Carbohydrate molecular recognition: A spectroscopic investigation of carbohydrate-aromatic interactions

Article abstract of DOI:10.1039/b704792d

The physical basis of carbohydrate molecular recognition at aromatic protein binding sites is explored by creating molecular complexes between a series of selected monosaccharides and toluene (as a truncated model for phenylalanine). They are formed at low temperatures under molecular beam conditions, and detected and characterized through mass-selected, infrared ion depletion spectroscopy - a strategy which exploits the extraordinary sensitivity of their vibrational signatures to the local hydrogen-bonded environment of their OH groups. The trial set of carbohydrates, α- and β-anomers of glucose, galactose and fucose, reflects ligand fragments in naturally occurring protein-carbohydrate complexes and also allows an investigation of the effect of systematic structural changes, including the shape and extent of 'apolar' patches on the pyranose ring, removal of the OH on the exocyclic hydroxymethyl group, and removal of the aglycon. Bound complexes invariably form, establishing the general existence of intrinsic intermolecular potential minima. In most of the cases explored, comparison between recorded and computed vibrational spectra of the bound and free carbohydrates in the absence of solvent water molecules reveal that dispersion forces involving CH-π interactions, which promote little if any distortion of the bound carbohydrate, predominate although complexes bound through specific OH-π hydrogen-bonded interactions have also been identified. Since the complexes form at low temperatures in the absence of water, entropic contributions associated with the reorganization of surrounding water molecules, the essence of the proposed 'hydrophobic interaction', cannot contribute and other modes of binding drive the recognition of sugars by aromatic residues. Excitingly, some of the proposed structures mirror those found in naturally occurring protein-carbohydrate binding sites. the Owner Societies.

Full text of DOI:10.1039/b704792d

View Article Online / Journal Homepage / Table of Contents for this issue
This paper is published as part of a
PCCP themed issue on
Spectroscopic probes of
molecular recognition
Guest edited by Martin Suhm
(Universität Göttingen)
Image reproduced by permission of Dr
Michel Mons from Phys. Chem. Chem.
Phys., 2007, 9, 4491
Image reproduced by permission of
Professor Takayuki Ebata from Phys.
Chem. Chem. Phys., 2007, 9, 4452
Published in issue 32, 2007 of PCCP
Other papers in this issue:
Carbohydrate molecular recognition: a spectroscopic investigation of carbohydrate–aromatic
interactions
John P. Simons et al., Phys. Chem. Chem. Phys., 2007, 9, 4444 (DOI: 10.1039/b704792d)
Laser spectroscopic study on the conformations and the hydrated structures of benzo-18-crown-6-ether
and dibenzo-18-crown-6-ether in supersonic jets
Takayuki Ebata et al., Phys. Chem. Chem. Phys., 2007, 9, 4452 (DOI: 10.1039/b704750a)
Conformational preferences of chiral molecules: free jet rotational spectrum of 1-phenyl-1-propanol
Walther Caminati et al., Phys. Chem. Chem. Phys., 2007, 9, 4460 (DOI: 10.1039/b705114j)
Electronic and infrared spectroscopy of jet-cooled (±)-cis-1-amino-indan-2-ol hydrates
Anne Zehnacker-Rentien et al., Phys. Chem. Chem. Phys., 2007, 9, 4465 (DOI: 10.1039/b705650h)
A peptide co-solvent under scrutiny: self-aggregation of 2,2,2-trifluoroethanol
Martin A. Suhm et al., Phys. Chem. Chem. Phys., 2007, 9, 4472 (DOI: 10.1039/b705498j)
Intramolecular recognition in a jet-cooled short peptide chain: -turn helicity probed by a neighbouring
residue
M. Mons et al., Phys. Chem. Chem. Phys., 2007, 9, 4491 (DOI: 10.1039/b704573e)
NMR studies of double proton transfer in hydrogen bonded cyclic N,N -diarylformamidine dimers:
conformational control, kinetic HH/HD/DD isotope effects and tunneling
Hans-Heinrich Limbach et al., Phys. Chem. Chem. Phys., 2007, 9, 4498 (DOI: 10.1039/b704384h)
Molecular recognition in 1 : 1 hydrogen-bonded complexes of oxirane and trans-2,3-dimethyloxirane with
ethanol: a rotational spectroscopic and ab initio study
Nicole Borho and Yunjie Xu, Phys. Chem. Chem. Phys., 2007, 9, 4514 (DOI: 10.1039/b705746f)
Conformational study of 2-phenylethylamine by molecular-beam Fourier transform microwave
spectroscopy
Jose L. Alonso et al., Phys. Chem. Chem. Phys., 2007, 9, 4521 (DOI: 10.1039/b705614a)
Raman jet spectroscopy of formic acid dimers: low frequency vibrational dynamics and beyond
P. Zielke and M. A. Suhm, Phys. Chem. Chem. Phys., 2007, 9, 4528 (DOI: 10.1039/b706094g)
Infrared spectroscopy of acetic acid and formic acid aerosols: pure and compound acid/ice particles
