Effect of the substituents of the neighboring ring in the conformational equilibrium of iduronate in heparin-like trisaccharides

Article abstract of DOI:10.1002/chem.201202770

Based on the structure of the regular heparin, we have prepared a smart library of heparin-like trisaccharides by incorporating some sulfate groups in the sequence α-D-GlcNS- (1-4)-α-L-Ido2S-(1-4)-α-D-GlcN. According to the 3Dstructure of heparin, which features one helix turn every four residues, this fragment corresponds to the minimum binding motif. We have performed a complete NMR study and found that the trisaccharides have a similar 3Dstructure to regular heparin itself, but their spectral properties are such that allow to extract very detailed information about distances and coupling constants as they are isotropic molecules. The characteristic conformational equilibrium of the central iduronate ring has been analyzed combining NMR and molecular dynamics and the populations of the conformers of the central iduronate ring have been calculated. We have found that in those compounds lacking the sulfate group at position6 of the reducing end glucosamine, the population of ²S₀ of the central iduronate residue is sensitive to the temperature decreasing to 19 % at 278K. On the contrary, the trisaccharides with 6-O-sulfate in the reducing end glucosamine keep the level of population constant with temperature circa 40% of ²S₀ similar to that observed at room temperature. Another structural feature that has been revealed through this analysis is the larger flexibility of the L-IdoAS- D-GlcN glycosidic linkage, compared with the D-GlcNS-L-IdoA. We propose that this is the point where the heparin chain is bended to form structures far from the regular helix known as kink that have been proposed to play an important role in the specificity of the heparin-protein interaction. The sulfation pattern of the neighboring ring drives the conformational equilibrium of iduronate in heparin trisaccharides. A library of heparin-like trisaccharides incorporating sulfate groups in key positions has been prepared. The conformational equilibrium of the central iduronate ring has been analyzed and the populations were calculated by combining NMR and molecular dynamics (see figure).

Full text of DOI:10.1002/chem.201202770

FULL PAPER
DOI: 10.1002/chem.201202770
Effect of the Substituents of the Neighboring Ring in the Conformational
Equilibrium of Iduronate in Heparin-like Trisaccharides
Juan Carlos MuÇoz-Garcꢀa,[a] Javier Lꢁpez-Prados,[a] Jesffls Angulo,[a]
Irene Dꢀaz-Contreras,[a] Niels Reichardt,[b] JosØ L. de Paz,[a] Manuel Martꢀn-Lomas,[b] and
Pedro M. Nieto*[a]
Abstract: Based on the structure of the
regular heparin, we have prepared a
smart library of heparin-like trisacchar-
ides by incorporating some sulfate
groups in the sequence a-d-GlcNS-
conformational equilibrium of the cen-
tral iduronate ring has been analyzed
combining NMR and molecular dy-
namics and the populations of the con-
formers of the central iduronate ring
have been calculated. We have found
that in those compounds lacking the
sulfate group at position 6 of the reduc-
ing end glucosamine, the population of
2S0 of the central iduronate residue is
sensitive to the temperature decreasing
to 19 % at 278 K. On the contrary, the
trisaccharides with 6-O-sulfate in the
reducing end glucosamine keep the
level of population constant with tem-
perature circa 40 % of 2S0 similar to
that observed at room temperature.
Another structural feature that has
been revealed through this analysis is
the larger flexibility of the l-IdoAS-
d-GlcN glycosidic linkage, compared
with the d-GlcNS-l-IdoA. We propose
that this is the point where the heparin
chain is bended to form structures far
from the regular helix known as kink
that have been proposed to play an im-
portant role in the specificity of the
heparin–protein interaction.
A
Ac-
cording to the 3D structure of heparin,
which features one helix turn every
four residues, this fragment corre-
sponds to the minimum binding motif.
We have performed a complete NMR
study and found that the trisaccharides
have a similar 3D structure to regular
heparin itself, but their spectral proper-
ties are such that allow to extract very
detailed information about distances
and coupling constants as they are iso-
tropic molecules. The characteristic
Keywords: carbohydrates · confor-
mation analysis · glycosaminogly-
cans · molecular dynamics · NMR
spectroscopy
Introduction
an enzymatic cascade that results in an anticoagulant activi-
ty.[3] Many other proteins require interactions with cell sur-
face glycosaminoglycans (GAGs) to exert their biologic ac-
tivity.[4] The effect of GAG binding on protein function
ranges from essential roles in development, cell growth, cell
adhesion, inflammation, tumorigenesis, and interactions with
pathogens.[5] Many other examples are not as evident and
when the interaction was treated in terms of pharmaco-
phore, it became imprecise and some authors have suggest-
ed a continuous range of affinities in function of the substi-
tution pattern or the population of a particular conforma-
tion of iduronate, or the presence of a kink in the carbohy-
drate chain.[6]
From the study of several well-known proteins, it is
known that the interactions of heparin or heparan sulfate
with proteins, apparently, do not include large conforma-
tional changes in the structure of the carbohydrate. There-
fore, owing to its nature and topology—a rigid helix with a
complete turn every four residues—[7] it can be expected
that the interactions with the same side of a protein surface
should be discontinued, grouping each three contiguous resi-
dues (Figure 1).[1b,8] Additionally, the number and distribu-
tion of sulfate groups should play some role in the specifici-
ty of the interaction.[9] Another structural key aspect of the
heparin or HS structure is that while it is very rigid from the
Heparin and heparan sulfate are glycosaminoglycans
(GAG) composed by an uronic acid and a glucosamine 1,4-
linked and strongly substituted by sulfate groups that regu-
late the activity of hundreds of proteins.[1] By means of
these interactions the activity of the proteins is regulated
and is able to exert a function. The best known example is
the interaction with antithrombin, in which a specific and
rare sequence of heparin is responsible for the interaction.[2]
The interaction occurs in a very selective mode, triggering
[a] J. C. MuÇoz-Garcꢀa,+ Dr. J. Lꢁpez-Prados,+ Dr. J. Angulo,
I. Dꢀaz-Contreras, Dr. J. L. d. de Paz, Dr. P. M. Nieto
Glycosystems Laboratory
Instituto de Investigaciones Quimicas, CSIC - US
Americo Vespucio, 49, 41092 Sevilla (Spain)
Fax : (+34)954-46-01-65
[b] Dr. N. Reichardt, Prof. M. Martꢀn-Lomas
Laboratory of Glyconanotechnology
CIC-BiomaGUNE, CIBER-BBN
Paseo Miramꢁn 182, Parque Tecnolꢁgico
20009 San Sebastiꢂn (Spain)
[+] These authors have contributed equally to this work
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202770.
Chem. Eur. J. 2012, 18, 16319 – 16331
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
16319

or shifted by two residues in a parallel fashion (Figure 3).
Therefore the complementary trisaccharide IdoA-GlcN-
IdoA should not be able to interact with a single binding
site within the protein, but could establish a bridge between
two receptors, as frequently found by X-ray crystallography.
Starting from the basic trisaccharide d-GlcN2S-l-IdoA2S-
d-GlcN, and assuming that the two sulfate groups are driv-
ing the main interaction, we have permutated the rest of the
potential sulfation sites except the position 3 of glucosamine
that has only been reported in the interaction with antith-
rombin.[12] We have replaced the sulfate groups by an un-
charged substituent present in the general structure of hepa-
rin of heparan sulfate. Thus, the N sulfamate in position 2
was substituted by an N-acetamido group as it is more fre-
quent in heparin than the amino group, additionally it does
not change the overall charge of the system, while the sul-
fate in position 6 of glucosamine was not replaced, thus
leaving an hydroxyl group. We have decided not to permu-
tate the carboxylate group of the iduronate because this re-
placement has not been reported in the regular region of
heparin and it is a key element in the control of the confor-
mational equilibrium of the iduronate ring.
Figure 1. Heparin structures calculated based on NMR restrains of a
dodecasaccharide representative of the regular region of heparin with all
1
2
the iduronates in C4 (top) and in S0 (bottom). Each group of three con-
tiguous sulfate groups is marked.[7]
backbone perspective, it is also quite flexible when the con-
formations of iduronate are considered. The internal iduro-
nate residues are in equilibrium between two conformations,
1
2
the chair C4 and the boat S0.[7,10] Such equilibrium, howev-
er, does not affect to the overall shape of the molecule,
which remains linear with the same helical pitch.[8] This is
because in spite of the apparent differences between confor-
mations, the relative orientations of the glycosidic bonds
linkages are very similar. The final result of the conforma-
tional equilibrium correspond
Herein we describe the synthesis, NMR analysis, and the
MD simulations of the library of trisaccharides. We have
found some interesting effects of the temperature on the
equilibrium of the central iduronate residue, as well as selec-
to the interchange in the rela-
tive positions of carbons C2
and C3 up and down the main
plane of the ring, while the
atoms in positions C1, O6, C5,
and C4 remain in the same rel-
ative positions. Thus, when C2
is up the mean plane of the
ring and C3 is down the con-
formation is the chair 1C4
while when the relative posi-
tions are the opposite the con-
2
formation is the skew boat S0.
Based on these premises we
have designed and synthesized
a small library of eight trisac-
charides, Tri1–Tri8 (Figure 2).
Assuming that the heparin
3D structure (Figure 1) has a
1808 turn every two disacchar-
ides,[7] a trisaccharide GlcN-
IdoA-GlcN
would
cluster
three key sulfate groups
toward the same side of the
molecule originating a mini-
mum recognition site,[11] while
a
potential adjacent second
site is necessarily directed to-
wards the opposite direction,
in an antiparallel orientation,
Figure 2. Library of heparin trisaccharides drawn displaying the 3D relative disposition up or down of the sub-
stituents. Nomenclature A, B, C is used for definition of the rings throughout the text.
16320
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16319 – 16331

