Dendrimer poly(ethylenimine)s linked to β-cyclodextrin

Article abstract of DOI:10.1006/bioo.1996.1054

β-Cyclodextrin was attached to two dendrimer poly(ethylenimine)s. The resulting cyclodextrin-containing dendrimers, CD-I and CD-II, can be considered either as dendrimers equipped with specific binding sites or as cyclodextrins containing amino groups around the cavities. Amines of CD-I and CD-II remarkably resisted protonation compared with those of the parent dendrimers. A compact conformation of CD-I or CD-II in which the dendrimer wraps itself around the cyclodextrin is proposed as a conformation consistent with the suppressed protonation. Esters containing t-butylphenyl groups were complexed by CD-I and CD-II and underwent fast deacylation. Kinetic data were obtained with several ester substrates, which revealed that two amino groups located in the vicinity of each cyclodextrin cavity of CD-I or CD-II participated as nucleophiles. In addition, optimum reactivity was attained when the spacer connecting the t-butylphenyl and the ester groups was -O- CH₂- or -CH=CH-. Structures of the active sites for the accelerated deacylation of esters were elucidated on the basis of the kinetic data.

Full text of DOI:10.1006/bioo.1996.1054

BIOORGANIC CHEMISTRY 25, 63–75 (1997)
ARTICLE NO. BH961054
Dendrimer Poly(ethylenimine)s Linked to ͱ-Cyclodextrin
JUNGHUN
S
UH,*,1 SANG
SOO
H
AH,* AND
SANG
H
EE
LEE
†
*Department of Chemistry and Center for Molecular Catalysis, Seoul National University,
Seoul 151-742; and †Department of Chemistry, Kunsan National University, Kunsan 573-360, Korea
Received November 20, 1996
ͱ-Cyclodextrin was attached to two dendrimer poly(ethylenimine)s. The resulting
cyclodextrin-containing dendrimers, CD-I and CD-II, can be considered either as dendrimers
equipped with specific binding sites or as cyclodextrins containing amino groups around the
cavities. Amines of CD-I and CD-II remarkably resisted protonation compared with those
of the parent dendrimers. A compact conformation of CD-I or CD-II in which the dendrimer
wraps itself around the cyclodextrin is proposed as a conformation consistent with the
suppressed protonation. Esters containing t-butylphenyl groups were complexed by CD-I and
CD-II and underwent fast deacylation. Kinetic data were obtained with several ester substrates,
which revealed that two amino groups located in the vicinity of each cyclodextrin cavity of
CD-I or CD-II participated as nucleophiles. In addition, optimum reactivity was attained
when the spacer connecting the t-butylphenyl and the ester groups was UOUCH2U or
UCHuCHU. Structures of the active sites for the accelerated deacylation of esters were
elucidated on the basis of the kinetic data.
1997 Academic Press
INTRODUCTION
Dendrimers are three-dimensional, highly ordered oligomeric and polymeric com-
pounds formed by reiterative reaction sequences starting from smaller molecules
(1, 2). Dendrimers can mimic various properties of large biomolecules and, thus,
numerous applications are conceivable (1). By attaching chemical moieties to den-
drimers, a variety of biomimetic functional molecules can be designed.
Dendrimers with hydrophilic exteriors and hydrophobic interiors may be regarded
as covalently fixed assemblies of amphiphiles (3). In view of intensive studies on
catalytic reactions of micelles (4, 5), dendrimers may be exploited as biomimetic
catalysts. In addition, dendrimers are free from the basic instability elements of
noncovalent self-assemblies like micelles and vesicles. Furthermore, dendrimers
possess well-defined structures unlike micelles or vesicles. To construct artificial
enzymes on dendrimers, creation of binding sites on the skeleton of a dendrimer
is desirable. Without specific binding sites, major principles of enzymatic catalysis
can be hardly reproduced.
Many cyclodextrin (CD) derivatives have been examined for their ability to
recognize guest molecules and to catalyze chemical transformation of the included
molecules (6–10). In order to devise effective biomimetic catalysts using CD deriva-
1 To whom correspondence and reprint requests should be addressed.
63
0045-2068/97 $25.00
Copyright 1997 by Academic Press
All rights of reproduction in any form reserved.

