Conformational choice and selectivity in singly and multiply hydrated monosaccharides in the gas phase

Article abstract of DOI:10.1002/chem.200800474

Factors governing hydration, regioselectivity and conformational choice in hydrated carbohydrates have been examined by determining and reviewing the structures of a systematically varied set of singly and multiply hydrated monosaccharide complexes in the gas phase. This has been achieved through a combination of experiments, including infrared ion-depletion spectroscopy conducted in a supersonic jet expansion, and computation through molecular mechanics, density functional theory (DFT) and ab initio calculations. New spectroscopic and/or computational results obtained for the singly hydrated complexes of phenyl β-D-mannopyranoside (β-D-PhMan), methyl α-D-gluco- and α-D-galactopyranoside (α-D-MeGlc and α-D-MeGal), when coupled with those reported earlier for the singly hydrated complexes of α-D-PhMan, β-D-PhGlc and β-D-PhGal, have created a comprehensive data set, which reveals a systematic pattern of conformational preference and binding site selectivity, driven by the provision of optimal, co-operative hydrogen-bonded networks in the hydrated sugars. Their control of conformational choice and structure has been further revealed through spectroscopic and/or computational investigations of a series of multiply hydrated complexes: they include β-D-PhMan-(H₂O)_2,3- which has an exocyclic hydroxymethyl group, and the doubly hydrated complex of phenyl α-L-fucopyranoside, α-L-PhFuc-(H₂O)₂, which does not. Despite the very large number of potential structurediggers and binding sites, the choice is highly selective with binding invariably focussed around the hydroxymethyl group (when present). In β-D-PhMan-(H₂O)_2,3, the bound water molecules are located exclusively on its polar face and their orientation is dictated by the (perturbed) conformation of the carbohydrate to which they are attached. The possible operation of similar rules governing the structures of hydrogen-bonded protein-carbohydrate complexes is proposed.

Full text of DOI:10.1002/chem.200800474

FULL PAPER
DOI: 10.1002/chem.200800474
Conformational Choice and Selectivity in Singly and Multiply Hydrated
Monosaccharides in the Gas Phase
Emilio J. Cocinero,[a] E. Cristina Stanca-Kaposta,[a] Eoin M. Scanlan,[b]
David P. Gamblin,[b] Benjamin G. Davis,*[b] and John P. Simons*[a]
Abstract: Factors governing hydration,
regioselectivity and conformational
choice in hydrated carbohydrates have
been examined by determining and re-
viewing the structures of a systemati-
cally varied set of singly and multiply
hydrated monosaccharide complexes in
the gas phase. This has been achieved
through a combination of experiments,
including infrared ion-depletion spec-
troscopy conducted in a supersonic jet
expansion, and computation through
molecular mechanics, density function-
al theory (DFT) and ab initio calcula-
tions. New spectroscopic and/or com-
putational results obtained for the
singly hydrated complexes of phenyl b-
when coupled with those reported ear-
lier for the singly hydrated complexes
of a-d-PhMan, b-d-PhGlc and b-d-
PhGal, have created a comprehensive
data set, which reveals a systematic
pattern of conformational preference
and binding site selectivity, driven by
the provision of optimal, co-operative
hydrogen-bonded networks in the hy-
drated sugars. Their control of confor-
mational choice and structure has been
further revealed through spectroscopic
and/or computational investigations of
a series of multiply hydrated com-
plexes; they include b-d-PhMan·
(H2O)2,3, which has an exocyclic hy-
droxymethyl group, and the doubly hy-
drated complexof phenyl a-l-fucopyr-
anoside, a-l-PhFuc·ACHTRE(UGN H2O)2, which does
not. Despite the very large number of
potential structures and binding sites,
the choice is highly selective with bind-
ing invariably “focussed” around the
hydroxymethyl group (when present).
In b-d-PhMan·CATHREU(NG H2O)2,3, the bound
water molecules are located exclusively
on its polar face and their orientation
is dictated by the (perturbed) confor-
mation of the carbohydrate to which
they are attached. The possible opera-
tion of similar rules governing the
structures of hydrogen-bonded pro-
tein–carbohydrate complexes is pro-
posed.
Keywords: carbohydrates · hydro-
gen bonds · IR spectroscopy · mo-
lecular recognition · solvent effects
d-mannopyranoside
(b-d-PhMan),
methyl a-d-gluco- and a-d-galactopyra-
noside (a-d-MeGlc and a-d-MeGal),
Introduction
ing selective carbohydrate–protein interactions are not
easily discerned. Selectivity must depend in part upon the
conformational structures of the interacting carbohydrate
ligand and its receptor. In addition, because of their flexibil-
ity and the universality of hydrogen-bonded interactions in
aqueous environments, their structures may be influenced
by explicit hydration, dehydration or by the direct involve-
ment of water at the receptor site.[2] Modelling and under-
standing protein–carbohydrate interactions presents a com-
Carbohydrate molecular recognition by proteins is a key
factor in many biological processes,[1] but the factors dictat-
[a] Dr. E. J. Cocinero, Dr. E. C. Stanca-Kaposta, Prof. J. P. Simons
Chemistry Department, University of Oxford
Physical and Theoretical Chemistry Laboratory
South Parks Road, Oxford OX1 3QZ (UK)
Fax: ( +44)1865-275410
[3]
plexchallenge.
