The complexation of peptides by aminocyclodextrins

Article abstract of DOI:10.1002/poc.348

Four β-cyclodextrins (CDs) were prepared bearing either an (N,N-dimethylamino)propylamino group (1), an (N-methyl)-piperazino group (2) or a benzylamino group (3), or seven methylamino susbtituents (4). Association constants K in water with di- and tripeptides reach up to 200 M^-1, and after protection at the N-terminus up to 680 M^-1. Appreciable binding occurs only in the presence of lipophilic amino acid side-chains, with preference for this at the C-terminus. A moderate sequence and side-chain selectivity is observed with 1, 2 and 3, but less so with the highly charged 4 where ion pairing dominates. Detailed NMR analyses with advanced techniques including T-ROESY and GHSQC allow full assigment of most ¹H and ¹³C signals, with extraction of many substituent and complexation induced shifts changes (SIS and CIS values, respectively). The CIS values and NOE cross peaks from ROESY experiments provide for insight into the binding modes of selected complexes, indicating, e.g., the simultaneous presence of complexes with a peptide phenyl unit approaching from both the narrow and the wide side of the CD cavity. With 3 one observes self-inclusion of the pendant phenyl ring within the cavity, and its replacement by analytes such as peptides, or by adamantanecarboxylic acid. The inclusion modes are illustrated with force field simulated structures and many NMR spectra, which are made available in electronic supplements. Copyright

Full text of DOI:10.1002/poc.348

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2001; 14: 159–170
The complexation of peptides by aminocyclodextrins
Frank Hacket, Svetlana Simova† and Hans-JoÈ rg Schneider*
FR Organische Chemie der UniversitaÈ t des Saarlandes, D-66041 SaarbruÈ cken, Germany
Received 24 August 2000; revised 5 November 2000; accepted 15 November 2000
ABSTRACT: Four b-cyclodextrins (CDs) were prepared bearing either an (N,N-dimethylamino)propylamino group
epoc
(1), an (N-methyl)-piperazino group (2) or a benzylamino group (3), or seven methylamino susbtituents (4).
Association constants K in water with di- and tripeptides reach up to 200 M 1, and after protection at the N-terminus
up to 680 M 1. Appreciable binding occurs only in the presence of lipophilic amino acid side-chains, with preference
for this at the C-terminus. A moderate sequence and side-chain selectivity is observed with 1, 2 and 3, but less so with
the highly charged 4 where ion pairing dominates. Detailed NMR analyses with advanced techniques including T-
1
ROESY and GHSQC allow full assigment of most H and 13C signals, with extraction of many substituent and
complexation induced shifts changes (SIS and CIS values, respectively). The CIS values and NOE cross peaks from
ROESY experiments provide for insight into the binding modes of selected complexes, indicating, e.g., the
simultaneous presence of complexes with a peptide phenyl unit approaching from both the narrow and the wide side
of the CD cavity. With 3 one observes self-inclusion of the pendant phenyl ring within the cavity, and its replacement
by analytes such as peptides, or by adamantanecarboxylic acid. The inclusion modes are illustrated with force field
simulated structures and many NMR spectra, which are made available in electronic supplements. Copyright 2001
John Wiley & Sons, Ltd.
Additional material for this paper is available from the epoc website at http://www.wiley.com/epoc: (NMR shift
tables; 3D coordinates for molecular structures; 1D- and 2D-NMR spectra)
KEYWORDS: peptides; complexation; aminocyclodextrins; NMR analyses; force field calculations
INTRODUCTION
underlying complex structures in the functional state of
solution.
The binding of lipophilic substrates inside the cavity of
cyclodextrins (CDs) has been the focus of many
investigations.1 Surprisingly, the complexation of
peptides has until now received little attention, in
contrast to many other studies with bioactive guest
compounds, including amino acids.2 Non-covalent
interactions with peptides are of paramount importance
for understanding mechanisms of protein folding, but
also for, e.g., pharmaceutical applications. The low
toxicity of CDs and their increasingly economic
availability makes these host compounds particularly
amenable for, e.g., drug protection and delivery. Also
for these reasons the design of suitable artificial peptide
receptors, which are less developed in comparison with
host compounds for nucleotides, should be pursued on
the basis of CDs. The present study addressed this need,
with an emphasis on a better understanding of the
Most of the literature on CD complexes is restricted to
measurements of association constants, including many
thermodynamic studies. The wealth of information from
these data, however, has not led to a consistent picture of
the relevant binding mechanisms with CDs, which are
more diverse than with other host compounds, and still a
matter of significant controvery.2,3 Often it is not even
known whether and which lipophilic parts of the substrate
are bound within the CD cavity. It has been shown that it
not necessarily the geometric fit that determines this,
against chemical intuition or predictions from computer-
aided molecular modelling.4 The obvious remedy here is
the use of NMR methods, which in view of the
underlying complex CD spin systems require application
of high fields and of advanced NMR techniques.5 The
complications are particularly large if the inherent
symmetry of the CD moiety is distorted by substitution;
such complexes are the main focus of the present study.
Complexes of native, unsubstituted CDs with peptides
are very weak, which might be one reason why they have
rarely been studied. We decided to use aminocyclodex-
trins, which have been found to be very effective host
compounds for nucleotides and even nucleosides.6 The
presence of a positively charged side-group in the CD can
*Correspondence to: H.-J. Schneider, FR Organische Chemie der
Universita¨t des Saarlandes, D-66041 Saarbru¨cken, Germany.
E-mail: ch12hs@rz.uni-sb.de
†On leave from the Institute of Organic Chemistry of the Bulgarian
Academy of Sciences.
Contract/grant sponsor: Deutsche Forschungegemeinschaft.
Contract/grant sponsor: Fonds der Chemischen Industrie.
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170

