Proline-rich proteins - Deriving a basis for residue-based selectivity in polyphenolic binding

Article abstract of DOI:10.1039/b800365c

¹H NMR titration experiments have been used to establish that minimal proline-based models show enhanced binding selectivity towards phenol in CDCl₃, relative to other similarly protected amino acid residues. Cooperative binding effects appear to play a role, with sarcosine models affording binding constants to phenol intermediate to those obtained from proline models and other amino acid models. The mechanism for binding, based on DFT calculations and the application of Hunter's molecular recognition toolbox model, cannot be solely attributed to hydrogen bond strength, and appears to be mediated through C-H-π bonds and the rotational freedom of the amide substrate. The Royal Society of Chemistry 2008.

Full text of DOI:10.1039/b800365c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Proline-rich proteins—deriving a basis for residue-based selectivity in
polyphenolic binding†
A. K. Croft* and M. K. Foley
Received 9th January 2008, Accepted 22nd February 2008
First published as an Advance Article on the web 19th March 2008
DOI: 10.1039/b800365c
1
H NMR titration experiments have been used to establish that minimal proline-based models show
enhanced binding selectivity towards phenol in CDCl , relative to other similarly protected amino acid
3
residues. Cooperative binding effects appear to play a role, with sarcosine models affording binding
constants to phenol intermediate to those obtained from proline models and other amino acid models.
The mechanism for binding, based on DFT calculations and the application of Hunter’s molecular
recognition toolbox model, cannot be solely attributed to hydrogen bond strength, and appears to be
mediated through C–H-p bonds and the rotational freedom of the amide substrate.
Introduction
confirmed that the main interaction of the phenolic rings of
tannins is primarily with the proline of PRPs.
35–39
Phenolic compounds are extremely important components of
plants, and are consumed in the diet through a number of sources,
including in high levels in both tea and wine. These molecules
have received much publicity due to their antioxidant nature,
1
and the perceived health benefits that arise from this activity.
However, free in solution, phenolics may interact adversely with
biomolecules because they are able to participate directly in
oxidative processes, as well as possessing one or more nucleophilic
sites in the reduced form and reactive electrophilic sites once
2
oxidised.
In nature, animals are able to compensate for the destructive
properties of phenolics through the selective binding of these
The study of minimal models provides a framework by which
to better understand residue-specific interactions and their con-
tribution to supramolecular assemblies. This can help, with the
aid of the studies already carried out on full protein–tannin
models, to establish firstly what are the underlying residue-specific
driving forces for interaction on an easily understood single-
molecule scale, and secondly where the gaps in understanding
are on going from intermolecular to supramolecular species.
In this context, this work examines a selection of N-acetylated
amino acid esters 1a–e and 2, as well as N-acetyl pyrrolidine
2,3
tannic components to proteins. The phenolics are then able
to impart their antioxidant ability, whilst being protected from
4–8
reaction with the general cellular environment. This is achieved
in humans and other animals, when they are exposed to a tannin-
9
rich diet, through the excretion of two classes of salivary proteins,
histidine-rich proteins (HRPs), which are present constitutively,
and proline-richproteins (PRPs) thatare presentin increased levels
3
,10–12
(
up to 70% of the parotid salivary protein).
The binding of proteins, particularly PRPs, to polyphenolics
1
13–39
3, and their interaction with phenol 4 by H NMR titration
has been extensively studied,
since protein–tannin complexes
40–46
methods. The thermodynamics of the interaction of N-acetyl
proline methyl ester 2 and phenol 4 are examined in further detail.
These results have been complemented by calculations of the
structures of relevant amino derivatives 1–3, with determination
of their electrostatic potentials to elucidate key contributions
to the binding mechanism, discussed in the context of Hunter’s
are key to the sensation of astringency.
to the proline features of PRPs, which constitute approximately
5% of the protein content, was postulated by Hagerman and
Specificity for binding
3
21
Butler from the results of competitive assays. In this work, all
polypeptides that showed good affinity for proanthocyanidin also
contained high amounts of proline. It was noted that polyproline
and polysarcosine showed similar affinities for proanthocyanidin.
Small peptides were unable to compete satisfactorily in this assay,
highlighting the requirement for several additive or synergistic
47
molecular recognition toolbox model. This combination allows
separate contributors to the binding to start to be understood.
1
interactions. Subsequent detailed H NMR studies have since
Experimental methods
Synthesis of compounds
School of Chemistry, University of Wales Bangor, Bangor, Gwynedd.
E-mail: a.k.croft@bangor.ac.uk; Fax: +44 1248 370 528; Tel: +44 1248
The L-amino acid derivatives 1a–e and 2 were prepared from
3
82 375
the corresponding L-amino acids as previously described in
†
Electronic supplementary information (ESI) available: Synthetic proce-
dures and characterisation, H NMR details and titration curves. See DOI:
48–50
51
1
the literature,
as was the N-acetyl pyrrolidine 3. Extended
1
0.1039/b800365c
experimental data is provided in the ESI.† Phenol 4 was purchased
1
594 | Org. Biomol. Chem., 2008, 6, 1594–1600
This journal is © The Royal Society of Chemistry 2008

View Article Online
from Sigma-Aldrich, and used as received. The structures and
purity of all compounds were verified by NMR prior to use.