Conformational Equilibrium of Heparin-Like Trisaccharides
FULL PAPER
Scheme 1. Retrosynthetic analysis.
Figure 3. Trisaccharide Tri1 in 1C4 chair conformation (up left) and in 2S0
skew boat conformation (up right), and complementary trisaccharide in
1C4 (down left) and in 2S0 (down right).
tal moiety in 9 as described above yielded the desired build-
ing blocks 3 and 4.
Trisaccharides 10, 11,[16] 12, and 13 were obtained in good
yield by glycosylation of the monosaccharide building
blocks 1–4 with the known disaccharide trichloroacetimidate
5 (Scheme 4). These fully protected intermediates were sub-
mitted to the deprotection/sulfation reaction sequence to
give the four heparin trisaccharides that do not contain a
sulfate group at position 6 of the non-reducing end.
tive hydration shell depending on the conformation of such
residue.
Results
The removal of methoxycarbonyl and acyl groups was
performed with lithium hydroperoxide and then aqueous/al-
coholic potassium hydroxide to minimize the b-elimination
of iduronic ester.[17] The resulting partially protected trisac-
charides 14–17 were O-sulfated by treatment with fresh
SO3·Me3N complex in dimethylformamide (DMF) at 508C.
Staudinger reduction of the azide protecting groups in 18–21
with Me3P and NaOH, followed by N-sulfation with fresh
SO3·Py complex afforded trisaccharides 22–25 in excellent
yield.[18] Reverse phase C-18 chromatography, using an
aqueous Et3NHOAc buffer (10 mm) at pH 7 as mobile phase
with a MeOH gradient,[19] was required to obtain high purity
Synthesis of a library of heparin-like trisaccharides: We car-
ried out the synthesis of the trisaccharide library following
the retrosynthetic analysis shown in Scheme 1. Monosac-
charide building blocks 1–4 were glycosylated with disac-
charide donor 5[13] to obtain the fully protected intermedi-
ates required for the synthesis of the final sulphated trisac-
charides.
Building blocks 1 and 2 containing a 2-acetamido func-
tionality were prepared from N-acetyl-d-glucosamine
(Scheme 2). Fischer glycosylation with isopropyl alcohol fol-
lowed by benzylidenation afforded 6 in good yield. Benzyla-
tion of the 3-hydroxyl group
gave 7, which was then submit-
ted to reductive benzylidenea-
cetal opening with NaCNBH3
and HCl[14] to obtain building
block 1. Alternatively, hydroly-
sis of the benzylideneacetal in
Scheme 2. Synthesis of building blocks 1 and 2. a) i) Amberlite IR-120, iPrOH, 708C, 48 h, ii) benzaldehyde di-
methyl acetal, pTsOH (cat.), DMF, 508C, 2 h, 77%. b) BnBr, TBAI, NaH, CH2Cl2, RT, 24 h, 86%.
c) NaCNBH3, HCl, THF, 4 ꢄ M.S., RT, 10 min, 90%. d) i) TFA, H2O, CH2Cl2, 1 h, RT, ii) BzCl, Et3N, CH3CN,
ꢀ308C, 2 h, 86%.
7, and then selective 6-OH
benzoylation gave building
block 2.
The synthesis of the two
monosaccharidic
blocks containing
functional group,
building
2-azido
and 4,
a
3
started with the glycosylation
of the known trichloroacetimi-
date 8[15] with isopropanol
under standard conditions
(Scheme 3). Further transfor-
mation of the benzylideneace-
Scheme 3. Synthesis of building blocks 3 and 4. a) iPrOH, TMSOTf, CH2Cl2, 08C, 30 min, 70%. b) NaCNBH3,
HCl, 4 ꢄ M.S., THF, RT, 15 min, 92%. c) i) TFA, H2O, CH2Cl2, RT, 30 min, ii) BzCl, Et3N, CH3CN, ꢀ308C,
2 h, 86%.
Chem. Eur. J. 2012, 18, 16319 – 16331
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16321

P. M. Nieto et al.
Scheme 4. Synthesis of trisaccharides Tri3, Tri7, Tri4, and Tri8. a) TMSOTf, CH2Cl2, 4 ꢄ M.S., RT, 1 h, 75–91%. b) i) 30% H2O2, 1 N LiOH, THF, RT,
1 day, ii) 3 N KOH, MeOH, RT, 1 day, 87–92%. c) SO3·Me3N, DMF, 508C, 24–72 h, 78–92%. d) i) 0.1 N NaOH, Me3P, THF, RT, 1–4 h, ii) SO3·Py,
Py/Et3N=5:1, RT, 30 min, 81–94%. e) H2, Pd(OH)2/C, MeOH/H2O=9:1, RT, 92–96%.
O- and N-sulfated products. Finally, global deprotection
with Pd(OH)2/C led to the desired trisaccharidesTri8, Tri4,
Tri7, and Tri3.
Interestingly, we found that the O sulfation of 31 using
SO3·Py complex in pyridine as solvent gave a mixture[21] of
the desired product 35a in 45% yield and the lactone[22] 35b
in 52% yield, while the use of SO3·Me3N complex in DMF
at 508C gave 35a in excellent 91% yield without detection
of the lactone side product (Scheme 6).
For the synthesis of the trisaccharides with a sulphate
group at the 6 position of the non-reducing end glucosamine
unit, trisaccharides 10–13 were first submitted to the reduc-
tive opening of the benzylideneacetal (Scheme 5). Et3SiH
was used as the reductive reagent and dichlorophenylborane
as the acid,[20] obtaining the corresponding 4-O-benzyl-6-hy-
droxy derivatives 26–29 in optimal yield with excellent selec-
tivity. These derivatives were then subjected to the depro-
tection/sulfation sequence as described above to give trisac-
charides Tri6, Tri2, Tri5, and Tri1.
Nuclear magnetic resonance: All the compounds have been
1H and 13C assigned at several temperatures (see the Sup-
porting Information). First, the spins systems resonances
from the three hexoses and the isopropyl group were as-
signed using COSY and TOCSY, thus obtaining spin systems
that were connected by interglycosidic NOE or ROE con-
tacts between the anomeric proton and H4 or H3 and H4,
Scheme 5. Synthesis of trisaccharides Tri6, Tri2, Tri5, and Tri1 a) Et3SiH, dichlorophenylborane, CH2Cl2, ꢀ788C (5 min)!ꢀ408C (30 min), 84–91%.
b) i) 30% H2O2, LiOH, THF, RT, 1–2 days, ii) 3 N KOH, MeOH, RT, 1–5 days, 72–91%. c) SO3·Me3N, DMF, 508C, 1–5 days, 88–92%. d) i) 0.1 N NaOH,
Me3P, THF, RT, 1–2 h, ii) SO3·Py, Py/Et3N=5:1, RT, 30–60 min, 85–89%. e) H2, Pd(OH)2/C, MeOH/H2O=9:1, RT, 95–97%.
16322
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16319 – 16331

Conformational Equilibrium of Heparin-Like Trisaccharides
FULL PAPER
At low temperatures, the in-
terglycosidic NOEs are indicat-
ing a syn rearrangement of the
glycosidic linkages. However,
at high temperature (298 K)
other minor NOE cross-peaks
(H1’–H5, H1’–H3, and H2’–
H5) corresponding to an anti
Scheme 6. Detection of a lactone side product in the O-sulfation of intermediate 31.
thus corresponding to a preferentially syn rearrangement of
arrangement for the first glycosidic linkage can be observed
together with a larger NOE (H1’–H4), exclusive of the syn
conformation (Figure 5).
1
the glycosidic linkage.[7] The effects of sulfonation on the H
and 13C chemical shifts correspond to the effects described,
1
observing a downfield shift of both the H adjacent and the
Regarding the conformation of the iduronate ring, the
analysis of the coupling constant and the observation of the
exclusive H2’–H5’ contact (NOE, ROE, T-ROE) indicate a
13C ipso (see the supporting information).[23] The larger
downfield effects are observed in positions 2 and 6 of ring A
and 6 of ring C. Interestingly, the H6 protons on ring A are
isochronous while those of ring C have different chemical
shift being the differences larger for the 6-O-sulfoderiva-
tives. When the different compounds were compared, all the
l-IdoA2S resonances were well clustered except those for
H5 that were dispersed without any apparent reason.
2
1
contribution of the S0 conformer in equilibrium with the C4
chair. (Figure 5 and Figure 6). This is a characteristic feature
NOESY, ROESY, and 1D-NOESY, 1D-ROESY, and
1D-T-ROESY experiments have been performed to build
the growth curves in order to quantify the experimental dis-
tances based on sigma NOE or sigma ROE.[24] The results at
room temperature are consistent with a molecule in a
motion regime close to the NOE zero-crossing point, thus
implying low intensities and difficulties to integrate because
of the low accuracy. At the same temperature ROE peaks,
although measurable, were too influenced for undesired ef-
fects of strong coupling and other artifacts. TOCSY transfer
of magnetization was frequently detected; in particular the
more interesting ones that correspond to the iduronate ring
as result of their minor chemical shift dispersion. Attempts
to solve the problem using 1D-T-ROESY did not give
better results. We decided to vary the temperature
Figure 4. 1D-NOESY growth curves (at 278 K) corresponding to the
H2–H5 distance of trisaccharides Tri1 (6OSO3 reducing end) and Tri5
(6OH reducing end).
ꢀ
in order to modify the correlation time, and the
best results were obtained at low temperatures
(278–283 K) using 1D-NOESY.[25] At high temper-
atures (308–313) the zero crossing was still a prob-
lem in some cases. Because of the comparative
nature of this work we decided to use low temper-
ature in this study as it allowed us to analyze all
the compounds in the same conditions. Based on
the 2D-NOESY spectra we constructed the grow-
ing-up NOE curves for those protons of particular
interest via mono dimensional selective analogues
of NOESY (dpfgse-NOESY) as they have a larger
linearity and better quality than the 2D ana-
logues[26] (Figure 4). Using the isolated spin-pair
approximation[27] we have calculated the interpro-
tonic distances by the ratio of their respective
sigma NOE (see below) obtained by two inde-
pendent methods: either fitting the whole curve to
a biexponential equation or to a linear growth
using a relative distance (see Table 1 and the Sup-
porting Information).
Table 1. Experimental and theoretical interglycosidic key distances. The theoretical
values were averaged over 40000 frames of MD simulation and weighted on the popu-
1
2
ꢀ
lations of chair C4 and skew boat S0 conformations at 278 K. The theoretical H1’ H6
distances represent the r-6 average over the pro R and pro S values. Linear regression
was used to calculate the experimental distances.
Method
Distance [ꢄ]
GlcN-IdoA2S
l-IdoA2S
IdoA2S-GlcN
H1’’–H3’ H1’’–H4’ H1’–H3’ H2’–H5’ H1’–H3 H1’–H4 H1’–H6
Tri1 MD
2.3
2.6
2.3
2.6
2.3
2.6
2.3
2.7
2.3
2.6
2.3
2.6
2.3
2.6
2.3
2.6
2.8
2.7
2.9
2.6
2.8
2.7
2.8
2.7
2.8
2.6
2.6
2.6
2.7
2.6
2.8
2.6
3.6
3.0
3.6
3.0
3.6
3.0
3.6
3.0
3.9
–
3.9
3.2
3.9
3.2
3.8
3.1
3.2
3.0
3.1
2.9
3.2
2.9
3.2
2.9
3.5
3.2
3.5
3.2
3.5
3.2
3.3
3.0
3.1
–
2.9
–
3.9
–
3.8
–
3.8
–
3.4
–
3.7
–
3.9
–
2.5
2.5
2.4
2.4
2.3
2.6
2.3
2.5
2.3
2.6
2.4
2.6
2.3
2.5
2.3
2.6
3.5
–
3.4
–
3.4
–
3.2
–
3.6
–
3.5
–
3.4
-
3.3
–
exptl
Tri2 MD
exptl
Tri3 MD
exptl
Tri4 MD
exptl
Tri5 MD
exptl
Tri6 MD
exptl
Tri7 MD
exptl
Tri8 MD
exptl
Chem. Eur. J. 2012, 18, 16319 – 16331
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16323