64
SUH, HAH, AND LEE
tives, it is necessary to introduce catalytic groups to CDs in positions suitable for
high catalytic efficiency. It is not easy to introduce a catalytic group in the productive
position by using a short spacer to connect CD and the catalytic group. When a
long spacer is used, however, conformational freedom of the resulting CD derivative
should be suppressed to freeze the molecule in the productive conformation.
In an attempt to construct artificial enzymes by using dendrimers as the molecular
skeleton, ͱ-CD, a cyclic heptamer of Ͱ- -glucose, was attached to dendrimer poly
D
(ethylenimine)s I and II in the present study. Dendrimer I containing ͱ-CD (CD-I)
and dendrimer II containing ͱ-CD (CD-II) can be regarded as either dendrimers
equipped with specific binding sites or ͱ-CD derivatives containing several amino
groups around the CD cavities. Thus, reactions of compounds included in the CD
cavity would be affected by the dendrimer moiety. In addition, the conformation
and the chemical behavior of the dendrimer would be in turn influenced by the
presence of ͱ-CD in the molecular framework. In this paper, conformation of CD-I
and CD-II in water and kinetic data for deacylation of carboxyl esters containing
t-butylphenyl residues in the presence of CD-I and CD-II are presented.
EXPERIMENTAL PROCEDURES
Synthesis of Dendrimers and Their Derivatives
Dendrimer poly(ethylenimine)s I and II were synthesized according to the proce-
dure reported in the literature (11) as shown in Scheme I. Dendrimers I and II
were examined by 1H NMR microscopy and GPC. The 1H NMR (300 MHz, DMSO-
d6) spectra of I and II were too broad to check the defect level. For the tosylated
derivatives of I and II (I-Ts and II-Ts), however, the ratio of the signal integration
of aromatic protons (ͳ 7.24–7.62) to ethylene protons (ͳ 2.26–2.64) agreed with

DENDRIMER POLY(ETHYLENIMINE)S LINKED TO ͱ-CYCLODEXTRIN
65
S
CHEME I
the theoretical value within 10% error. GPC analysis showed the molecular weight
distributions of ϭ 1.16 for I-Ts and ϭ 1.29 for II-Ts.
M
w
/
M
n
M
/
M
w n
ͱ-CD was linked covalently to I or II to obtain CD-I or CD-II, respectively, by
the reactions of I (0.45 g, 0.23 mmol) or II (0.93 g, 0.23 mmol) with mono-6-(p-
toluenesulfonyl)-ͱ-CD (12) (0.30 g, 0.26 mmol) in 100 ml dimethyl sulfoxide
(DMSO) at 60ЊC for 6 h followed by purification by ultrafiltration (Amicon Y1)
and gel filtration (Sephadex G25). The 1H NMR (300 MHz, D2O) spectra of CD-I
and CD-II showed that 0.87 and 1.05 residues, respectively, of CD were incorporated
in CD-I and CD-II.
Preparation of Substrates
4-Carboxy-2-nitrophenyl 4-t-butylbenzoate (S0-). This compound was prepared
according to the literature (13) by coupling the carboxylic acid and the phenol with
N,NЈ-dicyclohexylcarbodiimide (DCC), mp 199–200ЊC (lit. 199–200ЊC (10, 13)).
4-Nitrophenyl 4-t-butylbenzoate (S0). This compound was prepared by coupling
the carboxylic acid and the phenol with DCC and purified by chromatography
(silica, hexane/ethyl acetate ϭ 5 : 1) and recrystallization from hexane-ethyl acetate,
mp 124–126ЊC (lit. 123–125ЊC (14)).
4-Nitrophenyl (4-t-butylphenyl)acetate (S1). A solution of 4-t-butylbenzyl bromide
(1.5 g, 6.4 mmol), potassium cyanide (1.3 g, 20 mmol), and 18-crown-6 (0.2 g)
dissolved in tetrahydrofuran (30 ml) was refluxed for 1 day. The residue obtained
after evaporation of the solvent under reduced pressure was extracted with hexane.
The hexane solution was washed with water, dried with sodium sulfate, and evapo-
rated. The mixture of the resulting (4-t-butylphenyl)acetonitrile and 50 ml 4
NaOH was refluxed for 16 h and then acidified with 1 HCl. The residue was
extracted with diethyl ether and purified by chromatography (silica, hexane/ethyl
N
N