Ruth Signorell et al., Phys. Chem. Chem. Phys., 2007, 9, 4535 (DOI: 10.1039/b704600f)
Selectivity of guest–host interactions in self-assembled hydrogen-bonded nanostructures observed by
NMR
Hans Wolfgang Spiess et al., Phys. Chem. Chem. Phys., 2007, 9, 4545 (DOI: 10.1039/b704269h)
Molecular recognition in molecular tweezers systems: quantum-chemical calculation of NMR chemical
shifts
Christian Ochsenfeld et al., Phys. Chem. Chem. Phys., 2007, 9, 4552 (DOI: 10.1039/b706045a)
Molecular recognition in the gas phase. Dipole-bound complexes of benzonitrile with water, ammonia,
methanol, acetonitrile, and benzonitrile itself
David W. Pratt et al., Phys. Chem. Chem. Phys., 2007, 9, 4563 (DOI: 10.1039/b705679f)
Vibrational dynamics of carboxylic acid dimers in gas and dilute solution
Brooks H. Pate et al., Phys. Chem. Chem. Phys., 2007, 9, 4572 (DOI: 10.1039/b704900e)
IR-UV double resonance spectroscopy of xanthine
Mattanjah S. de Vries et al., Phys. Chem. Chem. Phys., 2007, 9, 4587 (DOI: 10.1039/b705042a)
Secondary structure binding motifs of the jet cooled tetrapeptide model Ac–Leu–Val–Tyr(Me)–NHMe
M. Gerhards et al., Phys. Chem. Chem. Phys., 2007, 9, 4592 (DOI: 10.1039/b706519a)
UV resonance Raman spectroscopic monitoring of supramolecular complex formation: peptide
recognition in aqueous solution
Carsten Schmuck et al., Phys. Chem. Chem. Phys., 2007, 9, 4598 (DOI: 10.1039/b709142g)
www.rsc.org/pccp

PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Carbohydrate molecular recognition: a spectroscopic investigation of
carbohydrate–aromatic interactions
E. Cristina Stanca-Kaposta,a David P. Gamblin,b James Screen,a Bo Liu,a
Lavina C. Snoek,a Benjamin G. Davisb and John P. Simons*a
Received 29th March 2007, Accepted 25th April 2007
First published as an Advance Article on the web 18th May 2007
DOI: 10.1039/b704792d
The physical basis of carbohydrate molecular recognition at aromatic protein binding sites is
explored by creating molecular complexes between a series of selected monosaccharides and
toluene (as a truncated model for phenylalanine). They are formed at low temperatures under
molecular beam conditions, and detected and characterized through mass-selected, infrared ion
depletion spectroscopy—a strategy which exploits the extraordinary sensitivity of their vibrational
signatures to the local hydrogen-bonded environment of their OH groups. The trial set of
carbohydrates, a- and b-anomers of glucose, galactose and fucose, reflects ligand fragments in
naturally occurring protein–carbohydrate complexes and also allows an investigation of the effect
of systematic structural changes, including the shape and extent of ‘apolar’ patches on the
pyranose ring, removal of the OH on the exocyclic hydroxymethyl group, and removal of the
aglycon. Bound complexes invariably form, establishing the general existence of intrinsic
intermolecular potential minima. In most of the cases explored, comparison between recorded and
computed vibrational spectra of the bound and free carbohydrates in the absence of solvent water
molecules reveal that dispersion forces involving CH–p interactions, which promote little if any
distortion of the bound carbohydrate, predominate although complexes bound through specific
OH–p hydrogen-bonded interactions have also been identified. Since the complexes form at low
temperatures in the absence of water, entropic contributions associated with the reorganization of
surrounding water molecules, the essence of the proposed ‘hydrophobic interaction’, cannot
contribute and other modes of binding drive the recognition of sugars by aromatic residues.
Excitingly, some of the proposed structures mirror those found in naturally occurring
protein–carbohydrate binding sites.
ring—the A and B faces (see Fig. 1). Some proteins also
1. Introduction
contain aromatic carbohydrate binding sites which engage
The molecular interactions involved in the recognition of
both faces of a pyranose ring, creating a sandwich or so-called
‘sugar tong’ motif5—notably glycosyl hydrolases responsible
carbohydrates by proteins mediate an extraordinarily wide
range of biological processes ranging from cell growth, adhe-
sion and death1–3 to the enzymatic recycling of photosynthe-
involvement of aromatic residues in carbohydrate recognition
for the degradation of plant cell wall carbohydrates. The
at protein binding sites is widespread.5 Sugar-aromatic recog-
tically generated plant cell-wall polysaccharides, such as
cellulose and xylan—the most abundant polymers on Earth.4,5
nition sites are therefore of primary importance in nature but
Non-covalent molecular interactions at protein–carbohydrate
the fundamentals of these interactions are still not well under-
binding sites may involve contributions from electrostatic
stood.