The great majority of monosaccharide carbohydrate struc-
tures have been determined in the condensed phase, either
in crystalline environments using X-ray crystallography[4–6]
or in solution through molecular dynamics (MD) calcula-
tions and NMR spectroscopy measurements.[7–9] Their inher-
ent, unperturbed conformational structures, free of any envi-
ronmental effects were until recently only accessible in
silico, computed through molecular mechanics methods
E-mail: John.Simons@chem.ox.ac.uk
[b] Dr. E. M. Scanlan, Dr. D. P. Gamblin, Prof. B. G. Davis
Chemistry Department, University of Oxford
Chemistry Research Laboratory
12 Mansfield Road, Oxford OX1 3TA (UK)
Fax: ( +44)1865-275674
E-mail: Ben.Davis@chem.ox.ac.uk
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200800474.
Chem. Eur. J. 2008, 14, 8947 – 8955
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8947

using a variety of force fields, through density functional
theory (DFT) and ab initio calculations, or through Car–Par-
rinello MD calculations.[8–10] But a start has now been made
in determining experimentally the inherent conformations
of bare carbohydrates and the preferred conformations and
structures of their singly hydrated complexes isolated in the
gas phase.[11–14] The vibrational spectroscopic signatures of
individually resolved conformers, initially isolated from the
environment and then interacting with it in a controlled way
within singly hydrated (and also molecular) complexes, have
been obtained through mass-selected ultraviolet and infra-
red laser spectroscopy conducted at low temperatures in
free jet expansions,[15] coupled with a combination of molec-
ular mechanics, DFT and ab initio calculations, to enable
conversion of their spectroscopic signatures into conforma-
tional and intermolecular structures.
In the gas phase, the preferred carbohydrate conforma-
tions, explicit water binding site(s) and their resulting inter-
molecular structures are dominated primarily by co-opera-
tive hydrogen bonding. Controlling factors that have been
identified include the flexibility of their exocyclic hydroxy-
methyl group (present, for example, in glucose, galactose
and mannose; Scheme 1), the relative orientations (axial
series of singly hydrated monosaccharides (see Scheme 1),
including the b-anomer of phenyl1 d-glucopyranoside (b-d-
PhGlc, in which all the peripheral hydroxyl groups are ori-
ented equatorially) and the b-anomer of phenyl d-galacto-
pyranoside[16] (b-d-PhGal, in which OH4 is oriented axially);
the a-anomer of phenyl d-mannopyranoside (a-d-PhMan, in
which OH2 is oriented axially); and the a- and b-anomers
of their truncated analogues, phenyl d-xylopyranoside[17] and
l-fucopyranoside[13] (in which the exocyclic hydroxymethyl
is altered to methyl, a/b-l-PhFuc, or to no substituent at all,
a/b-d-PhXyl).
The present work sets out, firstly, to test the predictive
power of these working rules by exploring experimentally
and also computationally their operation in governing the
preferred conformations and structures of a comprehensive
panel of systematically varied monosaccharides, including
the singly hydrated anomers b-d-PhMan, a-d-MeGal and a-
d-MeGlc together with a-d-PhMan, b-d-PhGlc and b-d-
PhGal; and secondly, to identify the working rules that may
govern multiply hydrated carbohydrates by exploring the
structural preferences of the two “trial” systems, b-d-
PhMan·
the exocyclic CH2OH group. Collectively this panel of
monosaccharides (Scheme 1)
ACHTRE(NUG H2O)2,3 and a-l-PhFuc·ACHTRE(UGN H2O)2, with and without
AHCTREUNG
provides a powerful, structural-
ly iterated series that can be
used to map out key aspects of
the hydration of these impor-
tant biomolecules.
Results and Discussion
Synthesis: Phenyl b-d-manno-
pyranoside (1), a key additional
monosaccharide, was prepared
in three steps starting from
commercially available d-man-
nose using a highly b-selective
Mitsunobu[18,19]
glycosylation
strategy (Scheme 2). Selective
protection of the 2, 3, 4 and 6
hydroxyl groups was achieved
through reaction of d-mannose
with two equivalents of 2-me-
Scheme 1. Schematic representation of the structural inter-relations in the conformational panel used to probe
the hydration of monosaccharide carbohydrates. R=phenyl, methyl.
versus equatorial) of their hydroxyl groups, and their
anomeric configuration.[13,16,17] These factors can operate
separately or collectively, adapting the carbohydrate confor-
mation to optimize the sequence of intra- and inter-molecu-
lar hydrogen-bonded interactions in the hydrated com-
plex.[11,13,16,17] The first investigations suggested a set of
“working rules” based upon the preferred structures of a
thoxypropene in the presence of para-toluenesulfonic acid
(p-TSA). The resulting kinetic diisopropylidene manno-
side[20] was then reacted with phenol using triphenylphos-
phine and diisopropyl azodicarboxylate (DIAD) to furnish
the O-phenyl compound as a mixture of a and b isomers
(7:93). The high b selectivity is notable given the difficulties
associated with the synthesis of b-mannosides and outstrip-
ped that associated with several other methods.[21] This first
example of Mitsunobu mannosylation may therefore have
general application and its wider synthetic utility is currently
being explored. Pure b-mannoside was readily obtained in
71% isolated yield following separation from the a-anomer
1
The phenyl group provides the ultraviolet chromophore required for
detection by means of the resonant two-photon ionization step in the
infrared ion-depletion (IRID) detection technique. DFT and ab initio
calculations have confirmed that its presence has a negligible influence
on the conformational landscape of small carbohydrates.[11]
8948
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8947 – 8955

Conformational Choice in Hydrated Carbohydrates
FULL PAPER
Scheme 2. Synthesis of phenyl b-d-mannopyranoside.
by flash chromatography. Hydrolytic deprotection in acetic
acid/water furnished the desired O-phenyl b-mannoside in
96% yield.