66
SUH, HAH, AND LEE
1
acetate ϭ 10 : 1) to give (4-t-butylphenyl)acetic acid. EIMS m/e 192(Mϩ). H NMR
(300 MHz, CDCl3): ͳ 1.29 (s, 9H), 3.56 (s, 2H), 7.18–7.34 (m, 8H).
A solution of (4-t-butylphenyl)acetic acid (0.34 g, 1.7 mmol), 4-nitrophenol
(2.3 g, 1.7 mmol), DCC (0.38 g, 1.8 mmol), and 4-N,NЈ-dimethylaminopyridine
(DMAP) (0.1 g) in methylene chloride (20 ml) was stirred for 2 h at room tempera-
ture. The solvent was evaporated under reduced pressure and the resulting mixture
was treated with ethyl acetate. After removal of the undissolved residue by filtration,
the product was purified by chromatography (silica, hexane/ethyl acetate ϭ 20 : 1)
and recrystallized from methylene chloride-hexane to give S1, mp 66–67ЊC. EIMS
1
m/e 313(Mϩ). H NMR (300 MHz, CDCl3): ͳ 1.33 (s, 9H), 3.87 (s, 2H), 7.25–7.32
(m, 4H), 7.39–7.43 (m, 2H), 8.83–8.88 (m, 2H). Anal. Calcd for C18H19N1O4: C,
68.99; H, 6.11; N, 4.47. Found: C, 68.58; H, 6.07; N, 4.63.
4-Nitrophenyl 3-(4-t-butylphenyl)propionate (S2A). Ethyl trans-4-t-butylcinna-
mate (see below) (0.50 g, 2.2 mmol) was hydrogenated with H2 (1 atm) and 10%
Pd/C (0.1 g) for 16 h at room temperature. Pd/C was removed through celite and
the filtrate was concentrated. The resulting ester and NaOH (170 mg, 4.3 mmol)
were dissolved in methanol (20 ml)–water (10 ml). After reflux for 2 h, methanol
was evaporated. The aqueous solution was washed with diethyl ether, acidified with
1
N
HCl, and extracted with methylene chloride. The organic layer was dried
with sodium sulfate and evaporated to give (4-t-butylphenyl)propionic acid, mp
116–117ЊC (lit. 116–117ЊC (16)). (4-t-Butylphenyl)propionic acid (100 mg, 0.49
mmol), 4-nitrophenol (68 mg, 0.49 mmol), and DCC (120 mg, 0.58 mmol) in methyl-
ene chloride (5 ml) were stirred for 6 h at room temperature. Hexane (20 ml) was
added to the reaction mixture and the resulting solid was removed by filtration.
The product thus obtained was purified by chromatography (silica, hexane/ethyl
acetate ϭ 5 : 1) and recrystallization from methylene chloride-hexane, to obtain
S2A, mp 74–76ЊC. 1H NMR (200 MHz, CDCl3): ͳ 1.30 (s, 9H), 2.85–3.11 (m, 4H),
7.15–8.23 (m, 8H). EIMS m/e 339(Mϩ). Anal. Calcd for C19H21N1O4: C, 69.71; H,
6.47; N, 4.28. Found: C, 69.82; H, 6.33; N: 4.39.
4-Nitrophenyl (4-t-butylphenoxy)acetate (S2B). This compound was prepared ac-
cording to the literature (15), mp 79–80ЊC (lit. 79–80ЊC (15)).
4-Nitrophenyl trans-4-t-butylcinnamate (S2C). A solution of 4-t-butylbenzalde-
hyde (1.2 g, 7.5 mmol) and (carboethoxymethylene)triphenylphosphorane (2.9 g,
8.3 mmol) in 20 ml of toluene was stirred for 6 h at room temperature. After
concentration, the residue was dissolved in diethyl ether (10 ml) and hexane (30
ml) was added slowly to precipate triphenylphosphine oxide. The resulting solid
was removed by filtration through celite. The product thus obtained was purified
by chromatography (silica, hexane/ethyl acetate ϭ 30 : 1) to give ethyl trans-4-t-
1
butylcinnamate. H NMR (200 MHz, CDCl3): ͳ 1.30 (s, 9H), 1.40 (t, 3H), 4.25 (q,
2H), 6.37 (s, 1H), 6.43 (s, 1H), 7.35–7.72 (m, 4H).
A solution of ethyl trans-4-t-butylcinnamate (0.50 g, 2.2 mmol) and NaOH (0.17
g, 4.3 mmol) in methanol (20 ml)–water (10 ml) was refluxed for 4 h. After methanol
was removed by evaporation, the aqueous solution was washed with diethyl ether,
acidified with 1
N
HCl, and extracted with methylene chloride. The organic layer
was dried with sodium sulfate and evaporated under reduced pressure. To a solution
of the resulting trans-4-t-butylcinnamic acid (100 mg, 0.49 mmol) in 10 ml methylene