160
F. HACKET, S. SIMOVA AND H.-J. SCHNEIDER
Table 1. Protonation degree and overall charges of
monoamino-CDs 1±3 (calculated from 1HNMR measure-
ments at pD = 2.4, 7.4 and 12.4)a±c
CD
Protonation degree (%)
Overall charge at pD 7.4
1
2
3
N-14, 50; N-3, 90
N-1, <10, N-4, 60
60
‡1.4
‡0.7
‡0.6
a At 298 K in D2O, c(NaCl) = 1 Â 10 2 mol l 1, pD = 7.4 corresponds
to pH = 7.0.
b
Literature values/overall charges at pH 7.0 of the underlying
unsubstituted amines from potentiometric measurements in water at
298 K (all values within 0.16 units)36. For 1 (1,3-diaminopropane),35
10.9, 9.0 /‡2; for 2 (dimethylpiperazine),35 8.54, 4.63 /‡1; for 3
(N-methylbenzylamine), 9.57, lowest literature value.36
Figure 1. Investigated amino-CDs
c At values of pD = 2.4 and 12.4 only the fully protonated, or the fully
deprotonated forms, respectively, exist. The protonation degree is
determined from the ratio the differences in the chemical shifts of
protons located close to the protonation site between pD = 7.4 and 12.4
to the chemical shift difference of the whole range pD = 2.4 and 12.4
(see supplementary table S_Table I).
lead to relatively strong salt bridges with the C-terminus
of peptides; together with the lipophilic side-chain part of
amino acids enhanced and eventually sequence-selective
recognition might be feasible. CDs with pendant
aromatic side-chains have been used in particular by
Ueno and co-workers for fluorimetric assays of many
analytes.7 Detailed NMR analyses7 with such host
compounds have already provided clear evidence for
self-inclusion of the pendant aromatic side-chain within
the CD cavity, which can be expelled by addition of
stonger binding analytes.
(see also S_Table I in the supplementary material at the
epoc website at http://www.wiley.com/epoc). The hepta-
substituted 4 was estimated to bear 4–5 positive charges
on the basis of potentiometric investigations by Hamelin
et al.16 with a comparable 6-heptaamino-b-CD.
RESULTS AND DISCUSSION
Association constants
The amino-CDs investigated are shown in Fig. 1.
K values were measured using a competition method with
the fluorescence dye 1-anilinonaphthalene-8-sulfonate
(ANS), which provides a sensitive change of fluorescence
emission upon addition of the substrates, also giving
evidence for an intracavity inclusion mode. Evaluation of
binding isotherms with known equations17 (see Experi-
Syntheses, protonation state
Monoaminocyclodextrins 1–3 were prepared from the
corresponding mono-6-O-tosylate by nucleophilic sub-
stitution (see Experimental section). The monotosylate
was obtained under aqueous conditions using toluene-
sulfonyl chloride (T. Ikeda, Tokyo Institute of Technol-
ogy, personal communication) giving the same yields and
purity reported before for pyridine as solvent.12 The
recently reported13 higher yields with p-toluenesulfonic
anhydride as reactant were not reproducible in our hands.
The heptaamino-CD 4 was prepared according to a
procedure by Breslow et al.14 Instead of the b-CD-6-
heptatosylate, the b-CD-6-heptaiodide was used for the
nucleophilic substitution. Following the synthesis of
Ashton et al.,15 the iodide was prepared and converted
into the desired amine 4 in a more reproducible way than
that with the use of the tosylate.
The charges for the monosubstituted CD derivatives 1–
3 were estimated from known pKA values of the
corresponding amines (see footnotes to Table 1). NMR
experiments at pD = 2.4, 7.4 and 12.4, however, showed
significant differences due to ligation of the amines to the
CD moiety, with protonation degrees as given in Table 1
Figure 2. Fluorescence competition of the complex Ac±Gly±
Phe, ANS and the cyclodextrin 4 in water, pH 7.0,
c(NaCl) = 1 Â 10 2 mol l 1, T = 298 K, ꢀex = 360 nm, ꢀem
=
500 nm (for ®tting equations and parameters, see Experi-
mental)
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170

COMPLEXATION OF PEPTIDES BY AMINOCYCLODEXTRINS
161
Table 2. Association constants, (Kass. (M 1) and free complexation enthalpies, DG (kJ/mol 1) of complexes with the
aminocyclodextrins 1±4 from ¯uorescence titrations with 1,8-ANS in water, [pH 7.0, c(NaCl) = 1 Â 10 2 mol l 1, T = 298 Ka,b
1
2
3
4
Complex
Kass.
DG
Kass.
DG
Kass.
DG
> 4
Kass.
DG
Gly–Gly–Glyc
Phe–Gly–Gly
Gly–Phe–Gly
Gly–Gly–Phe
Phe–Gly
Gly–Phe
Ac–Gly–Phe
Phe–Asp
<5
13
23
50
22
49
80
—
—
10
> 4.0
6.3
<10
19
37
48
—
—
> 5.6
7.3
<5
20
36
27
25
35
53
19
73
37
—
<50
5
—
56e
20
—
10.0
7.4
10.6
13.2
10.8
16.2
—
7.4
8.9
8.2
8.0
8.8
9.8
7.3
10.6
9.3
7.8
9.7
7.7
9.7
10.9
—
—
5.6
8.9
9.6
—
—
—
—
—
72
206
78e
680
—
—
52
—
<100
45
—
—
Asp–Phe
—
—
9.8
—
Trp–Gly–Gly
Gly–Trp–Glyd
Gly–Gly–Trpc
Leu–Gly–Gly
Gly–Leu–Gly
Gly–Gly–Leu
<20c
—
> 7.3
—
> 11.4
> 5.6
> 5.6
7.4
—
—
<50
<3
18
> 9.7
> 2.7
7.1
<100
<10c
<10c
21
> 9.7
4.0
> 11.4
9.5
12
21
6.2
7.5
13
31
6.4
8.5
19
7.4
a
Error in K <15%.
b Kass. (b-CD/Gly–Gly–Phe) <10 M 1, in line with literature37 data from UV–VIS measurements for the a-CD/Gly–Gly–Phe system (Kass. = 8.1–10-
1
4 M in water, pH 9.0).
c
No fluorescence change observed upon adding peptide to the system 1,8-ANS/CD.
d Too low solubility of the peptide (<5 Â 10 4mol l 1).
e
Error <20%.
mental) yielded satisfactory linear fitting for a 1:1
stochiometry (Fig. 2) and binding constants for the
amino-CDs 1–4 (Table 2). Limiting factors for the
determinations and the errors in K were often low
solubilities of the peptides; in several cases only
approximate values or upper or lower limits could be
given.
ammonium groups and the aromatic side-chain with the
CD cavity. NMR data with another peptide (see below)
support this inclusion mode. The alternative dipeptides
Phe–Gly and Gly–Phe show, in line with the discussed
inclusion mode, within error the same K values as the
tripeptide as long as the terminal sequence is Gly–Phe,
but a much larger constant with the heptamino-CD and
Phe–Gly, with a different sequence. In the structure with
this dipeptide, optimum ion pairing is not hindered by the
interactions between the lipophilic parts of the host and
guest. The presence of an additional anionic charge in the
Asp-containing peptides leads to a moderate affinity
increase in comparison with Phe–Gly or Gly–Phe, and
this only with the Asp–Phe-sequence.
Selectivity with respect to the nature of the amino acid
side-chain is again only moderate, but in line with
expectations based on the increasing lipophilicity. Again,
the dominating ion pairing with the highly charged host 4
leads to decreased selectivity.
Host 3 differs from the others by the possibility of
competing self-inclusion of the pendant phenyl unit, in
line with related literature systems of Ueno and co-
workers.7 This geometry is established by the observed
NMR shifts and NOEs, and also by an experiment with
added adamantanecarboxylic acid and with the added
peptides, which expel the phenyl unit from the cavity (see
below). NMR analyses also exclude the presence of
intermolecular associations at the measuring conditions
(as indicated in Fig. 3). The intramolecular occupation of
the CD cavity by the phenyl ring in 3 leads to an affinity
decrease towards peptides, which in view of the K values
Af®nity variations
Appreciable association of the peptides with the CDs
(Table 2) was only observed in the presence of lipophilic
amino acid side-chains. Thus, the tripeptide Gly–Gly–
Gly showed no affinity above the detection limit with any
CD host. Removal of the positive charge in a substrate by
converting it into the N-acetyl derivative leads, as
expected on the basis of ion pairing. to an increase in
the association constants. This is especially so with the
heptasubstituted 4, in which the many charges lead to ion
paring as the dominating factor. Here, the K value is
increased by a factor of almost 10 by neutralizing the
positive charge at the peptide, whereas the increase with
host 1 or 3 is only moderate.
With a series of tripeptides containing always two Gly
and either Phe, Trp or Leu units, we investigated the
possible sequence selectivity of the host compounds.
With all CD hosts except the benzyl derivative 3 the K
values, as far they could be measured, are larger if the
lipophilic amino acid is at the C-terminus—this location
allows simultaneous contact of the carboxylate to the CD
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170