Table 1 Association constants of selected amides 1–3 with phenol 4 in
1
CDCl
3
based on H NMR titration results
−
1
3
Compound
Ka 1 (mol dm )
1
H NMR titrations
N-Ac-Gly-OMe 1a
trans-N-Ac-Sar-OMe 1b
cis-N-Ac-Sar-OMe 1b
N-Ac-Ala-OMe 1c
N-Ac-Phe-OMe 1d
N-Ac-O-Bzl-Tyr-OMe 1e
trans-N-Ac-Pro-OMe 2
cis-N-Ac-Pro-OMe 2
N-Ac-Pyr 3
2.4 ± 0.6
6.0 ± 2.7
6.9 ± 1.1
2.4 ± 0.8
2.2 ± 1.1
2.7 ± 0.7
18.5 ± 2.9
28.0 ± 8.1
14.0 ± 8.5
1
H NMR titrations were performed on a 500 MHz Bruker
Avance spectrometer with a 5 mm broad band observation (BBO)
gradient probe with calibration from the position of residual non-
deuterated solvent, cross referenced to trimethylsilane (TMS).
Each measurement was taken on at least three freshly prepared
samples for each concentration point. Solutions (10 ml) of the
compounds 1–4 were prepared at known concentrations in CDCl .
For the direct titration, aliquots of 1.0 mol dm of phenol 4 were
added to 45 ll each of 0.1 mol dm of the amides 1–3, with
addition of extra CDCl to make up the total volume to 500 ll.
A H NMR spectrum was recorded after each aliquot addition
3
−
3
Electrostatic surface calculations
−
3
3
Density functional theory (DFT) calculations were performed
using the Spartan ‘02 programme.
1
53,54
The minimum energies of
and the chemical shifts of the proton signals of the amides 1–
3
(
both rotamer conformations of 2 were obtained by an initial
conformer distribution calculation, at the B3LYP/6-31(d) level
of theory. The generated conformations were then optimised to
obtain equilibrium geometry structures. The equilibrium geome-
try structures of 1–4 were also calculated at B3LYP/6-31(d). All
structures were confirmed as minima by the absence of negative
were monitored with each addition. The association constants
) and predicted change in chemical shift to produce saturation
K
a
of association (Ddpred) were derived by plotting non-linear titration
TM
curves using the computer graphing program, Mac-Curve-Fit
The association constant, K
a)–(c), which allow the observed chemical shift values of any
.
a
, is obtained by application of eqn
(
frequencies. Electrostatic potential surfaces of molecules 1–4 were
generated also using Spartan ‘02
given proton signal, dobs, to be fitted as a function of the initial
concentration of phenol 4. The association constant for each
experiment was evaluated as the weighted average on the observed
change in chemical shift for all signals monitored.
53,54
at the B3LYP/6-31G(d) level.
Results and discussion
1
H NMR titrations
NMR titrations provide one method to identify binding interac-
1
tions between species. For H NMR, the change in chemical shift
of protons of a molecule of interest can be monitored in response to
increasing quantities of a complexing agent, and usually indicate
a change in the chemical environment of species one initiated
by species two. From these changes in chemical shift, binding
(
a)
[
H] = [H] − [HG]
i
(b)
constants, K
amide derivatives 1–3 with phenol 4 are presented in Table 1.
The K values for the glycyl, alanyl, phenylalanyl and tyrosyl
a
, can be calculated. The K
a
values obtained for the
a
derivatives 1a and 1c–e, respectively, indicate negligible binding
within experimental error. A slightly more significant binding is
(c)
observed for the sarcosine derivative 1b, with a predicted K
a
of
−
1
3
Where d
the shift of the bound complex and [H]
concentrations of host and guest, respectively.
u
is the chemical shift of the unbound host and d
b
is
around 6 mol dm . For the proline and pyrrolidine derivatives 2
−
1
3
i
and [G]
i
are the initial
and 3, the measured K increases to around 14 mol dm in the
a
case of 3, 18 mol− dm for the trans-rotamer of 2 and 28 mol dm
1
3
−1
3
The stoichiometry of the association between the proline
for the cis-rotamer. Whilst these are small values compared with
typically reported binding constants, the trend and relative values
are reproducible for these systems.
derivative 2 and phenol 4 was ascertained by the continuous
52 1
variation method, otherwise known as Job plot analysis.
H
NMR spectra were obtained using a range of different ratios
of 2 and 4 and every proton signal of the proline derivative 2
was analysed for each rotamer. At the concentrations chosen to
Two rotamer conformations of the tertiary amides 1b and 2,
termed trans- and cis-rotamers (see e.g. Scheme 1), are observed
directly on the NMR time scale. This phenomenon is seen with
−
3
−3
55–59
perform the Job plot analysis (0.1 mol dm to 1.0 mol dm ), the
many tertiary amides.