UV and IR spectroscopy of b-d-PhMan: Figure 1 shows the
resonant two-photon ionization (R2PI) spectrum of b-d-
PhMan recorded in the argon expansion, non-solvated and
Figure 2. IRID spectra associated with the low-energy conformers of a-
and b-d-PhMan (with relative energies (MP2/6-311++G**/ZPE correct-
ed) in kJmolꢀ1 shown in square brackets). a) b-d-PhMan; experimental
and calculated (stick spectra) vibrational spectra for the two lowest
energy conformers, ccG+gꢀ [0.0] and ccGꢀg+ [1.2]; b) experimental vi-
brational spectra of the three lowest lying conformers of a-d-PhMan
(cGꢀg+) [0.0], (ccGꢀg+) [3.2] and (cTt) [3.7].[16] The dotted grey lines
ꢀ
indicate the positions of the O H vibrational bands, s3, s4 and s6 in a-
d-PhMan (ccGꢀg+).
Figure 1. R2PI (upper trace) and UV–UV hole-burn spectra (lower
*
traces) of b-d-PhMan; the symbols & and
identify the vibronic bands
selected for each hole-burn spectrum.
through DFT and ab initio calculations as the two lowest
energy conformers, ccG+gꢀ (Erel =0 kJmolꢀ1) and ccGꢀg+
(Erel =1.2 kJmolꢀ1), respectively; they differ only in the ori-
entation of the hydroxymethyl group. (See the Experimental
Section, last paragraph, for a definition of the conformation-
al nomenclature.) Each one presents a counter-clockwise
(“cc”) hydrogen-bonded chain OH4!OH3!OH2!O1,
encouraged by the relatively strong OH3(eq)!OH2(ax) in-
teraction, which shifts the vibrational mode s3 towards a
lower wavenumber, around 3600 cmꢀ1. The two IR spectra
of b-d-PhMan closely resemble the spectrum of the counter-
clockwise ccGꢀg+ conformer of the a-anomer, in which s3
free of any interaction with the environment, together with
its associated UV–UV hole-burn spectra, a double-reso-
nance pump–probe technique that selectively depletes each
selected conformer by transferring population out of its
ground electronic state.[11–14] The hole-burn spectra indicate
contributions from two distinct components with relative
populations of around 4:1. Their corresponding IR–UV in-
frared ion-depletion (IRID) spectra, shown in Figure 2, es-
tablish their association with two alternative conformations,
identified through comparison with the spectra computed
Chem. Eur. J. 2008, 14, 8947 – 8955
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8949

B. G. Davis, J. P. Simons et al.
appears at ꢁ3595 cmꢀ1, but they clearly differ from those of
the lowest or highest energy conformers of the a-anomer,
both of which adopt clockwise (“c”) conformations.
The monohydrate: The experimental vibrational spectrum
associated with the most strongly populated monohydrate
(shown in Figure 4b) is in excellent accord with the comput-
ed spectrum of its global minimum energy conformational
The structural origin of this difference, as might be ex-
pected, is the epimeric C1 centre. In the a-anomer its axial
orientation increases the OH2···O1 separation from around
2.3 to around 3.9 Š, and the most strongly shifted vibration-
al mode in its lowest energy conformation (cGꢀg+) be-
comes s2ꢁ3618 cmꢀ1, reflecting a strong (reversed) hydro-
gen-bonding interaction, OH2(ax)!OH3(eq). This, together
with the re-oriented hydroxymethyl group, facilitates the co-
operative “clockwise” sequence OH2!OH3!OH4!OH6.
The strong shift of s2 in the ccGꢀg+ conformer of a-d-
PhMan to around 3610 cmꢀ1 (cf. the counter-clockwise
ccGꢀg+ and ccG+gꢀ conformers of b-D-PhMan, in which
s2ꢁ3640 cmꢀ1) reflects the relative strengths of the OH2!
O5 and OH2!O1 interactions in the a- and b-anomers, re-
spectively.
structure, cGꢀg+·ins
ACHRTUNEG(4,6), in which insHCATRE(UGN 4,6) indicates a water
Hydrated mannoside complexes of b-d-PhMan·
R2PI spectra recorded in the ion mass channels correspond-
ing to b-d-PhMan·(H2O)n=1–3 are presented in Figure 3. One
ACTHERU(GN H2O)n=1–3:
ACHTREUNG
Figure 3. R2PI spectra of b-d-PhMan and its hydrated complexes recorded
in the singly, doubly and triply hydrated ion mass channels. The symbols
&
*
,
and + shown above the vibrational bands in the n=1 mass channel
identify spectra associated with three distinct monohydrate structures,
~
while the symbol identifies bands associated with the dihydrate.
major and two minor singly hydrated structures were distin-
guished through ion-depletion experiments, with relative
populations declining in the approximate ratio 10:3:1, but
the doubly and triply hydrated complexes populated essen-
tially single structures only; the weak features lying at wave-
numbers below 36800 cmꢀ1 in the dihydrate arise through
fragmentation of the trihydrate. Their associated IRID spec-
tra are presented in Figures 4, 5 and 6 (below).
Figure 4. Experimental and computed IRID spectra of singly hydrated a-
and b-d-PhMan; relative energies [kJmolꢀ1] are shown in square brack-
ets. a) a-d-PhMan·
A
,
cGꢀg+·ins
G
G
,
cGꢀg+·ins
A
A
,
ccG+gꢀ·ins
ACHTREUNG
(H2O)n=1, ccGꢀt·ins
(2,6). The symbols , + and identify the spectra as-
sociated with the three distinct monohydrate structures labelled in the
R2PI spectrum.