P. M. Nieto et al.
1
2
idues involve two conformations, C4 and S0. Thus, as the
characteristic interconversion rate within the millisecond
time scale is far from the MD accessible range, two alterna-
tive approaches can be used: first consider two different
starting geometries for the iduronate residues and run two
independent molecular dynamic simulations, and second use
experimental restrictions and run a single time averaged re-
strained molecular dynamic simulation. In this work we
have used the first approach and two alternative starting
conformations for the iduronate ring were considered to ac-
1
2
count for the conformational equilibrium: C4 and S0. The
global results were thus obtained by weighting both simula-
tions according to their respective populations calculated
from the experimental results about the conformational
equilibrium obtained from coupling constant analysis.[28]
Figure 5. Expansion of a NOESY experiment for Tri1 showing the signals
corresponding to an anti rearrangement together with the syn rearrange-
ment. Note that latter has the biggest intensity. Exclusive NOE peaks of
the anti-Y conformation are marked with a star symbol.
Global geometry: analysis of interglycosidic torsions: First,
to define the global geometries of trisaccharides, we ana-
lyzed the dynamic behavior of the two glycosidic linkages of
each trisaccharide by monitoring the torsions F (H1’-C1’-O-
C4) and Y (C1’-O-C4-H4) along the molecular dynamics
simulation. Qualitatively the effect of the conformational
1
2
change between C4 and S0 conformations is subtler than it
can be expected, affecting only to the relative position of
the carbons C2 and C3 of the iduronate ring, while the rela-
tive positions of the C1, C4, C5, and O6 remained the same.
Therefore for the change between both conformations only
the relative positions of C2 and C3 must be swapped.
In the case of the GlcN-IdoA linkages, the simulations
ꢀ
ꢀ
predict interglycosidic distances H1’’ H4’ and H1’’ H3’
within the range for NOE observation, in good agreement
with the experimental NMR data (Table 1). In general, the
distributions of F and Y were centered at ꢀ608 and ꢀ508,
respectively (see the Supporting Information). The distribu-
1
2
Figure 6. Iduronate conformational equilibrium between C4 and S0.
2
tions in the case of S0 conformation were narrower and in
the case of the C4 the data predict an increase in the local
1
of the internal iduronate ring in heparins and heparin-de-
rived oligosaccharides. In this case, we have been able to
measure with high precision the coupling constants as to
allow us to derive populations (See the Supporting Informa-
tion and the Molecular Dynamics section).[38,28] We have ob-
served a larger contribution of C4 for the compounds with-
out sulfate in position 6 of the reducing end glucosamine,
Tri5, Tri6, Tri7, Tri8, than for those with sulfate which is re-
flexibility of the GlcN-IdoA glycosidic linkage, principally
in the F angle although correlated with a minor effect in the
Y torsion (see the Supporting Information).
A parallel analysis for the IdoA-GlcN linkages reveals
that when the l-IdoA residues were in the 1C4 conformation,
the F torsion was distributed between 08 and 608, with a
1
2
maximum close to 608, whereas in the case of the S0 confor-
ꢀ
flected in a longer H2’ H5’ distance (3.0–3.2 ꢄ; Table 1)
mation, this interval was somewhat modified (from 08 to
758), slightly increasing the accessible conformational space.
The Y torsion was significantly more flexible and showed a
behavior sensitive to the conformational state of the l-IdoA
ring. In particular, for the 2S0 conformation of the central
ring some population of anti-Y conformers (Y= ꢁ1808),
over 30%, were predicted for three (Tri1, Tri5, and Tri6)
out of the eight trisaccharides (see the Supporting Informa-
tion), as well as significant differences in the relative popu-
lations of conformers at the two sub-minima (centered at
ꢀ608 and 08) for different trisaccharides. Yet, these differen-
ces could not be easily correlated to the sulfation pattern. In
the NMR spectroscopic study, the anti-Y conformation was
detected in some favorable cases (observation of the set of
than the sulfated ones (2.9–3.0 ꢄ; Table 1). In addition, we
have found differences in temperature dependence for those
lacking of sulfate into position 6 of reducing end glucosa-
2
mine. This behavior is consistent with a decrease of the S0
population from 40% to 35%–25%, when the temperature
is decreasing from 323 to 283 K. Interestingly, in trisacchar-
ides bearing sulfates on position 6 of reducing end glucosa-
mine this dependence is not observed and the skew boat
population remains constant above 40% along all the range
of temperatures.
Molecular dynamics (MD): The iduronate ring exhibits a
characteristic conformational equilibrium, that for inner res-
16324
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16319 – 16331

Conformational Equilibrium of Heparin-Like Trisaccharides
FULL PAPER
NOEs H1’–H3, H5’–H6a, and H5’–H6b; see Figure 5). The
simulations predicted interglycosidic distances H1’–H4,
H1’–H3, and H1’–H6 (pro R or pro S) within the range for
NOE observation, once more in agreement with the experi-
mental NMR data (Table 1).
GlcN-IdoA linkages, the weighted distributions of F and Y
were centered at ꢀ658 and ꢀ508, respectively (Figure 7).
Both torsions were quite rigid and there was no apparent in-
fluence of the sulfation pattern. For IdoA-GlcN linkages,
the weighted distributions were centered at 408 for the
F torsion, and 08 and ꢀ458 (two subminima) for the Y tor-
sion. Some minor contribution from anti-Y conformers was
A description of the global geometries of the trisacchar-
ides in solution must also take into account the different
populations of conformers of the central l-iduronate ring.
As the timescale for ring interconversion is out of the feasi-
ble current molecular dynamics capabilities (above micro-
seconds), we have included this effect by using the popula-
tions experimentally obtained, as described below
(Figure 7). With these data, we have weighted the ensembles
2
observed when the central IdoA residue was in S0 confor-
mation. As before, the sulfate pattern did not seem to
impose any noticeable effect on the IdoA-GlcN conforma-
tional space (Figure 7).
Solvation and hydrogen bonds: The analysis of the radial
distribution functions (rdf) identified a generalized effect of
the l-IdoA2S conformation on the solvation around the gly-
cosidic oxygen atoms of the IdoA2S-GlcN linkages (see the
Supporting Information). Specifically, when the l-IdoA2S
1
2
of F and Y values by the populations of C4 and S0 con-
formers at 298 K (Figure 7). The results indicated that for
2
residue adopts the S0 conformation, two solvation shells at
2 and 3 ꢄ appear around this glycosidic oxygen, which are
absent in the 1C4 conformation. Thus, these two solvation
2
shells are skew boat S0 exclusive. This observation is inter-
esting in the context of l-IdoA2S interactions with proteins,
as it indicates that the structuration of the molecular solva-
tion shell is sensitive to the conformational state of the idur-
onate ring.
We also analyzed the presence of inter-residual hydrogen
bonds. For that purpose we have chosen a distance and
angle cutoff of 3.0 ꢄ and 1208, respectively (see the Experi-
mental Section). The results allowed us to identify a con-
served hydrogen bond involving the oxygen ring of the
l-IdoA2S residue and the hydrogen H3O of the reducing
end GlcN (see Figure 19S in the Supporting Information). It
appears with an average occupancy of 57%, an average life-
time of 2.6 ps, an average distance of 2.4 ꢄ, and an average
angle of 328. Furthermore, it is not influenced by the confor-
mation of the l-IdoA2S residue or the sulfation pattern.
Other conserved hydrogen bond found was that between
oxygen O5 of the non-reducing end GlcN and the hydrogen
H3O of the l-IdoA2S residue. Nevertheless, this hydrogen
bond presented a low average lifetime (below 1 ps) and a
low percentage of occupancy (below 20%; see the Support-
ing Information).
Populations of central iduronate conformers: We have moni-
tored the four inter-proton dihedral angles (H1’-C1’-C2’-H2’,
H2’-C2’-C3’-H3’, H3’-C3’-C4’-H4’, H4’-C4’-C5’-H5’) of the
l-IdoA2S residue for both 1C4 and 2S0 starting conforma-
tions along the 40000 frames of each trajectory obtained.
From these values (see the supporting information), vicinal
3
proton–proton coupling constants, JHH, for each frame and
conformation have been calculated by using the Haasnoot–
Altona equation, which takes into account both the electro-
negativity and the orientation of the substituents on the
fragment H-C-C-H.
Figure 7. Population-weighted distribution curves for the glycosidic tor-
sions
F and Y of the glycosidic linkages a, b) GlcN-IdoA2S and
3
1
2
The final theoretical JHH values for C4 and S0 have been
obtained as an average for each of the models (l-IdoA2S in
c, d) IdoA2S-GlcN for the whole library of trisaccharides. The popula-
tions of conformers at 298 K have been taken for the weighting. A bin
size of 308 has been used.
2
1C4 or in S0) from the torsional angles calculated from the
Chem. Eur. J. 2012, 18, 16319 – 16331
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16325