The trans-rotamer is more prevalent
proton signals of the proline derivative 2 were generally easier to
resolve than when performing the H NMR titration experiments,
than the cis-rotamer and the rotamer conformations of both
1
sarcosine 1b and proline 2 were estimated to be at approximately
1
especially for the trans-rotamer.
Standard spectra were analysed at 20 C, with probe tempera-
a 4 : 1 ratio in a CDCl
integrals.
3
solution from analysis of the H NMR
◦
ture maintained by electronic heating and liquid nitrogen cooling.
In a similar fashion, all temperature dependent plots were cooled
or heated as required to the specific temperature with equilibration
of samples prior to measurement.
The two rotamer conformers of both sarcosine 1b and proline
2 were analysed separately by H NMR titration. The changes in
chemical shift of the a-protons of proline 2 (proton 1, Scheme 1),
with addition of phenol 4, are illustrated in Fig. 1. The a-proton
1
This journal is © The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 1594–1600 | 1595

View Article Online
Scheme 1 The two rotamer conformations of N-Ac-Pro-OMe 2. The
amide bond rotation equilibrium is observed directly on the NMR time
1
scale and both conformers are observed on a H NMR spectrum.
1
Fig. 2 (a) trans- and (b) cis-N-Ac-Pro-OMe 2 a- H NMR titration
curves with addition of phenol 4. The curves indicate that the trans-rotamer
associates at a 1 : 1 ratio that results in a down-field change in chemical
shift of the a-proton, but the cis-rotamer appears to produce a secondary
association with 4, and an up-field shift is observed after the initial
down-field change in chemical shift.
Fig. 1 A stack plot showing the a-proton signals of N-Ac-Pro-OMe 2
as a gradually increasing concentration of phenol 4 is added, shown in
−
3
3
mol dm (M) to the left. The solutions were made up in CDCl solution,
−
3
and the concentration of 2 is 0.009 mol dm . When no phenol is present
H
the signals correspond to (from left to right): d 4.49, doublet of doublets
signal is most probably due to shielding caused by the ring
current of the phenolic aromatic ring. The observed chemical
shift changes of the cis-rotamer proton signals are significantly
different. The initial association results in the chemical shift of the
protons moving down-field, apart from one of the ring protons,
although this observation is confounded by overlapping proton
signals. The secondary association causes the chemical shift of
most of the protons to move up-field, with the most significant
differences being observed for the ring protons. Although the
low concentration of the cis-rotamer makes the observations less
reliable, the results suggest that a very different mode of association
occurs than is seen for the trans-rotamer of 2.
The difference in association mechanism of the two rotamers is
further highlighted when continuous variation (Job plot) analysis
is carried out. This method allows the stoichiometry of binding
to be determined through variation in the concentration of
each analyte whilst maintaining a constant concentration of the
combined analytes. The resulting Job plots are presented in Fig. 3.
Fig. 3 indicates that for both rotamers of the proline derivative
2, the largest changes in shift are associated with the a-proton
in each case. For the trans-rotamer the maxima for the a-proton,
(
trans) a-CH; 4.38 ppm, doublet of doublets (cis) a-CH. Both a-proton
signals appear to change in chemical shift with addition of phenol.
signal was monitored due to the relatively large change in chemical
shift compared to the other signals, and also the ease of resolving
both rotamer conformation signals.
The association of the trans-rotamer of 2 to phenol 4 resulted in
one saturation point, but association to the cis-rotamer produced
two saturation points (Fig. 2). This suggests that a secondary
association occurs, unique to the cis-rotamer, with a quantifiable
K
a
value. However, the K values obtained for the secondary
a
association are appreciably smaller than the initial association
with values around 7 ± 1 calculated.
With association of the proline derivative 2 to phenol 4, the
largest changes in chemical shift observed for the signals of the
trans-rotamer protons are those corresponding to the a-proton
and one specific NCH
assign the facial locality of the protons, the larger change in
chemical shift of a specific NCH proton does suggest specific
facialassociation. The direction inchemical shift is also significant;
the down-field shift observed for the a-proton signal is presumably
due to deshielding by either an electron-withdrawing group, for
example the phenolic oxygen, or due to the ring current of the
2
proton. Although it is not possible to
2
OCH protons and NAc protons are around 0.45 indicating that
3
the stoichiometry is likely to be 1 : 1, although this maximum
value could suggest that a second stoichiometry is possible.
aromatic ring. The up-field shift observed by the NCH
2
proton
1
596 | Org. Biomol. Chem., 2008, 6, 1594–1600
This journal is © The Royal Society of Chemistry 2008

View Article Online
Fig. 4 (a) trans- and (b) cis-N-Ac-Pro-OMe 2 Ka1 values for the
2
association to phenol 4 vs. temperature. R = 0.96 (trans) and 0.92 (cis).
in a similar manner as observed with the cis-rotamer at room
temperature. As the temperature is increased, this secondary
association to the trans-rotamer disappears. Fig. 4 illustrates how
the Ka1 values correlate with temperature. A linear relationship fits
2
both sets of data relatively well, with R values of 0.96 and 0.92 for
the trans- and cis-rotamers respectively. The lower correlation of
the cis-rotamer association can be accounted for by larger errors
associated with the measurements than for the trans-rotamer. Both
correlations indicate that lower temperatures enhance the initial
association, as would be expected for a binding process and where
competitive interactions may occur.