forces such as those generated by coordinated Ca2+ ions; or
The local structures at carbohydrate–protein binding sites
specific hydrogen bonding to polar side chains; or dispersion
have been revealed primarily through X-ray crystallographic
forces generated by van der Waals interactions.1–3,6–9 The
or NMR spectroscopic investigations but identification of
latter typically involve aromatic residues, notably tyrosine
their structures does not establish the detailed nature of the
and tryptophan, which often provide platforms for stacking
interactions which create them. Binding to aromatic residues
interactions with so-called ‘apolar patches’ on the pyranose
can include dispersion interactions (CH–p bonding), or hydro-
ring and provide recognition sites exploited by Nature.5,10 In
gen bonding (OH–p), or a finely balanced complement of
both;7 there are arguments which suggest they might also be
glucose for example, a circumference of hydrophilic, equato-
rially oriented hydroxyl groups separates two axial apolar
driven by entropic effects associated with the hydrophobic
exclusion of water.11 Quantitative assessments of the electro-
hydrophobic faces, located above and below the pyranose
nic (enthalpic) contributions have been approached through
a Department of Chemistry, Physical and Theoretical Chemistry
theoretical investigation, using force field, hybrid density
Laboratory, South Parks Road, Oxford, UK OX1 3QZ
b Department of Chemistry, Chemistry Research Laboratory,
Mansfield Road, Oxford, UK OX1 3TA
functional (DFT) and ab initio computation to determine
the preferred structures and binding energies of model
ꢀc
This journal is the Owner Societies 2007
4444 | Phys. Chem. Chem. Phys., 2007, 9, 4444–4451

Fig. 1 Carbohydrate structures revealing the apolar patches in carbohydrates exemplified on b-D-glucopyranoside and a-D-galactopyranoside
(top) and the monosaccharide derivatives chosen for this study (bottom).
bimolecular carbohydrate-aromatic complexes.7,9,12,13 These
calculations, generally based upon crystallographic protein–
carbohydrate stacking structures collected in the protein
structure database14 do reassuringly, indicate their association
with attractive carbohydrate-aromatic interactions (and hence
negative enthalpy changes) generated through dispersive
(CH–p) bonding and hydrogen (OH–p) bonding.12 This ap-
proach has recently been extended to characterize some intra-
molecular complexes formed within model carbohydrate-
aromatic molecular scaffolds and probed via NMR NOESY
spectroscopy in aqueous solution.15
The sequence of monosaccharides, selected to reflect naturally
occurring protein–carbohydrate complexes, was also chosen to
explore systematically, the effect of structural changes. These
include the shape and extent of the apolar patches on the
pyranose ring, affected by axial 2 equatorial configurational
changes at the C1 (a 2 b anomers) and C4 (MeGlc 2
MeGal, MeFuc) sites; removal of the OH group at the
exocyclic carbon site, C6 (MeGal 2 MeFuc); and removal
of the aglycon (MeFuc - Fuc).
2. Experimental
Very recently, the existence of such attractive interactions
has been established unequivocally through the experimental
creation of monosaccharide-aromatic molecular complexes at
low temperatures under molecular beam conditions, detected
and characterized through mass-selected vibrational spectro-
scopy.16 Comparisons between the recorded or computed
infrared (IR) spectra of the bound and free carbohydrates
revealed the contributions made by dispersive, CH–p and in
some cases, specific OH–p H-bonded interactions—in the
absence of solvent water molecules. They also reflected the
incidence, or absence of any significant conformational or
structural changes imposed on the bound carbohydrate. The
strategy exploits the extraordinary sensitivity of their vibra-
tional signatures to the local, intra- and intermolecular hydro-
gen-bonded environment of their OH groups.17 The presence
and strength, or absence of H-bonded interactions is signaled
by the degree of displacement, or not of their associated
infrared bands towards lower wave number.
2.1 Spectroscopy and complex formation
A detailed description of the molecular beam experimental
method has been published previously.18 The carbohydrate
samples were vaporized at temperatures in the range 70–
140 1C, dependent upon the sugar, from an oven mounted
directly in front of the nozzle of a pulsed valve (General Valve,
0.8 mm orifice) into an argon free jet expansion (stagnation
pressure, 3–6 bar). The aromatic complexes were formed and
stabilized in the high collision frequency region of the expand-
ing jet by seeding the argon with toluene vapour (1–3%). The
free jet expansion passed through a 3 mm skimmer to form a
collimated molecular beam which was subsequently crossed by
one, or two tunable laser beams in the extraction region of a
linear time of flight mass spectrometer (Jordan).
The frequency-doubled output of a pulsed Nd:YAG-
pumped dye laser (Spectron 810/LambdaPhysik FL2002,
0.5 mJ/pulse UV, operating at 10 Hz), excited resonant two
photon ionization (R2PI) spectra of each complex, which were
recorded in the parent ion mass channel, the toluene partner
providing the UV chromophore; the mass spectra also re-
vealed the formation of toluene polymer ions. Conformer
specific UV spectra were recorded using IR-UV hole-burning;
R2PI and IR-UV hole-burn spectra associated with the com-
plex a-MeFuc Á toluene are shown, for illustration in Fig. 2.
(Although these were well resolved this was not always the
case, primarily because of saturation). Corresponding IR
The present article continues the exploration of intermole-
cular bonding, selectivity and structural change in carbo-
hydrate-aromatic complexes, through comparisons between
the experimental IR spectra of a- and b-anomers of methyl
D-gluco- and galacto-pyranoside (MeGlc and MeGal); methyl
L-fuco-pyranoside (MeFuc); and L-fucose itself (Fuc), see
Fig. 1. The experimental spectra have also been compared
with those of the lowest lying conformers of the uncomplexed
carbohydrates. Toluene, representing a truncated phenyl-
alanine protein residue, was selected as the aromatic partner.
ꢀc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 4444–4451 | 4445

activated with methyl triflate in the presence of anhydrous
methanol and the proton scavenger 2,6-di-tert-butyl-4-methyl-
pyridine (DTBMP). The use of the acetyl protection ensured
the desired 1,2-trans stereochemistry and subsequent deacety-
lation gave the desired b-fucoside in excellent yield. The
remaining mono-saccharides were used as supplied by Fluka.