P. M. Nieto et al.
frames obtained along the trajectories. Next, to obtain the
tempting to speculate on the importance of the electrostatic
compensation of ligand “internal energy stresses” (as those
present in the triad of sulfate groups) by the positively
charged sidechains of amino acid typically found in the
binding sites of heparin-binding proteins, on the modulation
of the l-iduronate conformations in the bound state.
2
populations of conformers (1C4 and S0) of the l-IdoA ring
in each trisaccharide, we carried out an iterative fitting of
theoretical and experimental J-coupling data [Eq. (1)]. At
4
this point the C1 conformation was disregarded as no exper-
imental support was obtained for it, particularly the exclu-
sive GlcN-IdoA H5’-H5 NOE was not observed.[29]
3
Therefore, the experimental JHH coupling constants were
Correlation times: Finally, we have performed the analysis
of the global time correlation function from the molecular
dynamics data. Overall correlation times (t0) for the eight
trisaccharides were calculated according to the model-free
approach[30] by first calculating the auto correlation function
for each H–H vector with AMBER, and considering the
conformational equilibrium and the conformers populations.
The results (see Figure 9) show a very significant increase in
3
considered as averages of the MD-derived JHH for each
conformer weighted on the molar fraction of each one.[28]
As the theoretical values were averages from MD simula-
tions, they implicitly reflected the fluctuations around can-
onical conformations, which must be considered for this
flexible hexopyranose ring[28] particularly to account for the
pseudorotational conformational space in the case of the
skew boat conformer (2S0).
The results are shown in Figure 8, classified by the two
series of trisaccharides. Those with a sulfate group in posi-
tion 6 of the glucosamine at the reducing terminal Tri1, Tri2,
Figure 9. Global time correlation (t0) at 298 K of each trisaccharide
3
weighted in populations determined by JHH analysis.
Figure 8. Variation of 2S0 conformer population as a function of the tem-
perature. Two different behaviors depending on the substitution on posi-
tion 6 at the reducing end glucosamine (Glc-A) are observed.
the global time correlation for the ensemble of trisacchar-
ides with sulfate groups in the position 6-O of the reducing
end glucosamine compared with those with a hydroxyl
group in that position. This is evident when compare the
pair Tri3 and Tri5, which have the same formula differing in
the position of the 6-O-sulfate group. Thus, the removal of
the 6-O-sulfate group on the reducing GlcN ring gives rise
to the reduction in the global time correlation, and there-
fore, a faster reorientation in solution for the trisaccharides
belonging to this ensemble.
Tri3, and Tri4 (dotted lines) and those lacking this group
Tri5, Tri6, Tri7, and Tri8 (full lines). The data revealed clear
differences between both series in the behavior of the con-
former populations of the l-IdoA2S ring as a function of
temperature. Explicitly, we have observed conformational
differences (Figure 8) that were specific to the presence or
absence of the 6-O-sulfate group on the reducing GlcN ring.
Particularly, the absence of this bulky charged group made
2
the S0 population very sensitive to temperature (Figure 8).
Discussion
1
At low temperature (278 K) the C4 conformer was favored
in second ensemble (65–80%), while it remained basically
unchanged in the first one (53–57%). As the temperature
increased, the differences between both ensembles started
to vanish.
The combination of NMR spectroscopy and MD simula-
tions (Figure 8) demonstrate a significant influence of the
sulfation pattern on the conformational equilibrium of an in-
ternal l-IdoA2S residue. As one major energy component
in protein–heparin interactions is the electrostatic term, it is
For the synthesis of the eight heparin-like trisaccharides, we
have followed a 2+1 glycosylation approach that used disac-
charide 5 and monosaccharides 1–4 as key building blocks.
We have employed an appropriate set of protecting groups
to obtain the required stereochemistry of the glycosidic link-
ages and introduce the sulfate groups at the desired posi-
tions. This strategy has allowed the preparation of the trisac-
charides in good overall yield.
16326
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16319 – 16331

Conformational Equilibrium of Heparin-Like Trisaccharides
FULL PAPER
The minimum size of these trisaccharides, lead to the ab-
sence of a preferential axis of rotation and the molecules
behave as spherical particles in isotropic motion, allowing us
to quantify the distances from NOESY build up curves by
assuming an isotropic behavior. In case of larger glycosami-
noglycans, longer than 5–6 residues, the rigidity of the glyco-
sidic linkage causes a top rotor behavior featuring two per-
pendicular axis of rotation with two different correlation
times[31] and as a consequence of the effect of the anisotropy
a more complex and inaccurate method of distance evalua-
tion is needed as the intensity of the NOE depends not only
on the distance but also on the angle between the interpro-
tonic distance and the axis of the molecule.[32] Therefore, the
distances obtained from the analysis of NOESY cross-peaks
in from trisaccharides should be more accurate in our
library than those obtained from larger compounds.
over, the shorter values obtained for both distances in the
Tri1–4 ensemble demonstrate the existence of a higher pop-
ulation of the 2S0 conformer in solution when the
6-OSO3 group is present in the reducing end GlcN residue.
However, the most interesting feature of this conforma-
tional equilibrium is the influence of the temperature and
substitution pattern. The trisaccharides lacking the sulfate
group in position 6 of the reducing end glucosamine (Tri5–
Tri8) have proportions of 2S0 skew boat smaller that de-
crease with temperature down to 19–28% at 283 K (see
Figure 8). On the contrary, the trisaccharides having a sul-
fate in the primary position (Tri1–Tri4) did not show any in-
fluence of temperature from 283 to 313 K in the conforma-
2
tional equilibrium with a constant population of S0 of 40%.
Considering that the main effect caused by the change from
2
1C4 to S0 is the exchange between the relative positions up
The smaller size also affords narrower signals, allowing
the extraction of reliable coupling constants, and therefore
to analyze more accurately the conformational equilibrium
of the central l-Ido2S ring, which is in equilibrium between
and down the mean plane of the ring of carbons 2 and 3, a
possible explanation for this behavior could be the electro-
static repulsion between the charges of sulfate in position 2
of l-Ido2S and the adjacent glucosamines (the sulfate in po-
sition 6 of glucosamine of the reducing end and the N-sul-
fate of glucosamine at the other end). The three groups
form a cluster of charges that in the case of the compounds
with sulfate in the position 6 of the reducing end the only
possibility to release the electrostatic energy is to slide to
1
2
the C4 and S0 conformations. This is a characteristic feature
of the internal iduronate residues in heparin, and has been
thoughtfully studied.[33] Chemical purity of these synthetic
molecules and their designed patterns of substitutions make
this set of eight trisaccharides an excellent framework to
study the effect of differences in sulfation on the conforma-
tional behavior of the l-Ido2S rings. As this ring holds a sul-
fate group at position 2 in all the trisaccharides, we are ex-
clusively analyzing the effects of sulfation of adjacent resi-
dues on the conformation of the central l-iduronate 2-O-sul-
fate.
2
the S0 conformation were the charges are further dispersed.
On the contrary, in the case of the trisaccharides without
this charge where there are only two charged groups in that
side of the molecule, the electrostatic interactions may be
relieved by rotating the sulfate group to a more favorable
position.
From the experimental determination of proton–proton
distances (Table 1) we firstly focused our analysis on those
defining the local conformation of the l-IdoA2S residue.
The observation of a NOE signal between protons H2’–H5’
Considering now the global conformation, this is going to
be directed by the glycosidic linkages, the interglycosidic
distances defining the GlcNS-IdoA2S linkage (H1’’–H3’ and
H1’’–H4’) are very similar for all the trisaccharides (from
2.6 ꢄ to 2.7 ꢄ), showing a more rigid behavior for this link-
age. This observation is in good agreement with the results
obtained from the MDs. Finally, referring to the l-IdoA2S-
GlcN glycosidic linkage, due to overlapping we have only
been able to quantify the H1’–H4 distance which ranks from
2.4 to 2.6 ꢄ. However, does not show correlatable differen-
ces with any structural feature. Furthermore, we have identi-
fied NOE peaks corresponding to the H3 and both H6 pro-
tons of the reducing end glucosamine from the anomeric po-
sition. However, in four of the eight trisaccharides (Tri1,
Tri3, Tri5, Tri7) the observation of the NOE intensity for
the H3 proton could belong to either a inter-glycosidic
H1’–H3 short distance or an intra-residue (IdoA2S) H1’–H3’
distance, thus not being able to confirm its existence. In any
case, of the trisaccharides Tri2, Tri4, Tri6, and Tri8, both
H1’–H3 and H5’–H6 NOE signals have been clearly identi-
fied (from weak to medium intensities, but are not quantifia-
ble due to overlapping). The simultaneous presence of both
NOE intensities confirms the existence of anti-Y conformer
population for the l-IdoA2S-GlcN glycosidic linkage, indi-
cating a larger flexibility of this terminal linkage as com-
pared with a-d-GlcN-(1-4)-a-l-IdoA.
2
reveals the presence of the conformation S0, however, from
the analysis of coupling constants is clearly deduced the par-
1
ticipation of other conformations, particularly C4. The de-
tailed analysis of the distances H1’–H2’, H1’–H3’, H2’–H5’
and H4’–H5’ for l-IdoA2S residue shed light on its confor-
mation. Thus, the calculated values for the H4’–H5’ distance
(between 2.4 and 2.6 ꢄ) are in agreement with a equilibrium
1
2
between the canonical C4 and S0 conformations. Further-
more, the absence of H5’’–H5’ NOE cross-peak for the
GlcN-IdoA2S glycosidic linkage confirmed that the chair
4C1 conformation for the l-IdoA2S residue is not present in
solution. Regarding the H1’–H2’ distance, the results show
consistently shorter values in the trisaccharides without sul-
fate group in position 6 of the reducing end glucosamine,
compounds Tri1–4. Since this distance (according to the can-
2
onical conformations) is longer in the S0 conformer, this ob-
2
servation agrees with a higher S0 population in the Tri1–4
ensemble (see Table 1). The H1’–H3’ and H2’–H5’ NOESY
signals, which correspond to exclusive NOEs of the
2S0 conformation of the l-IdoA2S residue, are also clearly
observed. Thus, the results confirm the presence of the
2S0 conformer in the whole library of trisaccharides. More-
Chem. Eur. J. 2012, 18, 16319 – 16331
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16327