3
Fig. 3 Job plots monitoring the a-proton (᭺), the OCH protons
(
ꢀ), the N-AcCH
3
protons (♦), and the NCH protons (D and ×) of
2
trans-N-Ac-Pro-OMe (a), and cis-N-Ac-Pro-OMe (b) at different ratios of
phenol 4.
The value of the change in enthalpy when association occurs
−1
(
DH) was calculated as −10.3 ± 3.9 kJ mol for the trans-rotamer
61
using the van’t Hoff equation, from the gradient of a plot of lnK
vs. 1/T. This value is consistent with a moderate hydrogen-bonding
a
2
However, when monitoring the NCH proton signals, a maximum
62
of around 0.40 was obtained. This can be attributed due to errors
in measurement arising from signals corresponding to the cis-
rotamer overlapping with the trans-rotamer proton signals. The
maxima for the cis-rotamer is about 0.3 and indicates that a
different stoichiometry is prevalent, probably 2 : 1. However,
with such small changes in chemical shift for the cis-rotamer
interaction.
d -DMSO typically competes for hydrogen-bonding
6
47
sites and addition (0.1 M) to the system commensurately reduced
the Ka1 to a very low level, namely from 18.5 ± 2.9 to 0.43 ± 0.06
for the trans-rotamer and from 28.0 ± 8.1 to 0.93 ± 0.30 for the cis-
rotamer. This suggests that hydrogen bonding may be important
in maintaining the complex.
60
association, the results become difficult to interpret.
The DG values of the association of the trans-rotamer of N-Ac-
Pro-OMe 2 with phenol 4 were calculated from the experimental
The association of both rotamers 2 with phenol 4 was examined
at different temperatures to extract thermodynamic information.
When the temperature of the system is decreased, the initial
association (Ka1) to both of the rotamers is strengthened (Table 2).
A secondary association (Ka2) to the trans-rotamer was also seen,
K
a
values at each temperature and the value of DH derived
above (Table 2). These values are all consistent, all being around
−
1
7 kJ mol , reflective of a process that is favoured at the
temperatures examined.
This journal is © The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 1594–1600 | 1597

View Article Online
1
Table 2 a- H NMR titration results for trans- and cis-N-Ac-Pro-OMe 2—phenol 4 association at different temperatures with corresponding
thermodynamic data
−
1
3
−1
3
DG/kJ mol−1
DH/kJ mol−1
−10.3 ± 3.9
DS/J K−1
Temperature/K
253
K
a1/mol dm
K
a2/mol dm
trans-Rotamer
cis-Rotamer
31.3 ± 4.9
21.4 ± 3.6
18.5 ± 2.9
11.5 ± 4.0
39.2 ± 10.9
31.9 ± 9.1
28.0 ± 8.1
11.1 ± 8.5
5.3 ± 1.7
6.8 ± 3.4
—
−7.2 ± 0.7
−7.0 ± 0.8
−7.1 ± 0.8
−6.4 ± 1.9
2
2
3
73
93
13
−10.8 ± 5.9
—
253
3.4 ± 1.4
2.2 ± 4.5
7.0 ± 1.4
—
2
2
3
73
93
13
Electrostatic surface calculations
The calculated electrostatic surface properties of the molecules
allow the analysis of the most likely modes of association. Hunter
has recently postulated a simplified model for molecules whereby
interactions between two species in solution can be predicted
47
primarily by their electrostatic charges. Using this model, the
calculated electrostatic surface charges of the amide 2 and phenol
4
predict that the most likely interaction would be a hydrogen
bonding interaction between the Emax of phenol, namely the
phenolic hydrogen, and the Emin of the amide, viz the amide
carbonyl oxygen (Fig. 5). This observation is consistent with
63–66
previous literature on the binding of phenols with amides.
Hydrogen-bond donor and acceptor constants (a and
67
b, Table 3) were calculated for the molecules 1–4 and the solvents
chloroform and water from the corresponding Emax and Emin values.
The free energy of hydrogen-bonding interaction (DDGH-bond) can
47
be crudely predicted from these numbers. This gives a rough
estimation of the K if the hydrogen bonding interaction were
a
the overwhelming driving force for binding. Since the carbonyl of
the ester could also act as an alternative hydrogen-bonding site,
the Eester value is also presented, along with the corresponding
hydrogen bond acceptor constant, denoted e here. The values of
E
ester are all significantly smaller than Emin, confirming that the
amide is the preferred binding site for the amino acid derivatives
investigated in this study. Calculating DDG and the corresponding
K for water (Table 4), indicates that hydrogen bonding is negligible
a
in aqueous solution, consistent with the assertion that hydrogen
bonding is not the main driving association between tannins and
2
2
proteins in solution. The same calculation for association in CCl
affords K values highly consistent with the magnitude expected
for phenol–amide complexation, based on reported values in
4
Fig. 5 Molecular electrostatic surfaces of: (a) trans-N-Ac-Pro-OMe 2,
and (b) phenol 4, plotted on a van der Waals surface calculated using
ꢀ
a
B3LYP/6-31G(d). Positive regions are shown in blue, negative regions
are shown in red, and neutral regions are shown in green. The locations
of the maximum electrostatic potential, Emax, the minimum electrostatic
potential, Emin, and the electrostatic potential of the aromatic ring (Ear) are
labelled.
