Studies on the conformational flexibility of α-l-rhamnose-containing oligosaccharides using 13C-site-specific labeling, NMR spectroscopy and molecular simulations: Implications for the three-dimensional structure of bacterial rhamnan polysaccha

Article abstract of DOI:10.1039/c2ob06924e

Bacterial polysaccharides are comprised of a variety of monosaccharides, l-rhamnose (6-deoxy-l-mannose) being one of them. This sugar is often part of α-(1 → 2)- and/or α-(1 → 3)-linkages and we have therefore studied the disaccharide α-l-Rhap-(1 → 2)-α-l

Full text of DOI:10.1039/c2ob06924e

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 2453
www.rsc.org/obc
PAPER
Studies on the conformational flexibility of α-L-rhamnose-containing
oligosaccharides using 13C-site-specific labeling, NMR spectroscopy and
molecular simulations: implications for the three-dimensional structure of
bacterial rhamnan polysaccharides
K. Hanna M. Jonsson, Elin Säwén and Göran Widmalm*
Received 15th November 2011, Accepted 6th January 2012
DOI: 10.1039/c2ob06924e
Bacterial polysaccharides are comprised of a variety of monosaccharides, L-rhamnose
(6-deoxy-L-mannose) being one of them. This sugar is often part of α-(1 → 2)- and/or α-(1 → 3)-linkages
and we have therefore studied the disaccharide α-L-Rhap-(1 → 2)-α-L-Rhap-OMe to obtain information
on conformational preferences at this glycosidic linkage. The target disaccharide was synthesized with
13C site-specific labeling at C1′ and at C2′, i.e., in the terminal group. 2D 1H,13C-HSQC-HECADE and
1H,13C-J-HMBC NMR experiments, 1D 13C and 1H NMR spectra together with total line-shape analysis
were used to extract conformationally dependent hetero- and homonuclear spin–spin coupling constants.
This resulted in the determination of 2JC2′,H1′, 3JC1′,C1, 3JC1′,C3, 3JC2′,C2, 2JC1′,C2, 1JC1′,C2′, and 1JC1′,H1′
.
These data together with previously determined JCH and 1H,1H NOEs result in fourteen conformationally
dependent NMR parameters that are available for analysis of glycosidic linkage flexibility and
conformational preferences. A 100 ns molecular dynamics (MD) simulation of the disaccharide with
explicit water molecules as solvent showed a major conformational state at ϕH ≈ 40° and ψH ≈ −35°,
consistent with experimental NMR data. In addition, MD simulations were carried out also for α-L-Rhap-
(1 → 3)-α-L-Rhap-OMe and a rhamnan hexasaccharide. The gathered information on the oligosaccharides
was used to address conformational preferences for a larger structure, a 2- and 3-linked nonasaccharide,
with implications for the 3D structure of rhamnan polysaccharides, which should be regarded as flexible
polymers.
plants.8 The LPS then act as part of the barrier that protects the
Introduction
bacteria from plant-derived antimicrobial substances. These
polysaccharide structures are conversely recognized as non-self
by the plant innate immune system and coiled polysaccharide
structures may be of importance in interactions between bacteria
and plants.
In carbohydrate structures from humans the number of different
monosaccharides is quite limited; typically seven different
sugars are present in glycoproteins and glycolipids.1 Constituents
of polysaccharides in man add a few more monosaccharides to
the repertoire. In bacteria, however, more than 100 different
monosaccharide components have been found.2 One of them, L-
rhamnose (6-deoxy-L-mannose) is present as a major constituent
of the O-antigen polysaccharides from Shigella flexneri3,4 and is
the sole monosaccharide in the repeating unit of an O-antigen
from a Klebsiella pneumoniae strain.5 It is the constituent sugar
of the O-glycan part of the S-layer glycoprotein in Geobacillus
stearothermophilus6 and the backbone of the O-specific polysac-
charide from the lipopolysaccharide of Aeromonas bestiarum
strain 207 is comprised of L-rhamnose residues.7 Furthermore, it
is often a constituent of the O-antigen of lipopolysaccharides
(LPS) from Gram-negative bacteria that are associated with
In many of the above mentioned polysaccharides L-rhamnose
residues are joined by α-(1 → 2)- and/or α-(1 → 3)-linkages and
conformational analysis of polysaccharides or oligosaccharides
being models for these polymers is often based on measurements
of the nuclear Overhauser effect (NOE) between protons in
different sugar residues.9–12 The experimental NMR data are
subsequently analyzed in the light of a molecular simulation
which employs a certain molecular mechanics force field.13,14
Additional information on conformational preferences in rham-
nose-containing oligosaccharides have been obtained from trans-
3
glycosidic heteronuclear coupling constants.15,16 These JCH are
important and to date a large number of NMR experiments have
been developed to this end.17–26 The long-range heteronuclear
coupling constants report on conformational averaging in a
different way compared to the NOEs which are inversely
Department of Organic Chemistry, Arrhenius Laboratory, Stockholm
University, S-106 91 Stockholm, Sweden. E-mail: gw@organ.su.se
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2453–2463 | 2453

View Online
Fig. 1 Schematic of (a) α-L-Rhap-(1 → 2)-α-L-Rhap-OMe and (b) α-L-
Rhap-(1 → 3)-α-L-Rhap-OMe with torsion angles ϕ and ψ indicated and
(c) α-L-Rhap-(1 → 3)-α-L-Rhap-(1 → 2)-α-L-Rhap-(1 → 3)-α-L-Rhap-(1
→ 3)-α-L-Rhap-(1 → 2)-α-L-Rhap-OMe with sugar residues I–VI
indicated.
Fig. 2 Ramachandran map of α-L-Rhap-(1 → 2)-α-L-Rhap-OMe using
the PARM22/SU01 force field, a dielectric constant of 3, and a grid size
of 15°. Contour lines are drawn at 0.3 kcal mol−1 increments above the
global energy minimum. Low energy conformational states are denoted
by A, B and C.
proportional to a distance averaging according to riÀj 6, where rij
is the internuclear distance between protons i and j. However,
3
the JCH have to be analyzed by a Karplus-type relationship in
which several solutions to the conformational averaging process
may be possible for a given coupling constant. Furthermore, the
Karplus-curves need to be adequately parameterized for the
coupling pathway under study.
the results of which have implications for the 3D structure of
rhamnan polysaccharides.