3. Results and discussion
3.1 MeGlc
The IRID spectra of the b- and a-MeGlc Á toluene complexes
are presented in Figs. 3 and 4, where they can be compared
Fig. 2 R2PI (a) and IR-UV hole burn (b, c) spectra of the complex
between a-MeFuc and toluene indicating the presence of two different
structures. The IR-UV hole burn spectra (b and c) were recorded with
the IR burn laser tuned to 3609.48 cmÀ1 and 3618 cmÀ1, respectively.
The dotted and dash-dotted lines indicate the bands in the R2PI
spectrum which correspond to one or other of the two contributing
complex structures. The dip in the spectrum at 37 477 cmÀ1, labeled by
an asterisk is due to UV absorption by the origin band of uncomplexed
toluene.
spectra were recorded through IR ion-dip (IRID) measure-
ments conducted in the range 3400–3800 cmÀ1 18
the ion dip
;
depletions were generally Z 50%, more than enough to
compensate for the low R2PI signals recorded in the parent
ion mass channels. The IR radiation (2–4 mJ/pulse, B10 ns
duration) was generated from the output of a Nd:YAG-
pumped IR dye laser (Continuum 8010/ND6000) by difference
frequency generation in a lithium niobate crystal. A delay
B150 ns between the IR pump and UV probe laser pulses
restricted measurements to the ground electronic state only.
2.2 Computation
Quantum chemical calculations of the structures of the lowest-
lying conformers of the uncomplexed methyl pyranosides were
performed (following a random MM conformational search20)
using the Gaussian 03 package.19 Initial structures were based
on those of the corresponding phenyl tagged carbohydrates
determined previously.17 Geometry optimizations and calcula-
tions of harmonic vibrational frequencies and intensities were
conducted using density functional theory (the B3LYP func-
tional) with a 6-31+G* basis set. Single point MP2 calcula-
tions with a larger basis set (6-311++G**) provided relative
energies. For comparison with observation, the OH stretch
harmonic frequencies were scaled by the factor 0.9734,
which provides excellent agreement with experimental data
for carbohydrates.18 Zero-point and free energy corrections to
the electronic energy were calculated using the harmonic
frequencies.
Fig. 3 (a) and (d): IRID spectra associated with two b-MeGlc Á
toluene complex structures. (b) and (e): IRID spectra and peak
assignments, based on previous calculations,21 of the two lowest lying
conformers of uncomplexed b-PhGlc. (c) and (f): predicted vibrational
frequencies, assignments and calculated structures for the two lowest-
lying conformers of uncomplexed b-MeGlc. Note the counter-clock-
wise orientation of the co-operative hydrogen bonds. Relative energies
in kJ molÀ1 are shown in brackets. Also shown: a suggested ‘three
point landing surface’ for b-MeGlc bound to toluene via a CH1,3,5–p
interaction together with the local structure at a tryptophan binding
site in the protein cellobiohydrolase I.22 The dashed lines indicate the
hydrophobic interactions between the sugar (CH1,3,5) and the aro-
matic residue.
2.3 Synthesis
Methyl a-L-fucopyranoside was obtained in near quantitative
fashion following a standard Fischer glycosylation using a
methanolic solution of HCl, generated from the addition of
acetyl chloride to methanol. Methyl b-L-fucopyranoside was
synthesised from a peracetylated thiofucoside which was
ꢀc
This journal is the Owner Societies 2007
4446 | Phys. Chem. Chem. Phys., 2007, 9, 4444–4451

3642 cmÀ1. The spectrum does not resemble either of the two
lowest-lying conformers of the uncomplexed anomer, each of
which presents an OH band located at B3600 cmÀ1 and a
cluster of bands lying between 3625 cmÀ1 and 3645 cmÀ1
,
associated with the H-bonded counter-clockwise orientation,
OH4 - OH3 - OH2 - O1, see the structures in Fig. 4. The
calculated band shifted to 3600 cmÀ1, corresponding to s2,
reflects the characteristic vibrational signature associated with
the OH2 (equatorial) - O1 (axial) bonding in the a-anom-
ers.18 Strikingly, this band does not appear in the experimental
spectrum of the complex.
The intense band at 3560 cmÀ1 indicates a specific, OH–p
hydrogen bond—but which of the OH groups is involved?
Two co-operatively bonded scenarios can be considered. In the
first, the counter-clockwise orientation of the peripheral OH
groups is retained to create the chain OH4 - OH3 - OH2 -
p. The strongly displaced band at 3560 cmÀ1 would then
correspond to the ‘missing’ OH2 band at 3600 cmÀ1 and the
bands at 3613 cmÀ1 and 3630 cmÀ1 could be identified with
OH3 and OH4, both displaced to lower wavenumber through
co-operativity. Alternatively, if the counter-clockwise orienta-
tion were reversed in the complex, as it is, for example, in the
mono-hydrated complex of b-PhGlc,23 to create the chain
sequence OH2 - OH3 - OH4 - p the strongly displaced
band would then correspond to OH4. The band associated
with OH2, released from its bonding to O1, would then shift
from B3600 cmÀ1 towards higher wavenumber, because of a
weaker acquired (equ - equ) interaction with OH3. Thus the
issue of identity remains open although a comparison between
the IR signatures of the complexes of a-MeGlc and a-MeGal,
see below, provides evidence favouring OH4. It should be
possible to resolve the issue unequivocally through 18O label-
ing, either of OH2 or OH4. In either case, the most strongly
shifted vibration would correspond to the final OH - p stage
of a co-operatively linked H-bonded chain. Indeed there are
strong analogies between the spectral shifts associated with
Fig. 4 (a) IRID spectrum of a-MeGlc Á toluene and (b), (c) predicted
IR spectra and assignments of the two lowest lying conformers of
uncomplexed a-MeGlc. Corresponding structures indicating the
counter-clockwise orientation of the co-operative hydrogen bonds
similar to b-MeGlc are shown as well. Relative energies in kJ molÀ1
are indicated in brackets.