162
F. HACKET, S. SIMOVA AND H.-J. SCHNEIDER
precluding determination of the preferred conformation
of the substituents with respect to the glucose ring. Well
separated signals are only observed with substituents
with large anisotropy or field effects. For the assignment
of these protons measurement of a 2D C–H correlation
such as HSQC or GHSQC is more advisable than the
3
usually first acquired COSY. Determination of the J5,6
coupling constants could be achieved in strongly coupled
systems with the aid of the coupled version of the HSQC
experiment.18 In the hetero-correlation experiment the
carbon NMR spectrum is obtained in addition; with this,
the assignment of the closely resonating C-2, C-3 and C-5
atoms could be made unambiguously.
Figure 3. Complex geometry for intermolecular inclusion of
the benzyl ring in 3
obviously is only partially compensated by interactions
between the expelled phenyl ring outside the CD cavity
and lipophilic parts of the peptides.
A particular problem with the heptasubstituted com-
pounds is their low solubility in water. Protonation of the
nitrogen atoms leads to better solubility. However,
complicated spectra are obtained at the pH values
between 5 and 7 usually necessary for complexation
studies, owing to the different protonation degree and the
pH dependence especially of the H-5 and H-6 signals.
DMSO is usually a better solvent in this case for
identification purposes.
Observation of different shifts in monosubstituted CDs
does not necessarily indicate restricted motions, but is in
line also with the existence of predominant conforma-
tions.16,19 For all compounds and complexes acquisition of
2D TOCSY, GHSQC and T-ROESY was necessary. The
area between 4.1 and 3.5 ppm is always a superposition of
many signals and only occasionally contains a signal with
the necessary shielding separation. For the investigated
compounds this was the case, when either a phenyl ring is
fully or partially immersed in the CyD cavity, owing to the
anisotropy effects of the corresponding ring current and/or
when the 6-amino nitrogen is fully protonated.
NMR analyses
Assignment and measuring methods. Owing to the
complexation the chemical shifts of both cyclodextrin
and peptide protons change, so that especially in the case
of the mono-substituted derivatives separate assignment
for the complex has to be made, taking into account the
stoichiometry and the binding energetics. Characteristic
patterns for a-glucosidic protons are a doublet with a
constant around 4 Hz for H-1, a double doublet (10 and
4 Hz) for H-2 and triplets with a coupling constant around
10 Hz for H-3 and H-4, the former usually resonating at a
0.3–0.4 ppm lower field. Only the pattern of the spin
system comprising H-5 and H-6a,b protons usually
changes owing to substitution and/or inclusion shielding
effects. Unresolved or second-order spectra for these
protons were observed even at 500 MHz, depending on
the chemical shift difference between them. These
signals often overlap with some others; determination
of the coupling constants (3J5,6) is difficult, thus
The 2D TOCSY spectra 4 afford separate sets of signals
for the different glucose units. From the traces of the
corresponding anomeric protons, assignment of the
remaining signals can be made, taking into account the
Figure 4. TOCSY NMR spectra of the 1±3 (only CyD protons are shown), illustrating the proton shift
dispersion due to different substitution
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170