P. M. Nieto et al.
Conclusion
conformational properties of the iduronate by the adjacent
rings comes mainly from the residue at the 4 position.
Another interesting observation was the increase of the
correlation time caused by the presence of a sulfate group
in position 6 of reducing end glucosamine. A feasible ex-
planation is the increase on rigidity caused by charge repul-
sion in that part of the molecule.
In summary, the study of the influence of the sulfation pat-
tern on the conformational equilibrium of an internal
l-IdoA2S has allowed us to confirm that the presence of
6-O-sulfation on the reducing end GlcN pushes the equili-
brium of the l-IdoA2S ring towards the skew boat
2S0 conformation. This result, which has never been reported
so far, is very important because it demonstrates that the
substitution of charged functional groups on heparin-like
oligosaccharides does not only alter its interaction energy
with a protein receptor (absence or presence of key func-
tional groups), but it can also cause a deep effect on the in-
trinsic conformational properties.
Experimental Section
Synthesis: As an example we describe the synthesis of Tri8, the rest of
the trisaccharides and the general procedures are described in the Sup-
porting Information.
Related to the latter, the trisaccharides lacking the sulfate
group in position 6 of the reducing end glucosamine (Tri5–
Tri8) present a temperature-sensitive population of S0 skew
Isopropyl O-(2-azido-3-O-benzyl-4,6-O-benzylidene-2-deoxy-a-d-gluco-
pyranosyl)-(1!4)-O-(methyl 3-O-benzyl-2-O-pivaloyl-a-l-idopyranosyl-
uronate)-(1!4)-2-acetamido-2-deoxy-3,6-di-O-benzyl-a-d-glucopyrano-
side (10): TMSOTf (3 mL, 0.017 mmol) was added to a solution of 1
(74 mg, 0.17 mmol) and 5 (177 mg, 0.198 mmol) in dry CH2Cl2 (4 mL),
containing 4 ꢄ molecular sieves and under argon atmosphere. After stir-
ring for 30 min at RT, Et3N (200 mL) was added and the solvent was
evaporated. The residue was purified by flash chromatography (toluene/
EtOAc=4:1, 2:1, and 1:1) to yield 10 (175 mg, 75%). [a]2D0 =+24.58 (c=
0.8, CHCl3); TLC (toluene/EtOAc=2:1) Rf =0.20; 1H NMR (500 MHz,
CDCl3): d=7.44–7.40 (m, 2H; PhH), 7.37–7.24 (m, 19H; PhH), 7.23–7.17
(m, 4H; PhH), 5.50 (s, 1H; PhCHO2), 5.26 (d, J1,2 =4.0 Hz, 1H; H1b),
5.22 (d, JNH,2 =9.5 Hz, 1H; CH3CONH), 4.98 (t, J2,1 =J2,3 =4.0 Hz, 1H;
H2b), 4.93 (d, J1,2 =3.5 Hz, 1H; H1c), 4.88 (d, J=11.0 Hz, 1H; PhCH),
4.86 (d, J1,2 =4.0 Hz, 1H; H1a), 4.78 (d, J5,4 =4.0 Hz, 1H; H5b), 4.75 (AB-
system, 2H; PhCH2), 4.74 (d, J=11.5 Hz, 1H; PhCH), 4.68 (d, J=
11.0 Hz, 1H; PhCH), 4.57 (ABsystem, 2H; PhCH2), 4.50 (d, J=11.5 Hz,
2
boat conformers, which decreases from 313 to 283 K. On the
contrary, the conformational equilibrium of trisaccharides
presenting a sulfate group in the primary position (Tri1–
Tri4) did not show any influence with temperature (range
283–313 K). Thus, we have hypothesized that the electrostat-
ic repulsion between the charges of the sulfate group in po-
sition 2 of l-IdoA2S residue and the adjacent sulfates
groups of both glucosamines residues forces the l-IdoA2S
2
ring to slide to a skew boat S0 conformation, and thus, some
repulsion electrostatic energy is released since charges
become further dispersed.
The use of the smaller possible model of heparin, a trisac-
charide, has allowed us to determine that the behavior of
the two types of glycosidic linkages, GlcN-IdoA and IdoA-
GlcN, is not equivalent. We have detected weak NOE peaks
corresponding to minor anti-Y rotameric populations in the
case of the IdoA-GlcN linkage that were not observed for
the GlcN-IdoA union. We propose that this is the flexibility
point where the heparin chain can bend to form a kink, in
order to adapt to the shape of the protein binding site.
Moreover, it is noticeable that, in the case described
herein, the sulphation pattern significantly affects the
“local” conformational properties (central ring) but not the
“global” ones (determined by the F and Y torsions). This is
one of the features that enhances heparin-like GAGs as bio-
molecules with a “special plasticity”, which apparently play
a key role in their bioactivities.
1H; PhCH), 4.25 (dd, J6,5 =5.5 Hz, J6,6’ =10.0 Hz, 1H; H6c), 4.23 (td, J2,1 =
3.5 Hz, JNH,2 =J2,3 =10.0 Hz, 1H; H2a), 4.08 (t, J4,3 =J4,5 =9.0 Hz, 1H; H4a),
4.03 (t, J4,3 =J4,5 =4.5 Hz, 1H; H4b), 4.00 (t, J3,2 =J3,4 =4.5 Hz, 1H; H3b),
3.93 (t, J3,2 =J3,4 =9.5 Hz, 1H; H3c), 3.87 (td, J5,6 =5.5 Hz, J5,4 =J5,6’
10.0 Hz, 1H; H5c), 3.82 (hept, J=6.5 Hz, 1H; CH iPr), 3.81 (brd, J6,6’
=
=
9.5 Hz, 1H; H6a), 3.78 (m, 1H; H5a), 3.66 (brd, J6’,6 =9.5 Hz, 1H; H6a’),
3.61 (brt, J4,3 =J4,5 =J6’,5 =J6’,6 =9.5 Hz, 2H; H4c +H6c’), 3.57 (t, J3,2 =J3,4
10.0 Hz, 1H; H3a), 3.42 (s, 3H; CH3OCO), 3.31 (dd, J2,1 =4.0 Hz, J2,3
=
=
9.5 Hz, 1H; H2c), 1.75 (s, 3H; CH3CONH), 1.16 (d, J=6.5 Hz, 3H;
CH3 iPr), 1.15 (s, 9H; CH3 Piv), 1.07 ppm (d, J=6.0 Hz, 3H; CH3 iPr);
13C NMR (125 MHz, CDCl3): d=177.59, 169.51, 169.13 (C=O), 138.63,
138.18, 137.73, 137.53, 137.34 (C Ph), 128.96, 128.39, 128.25, 128.18,
128.09, 127.85, 127.70, 127.59, 127.34, 127.25, 126.08 (CH Ph), 101.49
(PhCHO2), 99.50 (C1c), 97.80 (C1b), 95.74 (C1a), 82.47 (C4c), 78.50 (C3a),
76.06 (C3c), 75.91 (C3b), 75.20 (C4b), 74.80 (PhCH2), 74.68 (C4a), 73.63,
73.28, 73.15 (PhCH2), 71.08 (C5a), 70.01 (CH iPr), 69.81 (C2b), 69.34 (C5b),
68.47 (C6c), 68.34 (C6a), 63.22 (C5c), 63.00 (C2c), 52.12 (C2a), 51.84
(CH3OCO), 38.78 (C Piv), 27.15 (CH3 Piv), 23.26 (CH3CONH), 23.22,
21.57 ppm (CH3 iPr); HRMS: m/z calcd for C64H76N4O17+Na+: 1195.5103
[M+Na]+; found: 1195.5029.
Owing to the asymmetry of the compounds some conclu-
sions can be extracted about the factors that rule the confor-
mational properties of the iduronate ring. Most of the previ-
ous data on this conformational equilibrium come from reg-
ularly substituted fragments of heparin, finding that in some
cases the substitution of the neighbor ring affects clearly the
iduronate conformational equilibrium.[34] We have found
however an almost constant behavior with, at room temper-
ature, a participation of the C4 and S0 in a 60–40 approxi-
mate ratio, nearly independently on the pattern of substitu-
ents of the ring at the reducing end terminus. From this ob-
servation it can be concluded that the modulation of the
Isopropyl O-(2-azido-3-O-benzyl-4,6-O-benzylidene-2-deoxy-a-d-gluco-
pyranosyl)-(1!4)-O-(3-O-benzyl-a-l-idopyranosyluronic acid)-(1!4)-2-
acetamido-2-deoxy-3,6-di-O-benzyl-a-d-glucopyranoside
(14):
H2O2
(0.9 mL) and an aqueous LiOH solution (1.3 mL; 1 N) were added to a
solution of 10 (70 mg, 0.060 mmol) in THF (3.1 mL) at 08C, 30%. After
stirring for 24 h at RT, the mixture was cooled at 08C, MeOH (4.8 mL)
and an aqueous KOH solution (2.6 mL; 3 N) were added, and the mix-
ture was stirred at 08C for 10 min. After stirring for 24 h at RT, the reac-
tion mixture was acidified with IR-120-H+ amberlite resin, filtered and
concentrated. The residue was eluted from a Sephadex LH-20 chroma-
tography column (MeOH/CH2Cl2 =1:1) to afford 14 (56 mg, 87%).
[a]2D0 =+8.368 (c=1.1, MeOH); TLC (CH2Cl2/MeOH=25:1) Rf =0.15;
1H NMR (500 MHz, CD3OD): d=7.51–7.45 (m, 5H; PhH), 7.42–7.21 (m,
20H; PhH), 5.61 (s, 1H; PhCHO2), 5.22 (d, J1,2 =2.0 Hz, 1H; H1b), 5.15
(d, J1,2 =4.0 Hz, 1H; H1c), 5.00 (d, J=11.5 Hz, 1H; PhCH), 4.95 (d, J=
1
2
16328
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16319 – 16331