65,66
the literature.
hydrogen bonding as the major interaction in this solvent.
The predicted K values for the derivatives 1a–e, 2 and 3
in CHCl are all significantly smaller than the measured K
values. Furthermore, the predicted binding order contrasts with
the trends seen for the measured K values for these species. To
rationalise the observed binding, an interaction of the phenol ring
with the hydrogen atoms of either the NCH of the sarcosine
b or the NCH of the pyrrolidine derivatives 2 and 3 can be
proposed. Related CH–p interactions have been recognised in
This correlates with the evidence supporting
a
3
a
already well-established work that points to the influence of a
strong ‘hydrophobic effect’ in the binding of larger tannins to
a
22
polypeptides, and particularly to a specific interaction with the
37–39
3
CH groups of the proline.
1
2
A small increase in the DDG of binding, to account for CH–
p and related interactions, is able to account for the observed
binding constants satisfactorily. The higher binding observed for
the pyrrolidine derivatives 2 and 3, relative to the sarcosine 1b can
be rationalised through an increased number of possible CH–p
interactions.
68
peptides in the form of phenylalanine–proline interactions, and
may contribute enough of either an additive or synergistic effect to
improve the binding to a measurable level, relative to a derivative
with only a secondary amide. This would be consistent with
1
598 | Org. Biomol. Chem., 2008, 6, 1594–1600
This journal is © The Royal Society of Chemistry 2008

View Article Online
47
Table 3 Electrostatic surface charges for compounds 1–4 obtained at the B3LYP/6-31G(d) level with Spartan and derived a, b and e constants, with
calculated DDG for the hydrogen bond interaction and predicted and experimental K
a
s
h
DDGH-bond
/
Predicted
−
1
−1
−1
−1
−1
3
Molecule
E
min/kJ mol
E
max/kJ mol
E
ester/kJ mol
kJ mol
K
a
/mol dm
a
b
N-Ac-Gly-OMe 1a
trans-N-Ac-Sar-OMe 1b
cis-N-Ac-Sar-OMe 1b
N-Ac-Ala-OMe 1c
N-Ac-Phe-OMe 1d
N-Ac-O-Bzl-Tyr-OMe 1e
trans-N-Ac-Pro-OMe 2
cis-N-Ac-Pro-OMe 2
N-Ac-Pyr 3
−211.9
−209.5
−211.5
−203.3
−211.0
−210.2
−209.0
−217.6
−219.2
−133.4
−43.6
171.6 c
123.6
124.2
174.6
184.0
−146.4
−146.4
−143.2
−133.9
−125.5
−131.2
−158.9
−152.3
—
—
—
—
—
3.3
2.4
2.4
3.4
3.5
3.4
1.9
2.1
1.7
5.0
3.2
4.2
1.9
4.1
2.0
4.1
3.9
4.1
4.0
4.0
4.2
4.2
2.6
0.8
3.6
0.4
2.8
2.8
2.8
2.6
2.4
2.5
3.1
2.9
−5.7
−5.6
−5.7
−5.4
−5.6
−5.6
−5.6
−5.9
−5.9
−3.0
0.74
0.68
0.72
0.56
0.71
0.70
0.67
0.88
0.92
0.06
a
a
a
a
a
a
a
a
f
c
b
b
b
178.9
d
100.1 e
107.8
86.6
258.0
167.7
216.0
96.5
d
g
Phenol 4
—
—
—
—
Chloroform
Water
Carbon tetrachloride
−189.1
−22.6
a
b
c
d
Located over the amide carbonyl oxygen. Located over the amide NH hydrogen. Located on the CH
2
hydrogens. Located over the ring hydrogens.
e
f
g
h
Located over the OMe protons. Located over the OH oxygen. Located over the OH hydrogen. Calculated with chloroform as solvent at 298 K.
Table 4 Predicted DDG and association constants (K
molecules with phenol 4 in H O and CCl , based on the B3LYP/6-31G(d)
derived a and b values (Table 3)
a
) of selected
Conclusions
2
4
1
H NMR titrations of minimal peptide-based models with phenol
4
have established that, in CHCl solution, there is a selective
3
DDG
(H O)
K
a
DDG
K
a
binding of proline-based peptides, relative to other similarly
protected amino acid residues. This binding appears to be medi-
ated by interactions with the tertiary amide structure primarily,
with pyrrolidine-based structures 2 and 3 having the largest
interactions, followed by the less-substituted sarcosine derivative
1b and negligible binding for the secondary amide derivatives 1a
and 1c–e.