8950
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8947 – 8955

Conformational Choice in Hydrated Carbohydrates
FULL PAPER
molecule inserted between OH4 (acting as a hydrogen-bond
donor) and OH6 (the acceptor). It is also very similar to the
corresponding spectrum of the singly hydrated a-anomer
shown in Figure 4a. The strongly displaced doublet at low
wavenumber, which is characteristic of an insertion com-
plex,[11–13,16] is associated with the symmetric and anti-sym-
metric modes involving OH4 and the bound water molecule
(W). Its structure requires a conformational switch of the
co-operative hydrogen-bonding network from counter-clock-
wise (ccG+gꢀ) in the lowest energy conformer of the free
carbohydrate to clockwise (cGꢀg+) in the hydrate, OH2!
OH3!OH4!W!OH6. The switch is facilitated by the ro-
tated orientation of the exocyclic CH2OH group, the inser-
tion of the water molecule at the OH4–OH6 site, and the re-
versed OH2(ax)!OH3(eq) interaction. An identical confor-
mational change is also favoured in the monohydrates of the
b-anomers of glucose and galactose, b-d-PhGlc and b-d-
PhGal.[16] In the a-anomer of mannose, however, in which
the OH2(ax)···O1 distance is far greater (ꢁ3.9 cf. ꢁ2.2 Š)
and an OH2!O1 interaction is not possible, the clockwise
cGꢀg+ conformation is already favoured in the free carbo-
hydrate. Therefore this conformation, which corresponds to
the global minimum, is already set up to accommodate the
insertion of a first water molecule at the OH4–OH6 site.
Indeed it is the only monohydrate structure that is signifi-
cantly populated in a-d-PhMan.[11,16]
The dihydrate: In marked contrast to the monohydrate, the
relative energy of the doubly hydrated complex, b-d-
PhMan·ACTHER(UGN H2O)2 was calculated to be approximately
15 kJmolꢀ1 below that of its nearest neighbour. Not surpris-
ingly, its IRID spectrum is now free of “interfering” bands,
indicating the population of only one predominant structure
ꢀ
(Figure 5). All of its eight O H vibrational bands are clearly
Figure 5. Experimental and computed IRID spectrum of the dihydrate b-
~
d-PhMan·
N
(4,6; 6,2). The symbol
identifies the
spectrum associated with the dihydrate structure labelled in the R2PI
spectrum.
The more-weakly populated hydrates associated with the
IRID spectra (c) and (d) also contain “intruder bands” that
appear because of the difficulty in achieving a complete sep-
aration of individual conformers when their UV bands over-
lap or are congested. The “unfilled” stick diagram shown in
(d) corresponds to the IR spectrum computed for the inter-
resolved and the experimental spectrum is in remarkably
good accord with that computed for the minimum energy
structure cGꢀg+·ins
ACHTRE(UGN 4,6; 6,2). Like the most favoured
mono
A
conformation, and the two water molecules, inserted sepa-
rately and not as a water dimer, are located on each side of
the hydroxymethyl group. The first water molecule occupies
the (4,6) site favoured in the monohydrate and the second
one bridges across the pyranose ring to the axial OH2 group
to create the extended “virtuous circle” of hydrogen bond-
ing, OH2!OH3!OH4!W1!OH6!W2!OH2. The
powerful role of co-operativity in determining the conforma-
tional choice can be gauged by the very large spectral shift
of s6 from around 3640 to around 3300 cmꢀ1 and the fav-
oured population of a single structure. Its fragment ion
R2PI peaks, recorded in the singly hydrated ion mass chan-
nel, were more intense than those associated with parent
ions of the singly hydrated complexes (see Figure 3).
mediate energy structure, cGꢀg+·ins
ACHTRE(UGN 6,2) [1.2].
Comparisons between the experimental and computed vi-
brational spectra of the minor, more weakly populated mon-
ohydrate structures of b-d-PhMan (Figure 4c and d), al-
though complicated by the appearance of “intruder” bands,
predominantly support their association with the two coun-
ter-clockwise structures ccG+gꢀ·ins
A
ACHTRE(UNG 2,6).
The first assignment is favoured by the absence of any fea-
tures at wavenumbers below that of the intense band at
3510 cmꢀ1, the appearance of the closely spaced cluster of
features located between around 3600 and 3650 cmꢀ1, and
the “free” OHW band at ꢁ3720 cmꢀ1. Assignment of the
other hydrate to the structure ccGꢀt·ins
two distinguishing “fingerprints”: the strongly displaced
bands at around 3350 and 3510 cmꢀ1 and the widely spaced
quartet lying between around 3600 and 3730 cmꢀ1. The
strong shift of s2 to around 3350 cmꢀ1 reflects the asymmet-
ric disposition of the bound water molecule, which lies
closer to OH2 than O6 (OH2···OW =1.79 Š, OHW···O6=
1.89 Š). A possible contribution from a fourth hydrate
The trihydrate: When a third water molecule binds to b-d-
PhMan, the structure it adopts is a logical extension of the
structures assumed by the singly and doubly hydrated sugar.
The first water molecule continues to insert into the OH4–
OH6 site while the second and the third are inserted be-
tween OH6 and OH2, now as a water dimer, to conserve
the virtuous cycle of hydrogen bonds adopted in the dihy-
structure, cGꢀg+·ins
A
drate
OH2!OH3!OH4!W1!OH!W2!W3!OH2
also suggested by the correspondence between its computed
spectrum (unfilled bars) and some of the intruder bands.
and retain the cGꢀg+ conformation of the hydrated mono-
saccharide. This structure has the lowest relative energy and
generates an IR spectrum in good accord with the experi-
mental spectrum shown in Figure 6a, particularly when the
Chem. Eur. J. 2008, 14, 8947 – 8955
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8951

B. G. Davis, J. P. Simons et al.
cose—the water molecule bound in the singly hydrated com-
plexshifts further round the pyranose ring, occupying the
(3,2) site in both a- and b-l-PhFuc,[13] the (3,2) or (4,3) sites
in a-d-PhXyl, and the (2,1) site in b-d-PhXyl,[17] whereas the
counter-clockwise hydrogen-bonding orientation in the host
carbohydrate is retained. This continues to be the case in
doubly hydrated a-l-PhFuc, but rather than binding the two
water molecules individually (on each side of the hy
methyl group) as in b-d-PhMan, in the absence of a hy-
droxymethyl group they now prefer to bind as a water
ACHTREUNGdroxy-
AHCTREUNG
A
ACHTREUNG
dimer, again inserting at the (3,2) site, to create the counter-
clockwise hydrogen-bonded chain OH4!OH3!W1!