Conformational Equilibrium of Heparin-Like Trisaccharides
FULL PAPER
11.5 Hz, 1H; PhCH), 4.85 (d, J1,2 =4.0 Hz, 1H; H1a), 4.83 (brd, J5,4
2.0 Hz, 1H; H5b), 4.78 (d, J=11.5 Hz, 1H; PhCH), 4.75 (s, 2H; PhCH2),
4.60 (ABsystem, 2H; PhCH2), 4.57 (d, J=12.0 Hz, 1H; PhCH), 4.40 (dd,
=
AcOH) Rf =0.61; MS (ESI): m/z positive mode: 1173.4 [M+Na]+, 1151.4
[M+H]+; m/z negative mode: 1149.4 [MꢀH]ꢀ, 1127.4 [MꢀNa]ꢀ). Trie-
thylamine (0.5 mL) and sulfur trioxide-pyridine complex (19 mg,
0.12 mmol) were added to a solution of the crude in pyridine (2.5 mL),
After stirring for 30 min at RT under argon atmosphere, the solution was
purified by Sephadex LH-20 (MeOH/CH2Cl2 =1:1), C-18 reverse phase
chromatography (0% to 100% of MeOH in a 10 mm Et3NHOAc aq. sol-
ution), again Sephadex LH-20 column (MeOH/CH2Cl2 =1:1), and finally
passed through a Dowex 50WX4-Na+ column (MeOH/H2O=9:1) to give
22 (24 mg, 82%). Characterization of 22 as triethylammonium salt:
[a]2D0 =+25.708 (c=1.0, MeOH); TLC (15/5/3/1 EtOAc/Py/H2O/AcOH)
Rf =0.30; 1H NMR (500 MHz, CD3OD): d=7.52 (brt, J=7.5 Hz, 4H;
PhH), 7.49 (brd, J=7.5 Hz, 2H; PhH), 7.44 (m, 2H; PhH), 7.38–7.16 (m,
14H; PhH), 7.18 (brt, J=7.5 Hz, 1H; PhH), 7.13 (brd, J=7.5 Hz, 2H;
PhH), 5.58 (s, 1H; PhCHO2), 5.57 (brs, 1H; H1b), 5.48 (d, J1,2 =3.5 Hz,
1H; H1c), 5.05 (d, J=12.5 Hz, 1H; PhCH), 4.96 (brs, 1H; H5b), 4.94 (d,
J=12.5 Hz, 1H; PhCH), 4.91 (d, J=12.0 Hz, 1H; PhCH), 4.84 (ABsys-
tem, 2H; PhCH2), 4.83 (d, J1,2 =3.5 Hz, 1H; H1a), 4.78 (d, J=11.5 Hz,
1H; PhCH), 4.72 (brs, 1H; H2b), 4.58 (d, J=11.5 Hz, 1H; PhCH), 4.52
(brs, 1H; H3b), 4.46 (dd, J6,5 =5.0 Hz, J6,6’ =10.0 Hz, 1H; H6c), 4.38 (d, J=
13.0 Hz, 1H; PhCH), 4.29 (brs, 1H; H4b), 4.28 (brd, J6,6’ =11.0 Hz, 1H;
J
6,5 =5.0 Hz, J6,6’ =10.0 Hz, 1H; H6c), 4.38 (brt, J4,3 =J4,5 =3.0 Hz, 1H;
H4b), 4.15 (td, J5,6 =5.5 Hz, J5,4 =J5,6’ =10.0 Hz, 1H; H5c), 4.14 (t, J3,2
3,4 =9.5 Hz, 1H; H3c), 4.10 (dd, J2,1 =3.5 Hz, J2,3 =10.5 Hz, 1H; H2a), 3.95
=
J
(m, 1H; H3b), 3.94 (m, 2H; H4a +H5a), 3.89 (m, 1H; H6a), 3.88 (hept, J=
6.0 Hz, 1H; CH iPr), 3.82 (m, 1H; H3a), 3.81 (brt, J2,1 =J2,3 =3.0 Hz, 1H;
H2b), 3.731 (dd, J2,1 =4.0 Hz, J2,3 =10.0 Hz, 1H; H2c), 3.728 (m, 1H; H6a’),
3.72 (t, J4,3 =J4,5 =9.5 Hz, 1H; H4c), 3.69 (t, J6’,5 =J6’,6 =10.0 Hz, 1H; H6c’),
1.77 (s, 3H; CH3CONH), 1.25 (d, J=6.0 Hz, 3H; CH3 iPr), 1.15 ppm (d,
J=6.0 Hz, 3H; CH3 iPr); 13C NMR (125 MHz, CD3OD): d=173.30,
173.18 (C=O), 140.15 (C Ph), 139.47 (3ꢅC Ph), 139.05 (C Ph), 129.89,
129.40, 129.27, 129.18, 129.12, 129.08, 129.05, 129.02, 128.84, 128.72,
128.66, 128.41, 128.18, 127.25 (CH Ph), 103.56 (C1b), 102.44 (PhCHO2),
97.16 (C1c), 96.87 (C1a), 83.57 (C4c), 80.67 (C3a), 79.00 (C3c), 78.67 (C3b),
76.09 (C4a), 75.87, 75.41 (PhCH2), 74.79 (C4b), 74.44, 73.18 (PhCH2), 72.24
(C5a), 71.34 (C5b), 71.31 (CH iPr), 70.09 (C6a), 69.59 (C2b +C6c), 65.13
(C2c), 64.22 (C5c), 54.50 (C2a), 23.67 (CH3 iPr), 22.60 (CH3CONH),
21.61 ppm (CH3 iPr). MS (ESI) m/z positive mode: 1197.3 [M+Na]+; m/z
negative mode: 1073.4 [MꢀH]ꢀ.
H6a), 4.10 (td, J5,6 =5.0 Hz, J5,4 =J5,6’ =10.0 Hz, 1H; H5c), 3.95 (dd, J2,1
=
Isopropyl O-(2-azido-3-O-benzyl-4,6-O-benzylidene-2-deoxy-a-d-gluco-
pyranosyl)-(1!4)-O-(3-O-benzyl-2-O-sulfo-a-l-idopyranosyluronic acid)-
(1!4)-2-acetamido-2-deoxy-3,6-di-O-benzyl-a-d-glucopyranoside triethyl-
ammonium salt (18): Sulfur trioxide-trimethylamine complex (46 mg,
0.33 mmol) was added to a solution of 14 (36 mg, 0.033 mmol) in dry
DMF (3.3 mL). After stirring for 72 h at 508C under argon atmosphere,
Et3N (100 mL) was added and the solution stirred at RT for 5 min. The
solution was purified by Sephadex LH-20 chromatography (MeOH/
CH2Cl2 =1:1), and C-18 reverse phase chromatography (0% to 100% of
MeOH in a 10 mm Et3NHOAc aq. solution) to give 18 (35 mg, 78%).
[a]2D0 =+16.148 (c=1.4, MeOH); TLC (60/5/3/1 EtOAc/Py/H2O/AcOH)
Rf =0.30; 1H NMR (500 MHz, CD3OD): d=7.53–7.46 (m, 6H; PhH),
7.41–7.23 (m, 16H; PhH), 7.21–7.17 (m, 3H; PhH), 5.62 (s, 1H;
PhCHO2), 5.44 (brs, 1H; H1b), 5.12 (brs, 1H; H1c), 4.96 (brs, 1H; H5b),
4.95–4.87 (m, 3H; 3ꢅPhCH), 4.84 (d, J1,2 =3.5 Hz, 1H; H1a), 4.80 (d, J=
11.5 Hz, 1H; PhCH), 4.77 (d, J=11.5 Hz, 1H; PhCH), 4.73 (d, J=
11.0 Hz, 1H; PhCH), 4.67 (brs, 1H; H2b), 4.60 (d, J=12.0 Hz, 1H;
PhCH), 4.51 (d, J=12.5 Hz, 1H; PhCH), 4.31 (brd, J6,6’ =10.0 Hz, 1H;
H6c), 4.30 (brt, J3,2 =J3,4 =4.0 Hz, 1H; H3b), 4.22 (brs, 1H; H4b), 4.13
(brd, J6,6’ =11.0 Hz, 1H; H6a), 4.08 (m, 1H; H2a), 4.01 (m, 1H; H3c), 4.00
(m, 1H; H5c), 3.98 (t, J4,3 =J4,5 =9.5 Hz, 1H; H4a), 3.94 (brd, J5,4 =10.0 Hz,
1H; H5a), 3.89 (hept, J=6.0 Hz, 1H; CH iPr), 3.76 (t, J3,2 =J3,4 =9.5 Hz,
1H; H3a), 3.75 (brd, J6’,6 =11.0 Hz, 1H; H6a’), 3.73 (t, J4,3 =J4,5 =10.0 Hz,
1H; H4c), 3.70 (brd, J6’,6 =10.0 Hz, 1H; H6c’), 3.53 (brd, J2,3 =10.0 Hz,
1H; H2c), 2.99 (brq, J=7.0 Hz, 12H; CH2 Et3NH+), 1.68 (s, 3H;
CH3CONH), 1.26 (d, J=6.5 Hz, 3H; CH3 iPr), 1.18 (brt, J=7.0 Hz, 18H;
CH3 Et3NH+), 1.16 ppm (d, J=6.0 Hz, 3H; CH3 iPr); 13C NMR
(125 MHz, CD3OD): d=174.91, 173.13 (C=O), 140.65, 140.02, 139.87,
139.62, 139.27 (C Ph), 129.74, 129.39, 129.33, 129.26, 129.11, 129.04,
128.95, 128.71, 128.62, 128.54, 128.35, 128.03, 127.16 (CH ꢆ Ph), 102.18
(PhCHO2), 100.61 (C1b), 97.61 (C1c), 96.97 (C1a), 83.67 (C4c), 80.64 (C3a),
78.42 (C3c), 76.64 (C4a), 75.71, 75.42, 74.20 (PhCH2), 73.28 (C3b), 73.08
(C4b), 73.02 (PhCH2), 72.51 (C5a), 72.04 (C2b), 71.35 (CH iPr), 69.83 (C6a),
69.67 (C6c), 69.13 (C5b), 64.92 (C2c), 64.21 (C5c), 54.82 (C2a), 47.41
(CH2 Et3NH+), 23.65 (CH3 iPr), 22.46 (CH3CONH), 21.66 (CH3 iPr), 9.16
3.5 Hz, J2,3 =10.5 Hz, 1H; H2a), 3.91 (m, 2H; H4a +H5a), 3.87 (hept, J=
6.0 Hz, 1H CH iPr), 3.75 (t, J3,2 =J3,4 =9.5 Hz, 1H; H3c), 3.71 (m, 2H;
H
3a +H6a’), 3.64 (t, J6’,6 =J6’,5 =10.5 Hz, 1H; H6c’), 3.62 (t, J4,3 =J4,5 =
10.5 Hz, 1H; H4c), 3.59 (dd, J2,1 =3.5 Hz, J2,3 =10.0 Hz, 1H; H2c), 2.99 (q,
J=7.5 Hz, 18H; CH2 Et3NH+), 1.55 (s, 3H; CH3CONH), 1.25 (d, J=
6.0 Hz, 3H; CH3 iPr), 1.20 (t, J=7.5 Hz, 27H; CH3 Et3NH+), 1.13 ppm
(d, J=6.0 Hz, 3H; CH3 iPr); 13C NMR (100 MHz, CD3OD): d=176.17,
173.06 (C=O), 141.08, 140.94, 140.45, 140.24, 139.72 (C Ph), 129.50,
129.39, 129.28, 129.18, 129.10, 129.06, 129.03, 128.96, 128.91, 128.69,
128.53, 128.35, 128.09, 127.98, 127.12 (CH Ph), 102.35 (C1c), 101.77
(PhCHO2), 100.00 (C1b), 96.90 (C1a), 83.27 (C4c), 80.84 (C3a), 78.47 (C4b),
78.25 (C3c), 76.57 (C3b +C4a), 75.95, 75.60, 74.23, 73.02 (PhCH2), 72.68
(C2b +C5a), 71.30 (CH iPr), 69.97 (C6a), 69.91 (C5b), 69.85 (C6c), 64.34
(C5c), 60.78 (C2c), 55.06 (C2a), 47.41 (CH2 Et3NH+), 23.64 (CH3 iPr), 22.45
(CH3CONH), 21.65 (CH3 iPr), 9.42 ppm (CH3 Et3NH+); MS (ESI): m/z
positive
mode:
1291.3
[Mꢀ3Et3NH+2Na+K+H]+,
1247.7
[Mꢀ3Et3NH+K+3H]+, 1231.8 [Mꢀ3Et3NH+Na+3H]+.
Isopropyl
O-(2-deoxy-2-sulfamido-a-d-glucopyranosyl)-(1!4)-O-(2-O-
sulfo-a-l-idopyranosyluronic acid)-(1!4)-2-acetamido-2-deoxy-a-d-glu-
copyranoside sodium salt (Tri8): A solution of 22 (24 mg, 0.019 mmol) in
a MeOH/H2O=9:1 mixture (3.8 mL) was hydrogenated in the presence
of 20% Pd(OH)2/C. After 18 h, the suspension was filtered through a
pad of celite and lyophilized to give Tri8 (15 mg, 96%). [a]2D0 =+66.08
(c=0.5, H2O); TLC (5/5/3/1 EtOAc/Py/H2O/AcOH) Rf =0.29; 1H NMR
(500 MHz, D2O): d=5.32 (d, J1,2 =3.5 Hz, 1H; H1c), 5.15 (d, J1,2 =3.0 Hz,
1H; H1b), 4.92 (d, J1,2 =3.5 Hz, 1H; H1a), 4.69 (d, J5,4 =2.5 Hz, 1H; H5b),
4.25 (dd, J2,1 =3.0 Hz, J2,3 =5.5 Hz, 1H; H2b), 4.16 (dd, J3,2 =4.0 Hz, J3,4
=
5.5 Hz, 1H; H3b), 4.03 (t, J4,3 =J4,5 =3.5 Hz, 1H; H4b), 3.84 (hept, J=
6.0 Hz, 1H; CH iPr), 3.83 (m, 3H, H5a +H6a +H6a’), 3.82 (dd, J2,1 =4.0 Hz,
J
2,3 =9.5 Hz, 1H; H2a), 3.80 (dd, J6,5 =2,0 Hz, J6,6’ =12.0 Hz, 1H; H6c), 3.77
(m, 2H; H4a +H5c), 3.72 (dd, J6’,5 =4.5 Hz, J6’,6 =12.0 Hz, 1H; H6c’), 3.65
(t, J3,2 =J3,4 =9.0 Hz, 1H; H3a), 3.58 (dd, J3,4 =9.5 Hz, J3,2 =10.5 Hz, 1H;
H3c), 3.39 (t, J4,3 =J4,5 =9.5 Hz, 1H; H4c), 3.15 (dd, J2,1 =3.5 Hz, J2,3
=
10.5 Hz, 1H; H2c), 1.97 (s, 3H; CH3CONH), 1.15 (d, J=6.0 Hz, 3H;
CH3 iPr), 1.07 ppm (d, J=6.0 Hz, 3H; CH3 iPr); 13C NMR (125 MHz,
D2O): d=176.26, 175.75 (C=O), 100.63 (C1b), 98.14 (C1c), 96.25 (C1a),
78.94 (C3a), 77.20 (C2b), 77.07 (C4b), 73.09 (C5c), 72.43 (C3c), 72.27
(CH iPr), 71.31 (C4c), 70.91 (C4a +C5a), 70.36 (C5b), 70.01 (C3b), 61.68
(C6c), 61.39 (C6a), 59.44 (C2c), 55.40 (C2a), 23.65 (CH3 iPr), 23.19
(CH3CONH), 21.86 ppm (CH3 iPr); MS (ESI): m/z positive mode: 849.1
[M+Na]+, 843.1 [MꢀNa+K+H]+; m/z negative mode: 803.1 [MꢀNa]ꢀ,
(CH3 Et3NH+);
MS
(ESI):
m/z
positive
mode:
1215.4
[Mꢀ2Et3NH+Na+K+H]+, 619.2 [Mꢀ2Et3NH+2Na+K+H]2+; m/z nega-
tive mode: 1153.5 [Mꢀ2Et3NH+H]ꢀ.
Isopropyl O-(3-O-benzyl-4,6-O-benzylidene-2-deoxy-2-sulfamido-a-d-glu-
copyranosyl)-(1!4)-O-(3-O-benzyl-2-O-sulfo-a-l-idopyranosyluronic
acid)-(1!4)-2-acetamido-2-deoxy-3,6-di-O-benzyl-a-d-glucopyranoside
sodium salt (22): An aqueous NaOH solution (1.9 mL; 0.1 N) and then a
trimethylphosphine THF solution (184 mL; 1.0m) were added dropwise to
a solution of 18 (31 mg, 0.023 mmol) in THF (2.3 mL) at 08C and the
mixture was stirred at RT. After 4 h, the solution was neutralized with an
aqueous HCl solution (0.1 N), the solvents were evaporated until dryness
and the resulting crude was suspended in MeOH, filtered, concentrated,
and dried under high vacuum overnight (TLC (60/5/3/1 EtOAc/Py/H2O/
797.1 [Mꢀ3Na+K+H]ꢀ, 781.1 [Mꢀ2Na+H]ꢀ, 390.0 [Mꢀ2Na]2ꢀ
.
NMR: NMR experiments were performed on Bruker DRX 500 MHz
spectrometer equipped with 5 mm inverse triple-resonance probe. NMR
samples were prepared at pH*’7 in 500–600 or 200 mL in 5 mm or 3 mm
tubes, at 2 mm and 6 mm, respectively, in 99.9% D2O and at several tem-
peratures varying from 278 to 318 K. Sizes of acquisition matrices were
Chem. Eur. J. 2012, 18, 16319 – 16331
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16329

