Molecular recognition: II-Discrimination of specific and non-specific intermolecular interactions by means of magnetic resonance spectroscopy

Article abstract of DOI:10.1002/(SICI)1097-458X(199803)36:3<205::AID-OMR237>3.0.CO;2-L

The interactions of adenine- and benzimidazole-type substrates with a model receptor (Rebek's cleft) and benzoic acid were studied by magnetic resonance spectroscopy. The spectra reveal dynamic phenomena which are analysed in terms of static, two-jump and

Full text of DOI:10.1002/(SICI)1097-458X(199803)36:3<205::AID-OMR237>3.0.CO;2-L

MAGNETIC RESONANCE IN CHEMISTRY, VOL. 36, 205È210 (1998)
Molecular Recognition
II—Discrimination of Speciüc and Non-Speciüc Intermolecular Interactions
by Means of Magnetic Resonance Spectroscopy¤
Martin Jager,1 Paul Schuler,1 Hartmut B. Stegmann1* and Antal Rockenbauer2
1 Institut fur Organische Chemie der Universitat, Auf der Morgenstelle 18, D-72076 Tubingen, Germany
2 Central Research Institute for Chemistry, Hungarian Academy of Sciences, P.O. Box 17, H-1525 Budapest, Hungary
Received 20 June 1997; revised 15 August 1997; accepted 19 August 1997
ABSTRACT: The interactions of adenine- and benzimidazole-type substrates with a model receptor (RebekÏs cleft)
and benzoic acid were studied by magnetic resonance spectroscopy. The spectra reveal dynamic phenomena which
are analysed in terms of static, two-jump and two-site models. The interpretation provides information on the inter-
and intramolecular structures and allows speciÐc interactions to be distinguished from non-speciÐc interactions.
KEYWORDS: molecular recognition; molecular dynamics; exchange
INTRODUCTION
confrontation of functional groups so as to establish
attractive forces of electrostatic, hydrogen bonding or
van der Waals type.2 Owing to the non-covalent char-
acter of the attractions, lock-and-key relations are
usually found to be equilibria depending on environ-
mental parameters.
The study of molecular recognition by spectroscopic
methods requires detectable probes at the species to be
observed. Concerning magnetic resonance, most organic
molecules can be readily investigated by NMR spectros-
copy, but for the application of EPR/ENDOR (EMR)
spectroscopy at least one of the molecules has to bear a
paramagnetic moiety reÑecting the changes in the spec-
troscopic properties caused by any type of interaction.
In the course of our investigations, the molecular cleft
R proved to be suitable, providing a lock-like cavity
with speciÐc affinity to purine and benzimidazole deriv-
atives.3 We therefore modiÐed these heterocycles with
pre-paramagnetic phenol-type substituents (Scheme 1)
Particular molecules recognize each other, associate and
form lock-and-key complexes. These interactions
between receptors and substrates, known collectively as
molecular recognition, govern basic principles of life
such as regulation, transport and catalysis. If two mol-
ecules are able to interact in a speciÐc manner, they are
said to be complementary in size and shape. The
resulting stereochemistry ensures the proximity and
* Correspondence to: H. B. Stegmann, Institut fur Organische Chemie
der Universitat, Auf der Morgenstelle 18, D-72076 Tubingen,
Germany.
E-mail: stegmann=uni-tuebingen.de
¤ For Part I, see Ref. 1.
Contract/grant sponsor: Fonds der Chemischen Industrie.
Contract/grant sponsor: Deutsche Forschungsgemeinschaft.
Contract/grant sponsor: Hungarian National Research Fund (OTKA);
contract/grant number: T-015841.
Scheme 1. Receptor R and substrates 1, 2 and 3.
( 1998 John Wiley & Sons, Ltd.