COMPLEXATION OF PEPTIDES BY AMINOCYCLODEXTRINS
163
coupling pattern of the signals. Owing to the different
resolution in both directions, traces and columns usually
give supplementary information for the overlapping
regions. When severe overlapping of more signals occurs
and/or when second-order effects due to small chemical
shift differences between vicinal protons (mostly H-4, H-5
and/or H-5, H-6) arise, 1D TOCSY experiments with
different mixing times facilitate unambiguous assign-
ments. For the compounds studied, COSY-type spectra
even in the DQF (double quantum filter) version of the
experiment gave severe overlapping, and were therefore of
little use.
In the second assignment step through-space connectiv-
ities between the different glucose units are established.
WefounditmostusefultomeasureT-ROESYspectrawith
an 800 ms mixing time, as with this version the artifacts (J-
cross talk, TOCSY transfer, etc.) are minimized and the
intensities of the cross peaks are maximized. The main
crosspeaksarealwaysobservedbetweentheanomericH-1
protons and H-2 protons, belonging to the same glucose
ring, and H-4 protons, belonging to adjacent glucose rings.
Theintensitiesofbothtypesofcrosspeaksarecomparable,
Figure 5. Inverse heteronuclear C±H correlation (GHSQC)
NMR spectrum of the complex of 3 with Ac±Phe±Gly,
acquired with high resolution in the second dimension
(SW1 = 10 ppm)
˚
corresponding to an average distance of 2.4 A, according
to the minimized structures (CHARMm/QUANTA V.
4.020). Large cross peaks are always observed between the
geminal non-equivalent H-6 protons (average distance 1.8
centered around 73 ppm, in the middle of the most
crowded region and the sampling frequency for the
carbons was adjusted to a spectral width of 10 ppm, with
a very high resolution (Fig. 5). Under these conditions the
C-1/H-1 pairs are folded three times. However, owing to
the separate region of the anomeric proton shifts, which
are very well resolved, they could be phased separately.
The signals for C-6/H-6 and C-4/H-4 are folded once, with
some overlap. In this case, recalculation of the region and
acquisition of a new spectrum are inevitable. Linear
prediction in F1 was always used for achieving of a better
resolution in the carbon dimension with a significant
signal-to-noise ratio increase.23
˚
A) and between the 1,3- diaxial H-3/H-5 and H-2/ H-4
˚
protonsofeach glucosering(average distance2.7A). Near
diagonal signals are often obscured, but the absence of a
cross peak of the above-mentioned type between peaks
withachemical shiftdifferenceofmorethan0.1 ppmhelps
to avoid erroneous assignments. In the ROESY spectra we
often observed also J-cross talk mainly between H-2 and
H-3 and between H-3 and H-4; these cross peaks were
used, however, also as aid for assignment.
Assignment of the 13C NMR spectra is done with the
help of inverse-detected heteronuclear correlation (HSQC
or GHSQC).21 Replacement of the oxygen by a nitrogen
atom (or by another substituent) in the glucose ring A
moves the C-6 and C-5 carbon chemical shifts to
considerably higher fields than the remaining glucose
carbon atoms. The correlation then allows an unambig-
uous assignment of the protons in the glucose ring A and
thus of the whole molecule, taking into account the
through-bond connectivity from the TOCSY and the
through-space connectivity from the ROESY spectra. The
heterocorrelation experiments are also useful for determi-
nation and/or confirmation of some difficult proton
assignments. Besides the nuclei in the substituted ring A,
which are usually easy to assign, the remaining carbon
atoms resonate in well-defined chemical shift ranges: 101–
103 ppm for C-1, 80–85 ppm for C-4 and around 61 ppm
for C-6; only the range between 71 and 75 ppm comprises
C-2, C-3 and C-5 carbon atoms. In order to achieve the
necessary high resolution, we found it easy and useful for
these systems to use folding22 in the carbon dimension.
For this purpose, the carbon excitation frequency was
Substituent-induced 1H NMR shifts (H-SIS). In the
supplementary Tables S_Table II and III the calculated
substituent-induced proton and carbon NMR shifts of
compounds 1–4 are given; Figure 4 shows the CD protons
parts of the TOCSY spectra of the compounds 1–3. The
dispersion of the anomeric protons is 0.09–0.14 -ppm,
although these protons are at a considerable distance
from the substitution site. The non-anomeric protons
show two different patterns depending on the substitu-
tion. For the fully protonated compounds 1 and 2 (at
pD = 2.4), having an aliphatic pendant group, distinct
signals only for the protons of rings A, B, G and F were
identified, whereas the protons of rings C, D and E
possess nearly identical chemical shifts. It should be
noted that when the nitrogen atom in ring A is not
protonated, no appreciable spread of the chemical shifts
for the individual glucose protons is observed. A
hydrogen bond between protons of the protonated
nitrogen at ring A and nearby oxygen atoms may be
responsible here.24 Our shift data and the minimized
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170

164
F. HACKET, S. SIMOVA AND H.-J. SCHNEIDER
Figure 6. Minimized structure of 1. The proximity of the
protonated nitrogen atom to O 6 of ring B and O in ring A
can be seen. Protons are omitted for a better view
Figure 8. Preferred conformation for the benzyl ring of 3 in
the CyD cavity according to the NMR data
structure (Hyperchem 5.02) show that the protonated
nitrogen is located near ring B, and hydrogen bonds could
exist between it and both the O atom in ring A and also
the O-6 of ring B (see Fig. 6). At pH 7.00, where this
nitrogen atom is protonated to a lesser extent, only a
small dispersion of shifts in the unsubstituted rings is
observed, precluding their unambiguous assignment.
Self-inclusion of the benzyl ring in the CD
cavity of 3
Dilution experiments at constant pD = 7.4 in the con-
4
centration range 1 Â 10 2–5 Â 10
M do not show any
chemical shift variation, indicating no intermolecular
cavity inclusion of the phenyl ring (Fig. 3). Figure 9
illustrates the characteristic dispersion of the chemical
shifts of all protons, especially of the H-5 and H-6
protons, which indicate inclusion of the phenyl ring from
the primary side. This is supported by the observed NOEs
between the phenyl protons and some H-5 and H-3
protons (rings A, B, E and F), disappearing on addition of
Figure 9. Shielding differences between 3 and 1 showing
the net chemical shift effect arising from the introduction of
the benzyl ring in the CD cavity
Figure 7. 1H NMR Spectra of 3 (b) and its complexes with (a)
ADCA and (c) with Ac±Gly±Phe
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170