W2!OH2. The computed structures and relative energies
of the two lowest energy structures of doubly hydrated
phenyl a-l-fucopyranoside are shown in Figure 7.
Figure 7. The computed (MP2/6-311++G**//B3LYP/6-311+G*) lowest
energy structures of the dihydrate a-Ph-l-FucACHTRE(UGN H2O)2; their relative ener-
gies in kJmolꢀ1 are shown in square brackets. Each one binds a water
dimer at the (3,2) site; there is also a structure at an intermediate energy
that differs only in the orientation of the phenyl “tag”. The l enantiomer
was selected because of its natural occurrence (cf. the d enantiomers of
Glc, Gal and Man).
Monosaccharide hydration, the working rules: In Glc, Gal
and Man, the hydroxymethyl group always provides the pre-
ferred hydration site. Water binding also greatly enhances
the relative stability of the clockwise conformation (cGꢀg+
in the a- and b-anomers of glucose and in the b-anomers of
galactose and mannose), and although structures retaining
the original counter-clockwise ccG+gꢀ carbohydrate confor-
mation can be populated, the singly hydrated cGꢀg+ struc-
ture either becomes one of the preferred conformations or
the preferred conformation (see Table 1). This selectivity
and conformational preference is displayed a fortiori in the
doubly and triply hydrated complexes of b-d-PhMan. In the
a-anomer of mannose, in which the axial orientation of
OH2 increases the OH2···O1 distance to around 3.9 Š (cf.
ꢁ2.2 Š in a-d-MeGlc, a-d-MeGal and b-d-PhMan or
ꢁ2.5 Š in b-d-PhGlc and b-d-PhGal), there is no longer any
possibility of a (counter-clockwise) OH2!O1 interaction.
The clockwise cGꢀg+ conformation (and binding site selec-
tivity) is adopted both in its singly hydrated complex, a-D-
PhMan (cGꢀg+·ins4,6), and in the isolated carbohydrate.[12]
Indeed, its hydration greatly reduces the relative stability of
the ccG+gꢀ conformation (see Table 1).
Figure 6. a) Experimental and computed IRID spectra of the trihydrate
b-d-PhMan·
C
,
cGꢀg+·ins
A
ACHTRE(UGN 6,2:6,2)). Relative energies
falling IR laser power at wavenumbers <3300 cmꢀ1 is taken
into account, and certainly in much better accord than those
associated with the next lowest energy alternatives, located
at energies around 5–6 kJmolꢀ1 higher, also shown for com-
parison. The side-on view of the trihydrate, shown in Fig-
ure 6b provides a striking illustration of the amphiphilic
character of the carbohydrate. Its hydrophobic, apolar face
remains “untouched” and the water molecules are all locat-
ed exclusively on the hydrophilic, polar face.
a-l-PhFuc·ACHTREUNG(H2O)2: The hydroxymethyl group clearly pro-
vides the favoured binding site in both the singly and multi-
ply hydrated complexes of b-d-PhMan; the same selectivity
also operates in the singly hydrated complexes of b-d-
PhGlc, b-d-PhGal and a-d-PhMan.[16] However, when the
hydroxymethyl is replaced by a methyl group or a hydrogen
atom with no change of relative configuration—for example,
in fucose rather than galactose, or in xylose rather than glu-
The behaviour of the a-anomer of galactose appears at
first sight to be anomalous, since the hydrated complexof a-
d-MeGal prefers to retain its ccG+gꢀ conformation, though,
like the hydrated b-anomer, the water molecule still inserts
8952
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8947 – 8955

Conformational Choice in Hydrated Carbohydrates
FULL PAPER
Table 1. Calculated relative energies (DE(0) in kJmolꢀ1) of the lowest lying cGꢀg+ and ccG+gꢀ conformers
of unhydrated (W0) and singly hydrated (W1) complexes of the a- and b-anomers of glucose, galactose and
mannose; energies determined through single-point calculations conducted at the MP2/6-311++G** level,
based upon DFT structures (B3LYP/6-31+G* and B3LYP/6-311+G*). Numbers in brackets indicate their rela-
tive energy ordering.
These working rules continue
to operate in multiply hydrated
b-d-PhMan. In the dihydrate
the second molecule binds on
the opposite side of the hy-
a-MeGlc
a-MeGal
a-PhMan
ACHRTEdUNG roxyACHTREmUNG ethyl group, at the (6,2)
W0
W1
W0
W1
W0
W1
site, and in the trihydrate the
(6,2) site is occupied by a water
dimer.