The disaccharide α-L-Rhap-(1 → 2)-α-L-Rhap-OMe (Fig. 1),
also referred to as R2R, is a model for the constituent disacchar-
ide in these polysaccharides and has been studied previously
using NMR spectroscopy, molecular mechanics calculations
and molecular simulations.27–29 To date four interresidue
1H,1H-NOEs, the two transglycosidic 3JCH coupling constants as
well as residual dipolar coupling data have been measured.30 To
generate additional experimental NMR data for its conformation-
al analysis we have now synthesized two isotopologues of R2R
in which the C1′ and the C2′ atoms in the terminal sugar residue
are 13C site-specifically labeled. This labeling scheme facilitates
Results and discussion
The synthesis pathway to the three target molecules, α-L-Rhap-
(1 → 2)-α-L-Rhap-OMe (1), α-L-[1′-13C]Rhap-(1 → 2)-α-L-
Rhap-OMe (1-c1′) and α-L-[2′-13C]Rhap-(1 → 2)-α-L-Rhap-
OMe (1-c2′), was executed as in the previously published syn-
thesis of 131 with one exception: the target molecules were to be
protected solely with benzoyl groups to carry out the deprotec-
tion in one step only. The donor, 2,3,4-tri-O-benzoyl-α-L-rham-
nopyranosyl bromide (2), and its 13C-isotopologues, were
prepared by an initial per-benzoylation and then treated with
HBr in HOAc, thereby forming the α-bromide product.32 The
acceptor, methyl 3,4-di-O-benzoyl-α-L-rhamnopyranoside (3),
was synthesized from methyl 4-O-benzoyl-2,3-O-isopropyli-
dene-α-L-rhamnnopyranoside33 using the procedure described by
Norberg et al.34 The glycosylation reaction to form the protected
disaccharide 4 was performed at low temperature with acceptor
in excess using silver triflate as a promotor.35 Finally, the fully
benzoylated disaccharide 4 was deprotected according to stan-
dard procedures. The target molecules 1, 1-c1′ and 1-c2′ were
obtained in 54, 60 and 66% yield, respectively, starting from L-
rhamnose or its isotopologues.
3
determination of conformationally dependent homonuclear JCC
coupling constants related to the glycosidic torsion angles ϕ and
ψ. The determination of heteronuclear coupling constants were
carried out by two-dimensional J-HMBC and HSQC-HECADE
NMR experiments using either natural abundance or site-specifi-
n
cally 13C-labeled disaccharides. In addition, JCH were obtained
by total-lineshape analysis NMR spectral simulations of the site-
specifically 13C-labeled R2R molecules. The establishment of
seven conformationally dependent coupling constants will be
described in this study. Furthermore, molecular mechanics meth-
odology and molecular dynamics (MD) simulations are used to
assess conformational flexibility and population distributions in
R2R. Comparisons of data from NMR experiments to those
obtained by the MD simulation ascertain whether the molecular
description of R2R is suitable as a basis for delineating oligo-
and polysaccharide conformation and dynamics. Additionally,
MD simulations were carried out for α-L-Rhap-(1 → 3)-α-L-
Rhap-OMe and a hexasaccharide and the simulation results were
used to address conformational preferences of a nonasaccharide,
The conformational space available to R2R may be investi-
gated by a Ramachandran map in which all degrees of freedom
are relaxed except for the glycosidic torsion angles ϕ and ψ
(Fig. 2). There are two low energy regions denoted A and B
where the conformation at the ϕ torsion angle is governed by the
exo-anomeric effect.36 An additional region, labeled C, known
as the non-exo-anomeric conformation can be identified. This
2454 | Org. Biomol. Chem., 2012, 10, 2453–2463
This journal is © The Royal Society of Chemistry 2012

View Online
B-conformer at the global energy minimum.38 A better descrip-
tion of the indicated conformational equilibrium will be possible
if additional experimental data are utilized, e.g. the hitherto
lacking transglycosidic 3JCC coupling constants.
The site-specific 13C-labeling, at C2′ and at C1′, facilitates
torsion angle information to be obtained via Karplus-type
relationships for ϕ and ψ, respectively. At the ψ torsion angle
three three-bond coupling constants are available if determined,
3
3
viz., JC1′,H2,
3JC1′,C1 and JC1′,C3 (Fig. 3). The heteronuclear
coupling constant was previously determined using a one-dimen-
sional long-range (1DLR) experiment.39 Additional determi-
nation of this heteronuclear J-coupling is herein carried out by
alternative methods. The two carbon-13 homonuclear coupling
constants have so far not been determined for R2R. At the ϕ
torsion angle three conformationally dependent coupling con-
3
stants are available; JC2,H1′ was estimated in a previous study
3
and JC2′,C2 determined herein. The L-[2-13C]Rhap-labeling aids
2
accurate determination of JC2′,H1′ by different techniques and
information on this coupling constant may be useful since
Klepach et al.40 recently showed that this heteronuclear coupling
constant has a conformational dependence on the ϕ torsion
angle. Three coupling constant relationships are thus accessible
2
3
also for this torsion angle, viz., JC2′,H1′, .
3JC2′,C2 and JC2,H1′
The transglycosidic homonuclear coupling constants were
readily determined in 13C-labeled R2R isotopologues from resol-
ution enhanced 1D proton-decoupled 13C NMR spectra as
3
shown in Fig. 4. The JCC coupling constants are compiled in
Ramachandran map, created using the CHARMM SU/01 force
field modified for carbohydrates,37 is similar to that previously
described14 employing an MM3 molecular mechanics force
field. However, the simplified HSEA force field36 results in a
Fig. 3 Schematic of (a) [1′-13C]-R2R site-specifically labeled with
carbon-13 for measurement of 3JCC and 3JCH related to the torsion angle
ψ; (b) [2′-13C]-R2R site-specifically labeled with carbon-13 for measure-
ment of JCC and JCH related to the torsion angle ϕ. Pertinent proton
and carbon atoms are highlighted in bold characters.
Fig. 4 Selected resonances from the 13C NMR spectra of (a) [1′-13C]-
R2R showing 3JC1′,C1 = 0.7 Hz; (b) [1′-13C]-R2R showing 3JC1′,C3 = 1.9
Hz; (c) [2′-13C]-R2R showing 3JC2′,C2 = 3.6 Hz.
3
2
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2453–2463 | 2455

View Online
n
3
Table 1 Two- and three-bond JCH (Hz) in R2R from 1DLR, J-HMBC, and HSQC-HECADE experiments, total line-shape analysis and JCC (Hz)
from 13C NMR experiments. Pertinent torsion angles are indicated
Coupling constant
1DLRa
4.2, 4.3
J-HMBC
HSQC-HECADE
Spin simulation
13C
3.6
Compound
Torsion angle
3JH1′,C2
3JC2′,C2
2JC2′,H1′
3JC1′,H2
3JC1′,C1
3JC1′,C3
4.12 (0.20)b
1
ϕH
ϕC2′
ϕH
ψH
ψC1
ψC3
1-c2′
1-c2′
1/1-c1′
1-c1′
1-c1′
–2.2
–2.33
4.59
4.7, 4.8
4.52 (0.07)
0.7
1.9
a From ref. 29. b Standard deviations are given in parenthesis.
Fig. 6 1H NMR analysis at 700 MHz of [1′-13C]-R2R: (a) simulated
spectrum by total-lineshape analysis using the PERCH NMR software to
n
obtain, in particular, JCH coupling constants; (b) experimental
spectrum.