DENDRIMER POLY(ETHYLENIMINE)S LINKED TO ͱ-CYCLODEXTRIN
67
chloride, 4-nitrophenol (68 mg, 0.49 mmol), DCC (120 mg, 0.59 mmol), and DMAP
(10 mg) were added. After stirring for 6 h at room temperature, the reaction mixture
was evaporated under reduced pressure. After the resulting mixture was treated
with diethyl ether, the undissolved portion was removed by filtration. The product
thus obtained was purified by chromatography (silica, hexane/ethyl acetate ϭ 5 : 1)
and recrystallization from methylene chloride–hexane to give S2C. mp 152–154ЊC.
1H NMR (200 MHz, CDCl3): ͳ 1.35 (s, 9H), 6.54 (s, 1H), 6.62 (s, 1H), 7.30–8.35
(m, 8H). EIMS m/e 337 (Mϩ). Anal. Calcd for C19H19N1O4: C, 70.14; H, 5.89; N,
4.30. Found: C, 70.29; H, 5.99; N, 4.38.
4-Nitrophenyl (4-t-butylbenzyloxy)acetate (S3). To a solution of 4-t-butylbenzyl
alcohol (3.3 g, 20 mmol) and NaH (1.3 g, 53 mmol) in 200 ml ethanol was added
chloroacetic acid (2.0 g, 21 mmol). After reflux for 6 h, the reaction mixture was
acidified with 3
N
HCl and extracted with methylene chloride. The organic layer
was dried with sodium sulfate and evaporated under reduced pressure. The resulting
residue was recrystallized from ethanol to give (4-t-butylbenzyloxy)acetic acid. A
solution of (4-t-butylbenzyloxy)acetic acid (0.59 g, 2.6 mmol), 4-nitrophenol (0.30
g, 2.6 mmol), DCC (0.65 g, 3.2 mmol), and DMAP (0.1 g) in tetrahydrofuran (20
ml) was stirred for 5 h at room temperature. The residue obtained after evaporation
of the solvent was treated with ethyl acetate. After removal of the undissolved
portion by filtration, the solvent was evaporated. The resulting residue was recrystal-
1
lized from diethyl ether-hexane to obtain S3, mp 87–88ЊC. H NMR (200 MHz,
CDCl3): ͳ 1.33 (s, 9H), 2.60–2.85 (t, 2H), 3.40–3.80 (q, 2H), 6.34 (s, 1H), 7.16–8.30
(m, 8H). Anal. Calcd for C19H21N1O5: C, 66.46; H, 6.16; N, 4.08. Found: C, 66.68;
H, 6.07; N, 3.93.
4-Nitrophenyl N-(4-t-butylbenzoyl)-3-aminopropionate (S4). ͱ-Alanine methyl
ester (1.1 g, 11 mmol) was added to a solution of t-butylbenzoic acid (1.9 g, 11
mmol), DCC (2.2 g, 11 mmol), and DMAP (0.1 g) in methylene chloride (100 ml).
After the mixture was stirred for 2 h at room temperature, precipitates formed by
addition of hexane were removed by filtration. The product thus obtained was
purified by chromatography (silica, hexane/ethyl acetate ϭ 4 : 1) and recrystalliza-
tion from ethyl acetate–hexane to give methyl N-(4-t-butylbenzoyl)-3-aminopropio-
nate. Following the procedure described above for S2C, the methyl ester was hy-
drolyzed and then coupled with 4-nitrophenol to obtain S4 which was purified by
1
recrystallization from ethyl acetate–hexane (mp 98–99ЊC). H NMR (300 MHz,
CDCl3): ͳ 1.33 (s, 9H), 2.58–2.87 (t, 2H), 3.38–3.81 (q, 2H), 7.16–8.30 (m, 8H).
Anal. Calcd for C20H22N2O5: C, 64.86; H, 5.98; N, 7.56. Found: C, 64.93; H, 6.03;
N, 7.47.
4-Nitrophenyl N-(4-t-butylbenzoyl)-4-aminobutyrate (S5). This compound was
prepared according to the literature (15), mp 103–104ЊC (lit. 103–104ЊC (15)).
4-Nitrophenyl trans-cinnamate (S2C*). This compound was prepared by using
DCC, mp 146–147ЊC (lit. 146.5–147.5ЊC (17)).
Measurements
All the kinetic measurements and titration were carried out at 25ЊC. Reaction
rates were measured spectrophotometrically with a Beckman DU 650 UV/Vis
spectrophotometer. pH measurements were carried out with a Dong-Woo pH meter