cGꢀg+
>10
0.0 (1) ins4,6
3.9 (3) ins3,2
7.1 (4)
0.0 (1)
6.9 (15) ins4,6
0.0 (1) ins6,5
0.0 (1)
5.7 (5)
0.0 (1) ins4,6
>19 ins2,1
ccG+gꢀ
0.4 (2)
b-PhGlc
b-PhGal
b-PhMan
W0
W1
W0
W1
W0
W1
cGꢀg+
9.7 (4)
0.0 (1)
0.0 (1) ins4,6
0.9 (2) ins6,5
6.6 (4)
0.0 (1)
0.3 (2) ins6,5
0.0 (1) ins6,5
12.0 (13)
0.0 (1)
0.0 (1) ins4,6
0.1 (2) ins6,5
Carbohydrate–protein binding:
Given the propensity rules gen-
erated by optimisation of hy-
drogen-bonding networks in hy-
ccG+gꢀ
at the (6,5) site. The explanation lies in the axial orientation
of OH4, which enhances the OH4(ax)!OH3(eq) interac-
tion. This, together with the enhanced OH2(eq)!O1(ax) in-
teraction in the a-anomer, favours retention of the “cc”
chain, OH4!OH3!OH2!O1. In addition, retention of
drated carbohydrate complexes, it is not unreasonable to an-
ticipate the possible operation of similar rules in hydrogen-
bonded protein–carbohydrate complexes. Indeed, an initial
survey of published X-ray crystal structures of simple mono-
saccharide a-d-mannoside ligands[16] identified the O4–O6
binding site as a favoured, though not exclusive recognition
point. The hydroxymethyl group also acts as a focal point in
many other monosaccharide ligands, which bind to the pro-
tein at their O4–O6 or O6–O5 sites. Examples include: a-d-
MeGlc bound at O6–O5 to Concanavalin A,[22] pea lectin,[23]
Lathyrus Ochrus isolectin I,[24] Pterocarpus angolensis
lectin[25] and to Jacalin;[26] a-d-MeGal bound at O6–O4 to
Jacalin,[26] Artocarpus hirsuta[27] and to Jacalin;[28] and b-d-
MeGal bound at O6–O5 to C-type animal lectins,[29] and
peanut lectin.[30] When explicitly bound water molecules are
conserved in protein–carbohydrate complexes, it may be be-
cause their presence maintains a favoured ligand conforma-
tion as well as providing a stronger overall protein–carbohy-
drate interaction.[31] When there are no conserved water
molecules, the interaction between the protein and the
ligand and between the displaced and surrounding water
molecules will be the deciding factor. Finally, the clear dis-
tinction between the hydrophilic and hydrophobic faces dis-
played in the multiply hydrated carbohydrate b-d-PhMan·
(H2O)3 raises interesting issues relating to molecular recog-
nition. The hydrophobic face is “ready” for stacking interac-
tions with aromatic residues[32,33] but the hydrophilic face is
covered by water molecules with a structure directed by the
conformation of the underlying carbohydrate, which may
itself be altered by hydration. Which structure will the pro-
tein recognize?
the G+gꢀ orientation favours the ins
ACHTRE(UNG 6,5) water binding
motif OH6!W!O5. In contrast, in a-d-MeGlc, in which
OH4 is oriented equatorially, the conformational switch
from ccG+gꢀ to cGꢀg+, which removes the OH2!O1 in-
teraction, is compensated by the insertion at the (4,6) site to
create the alternative strongly bound motif OH2!OH3!
OH4!W!O6.
The full set of preferred Glc, Gal and Man monohydrate
structures is displayed in Figure 8. It includes those identi-
Figure 8. The preferred structures of the singly hydrated complexes of a-
and b-anomers of Glc, Gal and Man.
fied experimentally through a combination of vibrational
spectroscopy coupled with DFT and ab initio calculations
(b-d-PhGlc, b-d-PhGal and a/b-d-PhMan) and those pre-
dicted theoretically (a-d-MeGlc and a-d-MeGal). A clear
pattern of selectivity and conformational preference is re-
vealed. When the OH4 group is oriented equatorially, as in
Glc and Man, the cGꢀg+ conformation creates an optimised
structure for reception of the bound water molecule at the
(4,6) hydroxymethyl site—the weakest link in the original
ccG+gꢀ configuration—to maximise the co-operative hydro-
gen bonding. When the OH4 group is oriented axially, as in
Gal, this is achieved by insertion of the water molecule at
the (6,5) site.
Conclusion
The factors that govern selectivity and conformational
choice in hydrated carbohydrates have been examined by
determining the structures of a systematically varied set of
singly and multiply hydrated monosaccharide complexes in
the gas phase by using ultraviolet and infrared ion depletion
spectroscopy conducted in a supersonic jet expansion and
coupled with DFT and ab initio calculations. New results ob-
tained for the singly hydrated complexes of phenyl b-d-man-
Chem. Eur. J. 2008, 14, 8947 – 8955
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8953

B. G. Davis, J. P. Simons et al.
311+G* for the hydrated systems) based upon their relative energies
(with a cut-off at ꢁ15 kJmolꢀ1) and comparisons with previous investiga-
tions of carbohydrate structures. More accurate energies were then calcu-
lated for these optimised structures at the MP2/6-311++G** level of
theory. Zero-point corrections and harmonic vibrational frequencies for
the energetically lowest lying conformers were obtained using the B3LYP
structures; the frequencies computed for the O–H stretch modes (ex-
pressed in wavenumbers) were scaled by the factor 0.9734 for compari-
son with the experimental spectra. The full list of energies and structures
for the lowest energy conformers is given as Supporting Information.
nopyranoside (b-d-PhMan), methyl a-d-gluco- and a-d-gal-
actopyranoside (a-d-MeGlc and a-d-MeGal), when coupled
with those reported earlier for the singly hydrated com-
plexes of a-d-PhMan, b-d-PhGlc and b-d-PhGal, have creat-
ed a comprehensive data set that reveals a systematic pat-
tern of conformational preference and regioselectivity,
driven by the provision of optimal, co-operative hydrogen-
bonded networks in the hydrated sugars. In addition, despite
the very large number of potential structures and binding
sites, the choice is highly selective with binding invariably
“focused” around the hydroxymethyl group. A survey of
some of the structures listed in the protein databank sug-
gests the possible operation of similar rules governing the
structures of hydrogen-bonded protein–carbohydrate com-
plexes.