An alternative approach to obtain the heteronuclear coupling
1
constants is to carry out a total line-shape analysis46 of the H
Fig. 5 Selected part from the 1H,13C-HSQC-HECADE NMR spectrum
of [2′-13C]-R2R demonstrating the tilting of the cross-peak and the 2JC2′,
H1′ coupling constant of −2.2 Hz measured in the F2 dimension.
NMR spectra of the site-specifically 13C-labeled compounds.
The 1H and 13C NMR chemical shifts in 1 were assigned at
310 K using standard 2D NMR experiments47 and H chemical
1
3
shifts and JHH were subsequently refined using the NMR spin-
simulation software PERCH. Further analysis was the performed
2
Table 1. The transglycosidic JCOC coupling constant may also
n
reveal some conformational information,41 i.e., in respect to the
ϕ torsion angle, and from the 13C NMR spectrum of 1-c1′ it
could be established. The two-bond homonuclear coupling con-
for 1-c1′ (Fig. 6) and 1-c2′ to obtain JC,H (Table 2). There are
in addition three 1J coupling constants that may contribute infor-
mation on conformational preferences through their torsion angle
1
1
1
2
dependence, viz., JC1′,C2′, JC1′,H1′ and JCn,Hn, where n is the
stant JC1′,C2 = −1.9 Hz in R2R, assuming that the sign of the
substitution position.48,49 In R2R 1JC1′,C2′ = 47.5 Hz, determined
two-bond coupling is negative.42
1
from 1-c1′ and 1-c2′. In the present study JC1′,H1′ = 170.49 Hz
Heteronuclear coupling constants over two and three bonds
were measured by two different 2D NMR experiments, viz., the
from the total line-shape analysis of 1-c1′ compared to the value
of 170.85 Hz obtained from J-modulated constant-time HSQC
1H,13C-HSQC-HECADE
experiment43,44
that
utilizes
experiments of 1,30 which also resulted in JC2,H2 = 149.20 Hz.
1
2
3
1H,1H-TOSCY transfer and the JCH and/or JCH are determined
from cross-peak separations in the F2-dimension; the J-HMBC
experiment45 employs a scaling factor κ in the F1 dimension and
The agreement between different NMR approaches is very good
and we have thus herein obtained and compiled eight coupling
constants related to the conformationally dependent torsion
n
for sufficiently large values of κ the JCH couplings are readily
3
angles ϕ and ψ in addition to the previously reported two JCH
determined from the cross-peak separation. The HSQC-HE-
CADE experiment employed with a 30 ms isotropic mixing
and four NOEs, that can be used in the conformational analysis
of R2R. Thus, no less than fourteen NMR parameters related to
a single glycosidic linkage are available in the conformational
analysis of the molecule in isotropic solution, i.e., it is not per-
turbed to any extent by an alignment medium which may
influence the distribution of conformational states such when
2
period resulted in a small tilt of the cross-peaks from JC2′,H1′
corresponding to a negative coupling constant as shown in
Fig. 5. The the J-HMBC experiment was used to measure the
3
3
transglycosidic JC1′,H2 and JC2,H1′ coupling constants and the
results are compiled in Table 1.
2456 | Org. Biomol. Chem., 2012, 10, 2453–2463
This journal is © The Royal Society of Chemistry 2012

View Online
1
Table 2 1H and 13C NMR chemical shifts of 1 in D2O at 37 °C. The heteronuclear coupling constants were obtained using 1-c1′ and 1-c2′. The H
3
1
NMR chemical shifts and coupling constants were refined using the PERCH NMR software. Coupling constants, JHH in parentheses, JCH in braces,
nJC1′,H in angle brackets, and nJC2′,H in square brackets, are given in Hz
Sugar residue
1
2
3
4
5
6
OMe
α-L-Rhap(1 →
4.962
(1.85)
{170.49}
4.065
3.796
(9.77)
3.451
(9.59)
3.751
(6.31)
〈1.19〉
[−0.15]
69.94
3.673
(6.27)
69.40
1.276
(3.40)
〈1.41〉
{149.90}
70.89
3.923
(3.42)
79.30
[1.27]
70.91
3.816
(9.79)
70.86
[0.84]
72.85
3.464
(9.56)
72.99
→ 2)-α-L-Rhap-OMe
103.01
4.787
(1.78)
100.42
17.43
1.306
3.401
55.71
17.46
Fig. 7 Scatter plots and time dependence of the glycosidic torsion angels ϕH and ψH in α-L-Rhap-(1 → 2)-α-L-Rhap-OMe (a–c) and in α-L-Rhap-(1
→ 3)-α-L-Rhap-OMe (d–f) from the MD simulations using the PARM22/SU01 force field. The conformational space is divided into three regions
denoted A, B and C.
residual dipolar couplings are measured and utilized in the con-
formational analysis.50
(Fig. 7a). A large number of transitions between conformations
A and B occurs during the 100 ns simulation; thus, this region
of conformational space is well sampled (Fig. 7b and 7c). The
extent to which the three regions are populated is for A: 90%, B:
9%, and C: 1%. The average conformation given by the glycosi-
dic torsion angles in R2R was ϕH = 39° and ψH = −38°, shown
in Fig. 8.
For interpretation of the experimentally derived NMR data in
a conformational model an MD simulation of R2R was carried
out with explicit water molecules as solvent. The simulation
closely mirrors the information that can be concluded from the
Ramachandran map, i.e., the major conformational state is
present for the A-region, and the minor one occurs for the B-
region; the third conformational state is hardly populated
From MD trajectories effective proton–proton distances can
be calculated and indirect spin–spin coupling constants predicted
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2453–2463 | 2457

View Online
to the ϕH and ψH torsion angles, respectively. These coupling
constants were compared to those determined previously using
the 1DLR experiment, viz., 4.1 and 5.1 Hz, respectively.29 The
excellent agreement between the two computed and the exper-
3
imentally determined JCH coupling constants leads credence to
the population distribution described. Thus, the α-(1 → 3)-
linkage should be regarded as flexible with two significantly and
close to equally populated conformational states, as well as some
population of a non-exo-anomeric conformation (conformation
C), at the disaccharide level of complexity.