with the computed spectra of the two lowest-lying, near iso-
energetic conformers of the free mono-saccharides, a- and
b-MeGlc (ccG-g+) and (ccG + gÀ)w and also, those observed
experimentally for the corresponding phenyl pyranoside,
b-PhGlc. The spectra display some striking differences and
some striking similarities. Spectral hole-burning measure-
ments for the b-anomeric complex identify two distinct struc-
tures but their associated IRID spectra are virtually identical
to those computed for the two lowest-lying conformers of the
uncomplexed carbohydrate, b-MeGlc and also those observed
experimentally for b-PhGlc.21 The absence of any significant
spectral shifts excludes any contribution from OH–p hydrogen
bonding and the b-MeGlc Á toluene complexes must be bound
by dispersive, CH–p-interactions. Furthermore, the close si-
milarity between their IR spectra and those of the two lowest
lying conformers of the free carbohydrate implies an absence
of any significant distortion of their structure. This would be
consistent with weak binding either to the A face via the axially
oriented H2 and H4 atoms and possibly an H6 atom on the
exocyclic CH2OH group, or alternatively to the B face via the
axially oriented H1, H3 or H5 atoms of the pyranose ring
depicted by the cartoon in Fig. 3—and displayed for example,
in the 3D structure of b-MeGlc at a tryptophan binding site in
its complex with cellobiohydrolase from Trichoderma reesei,22
also shown in Fig. 3.
OH
-
p
bonding that have been observed in
(CH3OH)n=1À3 Á benzene clusters24 and those observed here.
As the length of the co-operatively bonded linear (CH3OH)n
chain increased, the wavenumber of the IR band associated
with the terminal p-bonded OH decreased from 3681 cmÀ1
(n = 0) to 3639 cmÀ1 (n = 1) to 3605 cmÀ1 (n = 2) to 3589
cmÀ1 (n = 3).24 In the current carbohydrate complex, albeit
with toluene rather than benzene, an even greater shift could
be anticipated consistent with the appearance of a band at
3560 cmÀ1
.
One question remains: why is the behaviour of the a- and
b-anomers so different? The switch in the orientation of the
anomeric OCH3 group from equatorial (b) to axial (a) evi-
dently alters the character of the most favoured molecular
interaction in the isolated molecular complex from dispersive
to H-bonded. In the a-anomer, the axial anomeric OCH3
group, directed towards the B-face, and the equatorial orien-
tation of CH1 would both reduce the ‘attractiveness’ of
binding through multiple CH - p interactions on the B-face
by simply reducing the number of appropriately oriented C–H
groups. Perhaps it is this that tips the balance in favour of the
alternative interaction provided by co-operatively enhanced,
OH–p bonding.
The IR signature of the a-MeGlc Á toluene complex is com-
pletely different from that of the b-complex. Only one struc-
ture is seen to be populated and its IRID spectrum is far more
dispersed, displaying an intense and strongly displaced band at
3560 cmÀ1 and three bands located between 3613 cmÀ1 and
w The designations, ccG-g+, . . . , indicate the conformation of the
carbohydrate: briefly, ‘cc’ indicates a counter-clockwise orientation of
the peripheral OH groups, OH4 - OH3 - OH2 - O1 and G-g+ . . .
indicates the gauche orientation of the exocyclic hydroxymethyl group
and its terminal OH6 group, respectively.17
ꢀc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 4444–4451 | 4447

3.2 MeGal and MeFuc
neighbours). The a-MeFuc complex displays two distinct,
though very similar structural isomers, identified through
hole-burning experiments, but only one is observed for
a-MeGal. In every case however, the bound, and the corre-
sponding free carbohydrate spectra are almost identical and
the binding must be associated with ‘structurally benign’
interactions that conserve the conformation of the carbo-
hydrate, namely CH - p interactions—in sharp contrast to
the a-MeGlc Á toluene complex where the binding is provided
by an OH - p interaction. Switching the orientation at C4
from equatorial to axial creates a more ‘focused’, denser
apolar patch on the B face of the carbohydrate, associated
with the three neighbouring C–H bonds, CH3, CH4 and CH5,
which are located away from the axially oriented OCH3 at C1.
It also has the effect of amplifying binding in the OH chain
rather more strongly due to the presence of an enhanced OH4
(ax) - OH3 (eq) bond, further supported in the a-anomer, by
the OH2 (eq) - O1 (ax) bond. These effects would each
favour binding at the B face mediated by CH3,4,5 - p
interactions and decrease the likelihood of a specific OH -
p interaction. Excitingly, the structures displayed in the pro-
tein-bound complexes of a-MeFuc with Lectin 1 (from Ulex
europaeus),26 and a-MeGal (with the Winged Bean Acidic
Lectin25 shown in Fig. 4), show a similar contact with this
extended B face. In each case, the counterclockwise orienta-
tion OH4 - OH3 - OH2 - O1 is preserved but the
hydroxymethyl group in a-MeGal adopts the Tg+ conforma-
tion, associated in the free carbohydrate with its second lowest
energy conformer.