COMPLEXATION OF PEPTIDES BY AMINOCYCLODEXTRINS
165
Table 3. Shielding differences between 3 and 1 for the glucose units A±G (Dꢁ Â 10 2 in ppm) at T = 300 K in D2Oa,b
Proton
A
B
C
D
E
F
G
H-1
H-2
H-3
H-4
H-5
H-6a
H-6b
1.5
0.3
8.9
8.4
2.4
3.6
8.4
1.1
12.5
2.0
7.3
1.0
1.3
0.8
15.0
5.0
0.2
2.4
3.6
8.0
1.0
5.7
1.7
1.0
1.0
2.0
5.1
1.0
2.0
6.2
1.0
9.0
9.0
8.7
6.6
19.6
2.8
53.7
13.1
45.0
3.4
5.0
11.6
27.3
6.7
33.0
29.0
26.0
1.0
a
c(1, 3) = 2.0 Â 10 2 mol l 1, c(NaCl) = 1.0 Â 10 2 mol l 1, pD = 2.4.
b Dꢁ (Hz) = [ꢁ(3) (ꢁ(1)]; negative values correspond to shielding and positive to deshielding.
an equimolar quantity of adamantane-1-carboxylic acid
(ADCA) as competing guest (see also Fig. 7). More
detailed analyses of the chemical shift data reveal that the
benzyl ring has a preferred conformation in the CD cavity,
since the largest shieldings are observed for the protons of
the oppositely lying rings G and D (see Figs 7 and 8).
In this conformation, the protonated nitrogen atom
points in the direction of ring B and can be involved in
hydrogen bonding analogous to that in 1 and 2. However,
only the chemical shifts of the protons in ring A show a
pronounced pH dependence (see supplementary table
S_Table I), indicating that the preferred conformation of
the phenyl ring does not change with pH. From the
differences in the chemical shifts of 3 and 1 the net
chemical shift effect arising from the ring current effect
could be calculated, as shown in Table 3 and Fig. 9.
Literature data on compounds with an aromatic pendant
group at the primary CD side7,25–27 (see also supplemen-
tary table S_Table IV) show that the largest chemical
shift dispersion of identical atoms in the individual
glucose rings is observed for the H-6 atoms, located near
the substituent, and also for the inner-cavity H-5 protons.
A characteristic dispersion is also observed for the H-3
protons, whereas for the other more distant protons H-1,
H-2 and H-4, smaller but measurable values are also
detected. It should be noted that the values for the
differently substituted naphthyl (dansyl) groups25–27
show comparable values that are about 3–4 times larger
than those for the corresponding phenyl groups. The
smaller values in 3 correspond to the smaller shielding
cone of the phenyl group as compared with that of the
naphthlene ring, and to the fact that the phenyl ring is in
equilibrium between the two preferred conformations,
one inside and the other outside the cavity.
13C SIS values
Small but measurable dispersion of the carbon chemical
shifts in the substituted compounds is detected, allowing
assignment of the individual signals through hetero-
nuclear correlation. This was possible for all carbon
signals in 3, but for 1 and 2 individual signals could be
assigned only for the nuclei in rings A, B and G, which
Table 4. 1H CIS values (Dꢁ Â 10 2 in ppm) in the complex of 1 and Ac±Gly±Phe for the glucose units A±G in D2O at pD = 7.4 and
300 Ka
Proton
H-1
A
B
C
D
E
F
G
1.0
9.0 for B, C, D
and E
<4.0 for
B, C, D, E, F and
G
2.0
4.0
H-2
<5.0
H-3
H-4
1.0
3.0
4.0*
9.0*
10.0*
15.0*
1.0
2.0
<5.0 for B, C,
D and E
< 11.0* for B
C, D and E
<12.0*
< 4.0
< 9.0
H-5
14.0
14.0
10.0
H-6a
H-6b
CH2(Gly)
6.2
19.0
9.0
CH2(Phe)
12.5
<12.0*
<12.0*
C-b
<6.0*
ꢀ15.0–25.0
C-c
<8.0
<8.0
H-o
<10.0
<10.0
H-m
<12.0*
<6.0*
H-p
< 12.0
C-a
56.2
12.5
12.5
6.2
12.5
12.5
N(CH3)3
0.6
CH3
6.2
a
2
Determined from the shift difference between the complex [c(1) = 1.2 Â 10 2 mol l 1, 80% in complexed state, c(Ac–Gly–Phe) = 6 Â 10
mol l 1, c(NaCl) = 1.0 Â 10 2 mol l 1] and 1 at pD 7.4.
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170

166
F. HACKET, S. SIMOVA AND H.-J. SCHNEIDER
Table 5. 1H CIS values (Dꢁ Â 10 2 in ppm) in the complex of 3 and Ac±Gly±Phe for the glucose units A±G in D2O at pD 7.4 and
300 Ka
Proton
A
B
C
D
E
F
G
H-1
H-2
H-3
H-4
H-5
H-6a
H-6b
0.1
0.8
6.3
1.5
2.1
1.1
1.4
1.1
2.1
2.6
2.9
2.5
4.7
6.3
2.3
7.4
4.1
4.1
3.4
3.4
1.8
3.8
8.2
5.5
5.5
0.1
0.8
4.0
6.6
<3.2
<8.0
<2.3
2.7
3.7
3.3
0
20.1
2.7
2.7
4.4
1.6
8.1
1.6
17.0
17.9
24.2
< 5.5
< 5.5
< 5.5
a Determined from the shift difference between the complex [c(3) = 1.2 Â 10 2 mol l 1, 73% in the complexed state, c(Ac–Gly–Phe) = 6 Â 10 2mol
1
l
1, c(NaCl) = 1.0 Â 10 2 mol l and 3 at pD 7.4.
give separate signals especially for C-1, C-4 and C-6. It is
difficult to explain these small differences because the
effects due to conformational distortions, which are
visible particularly in 13C shift variations, are of the same
order of magnitude as field and anisotropy effects. It
should be noted that in all compounds the carbon atoms
C-3 resonate always at slightly lower fields than the
usually overlapped C-2 and C-5 carbons, which could be
used for differentiation of the corresponding protons in
the complexes (see below).
results (see below). The 13C CIS values were also
determined (Table 6), but seem to give insufficient
selectivity for structural elucidation.
In the peptide complexes of 1 and 3 we were not able
to use the CIS values for determination of the complex
mode. The CIS values could be determined only if the
chemical shifts of the free and of the complexed form are
known. This was not the case for 1; the shifts of the free
form could not be reliably determined because of the too
heavily overlapped proton spectrum at pD = 7,4, which
is the value in the complexation experiments. Large
shieldings are observed for protons H-3 and H-5, located
inside the cavity, and also for some of the H-6 protons,
but, the values obtained for H-6 are not reliable for the
above reason. The data indicate inclusion of the phenyl
ring in the CyD cavity, but no conclusion on the
inclusion mode could be drawn solely from shift
arguments.
NMR shift analyses of the complexes/inclu-
sion modes
Assignment of the signals in the complexes was done in
the same manner as with host compounds. The
complexation induced shifts (CIS), measured for the
complexes of 1, 3 and 4 with the acetylated GlyPhe, are
given in Tables 4, 5 and 6. Inclusion of the phenyl ring of
Ac–Gly–Phe in 4 is obvious from the complexation
induced proton shifts: the higher shielding observed for
H-5 than for H-3 protons indicates that the aromatic ring
is closer to the primary side, fully in line with the NOE
The CIS values for the complex of Ac–Gly–Phe with 3
(see Table 5) could be exactly determined; the observed
shielding and deshielding, however, do not allow
unambiguous conclusions to be drawn on the complexa-
tion mode.
NOE analyses of the complexes
1
2
Table 6. H and 13C CIS values (Dꢁ Â 10 in ppm) in the
complex of 4 and Ac±Gly±Phe for the glucose units A±G in
D2O at pD 7.4 and 300 Ka
If cavity inclusion of the peptide takes place, the ROESY
spectra of the complexes show intermolecular cross
peaks between the ortho- and/or meta-protons of the
phenyl ring and the H-3 and/or H-5 protons of the
cyclodextrin host. Owing to severe overlapping, it is
often not possible to separate all intracavity protons; the
ranges of the H-5 and the H-3 protons are then deduced
from the high-resolution HSQC experiments. The NOE
data for the complexes studied are given in Table 7. A
preferred inclusion of the phenyl ring from the primary
side could be detected only for the complex of Ac–Gly–
Phe with 4 (see Fig. 10). The ortho-protons in the phenyl
ring are in the vicinity of both H-3 and H-5 CyD protons,
supported also by the higher shielding of the H-5 proton
(see above), whereas the meta-protons are close only to
13C CIS
1H CIS
H(C)-1
H(C)-2
H(C)-3
H(C)-4
H(C)-5
H(C)-6a
H-6b
12.0
20.0
20.0
7.0
15.0
28.0
1.0
1.0
7.0
1.0
15.0
1.0
1.0
1.0
CH3
40.0
a
Determined from the shift difference between the complex
[c(4) = 5 Â 10 3 mol l 1, 80% in the complexed state, c(Ac–Gly–
Phe) = 1 Â 10 3 mol l 1, c(NaCl) = 1.0 Â 10 2 mol l and 4 at pD
1
7.4.
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170