P. M. Nieto et al.
3
exptl
mðmþ1Þ
MD
mðmþ1Þ
1
2
< 3J
ð C4Þ > þ fð
<
Þ
3JMmðDmþ1Þð S0Þ
2 Kꢅ512 for COSY-dqf, gradient selected, experiments and 1 Kꢅ256 for
TOCSY with mixing time of 80 ms. HSQC were recorded in gradient en-
hanced versions using echo-antiecho detection both with or without de-
coupling during acquisition. When it was required presaturation was ap-
plied by low power irradiation at water frequency.
J
¼ fð
1C4
Þ
2S0
ð1Þ
In this equation f
mðmþ1ÞA
er, <3JHH (1C4)> are the averages from MD, and m is an index that
runs from 1 to 4. Therefore, the experimental JHH coupling constants
(1C4) and f(2S0) are the molar fractions of each conform-
ACHUTGTNRENNUG CAHTUNGTRENNUGN
CTHUNGTRENNUNG
The preliminary results of the NOESY were unsatisfactory because the
molecules were close to the zero crossing point and give very weak
peaks. ROESY sequences were also applied but the strong coupling of
the protons of the l-IdoA2S residue biased the results. No better results
were obtained using T-ROESY. Therefore, we recorded all the NOESY
experiments at 278 K to increase the correlation time in order to obtain
negative NOE peaks. Additionally, the use of lower temperatures al-
lowed us to exploit the increase of the population of the lowest energy
conformations. The build-up experiments were acquired with 1D se-
quence selected with gradients spin echo (dpfgse)[25] and two spin echoes
flanked by bipolar gradients during the mixing time (200, 300, 400, 500,
600, 700, 800, 1000, 1300, 1500 ms).
3
3
were considered as averages of the MD-derived JHH for each conformer
weighted on the molar fraction of each one. As the theoretical values
were averages from MD simulations, they implicitly reflected the fluctua-
tions around canonical conformations, which must be considered for this
flexible hexopyranose ring,[28] particularly to account for the pseudorota-
tional conformational space in the case of the skew boat conformer (2S0).
Acknowledgements
Molecular dynamics: All the molecular dynamics simulations have been
carried out on the Finis Terrae cluster belonging to the Centro de Super-
computaciꢁn de Galicia (CESGA), Spain, taking advantage of the priori-
tized computing time we have been awarded (ICTS CESGA 2010 call).
We thank the CSIC (Grant No. 201180E021), the Spanish Ministry of Sci-
ence and Innovation (Grant No. CTQ2009–07168), the Junta de Andalu-
cꢀa (Grant No. P07-FQM-02969, and “Incentivo a Proyecto Internacio-
nal”) and the European Union (FEDER support and Marie Curie Rein-
tegration Grant) for financial support. We acknowledge CESGA and
MICINN for co-financing the prioritized computing time awarded
(ICTS-2009–40). J.C.M. acknowledges financial support from CSIC. J.A.
acknowledges financial support from the MICINN through the Ramꢁn y
Cajal program.
In all cases, the starting geometries were generated from the available
data deposited in the Protein Data Bank (pdb code 1 hpn) and modified
accordingly. The topologies were built with PREP LINK EDIT PARM
module of Amber 5.0, employing the residues and the set of partial
charges published by Pꢇrez et al.[35] (the latter developed under the con-
text of the set of parameters for carbohydrates PIM[36]) and the force
fields parm91[37] of Amber and glycam 93[38] together with the set of
Altona parameters for sulfates.[39] Two independent starting geometries
of each heparin-like trisaccharide structure were built, one with the
l-IdoA2S residue in the chair 1C4 conformation and one with the
l-IdoA2S in the 2S0 skew boat geometry. Each of these models was im-
mersed in a 41 ꢄ-sided cube with pre-equilibrated TIP3P water mole-
cules.
[1] a) B. Casu, U. Lindahl, Adv. Carbohydr. Chem. Biochem. 2001, 57,
159–206; b) B. Mulloy, R. J. Linhardt, Curr. Opin. Struct. Biol. 2001,
11, 623–628; c) I. Capila, R. J. Linhardt, Angew. Chem. Int. Ed.
2002, 41, 390; d) U. Lindahl, J. P. Li, Int. Rev. Cell Mol. Biol. 2009,
276, 105–159.
[2] J. Choay, M. Petitou, J. C. Lormeau, P. Sinay, B. Casu, G. Gatti, Bio-
chem. Biophys. Res. Commun. 1983, 116, 492–499.
To equilibrate the system we have followed a protocol consisting of
10 steps. Firstly, only the water molecules are minimized, and then
heated to 300 K. After, the water box together with the sodium ions are
minimized and then followed by a short MD simulation. At this point,
the system is minimized in the four following steps with positional re-
straints imposed on the solute, decreasing the force constant step by step
from 20 to 5 kcalmolꢀ1. Finally, a non-restraint minimization is carried
out.
[3] a) A. Dementiev, M. Petitou, J. M. Herbert, P. G. W. Gettins, Nat.
Struct. Mol. Biol. 2004, 11, 863–867; b) M. Hricovini, M. Guerrini,
A. Bisio, G. Torri, M. Petitou, B. Casu, Biochem. J. 2001, 359, 265–
272; c) C. Laguri, N. Sapay, J. P. Simorre, B. Brutscher, A. Imberty,
P. Gans, H. Lortat-Jacob, J. Am. Chem. Soc. 2011, 133, 9642–9645.
[4] a) H. E. Conrad, Heparin Binding Proteins, Academic Press, San
Diego, 1988; b) J. D. Esko, S. B. Selleck, Annu. Rev. Biochem. 2002,
71, 435–471.
The production dynamics simulations have been accomplished at a con-
stant temperature of 300 K (by applying the Berendsen coupling algo-
rithm for the temperature scaling)[40] and constant pressure (1 bar). The
Particle Mesh Ewald Method,[37,41] (to introduce long range electrostatic
effects) and periodic boundary conditions have also been used. The
SHAKE algorithm for hydrogen atoms, which allows using a 2 fs time
step, was also employed. Finally, a 9 ꢄ cutoff was applied for the Len-
nard-Jones interactions.
[5] J. R. Bishop, M. Schuksz, J. D. Esko, Nature 2007, 446, 1030–1037.
[6] a) R. Raman, S. Raguram, G. Venkataraman, J. C. Paulson, R. Sasi-
sekharan, Nat. Methods 2005, 2, 817–824; b) R. Raman, V. Sasise-
kharan, R. Sasisekharan, Chem. Biol. 2005, 12, 267–277; c) S. Khan,
J. Gor, B. Mulloy, S. J. Perkins, J. Mol. Biol. 2010, 395, 504–521;
d) S. Khan, E. Rodriguez, R. Patel, J. Gor, B. Mulloy, S. J. Perkins, J.
Biol. Chem. 2011, 286, 24842–24854.
[7] B. Mulloy, M. J. Forster, C. Jones, D. B. Davies, Biochem. J. 1993,
293, 849–858.
[8] B. Mulloy, M. J. Forster, Glycobiology 2000, 10, 1147–1156.
[9] J. Angulo, R. Ojeda, J. L. de Paz, R. Lucas, P. M. Nieto, R. M.
Lozano, M. Redondo-Horcajo, G. Gimenez-Gallego, M. Martin-
Lomas, ChemBioChem 2004, 5, 55–61.
MD simulations have been performed with the sander module of
Amber 6.0, with explicit treatment of the 10 12 hydrogen bond potential,
in agreement with the parameters set for sulphates we used.[42] A total of
16 MD simulations of 20 ns each were obtained. The trajectory coordi-
nates were saved each 0.5 ps.
[10] a) M. J. Forster, B. Mulloy, Biopolymers 1993, 33, 575–588; b) S.
Ernst, G. Venkataraman, V. Sasisekharan, R. Langer, C. L. Cooney,
R. Sasisekharan, J. Am. Chem. Soc. 1998, 120, 2099–2107.
[11] L. Pellegrini, Curr. Opin. Struct. Biol. 2001, 11, 629–634.
[12] M. Guerrini, T. Agulles, A. Bisio, M. Hricovini, L. Lay, A. Naggi, L.
Poletti, L. Sturiale, G. Torri, B. Casu, Biochem. Biophys. Res.
Commun. 2002, 292, 222–230.
[13] J. L. de Paz, J. Angulo, J. M. Lassaletta, P. M. Nieto, M. Redondo-
Horcajo, R. M. Lozano, G. Gimenez-Gallego, M. Martin-Lomas,
ChemBioChem 2001, 2, 673–685.
The data processing of the 16 generated trajectories has been done with
the ptraj module of Amber 9.0, except for the Cremer–Pople puckering
coordinates, which have been calculated with the Carnal module of
Amber 5.0.
3
The final theoretical JHH values were obtained as averages for each of
the models (l-IdoA in 1C4 or in 2S0) according to Equation (1). Thus, to
obtain the populations of conformers (1C4 and 2S0) of the l-IdoA ring in
each trisaccharide, we have performed an iterative fitting of theoretical
and experimental J-coupling data [(Eq. (1)] (the 4C1 conformation was
disregarded as no experimental support was obtained for it, particularly
the exclusive D-GlcN-l-IdoA2S H5’-H5 NOE was not observed).
[14] P. J. Garegg, H. Hultberg, S. Wallin, Carbohydr. Res. 1982, 108, 97–
101.
16330
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16319 – 16331