Some time ago rhamnan oligosaccharides related to the O-
antigen polysaccharide isolated from Pseudomonas syringae
pvs. coronafaciens IMV 903052 were synthesized and their con-
formational preferences investigated by NMR spectroscopy and
molecular modeling as well as MD simulations.53 The tri-, hexa-
, and nonasaccharides corresponding to the repeating unit →3)-
α-L-Rhap-(1 → 3)-α-L-Rhap-(1 → 2)-α-L-Rhap-(1→ investigated
by NMR NOEs were concluded to populate the conformational
states with a positive ψ torsion angle most of the time for both
the α-(1 → 2)- and α-(1 → 3)-linkages as deduced by molecular
modeling and the Ramachandran maps revealing regions of low
potential energy. Predicted NOEs using a full matrix relaxation
approach for the conformations deduced by molecular modeling
and MD simulations were found to be in good agreement with
those determined experimentally. These comprise e.g. for the
α-L-Rhap-(1 → 2)-α-L-Rhap structural component the dipolar
interactions between H1′ and H2 as well as H5′ and H1 interpro-
ton pairs. Consequently, the molecular modeling and MD simu-
lations were judged reliable. However, these two proton pairs’
effective distances in the disaccharide are equally well described
by a conformation with a negative ψ torsion angle,27 and it has
been shown that the major conformational state of R2R is that in
which the ψ torsion angle takes a negative value.30 In order to
differentiate between these two conformations, i.e., with a posi-
tive or a negative ψ torsion angle (cf. Fig. 7a, regions B and A,
respectively) one has to e.g. acquire information on the effective
1H,1H-distances between H1′ and H1 or H2′ and H2 which will
reveal which of the two conformational states that is the preferred
one or indicate the relative populations if there is a conformation-
al equilibrium skewed towards one of them, in a corresponding
way to the analysis performed for α-D-Manp-(1 → 2)-α-D-Manp-
(1 → O)-L-Ser.54
Fig. 8 Molecular model of R2R with glycosidic torsion angles ϕH
39°and ψH = −38° (average conformation from the MD simulation).
=
using pertinent Karplus-type relationships. The four interresidue
proton–proton distances obtained from the present MD simu-
lation were compared to those previously determined by
1H,1H-T-ROESY NMR experiments:30 H1′,H2 was 2.28 Å in
the MD simulation vs. 2.24 Å from experiment; H1′,H1 was
2.84 vs. 3.10 Å; H5′,H1 was 2.70 vs. 2.59 Å; H2′,H2 was 3.81
vs. 3.99 Å. Thus, for effective proton–proton distances the agree-
ment between simulation and experiment is indeed good.
However, in order to attain concordance between the two, small
changes to the force field would be needed such as a shift of the
average torsion angle ψH towards a less negative value. The MD
n
simulation of R2R may also be used to calculate JCH coupling
constants via Karplus-type relationships where averaging is
carried out over all saved conformations, i.e., 5 × 105 coordinate
sets over the 100 ns of simulation. The agreement between com-
n
puted JCH coupling constants and experimentally determined
ones, using a novel set of Karplus-type relationships51 denoted
3
JCX/SU09, is good for the J-couplings related to the ϕ torsion
angle whereas those related to the ψ torsion angle still deviate
(Table 3); future force field improvements will probably remedy
the remaining differences, in particular since there are now accu-
rate experimental data to compare to, besides the recently devel-
oped Karplus-type relationships for both 3JCH and 3JCC coupling
constants.
With the detailed information obtained for R2R we decided to
extend the study to larger structures with the aim of shedding
light on the 3D structures of larger oligosaccharides and polysac-
charides. Another common disaccharide structural component of
bacterial polysaccharides is α-L-Rhap-(1 → 3)-α-L-Rhap and a
corresponding 100 ns MD simulation was also carried out for
the methyl glycoside of this disaccharide, i.e., α-L-Rhap-(1 →
3)-α-L-Rhap-OMe (R3R); the scatter plot and glycosidic torsion
angle trajectories are shown in Fig. 7d–7f. Again the confor-
mational states denoted A, B and C are present (Fig. 7d), but the
populations in the conformational equilibrium are now compared
to R2R changed to 53 : 43 : 4, respectively, i.e., the two major
conformational states are approximately equally populated. The
plausibility of this population distribution was verified by com-
To investigate whether the intrinsic conformational popu-
lations at the α-(1 → 2)- and α-(1 → 3)-linkages were present
also for larger oligosaccharides we performed additionally an
MD simulation on a hexasaccharide with the use of explicit
water molecules. The structure of the hexasaccharide is shown in
Fig. 1. Analysis of the 100 ns MD simulation revealed that only
small conformational and population changes occurred (Fig. 9)
and this is summarized in Table 3 for torsion angles and hetero-
nuclear three-bond coupling constants compared to the disac-
charide constituents. A shift in the average conformations occurs
for some ψ torsion angles as a result of slightly larger popu-
lations of conformational states denoted B, having a positive ψ
torsion angle. Again some population of a non-exo-anomeric
conformation C is present at both the α-(1 → 2)- and α-(1 → 3)-
linkages.
3
puting the transglycosidic JCH coupling constants over the 100
ns simulation using the JCX/SU09 implementation leading to
The conformational preferences of the disaccharides and the
proposed torsion angle equilibria in the hexasaccharide can now
3
3JH1′,C3 = 3.7 Hz and JC1′,H3 = 4.7 Hz for the couplings related
2458 | Org. Biomol. Chem., 2012, 10, 2453–2463
This journal is © The Royal Society of Chemistry 2012

View Online
Table 3 Torsion angles, coupling constants (calculated with JCX/SU09) and distributions between the three states (A, B and C) from the 100 ns