COMPLEXATION OF PEPTIDES BY AMINOCYCLODEXTRINS
167
Table 7. Intermolecular cross peaks in the T-ROESY spectra of the complexes with 1, 3 and 4 between the aromatic protons of
Ac±Gly±Phe and the cyclodextrin protons H-3 and H-5a,b
CDProp (1)
CDBenz (3)
HMA (4)
H-3
H-5
H-3
H-5
H-3
H-5
H-ortho
H-meta
H-para
‡
‡
‡
‡
‡
‡
‡
‡
‡
‡
‡
‡
‡
‡
‡
a
2
In D2O, c(NaCl) = 1 Â 10 2 mol l 1, T = 300 K; c(1, 2) = 1.2 Â 10 2 mol l 1, c(4 = 5 Â 10 3 mol l 1), c(Ac–Gly–Phe) = 6 Â 10
(for the
2
complexes with 1 and 3); 1 Â 10 (for 4).
b
‡ = Intermolecular cross peak observed; = no cross peak.
the H-3 protons. This arrangement allows interaction of
the multiple positive charges of the CyD molecule with
the negative C-terminus of the peptide.
In contrast, in the complexes with only one positive
charge (1 and 3), no preferred inclusion geometry could
be deduced (Fig. 11). Ortho-, meta- and the methylene
protons from the Phe residue give cross peaks to both
intracavity H-3 and H-5 protons. Both arrangements,
inclusion from the primary and from the secondary side,
seem to exist in these complexes.
EXPERIMENTAL
Materials. b-Cyclodextrin was a gift from Wacker-
Chemie (Burghausen, Germany). With the exception of
Ac–Gly–Phe (for preparation see below) the peptides
were purchased from Bachem Biochemica (Heidelberg,
Germany). CM-Sephadex C-25 was purchased from
‡
Aldrich-Chemie (Steinheim, Germany) in the Na form,
swollen in distilled water overnight and converted to the
‡
NH4 form by washing with 0.5 M ammonium hydro-
Figure 10. Complex geometry for the complex of 4 with Ac±Gly±Phe
Figure 11. Schematic representation of the possible complex geometries of 1 and 3 with Ac±Ghy±Phe
Copyright 2001 John Wiley & Sons, Ltd. J. Phys. Org. Chem. 2001; 14: 159–170