The switch from MeGlc to MeGal changes the configuration
at C4 and the OH4 group is now directed axially, enhancing
the interaction OH4 (axial) - OH3 (equ). MeFuc (6-deoxy-L-
galactose) similarly contains an axial OH4. Like MeGlc, the
lowest energy conformers of isolated MeGal and MeFuc (and
also the corresponding phenyl pyranosides17) present counter-
clockwise oriented OH chains, OH4 - OH3 - OH2, further
promoted in the a-anomers, by a strong OH2 - O1 inter-
action.
The IRID spectra of the a-MeFuc Á toluene and a-MeGa-
l Á toluene complexes are shown in Fig. 5 together with the
computed IR spectra of the isolated carbohydrates in their
lowest energy conformations (calculated to lie at energies
B7 kJ molÀ1 and B4 kJ molÀ1 below those of their nearest
Figs. 6 and 7 display the experimental IRID spectra of the
b-MeFuc Á toluene and b-MeGal Á toluene complexes together
with the computed and observed spectra of the lowest lying
conformers of the uncomplexed methyl (and phenyl) pyrano-
sides. Binding to toluene slightly perturbs the IR spectrum of
b-MeFuc (for which only a single complex structure was
observed) and the more weakly populated of the b-MeGal
complexes, Fig. 7(b), where a second more strongly perturbed
structure was also detected; unfortunately, because of UV
spectral overlap it was not possible to separate its IRID
spectrum fully from the first.
Consider first, the clearer experimental data for b-Me-
Fuc Á toluene, Fig. 6(a). If the band at highest wavenumber,
3641 cmÀ1 (the wavenumber associated with s2 in the
b-anomer) is associated with the OH2 group, the two bands
at 3595 cmÀ1 and 3620 cmÀ1 are most likely to correspond
to OH4 and OH3, each displaced to lower wavenumber by
B15 cmÀ1 with respect to the corresponding bands in the
uncomplexed anomer. While this could reflect a weak hydro-
gen-bonded interaction involving OH4 or OH3, cf. the
a-MeGlc Á toluene complex, where the perturbed band is much
more strongly displaced, it might instead reflect some distor-
tion of the conformation of the bound pyranose ring. In the
b-anomer of MeFuc (and MeGal) the re-orientation at C1
extends the apolar patch on the B-face still further to include CH1
as well CH3, 4 and 5, and it also directs the OCH3 group away
from the B-face. In this situation binding to the B-face mediated
by CH–p interactions would seem the more favoured option.
The behaviour of b-MeGal is more complex. The IRID
spectrum of its more weakly populated complex (Fig. 7(b)) is
Fig. 5 IRID spectra of (a) a-MeFuc Á toluene and (c) a-MeGal Á to-
luene complexes. (b) and (d) show the predicted stick spectra for the
energetically lowest lying conformers of uncomplexed a-MeFuc and
a-MeGal together with the frequency assignments and their corre-
sponding structures. Also shown: a suggested ‘three point landing
surface’ for a-MeGal bound to toluene via a CH3,4,5–p interaction
together with the local structure at a phenylalanine-containing carbo-
hydrate binding site in Winged Bean Acidic lectin.25 The dashed lines
indicate the hydrophobic interactions between the sugar (CH3,4,5,6)
and the aromatic residue.
ꢀc
This journal is the Owner Societies 2007
4448 | Phys. Chem. Chem. Phys., 2007, 9, 4444–4451

Fig. 6 IRID spectra of (a) the b-MeFuc Á toluene complex, (b)
uncomplexed b-PhFuc and (c) the calculated stick spectrum of
b-MeFuc. Peak assignments are indicated for the experimental and
predicted bands of the free molecules. The calculated structure of
b-MeFuc shows a counter-clockwise orientation of the co-operative
H-bonds. The dotted lines indicate the position of the peaks in the
complex relative to the bare molecules.
Fig. 7 (a) and (b) IRID spectra associated with two b-MeGal Á
toluene structures. (c) and (d) experimental IRID spectra and assign-
ments based on previous calculation17 for the energetically two lowest
lying conformers of b-PhGal. (e) and (f)-predicted stick spectra for the
lowest conformers of the uncomplexed b-MeGal. The dotted lines
indicate the band positions in the more strongly populated complex.
Calculated structures for the uncomplexed b-MeGal are shown in the
bottom-left corner. Relative energies in kJ molÀ1 are indicated in the
brackets.
very similar to those of the two lowest lying conformers of the
free carbohydrates, b-MeGal and b-PhGal (Fig. 7(c,d)) and
there is no evidence of any contribution from OH–p bonding.
On the other hand, the IRID spectrum associated with its
more strongly populated complex, Fig. 7(a), is quite strongly
perturbed. Three of its four bands are displaced by some
30 cmÀ1 towards lower wavenumber and a contribution from
OH–p interaction(s) can no longer be discounted reflecting
perhaps, an alternative orientation of the aromatic and carbo-
hydrate partners.
in the b-anomer, but not the a-anomer, the spectra do indeed
reveal small shifts towards lower wavenumber.