simulations
MD hexasaccharide
MD disaccharide
3JC,C Torsion
Δhexa-di
Torsion
(°)
JC,H
(Hz) (Hz)
JC,H
3JC,C Torsion JC,X
Linkage
Distribution/% Torsion angle
(°)
(Hz) (Hz)
(°)
(Hz)
II → I
A 72 (90)a
ϕH
ψH
H1′–C1′–O2 –C2
C1′–O2 –C2 –H2
41 (15)
–23 (34)
4.0
3.8
39 (12)
–38 (26)
4.3
3.4
2
15
−0.3 ϕH
α-L-Rhap-(1 → B 26 (9)
2)-α-L-Rhap
0.4
ψH
C 2 (1)
ϕC2′ C2′–C1′–O2 –C2
ψC1 C1′ –O2 –C2 –C1 98 (33)
ψC3 C1′–O2 –C2 –C3
ϕO5′ O5′–C1′–O2 –C2
161 (14)
4.0b
1.5b
2.6
158 (11)
84 (25)
–157 (25)
–80 (12)
4.0b
1.2b
3.2
3
14
13
0.0
0.3
ϕC2′
ψC1
–144 (32)
−79 (15)
−0.6 ψC3
ϕO5′
III → II
α-L-Rhap-(1 → B 49 (43)
3)-α-L-Rhap
A 48 (53)
ϕH
ψH
H1′–C1′–O3 –C3
C1′–O3 –C3 –H3
44 (18)
0 (26)
3.7
5.2
41 (21)
–7 (32)
3.7
4.7
3
7
0.0
0.5
ϕH
ψH
C 3 (4)
ϕC2′ C2′–C1′–O3 –C3
ψC2 C1′ –O3 –C3 –C2 118(25)
ψC4 C1′–O3 –C3 –C4
ϕO5′ O5′–C1′–O3 –C3
163 (17)
4.0b
1.2
1.4
160 (21)
112 (30)
–128 (30)
–78 (21)
3.9b
1.2
1.8
3
6
6
0.1
0.0
−0.4 ψC4
ϕC2′
ψC2
−122 (24)
−76 (24)
ϕO5′
IV → III
A 35
ϕH
ψH
H1′–C1′–O3 –C3
C1′–O3 –C3 –H3
44 (20)
8 (27)
3.6
5.0
3
15
−0.1 ϕH
α-L-Rhap-(1 → B 62
3)-α-L-Rhap
C 3
0.3
ψH
ϕC2′ C2′–C1′–O3 –C3
ψC2 C1′ –O3 –C3 –C2 126 (26)
ψC4 C1′–O3 –C3 –C4
ϕO5′ O5′–C1′–O3 –C3
163(19)
4.0b
1.6
1.1
3
14
13
0.1
0.4
ϕC2′
ψC2
–115 (25)
−75 (19)
−0.7 ψC4
ϕO5′
V → IV
A 72
ϕH
ψH
H1′–C1′–O2 –C2
C1′–O2 –C2 –H2
42 (15)
–22 (32)
4.0
4.0
3
16
−0.3 ϕH
α-L-Rhap-(1 → B 26
2)-α-L-Rhap
C 2
0.6
ψH
ϕC2′ C2′–C1′–O2 –C2
ψC1 C1′ –O2 –C2 –C1 99 (30)
ψC3 C1′–O2 –C2 –C3
ϕO5′ O5′–C1′–O2 –C2
161 (14)
4.0b
1.4b
2.5
3
15
15
0.0
0.2
ϕC2′
ψC1
–142 (30)
−79 (15)
−0.7 ψC3
ϕO5′
VI → V
A 48
ϕH
ψH
H1′–C1′–O3 –C3
C1′–O3 –C3 –H3
41 (22)
–2 (27)
3.7
5.1
0
5
0.0
0.4
ϕH
ψH
α-L-Rhap-(1 → B 47
3)-α-L-Rhap
C 5
ϕC2′ C2′–C1′–O3 –C3
ψC2 C1′ –O3 –C3 –C2 117 (26)
ψC4 C1′–O3 –C3 –C4
ϕO5′ O5′–C1′–O3 –C3
160 (21)
3.9b
1.2
1.5
0
5
4
0.0
0.0
−0.3 ψC4
ϕC2′
ψC2
–124 (25)
−78 (21)
ϕO5′
a Distributions between the three states in the disaccharides are given in parentheses. b With constant in-plane effect.
be utilized to predict the 3D structure of shape of larger oligosac-
charides and subsequently polysaccharides. For a nonasaccharide
[→ 3)-α-L-Rhap-(1 → 3)-α-L-Rhap-(1 → 2)-α-L-Rhap-(1 → ]3
having ψ torsion angles corresponding to all-A conformations
may be contrasted to the one with all-B conformations (Fig. 10).
The all-B structure is conspicuously more extended. The herein
described all-B structure is similar to that previously reported;53
however, based on the results presented above in this study
neither the all-A nor the all-B conformations for the nonasac-
charide are populated to any significant extent. The 3D structure
should therefore be represented at each glycosidic linkage by a
conformational equilibrium of two significantly populated con-
formations having either a positive or a negative value at the ψ
torsion angles. Thus, the description of the canonical all-A and
all-B structures of the nonasaccharide may be useful for visual-
ization, but its 3D structure is dynamic in between the idealized
all-A and all-B conformations.
In conclusion, synthesis of 13C site-specific labeling in R2R
n
n
has facilitated the extraction of both JCH and JCC coupling
constants by 1D and 2D NMR experiments as well as by total
line-shape analysis. Coupling constants having torsion angle
dependences were used in the conformational analysis of the
disaccharide. For the disaccharide the fourteen hitherto exper-
imentally determined NMR parameters in isotropic solution cor-
respond to an unusually large number of conformationally
dependent data. The MD simulation of R2R, in conjunction with
experimental NMR data, indicates a major conformational state
having the most probable conformation at ϕH ≈ 40° and ψH
≈
−35°. The knowledge of conformational preferences and accessi-
ble conformational states of disaccharide models facilitate analy-
sis of larger oligosaccharides and polysaccharides containing
L-rhamnose residues, such as the nonasaccharide presented
which is proposed to exist as a quite flexible oligosaccharide
since a conformational equilibrium at each ψ torsion angle is
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2453–2463 | 2459

View Online
Fig. 10 Space-filling representation of the rhamnan nonasaccharide [→
3)-α-L-Rhap-(1 → 3)-α-L-Rhap-(1 → 2)-α-L-Rhap-(1 → ]3 in the all-A
(left) and the all-B (right) conformations.
Davisil silica medium (35–70 micron) was used. NMR spec-
troscopy was performed on Varian 400 MHz or Bruker AVANCE
500 MHz spectrometers at ambient temperature except otherwise
stated. The 1H and 13C chemical shifts are referenced to the
solvent CDCl3 (δH 7.26, δC 77.23) or external TSP in D2O (δH
0.00) and external dioxane in D2O (δC 67.4) for D2O-solutions.
High-resolution mass spectrometry analysis was performed on a
Bruker Daltonics ESI-TOF spectrometer in the positive mode.
The detailed synthesis procedures are described for non-labeled
materials; the 13C-labeled compounds were obtained under the
same conditions.
Glycosidic torsion angles are defined as follows: ϕ = O5′–
C1′–On–Cn, ϕH = H1′–C1′–On–Cn, ϕC2′ = C2′–C1′–On–Cn, ψ =
C1′–On–Cn–C(n−1), ψH = C1′–On–Cn–Hn, ψC3 = C1′–O2–
C2–C3, and ψC4 = C1′–O3–C3–C4, where n is the substitution
position and atoms in the non-reducing end residue of the glyco-
sidic linkage are denoted by a prime. The regions of confor-
Fig. 9 Representative scatter plots of the glycosidic torsion angels ϕH
and ψH in the hexasaccharide; (a) between residues IV and III, in an α-(1
→ 3)-linkage, and (b) between residues V and IV, in an α-(1 → 2)-
linkage.
present leading to a dynamic 3D structure where the bounds are
the canonical all-A and all-B conformations.
mational space spanning −180° < ϕH < 180° and −180° < ψH
180° are defined by the following states, A: ψH < 0° and ψH
6ϕH; B: ψH > 0° and ψH > − ϕH; C: ϕH < 0° and ψH < − ϕH and
ψH > 6ϕH.