Binding to the trans-rotamer of N-acetylproline methyl ester 2
appears to be in a roughly 1 : 1 ratio at 298 K, whereas titration of
the cis-rotamer with phenol shows a secondary, smaller binding
with different overall stoichiometry of around 2 : 1. This secondary
association is also observed with the trans-rotamer 2 in cooled
solutions. Thermodynamic analysis indicates that the association
for both rotamers is an enthalpically favoured process. Interactions
Molecule
2
(H
2
O)
(CCl
4
)
4
(CCl )
N-Ac-Gly-OMe 1a
trans-N-Ac-Sar-OMe 1b
cis-N-Ac-Sar-OMe 1b
N-Ac-Ala-OMe 1c
N-Ac-Phe-OMe 1d
N-Ac-O-Bzl-Tyr-OMe 1e
trans-N-Ac-Pro-OMe 2
cis-N-Ac-Pro-OMe 2
N-Ac-Pyr 3
−0.4
−0.3
−0.4
−0.2
−0.3
−0.3
−0.3
−0.5
−0.5
0.9
0.01
0.01
0.01
0.00
0.01
0.01
0.01
0.01
0.01
0.00
−11.4
−11.2
−11.3
−10.8
−11.3
−11.3
−11.2
−11.7
−11.8
−6.6
145.59
127.02
142.01
90.04
138.18
132.08
123.50
199.53
218.15
1.82
Phenol 4
The strength of these combined interactions is clearly mediated
by the orientation of the amide, as can be ascertained from the
differences seen between the cis and trans rotamers of the N-
acetylproline methyl ester 2. This is consistent with an underlying,
directing hydrogen-bonding interaction. An additional contribu-
with d -DMSO indicate that hydrogen bonding is likely to be
important in the association mechanism.
6
Electrostatic calculations, based on Hunter’s molecular recogni-
tion toolbox methods, were used to help identify the contribution
of hydrogen bonding. These calculations confirm that the pre-
ferred electrostatic interaction is that between the amide carbonyl
and the phenolic hydroxyl, in line with previous studies on H-
−
1
tion of around 3.6 kJ mol from CH–p interactions for each
rotamer is then sufficient to account for the observed K s and
a
is entirely consistent with the expected energetic contribution
from this class of interaction. This estimation also adequately
63–66
bonding carried out in CCl
4
.
The predicted K
a
s for CHCl
3
accounts for the effect of hydrogen-bond disruption seen in d
6
-
as solvent illustrate that this hydrogen bond alone is insufficient
to account for the binding observed in the experimental system.
However, an additional, small contribution from proposed CH–p
interactions is sufficient to explain the experimental observations.
This is consistent with the work on full protein–tannin systems
that indicate in polar solvents that hydrogen bonding cannot
be the major driving force and that ‘hydrophobic’ interactions
DMSO, as, even if these interactions are fully retained, the
energetic contribution is insufficient to maintain a measurable
binding.
The same magnitude of binding might be expected for the N-
acetylpyrrolidine 3, and, in concert with the predicted hydrogen
bonding, a K slightly larger than that for the cis-N-acetylproline
a
22,37–39
methyl ester 2 might be anticipated. The discrepancy in the
experimental observation could be attributed to error within the
experimental measurements, but might also be attributed either
to differences in the degree of rotational freedom, or the electron
withdrawing nature of the carboxylic side chain. Clearly, more
detailed studies would be required to delineate these effects.
direct binding,
illustrating that these are reasonable models
for the larger systems and the utility of combining experimental
observations with theoretical calculations.
Whilst it is recognised that H-bonding is crucial to the overall
association, it cannot work as the only factor in mediating the
association. CH–p interactions offer a clear residue-level driving
This journal is © The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 1594–1600 | 1599

View Article Online
force for the preferential binding of phenolics to proline residues
in PRPs, but in our systems are also unlikely to be strong enough
to act alone. Thus the two interactions must work together, and
in doing so afford the selectivity observed in larger systems. An
additive combination of interactions is sufficient to account for the
binding in the reported systems. Further work is ongoing within
our group to more clearly delineate the particular contributions
of each interaction.
31 Y. Chen and A. E. Hagerman, J. Agric. Food Chem., 2004, 52, 4008–
4
011.
3
2 A. J. Charlton, N. J. Baxter, M. L. Khan, A. J. G. Moir, E. Haslam,
A. P. Davies and M. P. Williamson, J. Agric. Food Chem., 2002, 50,
1593–1601.
33 K. Wroblewski, R. Muhandiram, A. Chakrabartty and A. Bennick,
Eur. J. Biochem., 2001, 268, 4384–4397.
3
4 N. Naurato, P. Wong, Y. Lu, K. Wroblewski and A. Bennick, J. Agric.
Food Chem., 1999, 47, 2229–2234.
35 A. J. Charlton, N. J. Baxter, T. H. Lilley, E. Haslam, C. J. McDonald
and M. P. Williamson, FEBS Lett., 1996, 382, 289–292.