Notation: The nomenclature of the hydrated and unhydrated monosac-
charide conformations has been described previously.[13] The designations,
ccG+gꢀ, cTt, and so forth, indicate the conformation of the carbohy-
drate: briefly, “cc” (“c”) indicates a counter-clockwise (clockwise) orien-
tation of the peripheral OH groups, OH4!OH3!OH2!O1 (O1!
OH2!OH3!OH4), and G+gꢀ (Tt) indicates the gauche (trans) orienta-
tion of the exocyclic hydroxymethyl group and its terminal OH6 group,
respectively. In the case of the hydrated structures the insertion position
of the water molecule is indicated by adding “ins
ACHTRE(UNG position)” to the no-
Similar rules also continue to operate in the multiply hy-
drated carbohydrate b-d-PhMan·ACHTREU(NG H2O)2,3, in which the
menclature of the bare molecule, for example, ins(4,6) indicates a water
AHCTREUNG
molecule inserted between OH4 (acting as a hydrogen-bond donor) and
OH6 (the acceptor).
bound water molecules form a cyclic structure that incorpo-
rates the hydroxymethyl group. They are located exclusively
on the polar face and their orientation is dictated by the
(perturbed) conformation of the carbohydrate to which they
are attached.
Acknowledgements
The authors thank Dr. Jann Frey, Dr. Lavina C. Snoek, Ms. Yoana Perez
Badel and Prof. Luis A. Montero for their contributions to the studies
presented here. We acknowledge the financial support provided by Engi-
neering and Physical Sciences Research Council (EPSRC), the Lever-
hulme Trust (grant F/08788G), the Spanish Ministry of Education and
Science (EJC), the Oxford Supercomputing Centre, the STFC Laser
Loan Pool and the Physical and Theoretical Chemistry Laboratory.
Experimental Section
Spectroscopy: Detailed descriptions of the experimental strategy have
been published previously.[13] The carbohydrate samples were vaporised
into a supersonic jet argon expansion using a laser desorption system
constructed in-house, and their hydrated complexes were formed by
seeding the argon carrier gas with water vapour (ꢁ0.25%) and then sta-
bilised in the high collision frequency region of the free jet expansion.
The expanding jet passed through a 2 mm skimmer to form a collimated
molecular beam which was crossed by one or two tunable laser beams in
[1] H.-J. Gabius, H.-C. Siebert, S. AndrØ, J. JimØnez-Barbero, H. Rüdig-
er, ChemBioChem 2004, 5, 740–764.
[2] C. Clarke, R. J. Woods, J. Gluska, A. Cooper, M. A. Nutley, G. J.
Boons, J. Am. Chem. Soc. 2001, 123, 12238–12247.
[3] A. Laederach, P. J. Reilly, Proteins: Struct., Funct., Bioinf. 2005, 60,
591–597.
[4] A. J. Petrescu, S. M. Petrescu, R. A. Dwek, M. R. Wormald, Glyco-
biology 1999, 9, 343–352, and references therein.
[5] F. C. Bernstein, T. F. Koetzle, G. J. Williams, E. F. Meyer, Jr., M. D.
Brice, J. R. Rodgers, O. Kennard, T. Shimanouchi, M. Tasumi, J.
Mol. Biol. 1977, 112, 535–542; http://www.rcsb.org/pdb/home/
home.do.
[6] F. H. Allen, O. Kennard, Chemical Design Automation News; 1993,
8, 31–37; http://www.ccdc.cam.ac.uk/products/csd/.
[7] M. R. Wormald, A. J. Petrescu, Y.-L. Pao, A. Glithero, T. Elliott,
R. A. Dwek, Chem. Rev. 2002, 102, 371–386.
[8] V. Kräutler, M. Müller, P. H. Hünenberger, Carbohydr. Res. 2007,
342, 2097–2124.
[9] O. Coskuner, J. Chem. Phys. 2007, 127, 015101.
[10] L. Hemmingsen, D. E. Madsen, A. L. Esbensen, L. Olsen, S. B. En-
gelsen, Carbohydr. Res. 2004, 339, 937–948.
[11] J. P. Simons, R. A. Jockusch, P. C¸ arÅabal, I. Hünig, R. T. Kroemer,
N. A. Macleod, L. C. Snoek, Int. Rev. Phys. Chem. 2005, 24, 489.
[12] P. C¸ arÅabal, I. Hünig, D. P. Gamblin, B. Liu, R. A. Jockusch, R. T.
Kroemer, L. C. Snoek, A. J. Fairbanks, B. G. Davis, J. P. Simons, J.
Am. Chem. Soc. 2006, 128, 1976.
the extraction region of
a linear time-of-flight mass spectrometer
(Jordan). Mass-selected resonant two-photon ionization (R2PI) spectra
were recorded using the frequency-doubled output of a pulsed Nd:YAG-
pumped dye laser (Spectron 810/LambdaPhysik FL2002, 0.5 mJpulseꢀ1
UV for the monosaccharides and Continuum Powerlite II/Sirah PS-G,
3 mJpulseꢀ1 UV for their hydrated complexes, both operating at 10 Hz).