68
SUH, HAH, AND LEE
F
IG. 1. Degree of protonation of amino groups at 25ЊC and various pH for I (a), II (b), CD-I (c),
CD-II (d), and NH2(CH2)2NH(CH2)2NH(CH2)2NH2 (e). Concentrations of I, II, CD-I, and CD-II were
ca. 50 m in terms of individual amines.
M
DP 880. Stock solutions of ester substrates were made in acetonitrile and the buffer
solutions for deacylation of the esters contained 9.8% (v/v) acetonitrile. Buffer
(0.020
M
) used at pH 7.57 was Na2HPO4.
RESULTS
One of the primary hydroxyl groups of ͱ-CD was tosylated, and the resulting
tosyl derivative was used in the alkylation of the amino groups of dendrimers I and
II to obtain CD-I and CD-II. On the average, CD-I and CD-II contained 0.87 and
1.05 residue of CD, respectively. It is expected that the primary amines, instead of
the tertiary amines, of I and II were alkylated due to the steric hindrance involved
in the alkylation of tertiary amines.
In order to obtain information on microenvironments of amino groups of the
dendrimers, basicity of the amino groups was examined by titration with HCl. As
summarized in Fig. 1, both I and II manifested large buffer capacity at pH 5–6. On
the other hand, CD-I and CD-II showed considerably weaker buffer capacity at
pH 3–8. At pH 5, about 40% of the amino groups of I and II were protonated
whereas only 5–10% of the amino groups of CD-I and CD-II were protonated. At

DENDRIMER POLY(ETHYLENIMINE)S LINKED TO ͱ-CYCLODEXTRIN
69
F
IG. 2. Plot of ko (pseudo-first-order rate constant) against Co for deacylation of S0- in the presence
of CD-I (᭺; curve a), CD-II (ࡗ; curve b), I (᭝), II (छ), or CD (ϫ) under the condition of Co ӷ So
at pH 7.57 and 25ЊC. For CD-I, CD-II, and CD, Co was expressed in terms of the concentration of CD
moiety. For I or II, Co was corrected so that it corresponds to the concentration of I or II in CD-I or
CD-II. Curves a and b were obtained by fitting the data to Eq. [2].
pH 3, almost all of the amino groups were protonated for I and II whereas less
than 10% were protonated for CD-II.
Several nitrophenyl esters (S0-, S0, S1, S2A, S2B, S2C, S3, S4, S5, and S2C*)
containing various structural features were prepared. As illustrated in Fig. 2, deacyl-
ation of S0- was much faster in the presence of CD-I or CD-II compared with that
in the presence of CD, I, or II. Deacylation by CD-II was considerably faster with
S2C than with S2C* at low Co concentrations (Fig. 3).
Saturation kinetic behavior observed with deacylation by CD-I or CD-II was
consistent with complex formation between the ester and CD-I or CD-II. The
pseudo-first-order rate constants (ko) measured under the conditions of So (initially
added substrate concentration) Ӷ Co (initially added concentration of CD-I or
CD-II) were analyzed according to Eqs. [1] and [2] by analogy with the Michaelis–
Menten scheme. In this paper, Co is expressed in terms of the concentration of the
CD moiety for CD-I or CD-II unless noted otherwise. The values of kcat and Km
estimated for deacylation of the ester substrates by CD-I and CD-II are summarized
in Table 1.
kcat
S ϩ C
CS —Ǟ products
[1]
[2]
Km
ko ϭ kcatCo/(Km ϩ Co)