<
<
Materials and methods
General
The site-specifically 13C-labeled starting materials, L-[1-13C]
rhamnose and L-[2-13C]rhamnose, both 99 atom% 13C, were pur-
chased from OMICRON Biochemicals, Inc., USA. TLC analysis
was performed on precoated Merck 60 F254 plates and developed
with 8% H2SO4. Solvents specified as dry were stored over mol-
ecular sieves and used without further purification. In the glyco-
sylation reactions dichloromethane from Fluka (puriss, absolute,
kept over molecular sieves) was used to avoid contamination
from ethanol. For purification via column chromatography
Experimental
2,3,4-Tri-O-benzoyl-α-L-rhamnopyranosyl bromide (2)
L-Rhamnose (254 mg, 1.4 mmol) was dissolved in dry pyridine
(20 mL) and benzoyl chloride (1.0 mL, 6 eq.) was added at
0 °C. The solution was stirred at 60 °C for 5 h and monitored by
TLC (Toluene : Ethyl acetate (T : E) 4 : 1). When complete, the
reaction mixture was allowed to attain r.t.; water (0.2 mL) was
2460 | Org. Biomol. Chem., 2012, 10, 2453–2463
This journal is © The Royal Society of Chemistry 2012

View Online
added and after 15 min the mixture was diluted with CH2Cl2.
The organic phase was washed with cold water, H2SO4 (1 M),
sat. NaHCO3 (aq.) and then water again. The organic phase was
dried over MgSO4, the solvent was evaporated under reduced
pressure and subsequently co-evaporated four times with
toluene. The obtained yellow amorphous material (750 mg,
1.3 mmol) was dissolved in 10 mL glacial HOAc at r.t. To the
solution was added 6 mL HBr in HOAc (12 mM) and the
mixture was stirred at r.t. When the reaction, monitored by TLC
(T : E 6 : 1), was complete after 6 h, ice and CH2Cl2 was added
and the organic phase was washed with ice/water, sat. NaHCO3
(aq.) and subsequently water. The organic solution was dried
over Na2SO4 and subsequently concentrated under reduced
pressure. Compound 2 was obtained in 84% yield (647 mg,
resonances, CDCl3) δ: 1.36 (H6′), 1.41 (H6), 4.12 (H5), 4.31
(H2), 4.34 (H5′), 4.90 (H1, JH1,H2 1.8 Hz), 5.16 (H1′, JH1′,H2′
1.8 Hz), 5.66 (2H, H4, H4′), 5.79 (H3), 5.86 (H2′), 5.97 (H3′),
7.2–8.1 (25H, aromatic H); 13C NMR (CDCl3) δ: 17.9 (2C, C6,
C6′), 55.4 (OMe), 67.0 (C5), 67.8 (C5′), 69.9 (C3′), 70.8 (C2′),
71.3 (C3), 72.0 (2C, C4, C4′), 76.8 (C2), 99.7 (C1′), 100.0 (C1),
128.4–133.6 (aromatic C), 165.3–166.1 (5 × CO).
Methyl 2-O-α-L-rhamnopyranosyl-α-L-rhamnopyranoside (1)
Compound 4 (176 mg, 0.21 mmol) was dissolved in NaOMe–
MeOH (15 mL, 0.1 M) and the solution was stirred at r.t. for
2.5 h. The reaction was monitored by TLC (E : MeOH : H2O
7 : 2 : 1). To neutralize, Dowex-50 (H+-form) was added until the
pH stabilized to approx. 7. The solution was filtered through a
glass filter funnel, then passed through a short plug of Bio-Gel
P-2 and freeze-dried. The crude material was purified on a
column of Bio-Gel P-2 using water containing 1% n-butanol as
eluent. Compound 1 was obtained in 79% yield (54 mg,
1
1.2 mmol). H NMR (selected resonance, CDCl3) δ: 6.20 (H1);
13C NMR (CDCl3) δ: 17.4 (Me), 69.1–73.6 (C2–C5), 84.0 (C1),
128.6–134.0 (aromatic C), 165.4–165.9 (3 × CO).
Methyl 3,4-di-O-benzoyl-α-L-rhamnopyranoside (3)
1
0.17 mmol) after lyophilization. H and 13C NMR spectral data
were in agreement with those previously published.38
A solution of methyl 4-O-benzoyl-2,3-O-isopropylidene-α-L-
rhamnopyranoside (1.1 g, 3.4 mmol) in acetic acid (40 mL, 70%
aq.) was heated to 80 °C. After 2.5 h, the mixture was allowed to
attain r.t., the solvent was evaporated and the resulting mixture
co-evaporated three times with toluene. The oily residue was dis-
solved in a mixture of dry pyridine and dry CH2Cl2 (1 : 2) and
cooled to −40 °C on a dry ice–acetone bath. Benzoyl chloride
(0.35 mL, 1 eq.) was dissolved in dry pyridine : dry CH2Cl2
(1 : 2) and added dropwise to the mixture. The solution was then
stirred at r.t. for 1 h, monitored by TLC (T : E 2 : 1), after which
water was added. The mixture was diluted with CH2Cl2 and the
organic phase washed with cold water, H2SO4 (1 M), sat.
NaHCO3 (aq.) and then water again. The organic phase was
dried over Na2SO4 and the solvent was evaporated. The crude
material was purified by flash chromatography on a column of
silica (gradient, T → T : E 4 : 1). Compound 3 was obtained in
69% yield (889 mg, 2.3 mmol). ESI-MS: m/z [M + Na]+
409.127, expected 409.126. 1H NMR (selected resonance,
CDCl3) δ: 4.80 (H1, JH1,H2 1.8 Hz); 13C NMR (CDCl3) δ: 17.8
(Me), 55.4 (OMe), 66.6–72.8 (C2–C5), 100.8 (C1), 128.6–133.5
(aromatic C), 165.8–166.0 (2 × CO).
α-L-[1′-13C]Rhap-(1 → 2)-α-L-Rhap-OMe (1-c1′)
Yield 60% (over 3 steps). ESI-MS: m/z [M + Na]+ 348.1338,
1
expected 348.1346. H NMR (selected resonance, D2O, 37 °C)
δ: 4.96 (H1′, JH1′,H2′ 1.8 Hz, JH1′,C1′ 170 Hz).
α-L-[2′-13C]Rhap-(1 → 2)-α-L-Rhap-OMe (1-c2′)
Yield 66% (over 3 steps). ESI-MS: m/z [M + Na]+ 348.1350,
expected 348.1346 1H NMR (selected resonance, D2O, 37 °C) δ:
4.96 (H1′, JH1′,H2′ 1.8 Hz).
NMR spectroscopy
NMR samples of α-L-Rhap-(1 → 2)-α-L-Rhap-OMe (1), α-L-
[1′-13C]Rhap-(1 → 2)-α-L-Rhap-OMe (1-c1′) and α-L-[2′-13C]
Rhap-(1 → 2)-α-L-Rhap-OMe (1-c2′) were prepared in D2O (pD
6) to concentrations of <100 mM. NMR experiments were
recorded at 37 °C on Bruker AVANCE 500 MHz and Bruker
AVANCE III 700 MHz spectrometers; both equipped with 5 mm
TCI Z-Gradient CryoProbes. The chemical shifts are referenced
to external sodium 3-trimethylsilyl-(2,2,3,3-2H4)-propanoate
(TSP) in D2O (δH 0.00) and external dioxane in D2O (δC 67.4).