Conformer-specific UV and IR spectra were recorded by employing UV–
UV and IR ion-depletion (IRID) double-resonance spectroscopy. The
UV–UV experiments utilised the frequency-doubled output of the pulsed
Nd:YAG pumped dye lasers described above. The IR spectroscopy ex-
periments employed radiation in the range 3100–3800 cmꢀ1, generated by
difference frequency mixing of the fundamental of a Nd:YAG laser with
the output of a dye laser in a LiNbO3 crystal (Continuum Powerlite 8010/
ND6000/IRP module) or by a tunable OPO/OPA laser system (LaserVi-
sion); all laser pulses were of an approximately 10 ns duration. The delay
between the pump and the probe laser pulses was approximately 150 ns
in both the IR ion-depletion and UV hole-burning experiments, whereas
the delay time between opening the pulsed valve and triggering the UV
laser was adjusted to optimise the generation of either “naked” or hy-
drated molecules in the supersonic expansion.
Computation: Conformational and structural assignments were made
through comparison with calculation by using a combination of molecular
mechanics, ab initio and density functional theory (DFT) methods as im-
plemented in the MacroModel software (MacroModel v.8.5, Schrçdinger,
LLC21),[34] and the Gaussian 03 program package.[35] Initial structures
were generated by a molecular mechanics conformational search using
the Monte Carlo multiple minimisation method. A much-reduced set of
conformational or hydrated complexstructures was selected for DFT ge-
ometry optimisation (B3LYP/6-31+G* for the naked and B3LYP/6-
[13] P. C¸ arÅabal, T. Patsias, I. Hünig, B. Liu, C. Kaposta, L. C. Snoek,
D. P. Gamblin, B. G. Davis, J. P. Simons, Phys. Chem. Chem. Phys.
2006, 8, 129.
[14] Carbohydrate structural library: http://physchem.ox.ac.uk/~jps.
[15] E. G. Robertson, J. P. Simons, Phys. Chem. Chem. Phys. 2001, 3, 1–
18.
8954
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 8947 – 8955

Conformational Choice in Hydrated Carbohydrates
FULL PAPER
[16] P. C¸ arÅabal, R. A. Jockusch, I. Hünig, L. C. Snoek, R. T. Kroemer,
B. G. Davis, D. P. Gamblin, I. Compagnon, J. Oomens, J. P. Simons,
J. Am. Chem. Soc. 2005, 127, 11414.
[17] I. Hünig, A. J. Painter, R. A. Jockusch, P. C¸ arÅabal, E. M. Marzluff,
L. C. Snoek, D. P. Gamblin, B. G. Davis, J. P. Simons, Phys. Chem.
Chem. Phys. 2005, 7, 2474.
[18] O. Mitsunobu, Y. Yamada, Bull. Chem. Soc. Jpn. 1967, 40, 2380–
2382.
[19] O. Mitsunobu, Synthesis 1981, 1–28.
[20] M. Mella, L. Panza, F. Ronchetti, L. Toma, Tetrahedron 1988, 44,
1673–1678.
[21] J. J. Gridley, H. M. I. Osborn, J. Chem. Soc. Perkin Trans. 1 2000,
1471–1491.
[22] S. J. Harrop, J. R. Helliwell, T. C. Wan, A. J. Kalb, L. Tong, J. Yariv,
Acta Crystallogr. Sect. A 1996, 52, 143–155; (PDB ID=1gic).
[23] M. B. Shevtsov, I. N. Tsygannik, to be published (PDB ID=1hkd).
[24] Y. Bourne, A. Roussel, M. Frey, P. RougØ, J. C. Fontecilla-Camps, C.
Cambillau, Proteins: Struct., Funct., Bioinf. 1990, 8, 365–376; (PDB
ID=1 Loa).
[25] R. Loris, A. Imberty, S. Beeckmans, E. V. Driessche, J. S. Read, J.
Bouckaert, H. D. Greve, L. Buts, L. Wyns, J. Biol. Chem. 2003, 278,
16297–16303; (PDB ID=1n3o).
[26] A. A. Jeyaprakash, G. Jayashree, S. K. Mahanta, C. P. Swaminathan,
K. Sekar, A. Surolia, M. Vijayan, J. Mol. Biol. 2005, 347, 181–188;
(PDB ID=1s4).
[27] K. N. Rao, C. G. Suresh, U. V. Katre, S. M. Gaikwad, M. I. Khan,
Acta Crystallogr. Sect. A 2004, 60, 1404–1412; (PDB ID=1tp8).
[28] R. Sankaranarayanan, K. Sekar, R. Banerjee, V. Sharma, A. Surolia,
M. Vijayan, Nat. Struct. Biol. 1996, 3, 596–603; (PDB ID=1jac).
[29] A. R. Kolatkar, W. I. Weis, J. Biol. Chem. 1996, 271, 6679–6685;
(PDB ID=1afa).
[30] R. Ravishankar, K. Suguna, A. Surolia, M. Vijayan, Acta Crystal-
logr. Sect. A 1999, 55, 1375–1382; (PDB ID=1qf3).
[31] S. Elgavish, B. Shaanan, J. Mol. Biol. 1998, 277, 917–932.
[32] J. Screen, E. C. Stanca-Kaposta, D. P. Gamblin, B. Liu, N. A. Ma-
cleod, L. C. Snoek, B. G. Davis, J. P. Simons, Angew. Chem. 2007,
119, 3718–3722; Angew. Chem. Int. Ed. 2007, 46, 3644–3648.
[33] E. C. Stanca-Kaposta, D. P. Gamblin, J. Screen, B. Liu, L. C. Snoek,
B. G. Davis, J. P. Simons, Phys. Chem. Chem. Phys. 2007, 9, 4444–
4451.
[34] F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M.
Lipton, C. Caufield, G. Chang, T. Hendrikson, W. C. Still, J.
Comput. Chem. 1990, 11, 440–467.
[35] Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomer-
y, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyen-
gar, 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, V. Bakken, 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. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004.
Received: March 14, 2008
Published online: August 20, 2008
Chem. Eur. J. 2008, 14, 8947 – 8955
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8955