3
3
3
3
4
6 C. Simon, K. Barathieu, M. Laguerre, J. M. Schmitter, E. Fouquet, I.
Pianet and E. J. Dufourc, Biochemistry, 2003, 42, 10385–10395.
7 A. J. Charlton, E. Haslam and M. P. Williamson, J. Am. Chem. Soc.,
2002, 124, 9899–9905.
8 N. J. Murray, M. P. Williamson, T. H. Lilley and E. Haslam,
Eur. J. Biochem., 1994, 219, 923–925.
9 N. J. Baxter, T. H. Lilley, E. Haslam and M. P. Williamson, Biochem-
istry, 1997, 36, 5566–5577.
0 J. L. Goldstein and T. Swain, Phytochemistry, 1963, 2, 371–383.
Acknowledgements
The authors would like to thank Dr J. B. Harper for extended
useful discussions and assistance with Mac-Curve-Fit.
References
1
K.-T. Chung, T. Y. Wong, C.-I. Wei, Y.-W. Huang and Y. Lin, Crit. Rev.
Food Sci. Nutr., 1998, 38, 421–464.
41 E. C. Bate-Smith, Phytochemistry, 1973, 12, 907–912.
42 L. J. Porter and J. Woodruffe, Phytochemistry, 1984, 23, 1255–1256.
43 T. Ozawa, T. H. Lilley and E. Haslam, Phytochemistry, 1987, 26, 2937–
2942.
44 G. Luck, H. Liao, N. J. Murray, H. R. Grimmer, E. E. Warminski,
M. P. Williamson, T. H. Lilley and E. Haslam, Phytochemistry, 1994,
37, 357–371.
2
3
H. M. Appel, J. Chem. Ecol., 1993, 19, 1521–1551.
H. Mehansho, L. G. Butler and D. M. Carlson, Annu. Rev. Nutr., 1987,
4
23–440.
4
5
Y. Chen and A. E. Hagerman, J. Agric. Food Chem., 2004, 52, 6061–
6
067.
M. J. T. J. Arts, G. R. M. M. Haeman, L. C. Wilms, S. A. J. N. Beetstra,
C. G. M. Heijnem, H.-P. Voss and A. Bast, J. Agric. Food Chem., 2002,
45 E. Jobstl, J. O’Connell, J. P. A. Fairclough and M. P. Williamson,
Biomacromolecules, 2004, 5, 942–949.
5
0, 1184–1187.
46 I. Lesschaeve and A. C. Noble, Am. J. Clin. Nutr., 2005, 81, 330S–335S.
47 C. A. Hunter, Angew. Chem., 2004, 43, 5310–5324.
48 R. G. Webb, M. W. Haskell and C. H. Stammer, J. Org. Chem., 1969,
34, 576–580.
6
7
K. M. Riedl and A. E. Hagerman, J. Agric. Food Chem., 2001, 49,
917–4923.
S. Rohn, H. M. Rawel and J. Kroll, J. Agric. Food Chem., 2004, 52,
725–4829.
4
4
49 A. G. Griesbeck, S. Bondock and J. Lex, J. Org. Chem., 2003, 68,
8
9
W. Wang and M. T. Goodman, Nutr. Res., 1999, 19, 191–202.
A. Bennick, Crit. Rev. Oral Biol. Med., 2002, 13, 184–196.
9899–9906.
50 M. Prein, P. J. Munley and A. Padwa, Tetrahedron, 1997, 53, 7777–7794.
51 S. Tsuboi, Y. Nooda and A. Takeda, J. Org. Chem., 1984, 49, 1204–
1208.
1
1
1
0 H. Mehansho, D. K. Ann, L. G. Butler, J. C. Rogler and D. M. Carlson,
J. Biol. Chem., 1987, 262, 12344–12350.
1 H. Mehansho, T. N. Asquith, L. G. Butler, J. C. Rogler and D. M.
Carlson, J. Agric. Food Chem., 1992, 40, 93–97.
2 H. Mehansho, S. Clements, B. T. Sheares, S. Smith and D. M. Carlson,
J. Biol. Chem., 1985, 260, 4418–4423.
3 E. Haslam, Biochem. J., 1974, 139, 285–288.
4 J. P. McManus, K. G. Davis, T. H. Lilley and E. Haslam, J. Chem. Soc.,
Chem. Commun., 1981, 309–311.
52 J. W. Steed and J. L. Atwood, in Supramolecular Chemistry, John Wiley
and Sons Ltd., Chichester, England, 2000, pp. 16–19.
53 C. A. Irvine, Spartan ‘02, Wavefunction, Inc., 2002.
54 J. Kong, C. A. White, A. I. Krylov, C. D. Sherrill, R. D. Adamson,
T. R. Furlani, M. S. Lee, A. M. Lee, S. R. Gwaltney, T. R. Adams,
C. Ochsenfeld, A. T. B. Gilbert, G. S. Kedziora, V. A. Rassolov, D. R.
Maurice, N. Nair, Y. Shao, N. A. Besley, P. E. Maslen, J. P. Dombroski,
H. Daschel, W. Zhang, P. P. Korambath, J. Baker, E. F. C. Byrd, T. V.
Voorhis, M. Oumi, S. Hirata, C.-P. Hsu, N. Ishikawa, J. Florian, A.