The 13C NMR experiments were recorded over 109.2 ppm with
76′920 points and the FIDs zero-filled to 2048k data points. To
Methyl 3,4-di-O-benzoyl-2-O-(2,3,4-tri-O-benzoyl-α-L-
rhamnopyranosyl)-α-L-rhamnopyranoside (4)
A flask with 2 (245 mg, 0.45 mmol), 3 (275 mg, 1.6 eq.) and
molecular sieves (4 Å) was placed under vacuum over night.
The solids were dissolved in dry CH2Cl2, sym-collidine (approx.
50 mg, 1 eq.) was added and the suspension stirred at r.t. for
15 min under argon atmosphere. The mixture was cooled to
−40 °C and AgOTf (125 mg, 1.1 eq.) was added in one portion.
The mixture was stirred for 2 h and monitored by TLC (T : E
10 : 1) during which the temperature was allowed to rise to r.t.
and pyridine was added to quench the reaction. The mixture was
filtered through a pad of Celite and subsequently concentrated
under reduced pressure. The crude material was purified on a
column of silica (gradient, T → T : E 30 : 1). Compound 4 was
measure 13C,13C coupling constants
a Lorentz–Gaussian
window function (lb = −0.5 to −1.5, gb = 0.5) was applied. The
1H,13C-HSQC-HECADE experiment44 was recorded with 160
scans per t1-increment, a DIPSI-2 spin-lock and a mixing time
of 30 ms. A 2D matrix of 256 × 2k data points in F1 and F2
dimensions, respectively, a t1Ã=t1 scaling factor of unity and the
echo-antiecho method were used. Prior to Fourier transformation
zero-filling was carried out to 4k × 8k data points and 90°
shifted squared sine-bell functions were applied in both dimen-
sions. For compound 1 the 1H,13C-J-HMBC NMR experiments45
were recorded as described55,56 using five different values of the
1
obtained in 82% yield (311 mg, 0.37 mmol). H NMR (selected
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2453–2463 | 2461

View Online
1
6 K. Steiner, R. Novotny, D. B. Werz, K. Zarschler, P. H. Seeberger,
A. Hofinger, P. Kosma, C. Schäffer and P. Messner, J. Biol. Chem., 2008,
283, 21120–21133.
7 A. Turska-Szewczuk, A. Kozinska, R. Russa and O. Holst, Carbohydr.
Res., 2010, 345, 680–684.
scaling factor κ between 10 and 34. H NMR chemical shifts
n
and J coupling constants were refined from 1D spectra using
the PERCH NMR software (PERCH Solutions Ltd., Kuopio,
Finland).
8 A. Molinaro, M.-A. Newman, R. Lanzetta and M. Parrilli, Eur. J. Org.
Chem., 2009, 5887–5896.
9 P.-E. Jansson, L. Kenne and T. Wehler, Carbohydr. Res., 1987, 166, 271–
282.
Computer simulations
10 J.-R. Brisson, H. Baumann, A. Imberty, S. Pérez and H. J. Jennings, Bio-
chemistry, 1992, 31, 4996–5004.
11 M. Michela Corsaro, C. De Castro, T. Naldi, M. Parrilli, J. M. Tomás and
M. Regué, Carbohydr. Res., 2005, 340, 2212–2217.
12 I. Sánchez-Medina, M. Frank, C.-W. von der Lieth and J. P. Kamerling,
Org. Biomol. Chem., 2009, 7, 280–287.
13 B. Coxon, N. Sari, G. Batta and V. Pozsgay, Carbohydr. Res., 2000, 324,
53–65.
14 M.-J. Clément, A. Imberty, A. Phalipon, S. Pérez, C. Simenel,
L. A. Mulard and M. Delepierre, J. Biol. Chem., 2003, 278, 47928–
47936.
15 V. Pozsgay, N. Sari and B. Coxon, Carbohydr. Res., 1998, 308, 229–238.
16 G. Batta and A. Lipták, J. Chem. Soc., Chem. Commun., 1985, 368–370.
17 A. Bax and R. Freeman, J. Am. Chem. Soc., 1982, 104, 1099–1100.
18 C. Bauer, R. Freeman and S. Wimpers, J. Magn. Reson., 1984, 58, 526–
532.
19 J.-M. Nuzillard and J.-M. Bernassau, J. Magn. Reson., Ser. B, 1994, 103,
284–287.
20 L. Poppe, S. Sheng and H. van Halbeek, J. Magn. Reson., Ser. A, 1994,
111, 104–107.
21 T. Nishida, G. Widmalm and P. Sándor, Magn. Reson. Chem., 1995, 33,
596–599.
22 R. T. Williamson, B. L. Márquez, W. H. Gerwick and K. E. Kövér,
Magn. Reson. Chem., 2000, 38, 265–273.
23 T. Parella and J. Belloc, J. Magn. Reson., 2001, 148, 78–87.
24 R. T. Williamson, B. L. Marquez, W. H. Gerwick, G. E. Martin and
V. V. Krishnamurthy, Magn. Reson. Chem., 2001, 39, 127–132.
25 D. Uhrín, J. Magn. Reson., 2002, 159, 145–150.
26 P. Vidal, N. Esturau, T. Parella and J. F. Espinosa, J. Org. Chem., 2007,
72, 3166–3170.
For the molecular dynamics (MD) simulations CHARMM57
software was used employing a CHARMM22 type of force field
modified for carbohydrates and referred to as PARM22/SU01.37
Initial conditions were prepared by placing each disaccharide in
a cubic water box of length 29.97 Å containing 900 modified
TIP3P water molecules.58 The hexasaccharide was placed in a
water box of length 50 Å containing 3921 modified TIP3P water
molecules. The solvent molecules that were closer than 2.5 Å to
any solute atom were removed resulting in 868 water molecules
for R2R, 864 water molecules for R3R, and 3838 water mol-
ecules for the hexasaccharide. Energy minimization was per-
formed with Steepest Descent, 200 steps, followed by Adopted
Basis Newton–Raphson until the root-mean-square gradient was
less than 0.01 kcal·mol−1·Å−1. The initial velocities were
assigned at 105 K, followed by heating with 5 K increments
during 8 ps to 310 K, where the systems were equilibrated for 1
ns. The production runs were performed for 100 ns at 310 K.
Simulations were carried out at the Center for Parallel Compu-
ters, KTH, Stockholm, using 1 node with two quad-core pro-
cessors per node. For the disaccharides parallel version C34b2
was used and for the hexasaccharide version C35b4 was used.
The CPU time was approximately 2.5 h per ns for the disacchar-
ides and 6.5 h per ns for the hexasaccharide.
The potential energy maps were computed from a 15° grid
search over the entire glycosidic torsion angle space using a
dielectric constant of 3. Energy minimization was performed
with Steepest Descent, 50 steps, followed by Adopted Basis
Newton–Raphson until the root-mean-square gradient was less
than 0.01 kcal·mol−1·Å−1. The nonasaccharide was built by
CarbBuilder59 as part of CASPER60 and visualization of oligo-
saccharides was made using PyMOL (The PyMOL Molecular
Graphics System, Version 0.99rc6, Schrödinger, LLC).