168
F. HACKET, S. SIMOVA AND H.-J. SCHNEIDER
gencarbonate solution followed by neutralization with
water. Deuterated solvents were purchased from Deutero
(Kastellaun, Germany) (deuteration 99.9%). All other
chemicals were purchased from Fluka-Chemie A (Buchs,
Switzerland) and used without further purification. With
the exception of methylamine, all amines were distilled
over KOH before use.
tate–water–ammonia (7:7:5:4, v/v)], Rf = (1) 0.40, (2)
0.61 and (3) 0.84. Elemental analysis: 1, calc. N 2.30,
C 46.30, H 6.78, found N 2.29, C 45.05, H 6.83%; 2, calc.
N 2.30, C 46.38, H 6.62, found N 2.51, C 45.22, H 6.61%;
3, calc. N 1.14, C 48.08, H 6.34, found N 1.08, C 44.04,
H 6.41%. For NMR spectra, see NMR section.
TLC was performed on silica-coated aluminium plates
(Alugram-SIL; Macherey–Nagel, Du¨ren, Germany) For
detection the plates were treated with a solution of 5%
sulfuric acid in ethanol and heated, resulting a black spot
for CD-containing fractions. Elemental analysis was
performed on an Elementar Analyser Modell 1106 from
Carlo Elba.
6-Heptakisdeoxy-6-heptakismethylamino-b-cyclodex-
trin (4). The heptaamine 4 was synthesized in two steps
from b-CD.14 After a persubstitution of the primary
hydroxy functions15 by iodide–triphenylphoshpine, the
amine was prepared with methylamine in dry ethanol and
purification over carboxymethyl-Sephadex. NMR data
and elemental analyses were identical with previously
published data.6
Mono-6-p-toluenesulfonyl-b-cyclodextrin. To a solu-
tion of 50.0 g ( 44.1 mmol) of b-CD in 1.5 l of 0.4 M
sodium hydroxide, 50.0 g ( 262 mmol) of p-toluene
sulfonyl chloride were added with vigorous stirring. The
resulting suspension was stirred for 3 h at 0°C. The
reaction mixture was filtered and neutralized with 1 M
hydrochloric acid to yield a white precipitate. After
filtration, the crude product was washed with acetone and
dried under vacuum. Recrystallization from water gave
12.2 g (22%) of mono-6-p-toluenesulfonyl-b-cyclodex-
trin. TLC [acetic acid–chloroform–water (8:1:1, v/v)],
Rf = 0.36.
Acetyl-glycyl-(L)-phenylalanin: Gly-(l)-Phe was pro-
tected similar to literature known synthesis.29 Gly-(L)-
Phe (1.0 g, 4.5 mmol) was dissolved in a minimum
amount of a saturated aqueous sodiumhydrogencarbo-
nate solution and acetic anhydride (0.85 ml, 9.0 mmol)
added. The reaction mixture was stirred at room
temperature for 10 min and refluxed for another 3 min
to destroy the remaining anhydride. The reaction
mixture was passed through an ion exchange column
‡
(Dowex 50W8 H -form). The resulting solution was
evaporated to dryness yielding 686 mg (58%) of Ac–
NMR analyses confirmed the monosubstitution by the
tosylate as described earlier for the same compounds.28
NMR also showed a purity of only 92%. Repeated
recrystallization from water did not lead to a improved
purity of the product. The product always shows some
impurities of unreacted CD tosyl chloride. As the
byproducts did not lead to problems in the further
preparation of the amines, no additional purification steps
were necessary with the tosylate.
Gly–(L)–Phe as a white solid. m.p. 169°C, TLC:
1
(Ethylacetate, DMF 9:1) Rf = 0.44, H NMR, DMSO-
1
d6, TMS internal reference, ꢁ [ppm]: 12.82 (s, H,)
COOH, 8.18 (d, 3J = 7.6 Hz, 1H) NH-Phe, 8.08 (t,
3J = 5.6Hz, 1H) NH-Gly, 7.32–7.21 (m, 5H) phenyl-H,
4.45 (m, 1H) CH-Phe, 3.72 (dd, 2J = 16.6 Hz,
3J = 5.8 Hz, 1H) CH2a-Gly, 3.63 (dd, 2J = 16.8 Hz,
3J = 6.0 Hz, 1H) CH2b-Gly, 3.06 (dd 2J = 13.4 Hz,
2
3J = 4.8 Hz, 1H) CH2a-benzyl, 2.89 (dd, J = 13.6 Hz,
3J = 8.8, 1H) CH2b-benzyl, 1.85 (s,3H) CH3, 13C-NMR,
DMSO-d6, ref = DMSO = 39.5 ppm, ꢁ [ppm]: 172.2
(Carbox.), 169.4, 168.9 (Amide), 137.4, 129.1, 128.2,
126.4 (phenyl), 53.3 (CH), 41.7 (CH2-Gly), 36.8 (CH2-
benzyl), 22.4 (CH3).
Monoamino-CDs 1±3. A 1.0 g (0.78 mmol) amount of
mono-6-p-toluene sulfonyl-b-cyclodextrin was dissolved
in 5 ml of the freshly distilled amine and heated at 70°C
in an atmosphere of nitrogen for 15 h. The amine was
evaporated, yielding a pale-yellow syrup that was
refluxed in ethanol for 30 min. The white precipitate that
formed was isolated by filtration and the procedure was
repeated. Purification was performed by ion-exchange
chromatography on carboxymethyl-Sephadex C-25
Fluorescence Spectroscopy. All fluorescence measure-
ments were carried out at 298 K in doubly distilled water
using a Hitachi fluorescence spectrophotometer F-2000
and Helma ‘Quarz-cells 111’. The samples were exited at
360 nm and emission intensities measured at 480, 500
and 520 nm.
‡
(NH4 form). The column was eluted initially with water
to replace impurities. In a second step the amino-CDs
bound to the column were removed by a gradient from 0
1
to 0.5 mol l aqueous ammonium hydrogencarbonate.
NMR-Spectroscopy: NMR spectra were recorded on
Bruker AM 400 (only for the routine analysis of the raw
products) and on AVANCE 500 NMR spectrometers,
operating at 400.1 (500.1) MHz for protons and 100.0
(125.0) MHz for carbon-13. A 5 mm dual (inverse)
1H-13C probehead was used at room temperature
300ÆK. Complexation-induced shifts (CIS) values
were calculated from the difference of the shifts in the
CD-containing fractions (assayed by TLC showing only
one spot) were collected and dried under vaccum. The
remaining white solid was dissolved three times in water
and evaporated to dryness to decompose ammonium
hydrogencarbonate.
Yields (after purification): 1, 695 mg (73%); 2, 708 mg
(75%); 3, 530 mg (42%). TLC (2-propanol–ethyl ace-
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170

COMPLEXATION OF PEPTIDES BY AMINOCYCLODEXTRINS
169
complex and the free state using the KA-values from the
where
fluorescence displacement titration. (Measurement
conditions see tables). 1 Â 10 2 mol/l NaCl concentra-
tion was present in all solutions for an equal ion strength
as in the binding constants measurements. Typical
measuring conditions for the 2D spectra were: 90°–>
pulses for protons and carbons, 9 and 12 s, gradient
pulse length 1 ms, sweep width 4000 (2000) Hz; data
size 2K/1K in the !2/ !1 direction, p/3 shifted squared
sine-bell windows in both directions; delay between the
scans 2 s, 8–64 scans, depending on concentration.
Phase-sensitive spectra using the TPPI method were
acquired. In the homonuclear experiments elimination
of the water signal was always made; using in the 2D
TOCSY spectra the 3–9–19 pulse sequence with
gradients16,30 and presaturation of the water signal at
a power of 50 db below the maximum output in all other
experiments. The 2D T-ROESY31 spectra were ac-
quired using for spin locking field a sequence of 180x
180 x pulses, each with a duration of 125 s. For a
minimum sweep width in the second dimension
application of quadrature detection was possible by
the use of two transmitter frequencies, one for
excitation in the middle of the spectrum and the other
for the spin locking field at the high-field end of the
spectrum. GHSQC21 experiment was used for hetero-
nuclear correlations. For good quality 1D TOCSY
spectra the z-filtering scheme32 with application of a
selective excitation pulse was used.
StK11
P ˆ
QK1 ‡ K11
ꢁe e1L†
Q ˆ
ꢁe1 e†
S ˆ ‰peptideŠ; St ˆ ‰peptideŠ
total
L ˆ CD; Lt ˆ ‰CDŠ
total
I ˆ ANS; It ˆ ‰ANSŠ
total
‰ILŠ
‰SLŠ
K1 ˆ
; K11
ˆ
‰IŠ‰LŠ
‰SŠ‰LŠ
e = measured fluorescence intensity
eI = fluorescence intensity of ANS without CD
eIL = fluorescence intensity of the ANS–CD complex
with K in M 1, S, P and L in M and other parameters
dimensionless.
Acknowledgements
This work was supported by the Deutsche Forschungs-
gemeinschaft, Bonn, and the Fonds der Chemischen
Industrie, Frankfurt.
Binding constants. All titrations were carried out at
298 K in water, pH 7.0, c(NaCl) = 1 Â 10 2 mol l 1. All
binding constants are average values calculated from
three emission wavelengths as mentioned above, differ-
ing by <5% from each other. In a first step the association
constants between the amino-CD and the fluorescence
indicator ANS were determined by a normal fluorescence
titration. A CD stock solution was added to a ANS
containing solution following the increase in fluorescence
intensity. The results were evaluated according to a non-
linear least-squares fitting procedure described earlier.33
The K values [Kass. (M 1) = 1, 251; 2, 239; 3, 156; 4,
2380] reflect stronger binding of the negatively charged
ANS to the positively charged amino CDs in comparison
with unmodified CD ( Kass. = 110 M 1), in line with
earlier studies in this field.34 In the second step stock
solutions of the investigated peptides were added to an
ANS–CD solution and the decrease in fluorescence
intensities was followed.
REFERENCES
1. (a) Szeijtli J, Osa T, (eds). In Comprehensive Supramolecular
Chemistry, Vol. 3: Cyclodextrins. Elsevier: Oxford, 1996; (b)
Duchene D, (ed). In New Trends in Cyclodextrins and Derivatives.
Editions de Sante´: Paris, 1991: 179ff; Djedaini F, Perly B. In New
Trends in Cyclodextrimes and Derivatives. Editions de Sante´:
Paris, 1991; 215ff; (c) Szeijtli J. Cyclodextrin Technology, 1988
Kluwer: Dordrecht; (d) Wenz G. Angew. Chem. 1994 33: 803;
Angew. Chem., Int. Ed. Engl. 1994; 106: 851; (e) Chem. Rev. 98
1998; 98: No. 5.
2. Connors K. J.Pharm. Sci. 1995; 84, 843 available on-line via
supplement at: http://www.pharmacy.wisc.edu/facstaff/sciences/
Connors.html.
3. Connors KA, Chem. Rev. 1997; 97: 1325.
4. Schneider H-J, Blatter T, Simova S. J. Am. Chem. Soc. 1991; 113:
1996.
5. Schneider H-J, Hacket F, Ru¨diger V, Ikeda H. Chem. Rev. 1998;
98: 1755.
6. Eliseev AV, Schneider H-J. J. Am. Chem. Soc. 1994; 116: 6081.
7. Hamasaki K, Usui S, Ikeda H, Ikeda T, Ueno A. Supramol. Chem.
1997; 8: 125; Ikeda H, Nakamura M, Ise N, Oguma N, Nakamura
A, Ikeda T, Toda F, Ueno A. J. Am. Chem. Soc. 1996; 118: 10980;
Usea S, Tanabe T, Ikeda H, Ueno A. J. Mol. Struct. 1998; 442: 161;
Hamai S, Ikeda H, Ueno A. J. Inclus. Phenom. Mol. Recogn. 1998;
31: 265.
8. Narita M, Hamada F, J. Chem. Soc., Perkin Trans. 2000; 2: 823.
9. Uccello-Barretta G, Balzano F, Cuzzola A, Menicagli R, Salvadori
P, Eur. J. Org. Chem. 2000; 3: 449.
The data were treated with a linear least-squares fit
according to Eqn. (1),17 yielding the peptide CD-binding
constants K.
10. Salvatierra D, Sanchez-Ruiz X, Garduno R, Cervello E, Jaime C,
Vergili A, Sanchez-Ferrando F. Tetrahedron 2000; 56: 3035.
11. Impellizzeri G, G Pappalardo, D’Alessandro F, Rizzarelli E,
St
P
K1
ˆ
ꢁQ ‡ 1†
ꢁ1†
K11
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170