70
SUH, HAH, AND LEE
F
IG. 3. Plot of ko against Co for deacylation of S2C (ࡗ) and S2C* (᭺) in the presence of CD-II under
the conditions of Co ӷ So at pH 7.57 and 25ЊC. The curves were obtained by fitting the data to Eq. [2].
TABLE 1
Values of Kinetic Parameters for Deacylation by n-Butylamine (BA) at pH 10.74 and
by CD-I and CD-II at pH 7.57 and 25ЊCa
BA
CD-I
CD-II
ester
kam
kcat
Km
kcat
Km
3
1
1
3
1
4
1
3
1
4
1
(10Ϫ MϪ sϪ
)
(10Ϫ sϪ
)
(10Ϫ MϪ
)
(10Ϫ sϪ
)
(10Ϫ MϪ
)
S0-
S0
S1
S2A
S2B
S2C
S3
S4
S5
S2C*
4.18
0.742
1.07
1.59
2.72
2.78
0.541
1.67
5.01
2.76
1.16
1.09
1.04
0.921
1.28
6.36
6.74
3.77
6.27
5.78
2.08
4.46
5.96
4.88
33.1
0.858
1.27
1.36
1.43
7.08
1.88
6.70
1.88
1.65
0.232
0.182
1.39
1.03
1.18
0.996
7.44
0.179
a See text for details of reaction conditions. Standard deviations were 1–3% of the parameter values
for kam, 1–10% of the parameter values for kcat, and 3–15% of the parameter values for Km. The values
of kam were obtained by normalizing the rate data to fully unprotonated BA.

DENDRIMER POLY(ETHYLENIMINE)S LINKED TO ͱ-CYCLODEXTRIN
71
Deacylation of S0- by CD-I and CD-II was examined under the conditions
of So Ͼ Co. As illustrated in Fig. 4 (not shown for CD-I), rapid initial burst followed
by slow release was observed for the appearance of 4-hydroxy-2-nitrobenzoate
during the deacylation of S0- by Cd-I or CD-II. The amount (⌬i of Fig. 4) of 4-
hydroxy-2-nitrobenzoate released during the initial burst stage can be estimated by
extrapolation of the curve. The value of ⌬i corresponds to 1.8 and 2.2 times Co for
CD-I and CD-II, respectively. Since ratio So/Co was 4.5 or 7.3 for CD-I or CD-II,
respectively, data of the initial burst experiment indicate that about 2 molecules of
the ester were deacylated for each CD residue of CD-I or CD-II.
In order to estimate the intrinsic reactivities of the ester substrates for aminolysis,
5
kinetics of deacylation of the esters (1–5 ϫ 10Ϫ
M) were measured in the presence
of n-butylamine (0.025–0.1
M
) at pH 10.74. At this pH, n-butylamine was half-
protonated as checked by potentiometric titration. The bimolecular rate constants
(kam) estimated for the aminolysis of the esters are summarized in Table 1.
DISCUSSION
In polyamines such as I and II, protonation of an amino group would suppress
subsequent protonation due to electrostatic interactions. In addition, the ethylene
moieties of the dendrimers may exert hydrophobic effects leading to suppression

72
SUH, HAH, AND LEE
of ionization of amines. The fractions of amino groups protonated at various pH
values are calculated for NH2(CH2)2NH(CH2)2NH(CH2)2NH2 by using the pKa
values of the amines reported in the literature (18) and are illustrated in Fig. 1
together with those measured with I, II, CD-I, and CD-II.
Amines of CD-I and CD-II show remarkably lower basicity compared with I and
II. Attachment of CD to I or II appears to induce changes in the conformation of
I or II, placing amino groups either in more hydrophobic microenvironments or in
closer proximity to other amino groups compared with those of I or II. If some of
the amino groups of CD-I or CD-II are hydrogen-bonded to hydroxyl groups of
CD, their basicity would be also reduced considerably. Although the conformations
of CD-I and CD-II are not clearly determined at present, the dendrimer
conformation is considerably altered upon attachment of CD. If CD does not
affect conformation of the dendrimer considerably (III), attachment of CD to the
dendrimers would not influence the protonation of amino groups of I or II apprecia-
bly. The compact conformation (IV) of CD-I or CD-II in which the dendrimer
wraps itself around CD is one of the possible conformations consistent with the
results of Fig. 1.
Deacylation of S0- is much faster in the presence of by CD-I or CD-II compared
with CD, I, or II. Comparison of deacylation of S2C and S2C* indicates that the