27 G. Widmalm, R. A. Byrd and W. Egan, Carbohydr. Res., 1992, 229,
195–211.
28 B. J. Hardy, W. Egan and G. Widmalm, Int. J. Biol. Macromol., 1995, 17,
149–160.
29 B. J. Hardy, S. Bystricky, P. Kovac and G. Widmalm, Biopolymers, 1997,
41, 83–96.
30 C. Landersjö, B. Stevensson, R. Eklund, J. Östervall, P. Söderman,
G. Widmalm and A. Maliniak, J. Biomol. NMR, 2006, 35, 89–101.
31 P. Söderman, S. Oscarson and G. Widmalm, Carbohydr. Res., 1998, 312,
233–237.
32 R. K. Ness, G. F. Hewitt Jr. and C. S. Hudson, J. Am. Chem. Soc., 1951,
73, 296–300.
33 H. Rainer, H.-D. Scharf and J. Runsink, Liebigs Ann. Chem., 1992, 103–
107.
34 T. Norberg, S. Oscarson and M. Szöni, Carbohydr. Res., 1986, 152, 301–
304.
35 P. J. Garegg and T. Norberg, Acta Chem. Scand., Ser. B, 1979, 33, 116–
118.
36 H. Thøgersen, R. U. Lemieux, K. Bock and B. Meyer, Can. J. Chem.,
1982, 60, 44–57.
Acknowledgements
This work was supported by grants from the Swedish Research
Council (VR), The Knut and Alice Wallenberg Foundation, and
Carl Tryggers Stiftelse för Vetenskaplig Forskning. The Center
for Parallel Computers (PDC), Stockholm, Sweden, is thanked
for computing resources.
37 R. Eklund and G. Widmalm, Carbohydr. Res., 2003, 338, 393–398.
38 P.-E. Jansson, L. Kenne and G. Widmalm, Acta Chem. Scand., 1991, 45,
517–522.
39 T. Nishida, G. Widmalm and P. Sándor, Magn. Reson. Chem., 1996, 34,
377–382.
40 T. E. Klepach, I. Carmichael and A. S. Serianni, J. Am. Chem. Soc.,
2005, 127, 9781–9793.
41 I. Carmichael and A. S. Serianni, Carbohydr. Res., 1996, 280, 177–186.
42 F. Cloran, I. Carmichael and A. S. Serianni, J. Am. Chem. Soc., 2000,
122, 396–397.
43 W. Koźmiński and D. Nanz, J. Magn. Reson., 1997, 124, 383–392.
44 W. Koźmiński and D. Nanz, J. Magn. Reson., 2000, 142, 294–299.
45 A. Meissner and O. W. Sørensen, Magn. Reson. Chem., 2001, 39,
49–52.
References
1 Essentials of Glycobiology ed. A. Varki, R. Cummings, J. Esko,
H. Freeze, G. Hart and J. Marth, Cold Spring Harbor Laboratory Press,
Cold Spring Harbor, 1999.
2 B. Lindberg, in Polysaccharides ed. S. Dumitriu, Marcel Dekker,
New York, 1998, pp. 237–273.
3 N. I. A. Carlin and T. Wehler, Eur. J. Biochem., 1988, 176, 471–476.
4 J. Kulber-Kielb, E. Vinogradov, C. Chu and R. Schneerson, Carbohydr.
Res., 2007, 342, 643–647.
46 R. Laatikainen, M. Niemitz, U. Weber, J. Sundelin, T. Hassinen and
J. Vepsäläinen, J. Magn. Reson., Ser. A, 1996, 120, 1–10.
5 M. Ansaruzzaman, M. J. Albert, T. Holme, P.-E. Jansson, M. M. Rahman
and G. Widmalm, Eur. J. Biochem., 1996, 237, 786–791.
2462 | Org. Biomol. Chem., 2012, 10, 2453–2463
This journal is © The Royal Society of Chemistry 2012

View Online
47 G. Widmalm, in Comprehensive Glycoscience, ed. J. P. Kamerling, Else-
vier, Oxford, 2007, pp. 101–132.
55 E. Säwén, M. U. Roslund, I. Cumpstey and G. Widmalm, Carbohydr.
Res., 2010, 345, 984–993.
48 I. Carmichael, D. M. Chipman, C. A. Podlasek and A. S. Serianni, J. Am.
Chem. Soc., 1993, 115, 10863–10870.
56 K. H. M. Jonsson, R. Pendrill and G. Widmalm, Magn. Reson. Chem.,
2011, 49, 117–124.
49 I. Tvaroska and F. R. Taravel, Carbohydr. Res., 1991, 221, 83–94.
50 P. Berthault, D. Jeannerat, F. Camerel, F. A. Salgado, Y. Boulard, J.-C.
P. Gabriel and H. Desvaux, Carbohydr. Res., 2003, 338, 1771–1785.
51 E. Säwén, T. Massad, C. Landersjö, P. Damberg and G. Widmalm, Org.
Biomol. Chem., 2010, 8, 3684–3695.
52 E. L. Zdorovenko, G. V. Zatonsky, G. M. Zdorovenko, L. A. Pasichnyk,
A. S. Shashkov and Y. A. Knirel, Carbohydr. Res., 2001, 336, 329–336.
53 E. Bedini, C. De Castro, G. Erbs, L. Mangoni, J. M. Dow,
A.-M. Newman, M. Parrilli and C. Unverzagt, J. Am. Chem. Soc., 2005,
127, 2414–2416.
57 B. R. Brooks, C. L. Brooks III, A. D. MacKerell Jr, L. Nilsson,
R. J. Petrella, B. Roux, Y. Won, G. Archontis, C. Bartels, S. Boresch,
A. Caflisch, L. Caves, Q. Cui, A. R. Dinner, M. Feig, S. Fischer, J. Gao,
M. Hodoscek, W. Im, K. Kuczera, T. Lazaridis, J. Ma, V. Ovchinnikov,
E. Paci, R. W. Pastor, C. B. Post, J. Z. Pu, M. Schaefer, B. Tidor, R.
M. Venable, H. L. Woodcock, X. Wu, W. Yang, D. M. York and
M. Karplus, J. Comput. Chem., 2009, 30, 1545–1614.
58 E. Neria, S. Fischer and M. Karplus, J. Chem. Phys., 1996, 105, 1902–
1921.
59 M. Kuttel, Y. Mao, G. Widmalm and M. Lundborg, Proc. 7th IEEE Int.
Conf. e-Science 2011, 395–402.
54 K. Lycknert, A. Helander, S. Oscarson, L. Kenne and G. Widmalm,
Carbohydr. Res., 2004, 339, 1331–1338.
60 M. Lundborg and G. Widmalm, Anal. Chem., 2011, 83, 1514–1517.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2453–2463 | 2463