CCC 0749-1581/98/030205È06 $17.50

206
M. JAGER ET AL .
in sites where they do not interfere with the intermolec-
ular interactions.1
The spectra were recorded with a Bruker ESP 300E
spectrometer equipped with an ENDOR unit (Bruker
ER 810). Typical instrumental parameters for ENDOR
investigations were microwave power 30 mW, r.f. power
7 dB, 500 W and modulation 70 kHz.
The EMR investigations reveal, owing to the better
time resolution compared with 250 MHz NMR, the
dynamic behaviour of all components: the global inter-
molecular equilibrium and the intramolecular Ñexibility.
Hence, di†erent phenomena are traced back to the
internal motions: superposition of hyperÐne structures
and exchange based upon a two-jump mechanism and
upon a two-site model.4 In this paper, we will show that
by means of EMR spectroscopy speciÐc lock-and-key
interactions can be distinguished from non-speciÐc
interactions.
RESULTS
The affinity of R towards 1 and 2 has already been
investigated quantitatively by 1H NMR spectroscopy in
CDCl (NMR titration) yielding K \ 1 ] 103 l mol~1
3
A
for R ] 1 H R1 and 3 ] 103 l mol~1 for R ] 2 H R2.1
As the addition of acids has had no signiÐcant inÑuence
upon the chemical shifts of the substrates 1, 2 and 3, the
EXPERIMENTAL
observed chemical shift di†erences *d \ d
[ d
compl
free
(CIS values8) are traced back to complexation.
The receptor R is well known.5 The syntheses of
phenols 1, 2 and 3 have been described in theses.6,7
2,6-Di-tert-butyl-4-(dimethylaminomethyl)phenol,
lead dioxide, benzoic acid and dichloromethane
(CH Cl ) are commercially available. Lead dioxide was
Hence samples of free radicals 1*, 2* and 3* were
prepared, together with mixtures with R and with
benzoic acid, according to Scheme 2.
2
2
further dried in vacuo. Benzoic acid was recrystallized
from toluene. CH Cl was redistilled and stored over
1* and 1* + R
2
2
molecular sieves.
Solutions of the diamagnetic species in CH Cl were
The phenoxyl 1* at 232 K presents an EPR spectrum
which is readily interpreted in terms of the hyperÐne (hf)
coupling parameters given in Table 1. Addition of
receptor R alters the hf structure. The spectrum record-
ed at a substrateÈreceptor ratio of 1 : 1 is presented in
Fig. 1(a). By means of ENDOR spectroscopy, resolution
enhancement can be achieved allowing the observation
of distinct signals of two species, one being 1*; the other
2
2
prepared containing substrate concentrations between 1
and 10 mM. The radicals 1*, 2* and 3* were generated
by monovalent oxidation of the corresponding phenols
with PbO . The receptor R or benzoic acid (BA) can be
2
added before or after radical generation according to
Scheme 2. Deoxygenation was performed by bubbling
argon through the samples.
Scheme 2. Equilibria of the substrates with BA and R and pathways of oxidative radical generation.
Table 1. hf parameters of 1*, 1* ] BA and 1* ] R H R1*
1*
(232 K)
1* ] BA
(232 K)
R1*
(203 K)b
R1*
(232 K)c
Parameter
a
a
a
a
(mT)
(mT)
(mT)
0.684
0.178
0.178
0.051
1(a), (b)
0.695a
È
0.181a
0.049a
0.645
0.210
0.186
0.044
0.515
0.139
0.176
0.078
0.530
0.142
0.179
0.071
1(a), (b)
0.537a
È
0.181a
0.071a
H
N
Hm (mT)
Hc
Fig.
a hf constants obtained from ENDOR experiments at T \ 243 K.
b Original ratio: R : 1* \ 3 : 1.
c Original ratio: R : 1* \ 1 : 1.
( 1998 John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 36, 205È210 (1998)

DISCRIMINATION OF SPECIFIC AND NON-SPECIFIC INTERMOLECULAR INTERACTIONS
207
Figure 1. EPR spectrum of the equilibrium 1* ] R H R1*
at T \ 232 K. (a) Experimental spectrum. Original
mixture 1* : R \ 1 : 1. (b) Simulation. Calculated mixture
1* : R1* \ 0.55 : 0.45.