Warshel, B. G. Johnson, P. M. W. Gill, M. Head-Gordon and J. A.
Pople, J. Comput. Chem., 2000, 21, 1532–1548.
55 D. A. Block, D. Yu, D. A. Armstrong and A. Rauk, Can. J. Chem.,
1998, 76, 1042–1049.
56 F. X. Schmid, Adv. Protein Chem., 2002, 59, 243–282.
57 K. R. Williams, B. Adhyaru, I. German and E. Alvarez, J. Chem. Educ.,
2002, 79, 372–373.
1
1
1
1
5 J. P. McManus, K. G. Davis, J. E. Beart, S. H. Gaffney, T. H. Lilley and
E. Haslam, J. Chem. Soc., Perkin Trans. 2, 1985, 1429–1438.
6 C. M. Spencer, Y. Cai, R. Martin, S. H. Gaffney, P. N. Goulding, D.
Magnolato, T. H. Lilley and E. Haslam, Phytochemistry, 1988, 27,
2
397–2409.
1
1
7 J. E. Beart, T. H. Lilley and E. Haslam, Phytochemistry, 1985, 24, 33–38.
8 K. S. Feldman, A. Sambandam, S. T. Lemon, R. B. Nicewonger, G. S.
Long, D. F. Battaglia, S. M. Ensel and M. A. Laci, Phytochemistry,
1
999, 51, 867–872.
1
2
9 R. A. Frazier, A. Papadopoulou, I. Mueller-Harvey, D. Kisoon and
R. J. Green, J. Agric. Food Chem., 2003, 51, 5189–5195.
58 W. J. Wedemeyer, E. Welker and H. A. Scheraga, Biochemistry, 2002,
41, 14637–14644.
59 C. Dugave and L. Demange, Chem. Rev., 2003, 103, 2375–2532.
60 J. M. Sanderson and E. J. Whelan, Phys. Chem. Chem. Phys., 2004, 6,
1012–1017.
0 A. E. Hagerman and L. G. Butler, J. Agric. Food Chem., 1980, 28,
9
44–947.
2
2
1 A. E. Hagerman and L. G. Butler, J. Biol. Chem., 1981, 256, 4494–4497.
2 H. I. Oh, J. E. Hoff, G. S. Armstrong and L. A. Haff, J. Agric. Food
Chem., 1980, 28, 394–398.
61 P. W. Atkins, The Elements of Physical Chemistry, Oxford University
Press, Oxford, 1996.
2
2
3 T. N. Asquith and L. G. Butler, Phytochemistry, 1986, 25, 1591–1593.
4 M. Zhu, J. D. Phillipson, P. M. Greengrass, N. E. Bowery and Y. Cai,
Phytochemistry, 1997, 44, 441–447.
62 J. W. Steed and J. L. Atwood, Supramolecular Chemistry, John Wiley
& Sons, Ltd, Chitchester, 2000.
63 J.-Y. Le Questel, C. Laurence, A. Lachkar, M. Helbert and M.
Berthelot, J. Chem. Soc., Perkin Trans. 2, 1992, 2091–2094.
64 M. D. Joesten and R. S. Drago, J. Am. Chem. Soc., 1962, 84, 2696–
2699.
2
2
5 A. E. Hagerman, M. E. Rice and N. T. Ritchard, J. Agric. Food Chem.,
1
998, 46, 2590–2595.
6 K. J. Siebert, N. V. Troukhanova and P. Y. Lynn, J. Agric. Food Chem.,
996, 44, 80–85.
1
65 T. Gramstad and W. J. Fuglevik, Acta Chem. Scand., 1962, 16, 1369–
2
2
7 V. A. P. deFreitas and N. Mateus, J. Sci. Food Agric., 2002, 82, 113–119.
8 V. A. P. deFreitas and N. Mateus, J. Agric. Food Chem., 2001, 49,
40–945.
9 J. R. Bacon and M. J. C. Rhodes, J. Agric. Food Chem., 1998, 46,
083–5088.
1377.
66 E. Kwiatkowski, K. Kozubek and Z. Pepli n˜ ski, Z. Naturforsch., B:
Anorganische, organische und biologische Chemie, Botanik, Zoologie und
verwandte Gebiete, 1978, 33, 230–236.
67 M. H. Abraham and J. A. Platts, J. Org. Chem., 2001, 66, 3484–3491.
68 Z. Ciunik, S. Berski, Z. Latajka and J. Leszczynski, J. Mol. Struct.,
1998, 442, 125–134.
9
2
3
5
0 S. Verge, T. Richard, S. Moreau, S. Richelme-David, J. Vercauteren,
J.-C. Prome and J.-P. Monti, Tetrahedron Lett., 2002, 43, 2363–2366.
1
600 | Org. Biomol. Chem., 2008, 6, 1594–1600
This journal is © The Royal Society of Chemistry 2008