170
F. HACKET, S. SIMOVA AND H.-J. SCHNEIDER
Savioano M, Iacovino R, Benedetti E, Pedone C. Eur. J. Org.
Chem. 2000; 6: 1065.
12. Matsui Y, Okimoto A. Bull. Chem. Soc. Jpn. 1978; 51: 3030.
13. Zhong N, Byup HS, Bittman R. Tetrahedron Lett. 1998; 39: 2919.
14. Breslow R, Czarniecki MF, Emert J, Hamaguchi H, J. Am. Chem.
Soc. 1980; 102: 762.
25. Hamasaki K, Ikeda H, Nakamura A, Ueno A, Toda F, Suzuki I, Osa
T. J. Am. Chem. Soc. 1993; 115: 5035.
26. Corradini R, Dossena A, Marchelli R, Panagia A, Sartor G,
Saviano M, Lombardi A, Pavone V. Chem. Eur. J. 1996; 2: 373.
27. Ikeda H, Nakamura M, Ise N, Toda F, Ueno A. J. Org. Chem.
1997; 62: 1411.
15. Ashton PR, Ko¨niger R, Stoddart JF, J. Org. Chem. 1996; 61: 903.
16. Hamelin B, Jullien L, Guillo F, Lehn JM, Jardy A, Robertis LDe,
Driguez H. J. Phys Chem. 1995; 99: 17877.
17. Connors K. Binding Constants. John Wiley & Sons: New York,
1987.
18. Simova S. Magn. Reson. Chem. 1998; 36: 505.
19. Chen W, Fukudome M, Yuan D, Fujioka T, Mihashi K, Fujita K. J.
Chem Soc., Chem. Commun. 2000; 541.
20. Brooks CL, Karplus M. Methods Enzymol. 1986; 127: 369;
Bru¨nger AT, Karplus M. Acc. Chem. Res. 1991; 24: 54.
21. Bruker Pulse Library 98/08/04, 1998.
22. Schmieder P, Zimmer S, Kessler H. Magn. Reson. Chem. 1991; 29:
375.
23. Reynolds WF, Yu M, Enriquez RG, Leon I, Magn. Reson. Chem.
1997; 35: 505.
24. Bonomo RP, Cucinotta V, D’Allessandro F, Impellizzeri G,
Maccarrone G, Rizzarelli E, J. Inclus. Phenom. 1993; 15: 167.
28. Bonomo RP, Cucinotta V, D’Alessandro F, Impellizzeri G,
Maccarrone G, Vecchio G, Rizzarelli E. Inorg. Chem. 1991; 30:
2708.
29. Ki-Hwan K, Martin Y, Otis E, Mao J. J. Med. Chem. 1989; 31: 84.
30. Sklenar V, Piotto M, Leppik R, Saudek V. J. Magn. Reson. A 1993;
102: 241.
31. Hwang T-L, Shaka AJ, J. Am. Chem. Soc. 1992; 114: 3157.
32. Simova S, Dimitrov V. J. Magn. Res. A, 1995; 117: 292.
33. Schneider H-J, Kramer R, Simova S, Schneider U, J. Am. Chem.
Soc. 1988; 110: 6442.
34. Ito N, Yoshida N, Ichikawa K. J. Chem. Soc., Perkin Trans. 1996;
2: 965.
35. Dean JA. Handbook of Organic Chemistry. 1987; McGraw Hill:
New York.
36. Wang F, Sayre LS. J. Am. Chem. Soc. 1992; 114: 248.
37. Chokchainarong S, Fennema OR, Connors KA. Carbohydr. Res.
1992; 232: 161.
Copyright 2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 159–170