Figure 2. (a) EPR spectrum of free 2* at T \ 293 K. (b)
EPR spectrum of 2* ] BA at T \ 293 K.
1* + BA. Addition of benzoic acid in the ratio 5 : 1 to a
sample of 1* leads to a simple spectral alteration: in this
case, the spectrum is due to a single paramagnetic
species. The value of the nitrogen coupling constant
increases compared with free 1* whereas the b-proton
splitting decreases (see Table 1). Both hf parameters
decrease on lowering the temperature, a behaviour
observed both for the free and the complexed species.
is attributed to R1* (for the corresponding ENDOR
spectra, see Ref. 1). Superimposing the hyperÐne pattern
based on coupling constants observed in ENDOR
experiments with the hyperÐne structure of free 1* at a
1* : R1* ratio of 0.55 : 0.45 leads to a good agreement
with the 1* ] R H R1* spectrum in Fig. 1(a). The simu-
lation is shown in Fig. 1(b).
At a receptor : substrate ratio of 3 : 1 and a tem-
perature of 203 K, the EPR spectrum of R1* is charac-
terized by the parameters given in Table 1. No free 1* is
present in the sample, as is also conÐrmed by ENDOR
experiments.
2* and 2* + R
The EPR spectra of the radical 2* show a hf structure
characterized by a large triplet remaining unchanged
Table 2. hf parameters of 2*, 2* ] BA and 2* ] R H R2*
R2*
2*
(293 K)
2* ] BA
(293 K)
Parameter
252 K
262 K
272 K
282 K
297 K
313 K
323 K
333 K
a
a
a
a
(mT)
(mT)
0.918
0.918
0.215
0.188
È
0.867
0.867
0.238
0.188
È
1.090
0.832
0.204
0.188
5.5
1.090
0.832
0.210
0.188
8.5
1.090
0.832
0.217
0.188
17.0
1.090
0.832
0.220
0.188
32
1.090
0.832
0.214
0.194
55.0
1.070
0.832
0.223
0.191
100
1.070
0.832
0.223
0.191
170
1.070
0.832
0.226
0.191
300
Hb
Hb(mT)
N
(mT)
Hm
k (MHz)
Fig.
2(a)
2(b)
Èa
Èa
Èa
a Published in Ref. 1.
( 1998 John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 36, 205È210 (1998)

208
M. JAGER ET AL .
over the experimentally accessible temperature range
between 252 and 333 K. The room temperature spec-
trum is presented in Fig. 2(a). It is easily explained with
the coupling constants given in Table 2. However, when
R is added and the temperature is lowered, the centre
lines of the triplet structure undergo strong line
broadening. Regaining intensity after coalescence, the hf
structure now exhibits a quartet pattern (the series of
spectra was published previously1). The complete set of
hf parameters is given in Table 2. A plot of the exchange
frequencies k vs. the reciprocal absolute temperature
according to Arrhenius yields a straight line leading to
an activation energy E \ 33.7 ^ 1.1 kJ mol~1
A
(8.1 ^ 0.3 kcal mol~1) and an intercept log K \ 13.7
0
^ 0.2 (the errors given refer to the linear regression
within the plot). Assuming a temperature-independent
*a observed at T \ 252 K allows the estimation of the
exchange frequency at coalescence via the approx-
imation k \ *a ] 62.2 MHz and, thus, of a coalescence
c
temperature T \ 271 K, which is in good agreement
c
with the experimental spectrum (the approximation
formula is adopted from NMR for use in EPR spectros-
copy, i.e. [*a] \ mT , and is discussed elsewhere7). Con-
sequently, the free activation enthalpy *G* can be
c
estimated by means of *G* \ 19.14 T* [10.32
c
c
] log(T /k )] kJ mol~1 (Refs 9 and 10) to be 28.8 kJ
c c
mol~1 (6.9 kcal mol~1).
