
used to estimate host-guest binding constants.35
(34) Shimizu, H.; Kaito, A.; Hatano, M. Bull. Chem. Soc. J pn. 1981,
54, 513. Kodaka, M.; Fukaya, T. Bull. Chem. Soc. J pn. 1989, 62, 1154.
Yoshida, N.; Yamaguchi, H.; Higashi, M. J . Chem. Soc., Perkin Trans.
2, 1994, 2507. Meyer, B.; Marconi, G.; Klein, C.; Ko¨hler, G.; Wolschann,
P. J . Incl. Phenom. 1997, 29, 79.
(35) Bowen, J . M.; Purdie, N. Anal. Chem. 1981, 53, 2239. Han, S.
M.; Atkinson, M.; Purdie, N. Anal. Chem. 1984, 56, 2827.
(30) Connors, K. A. Binding Constants: The Measurements of
Molecular Complex Stability; Wiley-Interscience: New York, 1987.
(31) (a) Buvari, A.; Barcza, L. Inorg. Chim. Acta 1980, 33, L179. (b)
Taguchi, K. J . Am. Chem. Soc. 1986, 108, 2705. (c) Buvari, A.; Barcza,
L.; Kajtar, M. J . Chem. Soc., Perkin Trans. 2 1988, 1687. (d) Sasaki,
K. J .; Christian, S. D.; Tucker, E. E. J . Colloid Interface Sci. 1990,
134, 412. (e) Gray, J . E.; MacLean, S. A.; Reinsborough, V. C. Aust. J .
Chem. 1995, 48, 551.