3.3 Fuc
In sharp contrast to the complexes of MeFuc, the IRID
spectrum of the Fuc Á toluene complex displays by far the
clearest signal of binding through a specific OH - p bonded
interaction, with one of its four bands displaced to 3530 cmÀ1
,
see Fig. 8. The obvious candidate for this is OH1, since the
interaction OH1 - p would be unavailable (blocked by a
methyl) in MeFuc. If the OH - p interaction lies at the
terminus of the extended co-operative, counter-clockwise se-
quence, OH4 - OH3 - OH2 - OH1 - p this assignment
would also be consistent with the large shift, B110–120 cmÀ1
below the corresponding band in the (computed) IR spectra of
the a- or b-anomers of uncomplexed fucose. In b-MeGlc the
sequence length corresponds to n = 3; in the analogous
Fuc Á toluene complex it would be n = 4. Co-operativity
should promote corresponding but much smaller shifts in
the remaining bands in the experimental spectrum. When
comparisons are made with the remaining (computed) bands
Fig. 8 IRID spectrum of L-Fuc Á toluene (a) and predicted vibrational
spectra, assignments and structures of uncomplexed b-L-Fuc (b) and
a-L-Fuc (c).
ꢀc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 4444–4451 | 4449

spectra of the resulting cluster(s). CH–p interactions can be
explored directly, by recording the vibrational spectra of
bound and free carbohydrates in the C–H region; deuteration
will shift the vibrational spectrum of the aromatic partner ‘out
of sight’.30 Assignment of specific OH–p interactions can be
facilitated by recording the IRID spectra of 18O labeled
carbohydrates. Moving beyond the simplest model systems,
the role of the phenolic OH group in tyrosine, which is more
common in protein binding sites than phenylalanine, can be
explored by substituting p-cresol for toluene. Does the phe-
nolic OH provide a hydrogen-bonding ‘anchor’ to help bind
the carbohydrate to the aromatic platform? 3-methyl indole
provides a truncated model for tryptophan. The influence of
back-bone interactions may be explored by substituting a
suitably protected phenylalanine or tyrosine for toluene. Chir-
al discrimination can be explored by comparing the complexes
of their D and L enantiomers with D and L sugars. The
carbohydrate partners need not be restricted to monosacchar-
ides. There is much to be done and much to learn.
4. Concluding remarks
The experimental programme set out to establish the general-
ity of intermolecular binding in carbohydrate-aromatic com-
plexes; to identify and characterize some of the interactions
which promote it, particularly the operation of hydrophobic
dispersion interactions vs. specific hydrogen bonded inter-
actions; and to explore their sensitivity to structural change.
The first aim has been achieved—at least for the ‘basis set’ of
carbohydrates explored to date. Each one is seen to bind to
toluene at low temperatures—there are without doubt, poten-
tial energy minima which support bound complexes. Since the
carbohydrate-aromatic complexes were formed at low tem-
peratures in the absence of water, entropic contributions
associated with the reorganization of surrounding water
molecules, the essence of the proposed hydrophobic interac-
tion11 cannot contribute—which is not to say that they might
not do so in aqueous media at more elevated temperatures.
In most of the cases explored, dispersion forces involving
CH–p interactions, which promote little if any distortion of
the bound carbohydrate, predominate though complexes
bound through specific OH–p H-bonded interactions have
also been identified.
Acknowledgements
We have appreciated helpful discussions and contributions
made by Drs N. A. Macleod, T. van Mourik and Dr J. Frey.
Financial support has been provided by the EPSRC, the
Leverhulme Trust (Grant F/08788G) and the Royal Society
(LCS, University Research Fellowship). Support from the
CLRC Laser Support Facility and Corpus Christi College
(LCS) is also acknowledged.
Qualitative attempts to correlate the balance of their relative
contributions with the conformational structures of the bound
carbohydrates, or the choice of possible orientations with
respect to the aromatic platform, have only been partially
successful. The factors which may favour localized H-bonding
over the more broadly based dispersion interaction and which
determine the global minimum energy structures of the iso-
lated complexes are delicately balanced. In a ‘real life’
situation, where a carbohydrate is bound at an aromatic
protein molecular recognition site, other interactions with
the protein backbone or neighbouring residues (determined
by the environment and protein architecture) will also be
present. Given the delicacy of the local p-electron interactions
it should not be difficult for other neighbouring interactions to
modulate the relative carbohydrate-aromatic dispositions and
orientations at individual binding sites. Indeed a survey of the
deposited crystallographic structures of carbohydrate (e.g. Gal
or Glc)– protein complexes listed in the Brookhaven PDB14
reveals a broad spread of sugar-aromatic interactions.13 These
suggest, at least in part, that the primary role of the aromatic
platform is the provision of an attractive but malleable con-
tribution to the net intermolecular potential energy that can be
modulated by the local protein environment to provide the
remarkable selectivity displayed by carbohydrate recognition
proteins. Attempts to calculate the interactions accurately
through ab initio computation7,12,19,27 will need to be con-
ducted at high levels of theory if they are to provide quanti-
tative agreement with experimental spectroscopic or structural
data and DFT methods will certainly fail without the inclusion
of dispersion interactions.28,29
References
1 W. I. Weis and K. Drickamer, Annu. Rev. Biochem., 1996, 65, 441.
2 R. A. Dwek, Chem. Rev., 1996, 96, 683.
3 A. Varki, Glycobiology, 1993, 3, 97.
4 A. B. Boraston, D. Nurizzo, V. Notenboom, V. Ducros, D. R.
Rose, D. G. Kilburn and G. J. Davies, J. Mol. Biol., 2002, 319,
1143.