2* + BA. The presence of acids has a strong inÑuence
on the hfs of phenoxyl 2*, as can be seen from Fig. 2(b):
the triplet splitting found with 2* remains dominant
whereas the nitrogen coupling is signiÐcantly altered
(see the data in Table 2). Within the experimentally
accessible temperature range, broadening of the central
lines could not be observed.
3* and 3* ] R
The observation of 3* over a temperature range of
about 100 K reveals that the spectral features change
from a large triplet to a broadening of the central lines,
illustrated in Fig. 3(a) for 232 K. The hf parameters at
232 and 293 K are given in Table 3. Addition of R
Figure 3. (a) EPR spectrum of free 3* at T \ 232 K. (b)
EPR spectrum of 3* ] R H R3* at T \ 293 K. (c) EPR spec-
trum of 3* ] R H R3* at T \ 232 K.
Table 3. hf parameters of 3*, 3* ] BA and 3* ] R H R3*
R3*
3*
232 K
293 K
3* ] BA
(293 K)
Parameter
232
293 K
A
B
A
B
a
(mT)
(mT)
1.144
1.273
0.116
0.179
È
1.144
1.214
0.125
0.182
È
0.816
1.308
0.179
0.179
È
1.462
0.700
0.178
0.177
0.51
1.259
1.054
0.141
0.177
0.49
1.424
0.757
0.177
0.180
0.44
1.188
1.116
0.140
0.180
0.56
Hb
a
a
a
Hb(mT)
(mT)
Hm
N
Population
k (MHz)
Fig.
11.1
3(a)
50.5
120
5.8
3(c)
4.8
3(c)
22.2
3(b)
17.5
3(b)
( 1998 John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 36, 205È210 (1998)

DISCRIMINATION OF SPECIFIC AND NON-SPECIFIC INTERMOLECULAR INTERACTIONS
209
changes the original hf structure of 3* to that illustrated
in Fig. 3(b). When the temperature decreases, the hfs in
Fig. 3(b) close to coalescence loses the broadening of the
central lines and reveals sharp lines regaining intensity
[Fig. 3(c)]. The interpretation can be achieved with the
data in Table 3, in good agreement with the experimen-
tal spectra.
the b-protons is also altered, lowering the average value
of their coupling constants. The electronic inÑuence
upon the nitrogen coupling constant cannot be
explained as easily.
The EPR spectra of R1*, R2* and R3* proved to be
more complicated and did not show a similarly uniform
behaviour: while both the nitrogen and the b-proton
coupling constant are decreased in R1* compared with
free 1*, a remains nearly unchanged in R2* but a
3* + BA. An excess of benzoic acid (ratio 5 : 1) a†ects
the hfs of 3* in a similar way as already discussed for
2*: the addition of BA leads to a strong increase in the
nitrogen coupling constant and a decrease in the
average value of the b-proton splitting, as indicated by
the hf parameters in Table 3. A systematic temperature
variation reveals the coalescence point and also the
splitting of the two b-protons inequivalent at low tem-
perature.
N
H
decreases. In the complex R3*, two species are present
with a larger nitrogen coupling constant and a smaller
average value of the b-proton hf components. Both
species show an exchange indicated by the broadening
of the central lines.
We attribute this non-uniform behaviour to di†erent
orientations of the substrates in the receptor and to dif-
ferent stabilities of the complexes formed. The e†ect of
the intramolecular dynamic Ñexibilities of 2* and 3*
within the complex might also be considered to inÑu-
ence the hf values observed. For purine derivatives such
as 1* and 2*, the existence of di†erent orientations
within the complex has been discussed:12,13 the lockÏs
acridine base and the purine moiety are oriented either
face-to-face allowing n-stacking interactions or edge-to-
face. Hydrogen bonds are likely to be formed between
the carboxyclic groups of the cleft R and both the imid-
azole and the pyrimidine heterocyclic units. Our EPR
results indicate that no hydrogen bonding exists to the
nitrogen atom in position N6 of the adenine, otherwise
its coupling constant would be increased (as is the case
when benzoic acid is added). The change in the b-
proton hf values can be explained as a consequence of
the McConnellÈHeller relationship due to the altered
orientation of the b-protons to the phenoxyl plane.