DENDRIMER POLY(ETHYLENIMINE)S LINKED TO ͱ-CYCLODEXTRIN
73
F
IG. 4. Release of 4-hydroxy-2-nitrobenzoate during the deacylation of S0- in the presence of CD-II
4
5
under the condition of So (3.86 ϫ 10Ϫ
M M) at pH 7.57 and 25ЊC. A control reaction
) Ͼ Co (5.31 ϫ 10Ϫ
was carried out with II. ⌬i is the absorbance increase due to 4-hydroxy-3-nitrobenzoate released during
the initial burst stage. See Fig. 2 for definition of Co.
t-butylphenyl portions of the esters are recognized by CD-I or CD-II. It is highly
likely that the t-butylphenyl portions are included in the cavity of CD portions of
CD-I or CD-II. If the esters are complexed to CD-I or CD-II of the extended
conformation such as III, amino groups on the dendrimer portion would have
limited access to the complexed ester. If the esters are complexed to CD-I or CD-
II of the compact conformations such as IV, some amino groups would be located
close to the ester linkage of the complexed substrate. Fast deacylation of the
complexed esters is, therefore, consistent with the compact conformation.
Nonpolar species tend to aggregate in water solution so as to decrease the hydrocar-
bon–water interfacial area (hydrophobic effect) (19). Substrate binding into the CD
cavity is driven by the hydrophobic effect (19). For example, inclusion of t-butylphe-
nyl rings of the ester substrates examined in the present study protects t-butylphenyl
rings from water. Since the exterior wall of CD is as hydrophobic as the interior
wall of CD, wrapping of CD with dendrimers I and II can also reduce the interfacial
area between the nonpolar surface and water. The hydrophobicity of the dendrimer
portion would decrease as the amino groups are protonated. At pH 7.5, the dendrimer
portions of CD-I and CD-II are mostly unprotonated and would exhibit near maximal
hydrophobicity. Wrapping of CD with the dendrimer may involve hydrogen bonding
between the amino groups of the dendrimer and the hydroxyl groups of CD and/
or hydrophobic interaction between the dendrimer and CD.
Results of the initial burst experiment (Fig. 4) indicate that CD-I or CD-II is
deactivated when about two molecules of S0- are deacylated for each CD moiety.
As illustrated in V, CD is attached to a peripheral amino group (ANH) of the

74
SUH, HAH, AND LEE
F
IG. 5. Relative magnitude of kcat/kam for the ester substrates in the deacylation by CD-I (light bars)
and CD-II (dark bars). The values of kcat and kam are listed in Table 1.
dendrimer. Examination by the CPK space filling model and by a molecular mechan-
ics calculation indicated that considerable amount of strain is introduced when the
nearest neighboring primary amine (BNH2) makes nucleophilic attack at the in-
cluded ester. When the next neighboring primary amines (CNH2) are the nucleo-
philes, the strain is not significant. As the number of CNH2 is two, results of the
initial burst experiment are consistent with the mechanism of V.
The Km value for the deacylation of S2C by CD-II is much smaller than that of
S2C*, in agreement with the inclusion of the cinnamoyl portions of the two esters
in the cavity of CD. Despite the large difference in Km values, kcat for S2C is not
much different from that of S2C*. Once the ester is bound to the host molecule,
therefore, the reactivity is not much different.
Reactivities of the ester substrates complexed to CD-I or CD-II are reflected by
kcat . Apparently, kcat varies widely for the ester substrates examined in the present
study (Table 1). To consider the differences in intrinsic reactivity toward aminolysis
of the ester substrates, kinetic data for aminolysis by n-butylamine were obtained.
The relative reactivities of the nucleophilic amino group toward various ester sub-
strates bound by CD-I or CD-II are best estimated by the values of kcat/kam, which