5 A. B. Boraston, D. N. Bolam, H. J. Gilbert and G. J. Davies,
Biochem. J., 2004, 382, 769.
6 H. Lis and N. Sharon, Chem. Rev., 1998, 98, 637.
7 J. Jimenez-Barbero, J. L. Asensio, F. J. Canada and A. Poveda,
´
Curr. Opin. Struct. Biol., 1999, 9, 549.
8 R. Palma, M. E. Himmel and J. W. Brady, J. Phys. Chem. B, 2000,
104, 7228.
´ ´
9 M. C. Fernandez-Alonso, F. J. Canada, J. Jimenez-Barbero and G.
Cuevas, J. Am. Chem. Soc., 2005, 127, 7379.
10 X. Robert, R. Haser, T. E. Gottschalk, F. Ratajczak, H. Driguez,
B. Svensson and N. Aghajari, Structure, 2003, 11, 973.
11 R. U. Lemieux, Acc. Chem. Res., 1996, 29, 373.
12 V. Spiwok, P. Lipopova
lek, J. Hasek and B. Kra
19, 887.
´
, T. Skalova, E. Vondrackova, J. Dohna-
´ ´ ´ ´ ´
´
lova, J. Comput.-Aided Mol. Des., 2006,
´
13 M. S. Sujatha, Y. U. Sasidhar and P. V. Balaji, Biochemistry, 2005,
44, 8554.
14 The protein data bank: url: www.rcsb.org.
15 G. Terraneo, D. Potenza, A. Canales, J. Jimenez-Barbero, K. K.
´
There are many other experimental approaches that may
now be explored. The nature of hydrophobic interactions—or
water-mediated binding—can be investigated by creating
hydrated aromatic complexes in a molecular beam, passing
them through a plume of carbohydrate vapour generated from
an oven or by laser ablation, and recording the vibrational
Baldridge and A. Bernardi, J. Am. Chem. Soc., 2007, 129, 2890.
16 J. Screen, E. C. Stanca-Kaposta, D. P. Gamblin, N. A. Macleod,
L. C. Snoek, B. G. Davis and J. P. Simons, Angew. Chem., Int. Ed.,
2007, 46, DOI: 10.1002/anie.200605116.
17 J. P. Simons, R. A. Jockusch, P. C¸ arc¸ abal, I. Hunig, R. T.
¨
Kroemer, N. A. Macleod and L. C. Snoek, Int. Rev. Phys. Chem.,
2005, 24, 489.
ꢀc
This journal is the Owner Societies 2007
4450 | Phys. Chem. Chem. Phys., 2007, 9, 4444–4451

18 P. C¸
arc¸
abal, T. Patsias, I. Hunig, B. Liu, C. Kaposta, L. C. Snoek,
¨
21 F. O. Talbot and J. P. Simons, Phys. Chem. Chem. Phys., 2002, 4,
3562.
22 C. Divne, J. Stahlberg, T. T. Teeri and T. A. Jones, J. Mol. Biol.,
1998, 275, 309: pdb 7CEL.
23 P. C¸ arc¸ abal, R. A. Jockusch, I. Hunig, L. C. Snoek, R. T.
¨
Kroemer, B. G. Davis, D. P. Gamblin, I. Compagnon, J. Oomens
and J. P. Simons, J. Am. Chem. Soc., 2005, 127, 11414.
24 R. N. Pribble, F. C. Hagemeister and T. S. Zwier, J. Chem. Phys.,
1997, 106, 2145.
25 N. Manoj, V. R. Srinivas, A. Surolia, M. Vijayan and K. Suguna,
J. Mol. Biol., 2000, 302, 1129: pdb 1F9K.
D. P. Gamblin, B. G. Davis and J. P. Simons, Phys. Chem. Chem.
Phys., 2006, 8, 129–136.
19 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, J. Montgomery, J. A. T. Vreven, K. N.
Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V.
Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A.
Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,
M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, C. Adamo,
J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma,
G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S.
Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A.
D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A.
G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A.
Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T.
Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challa-
combe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C.
Gonzalez and J. A. Pople, Gaussian, Inc., Wallingford CT, 2004.
20 G. Chang, W. C. Guida and W. C. Still, J. Am. Chem. Soc., 1989,
111, 4379.
26 G. F. Audette, D. J. H. Olson, A. R. Ross, J. W. Quail and L. T.
Delbaere, Can. J. Chem., 2002, 80, 1010: pdb 1JXN.
27 M. I. Cha
Groves, J. L. Asensio, M. Muraki, F. J. Canada and J. Jime
Barbero, Chem.–Eur. J., 2005, 11, 7060.
´
vez, C. Andreu, P. Vidal, N. Aboitiz, F. Freire, P.
´
nez-
28 M. Elstner, P. Hobza, T. Frauenheim, S. Suhai and E. Kaxiras,
J. Chem. Phys., 2001, 114, 5149.
29 T. van Mourik, P. G. Karamertzanis and S. L. Price, J. Phys.
Chem. A, 2006, 110, 8.
30 S. Morita, A. Fujii, N. Mikami and S. Tsuzuki, J. Phys. Chem. A,
2006, 110, 10583.
ꢀc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 4444–4451 | 4451