Within the accessible temperature range, EPR spectros-
copy cannot decide whether di†erent possible orienta-
tions of the keys in the lock are realized by a
dissociation mechanism or a rearrangement within the
complex takes place. However, superimposed spectra of
the complexed and the free species can be obtained in
addition to those which show merely the hfs of the
complex, but no evidence is found for an exchange
between free and complexed species for 1*, R1*, 2* and
R2*.
Neither free nor complexed 1* shows any intramole-
cular dynamic behaviour as far as the CHX-N substi-
tuent is concerned. Conformational exchange is found
in 2* when complexed. The two b-protons are assumed
to exchange by a two-jump mechanism,14 a model
which leads to good agreement between the experimen-
tal and calculated spectra. The two-jump model applies
also for the interpretation of the exchange observed
with free 3* and 3* ] BA. Concerning R3*, the
exchange observed within the spectra cannot be
explained by a two-jump model as the line broadening
of the outer quintets is due to a change in the nitrogen
splitting [cf Fig. 3(b)]. In addition, at low temperature
four di†erent b-proton and two nitrogen coupling con-
stants are present in the spectrum shown in Fig. 3(c).
Therefore, the degeneracy of the two-jump mechanism
DISCUSSION
When molecules interact, their spectroscopic properties
are known to change. Therefore, whether averaged or
distinct signals of the free and the inÑuenced species are
observed depends on the spectroscopic time-scale of the
method applied. In the course of our investigation, spe-
ciÐc lock-and-key interactions proved to be slow within
the EPR time-frame, and therefore superimposed
spectra of the di†erent species are obtained. On the
other hand, no evidence was found for superimposed
spectra when non-speciÐc interactions, i.e. mere proto-
nation equilibria, were studied. Nevertheless, both spe-
ciÐc and non-speciÐc interactions exhibit
a
characteristic inÑuence upon the intramolecular
dynamic behaviour of the substrates, a phenomenon
which is resolved by EPR spectroscopy.
All three substrates reÑect a general change of their
EPR spectra when BA is added: the nitrogen coupling
constant increases, whereas that of the b-proton for 1*
and the average value of the two b-protons for 2* and
3* decrease. Assuming that possible protonations at the
heterocyclic ring nitrogen atoms would not lead to a
signiÐcant e†ect upon the hf components observed, we
trace the reported spectral alterations back to proto-
nation of the nitrogen at the methylene group. The
same e†ect can be found with 2,6-di-tert-butyl-4-
(dimethylaminomethyl)phenoxyl when the solvent is
varied systematically: from apolar through polar to
protic solvents, the nitrogen coupling constant increases
dramatically, while the average value of the b-protons is
lowered. The meta-protons (H ) of the phenoxyl remain
m
nearly unchanged [a \ 0.188 mT (toluene), 0.203 mT
N
(acetone), 0.238 mT (ethanol); a \ 1.098 mT (toluene),
Hb
1.082 mT (acetone), 0.930 mT (ethanol)]. Hence the elec-
tronic and steric properties of the nitrogen are altered:
while a neutral nitrogen is known to prefer a position in
the vicinity of the phenoxyl plane,4 quaternization
directs it towards a perpendicular position leading to a
higher coupling constant according to the McConnellÈ
Heller equation.11 Consequently, the dihedral angle of
( 1998 John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 36, 205È210 (1998)