DENDRIMER POLY(ETHYLENIMINE)S LINKED TO ͱ-CYCLODEXTRIN
75
are compared in Fig. 5.2 Each ester substrate contains a different spacer connecting
the t-butylphenyl portion and the ester group. The data of Fig. 5 reveal that the
relative reactivity is considerably greater for S2B and S2C for CD-I and CD-II.
The spacer contained in S2B is UOUCH2U and that in S2C is UCHuCHU. As
the spacer becomes shorter or longer, the relative reactivity is reduced considerably.
Deacylation of the esters by CD-I and CD-II, therefore, manifests appreciable
selectivity with regard to the substrate structure with the optimum geometry of the
transition state being provied by S2B or S2C.
ACKNOWLEDGMENTS
This work was supported by grants from Basic Science Research Institute Program (1995), Ministry
of Education and Center for Molecular Catalysis.
REFERENCES
1. TOMALIA, D. A., NAYLOR, A. M., AND
GODDARD III, W. A. (1990) Angew. Chem. Int. Ed. Engl.
138, 175.
2. TOMALIA, D. A., AND
3. TOMALIA, D. A., HALL, B. M., AND
4. FENDLER, J. H. (1982) Membrane Mimetic Chemistry, Wiley-Insterscience, New York.
5. BUNTON, C. A., NOME, F., QUINA, F. H., AND OMSTED, L. S. (1991) Acc. Chem. Res. 24, 357.
6. BENDER, M. L., AND OMIYAMA, M. (1977) Cyclodextrin Chemistry, Springer-Verlag, New York.
DURST, H. D. (1993) Top. Current. Chem. 165, 193.
H
EDSTRAND, D. M. (1987) Macromolecules 20, 1167.
R
K
7. TABUSHI, I. (1982) Acc. Chem. Res. 15, 66.
8. BENDER, M. L. (1987) in Enzyme Mechanisms (Page, M. I., and Williams, A., Eds.), Chap. 4,
R. Soc. Chem., London.
9. BRESLOW, R., ZHANG, B. (1992) J. Am. Chem. Soc. 114, 5882.
10. SUH, J., LEE, S. H., AND
11. TOMALIA, D. A., AND
12. MATSUI, Y., YOKOI, T., AND
Z
OH, K. D. (1992) J. Am. Chem. Soc. 114, 7961.
EWALD, J. R. (1986) U.S. PATENT 4631337.
OCHIDA, K. (1976) Chem. Lett. 1037.
D
M
13. LEE, S. H. (1994) Ph.D. thesis, Seoul National University.
14. SILVER, M. S. (1966) J. Am. Chem. Soc. 88, 4247.
15. ZOH, K. D., LEE, S. H., AND
16. TSATSAS, G., P-SANDRIS, A., AND
17. BENDER, M. L., SCHONBAUM, G. R., AND
18. JENCKS, W. P., AND EGENSTEIN, J., in Handbook of Biochemistry and Molecular Biology (Fasman,
S
UH, J. (1994) Bioorg. Chem. 22, 242.
ANDRIS, C. (1964) Bull Soc. Chim. France 10, 2612.
ERNER, B. (1959) J. Am. Chem. Soc. 81, 2540.
S
Z
R
G. D., Ed.), 3rd ed., Vol. 1, p. 323, CRC Press, Cleveland.
19. BRESLOW, R. (1991) Acc. Chem. Res. 24, 159.
2 To achieve the deacylation rate (kcat) of S2B complexed to CD-II, 12
M of unprotonated n-butylamine
is needed in water. The values of kam measured for aminolysis of S2A by n-butylamine at 25ЊC were
4
1
1
4
1
1
(8.13 Ϯ 0.03) ϫ 10Ϫ MϪ sϪ in methanol, (5.52 Ϯ 0.10) ϫ 10Ϫ MϪ sϪ in ethanol, and (1.94 Ϯ 0.18) ϫ
4
1
1
10Ϫ MϪ sϪ in n-propanol, being smaller than that in water (Table 1). Thus, the aminolysis became
slower as the solvent was made less polar. Although the exact microenvironment of the active site for
ester deacylation in CD-I or Cd-II is not precisely known, it would be appreciably less polar than water.
If the microenvironment of the nucleophilic amino group (NHc2 of structure V) is similar to methanol
or n-propanol, its effective molarity (A. J. Kirby (1980) Adv. Phys. Org. Chem. 17, 183) toward the
bound ester would be ca. 24 or 100
M, respectively.