210
M. JAGER ET AL .
is removed and the description of the system has to be
extended to a two-site model where two sets of hf com-
ponents related to distinct positions of the nuclei inter-
change. The two sets of hf components refer to two
species with di†erent orientations and binding sites in
the receptorÈsubstrate complex related by an equi-
librium. Thus, two complexed species, denoted A and B,
exist within the equilibrium 3* ] R ¢ R3*. In detail, the
average value of the b-coupling constants decreases
from free 3* over B to A in line with the alteration of
pure static, a two-jump and a two-site mechanism, were
applied to the interpretation of the EPR spectra
observed. They are referred to the molecular level, pro-
viding information on the intermolecular interactions
and the resulting stereochemistry of the species.
Acknowledgements
This work was supported by the Fonds der Chemischen Industrie, the
Deutsche Forschungsgemeinschaft within the Graduiertenkolleg
“Analytische ChemieÏ and a grant from the Hungarian National
Research Fund (OTKA), T-015841.
a : A has a much larger nitrogen coupling than B,
N
which is slightly larger than that of free 3*. By analogy
with the observations when BA is added, we suggest
that A is attached to the receptor by a hydrogen bond
to the exocyclic nitrogen atom, whereas B is bound via
the intracyclic nitrogen atoms of the benzimidazole
moiety.
REFERENCES
1. M. Jager and H. B. Stegmann, Angew. Chem. 108, 1943 (1996);
Angew. Chem., Int. Ed. Engl. 35, 1815 (1996).
2. J.-M. Lehn, Supramolecular Chemistry. VCH, Weinheim (1995).
3. J. Rebek, Jr, Angew. Chem. 102, 261 (1990); Angew. Chem., Int. Ed.
Engl. 29, 245 (1990).
4. M. Jager, PhD Thesis, Universitat Tubingen, to be published.
5. J. Rebek, Jr, L. Marshall, R. Wolak, K. Parris, M. Killoran, B.
Askew, D. Nemeth and N. Islam, J. Am. Chem. Soc. 107, 7476
(1985).
6. U. HoÑer, Dissertation, Universitat Tubingen (1994).
7. V. Fischer, Dissertation, Universitat Tubingen (1981).
8. H.-J. Schneider, R. Kramer, S. Simova and U. Schneider, J. Am.
Chem. Soc. 110, 6442 (1988).
CONCLUSION
Molecular recognition processes have been analysed by
EPR spectroscopy. Intermolecular equilibria of 1*, 2*,
3* and their lock-and-key complexes with R, depending
on the concentration and temperature of the com-
ponents, are resolved within the EPR time-scale.
Non-speciÐc protonation can be distinguished from
speciÐc receptorÈsubstrate interaction. Whereas non-
speciÐc interaction exhibits a uniform inÑuence on the
substrates, speciÐc recognition does not have a gener-
alized e†ect on the molecules studied. This is interpreted
in terms of di†erent stereochemistry and rigidity of R1*,
R2* and R3*. The complexation also changes the intra-
molecular Ñexibility of the substrates. Three models, a
9. H. Kessler, Angew. Chem. 82, 237 (1970); Angew. Chem., Int. Ed.
Engl. 9, 219 (1970).
10. H. Friebolin, Ein- und Zweidimensionale NMR-Spektroskopie.
VCH, Weinheim (1988).
11. K. Scheffler and H. B. Stegmann, Elektronenspinresonanz.
Springer, Berlin (1970).
12. J. Rebek, Jr, and D. Nemeth, J. Am. Chem. Soc. 108, 5637 (1986).
13. J. Rebek, Jr, B. Askew, M. Killoran, D. Nemeth and F.-T. Lin, J.
Am. Chem. Soc. 109, 2426 (1987).
14. V. Fischer, W. Buhler and K. Scheffler, Z. Naturforsch., T eil A 38,
570 (1983).
( 1998 John Wiley & Sons, Ltd.
MAGNETIC RESONANCE IN CHEMISTRY, VOL. 36, 205È210 (1998)