Conformational Equilibrium of Heparin-Like Trisaccharides
FULL PAPER
[15] J. L. de Paz, R. Ojeda, N. Reichardt, M. Martin-Lomas, Eur. J. Org.
Chem. 2003, 3308–3324.
[30] G. Lipari, A. Szabo, J. Am. Chem. Soc. 1982, 104, 4546–4559.
[31] M. Hricovꢀni, G. Torri, Carbohydr. Res. 1995, 268, 159–175.
[32] a) R. Ojeda, J. Angulo, P. M. Nieto, M. Martin-Lomas, Can. J.
Chem. 2002, 80, 917–936; b) R. Lucas, J. Angulo, P. M. Nieto, M.
Martin-Lomas, Org. Biomol. Chem. 2003, 1, 2253–2266.
[33] a) B. Casu, J. Choay, D. R. Ferro, G. Gatti, J. C. Jacquinet, M. Peti-
tou, A. Provasoli, M. Ragazzi, P. Sinay, G. Torri, Nature 1986, 322,
215–216; b) D. R. Ferro, A. Provasoli, M. Ragazzi, G. Torri, B.
Casu, G. Gatti, J. C. Jacquinet, P. Sinay, M. Petitou, J. Choay, J. Am.
Chem. Soc. 1986, 108, 6773–6778; c) M. Ragazzi, D. R. Ferro, B.
Perly, G. Torri, B. Casu, P. Sinay, M. Petitou, J. Choay, Carbohydr.
Res. 1987, 165, c1–c5.
[16] J. L. de Paz, M. Martin-Lomas, Eur. J. Org. Chem. 2005, 1849–1858.
[17] H. Lucas, J. E. M. Basten, T. G. Vandinther, D. G. Meuleman, S. F.
Vanaelst, C. A. A. Vanboeckel, Tetrahedron 1990, 46, 8207–8228.
[18] a) S. Maza, G. Macchione, R. Ojeda, J. Lopez-Prados, J. Angulo,
J. L. de Paz, P. M. Nieto, Org. Biomol. Chem. 2012, 10, 2146–2163;
b) C. Noti, J. L. de Paz, L. Polito, P. H. Seeberger, Chem. Eur. J.
2006, 12, 8664–8686.
[19] A. Dilhas, R. Lucas, L. Loureiro-Morais, Y. Hersant, D. Bonnaffe, J.
Comb. Chem. 2008, 10, 166–169.
[20] M. Sakagami, H. Hamana, Tetrahedron Lett. 2000, 41, 5547–5551.
[21] S. Maza, J. L. de Paz, P. M. Nieto, Tetrahedron Lett. 2011, 52, 441–
443.
[34] B. Mulloy, M. J. Forster, C. Jones, A. F. Drake, A. E. A. Johnson,
B. D. B. Davies, Carbohydr. Res. 1994, 255, 1–26.
[22] Y.-P. Hu, S.-Y. Lin, C.-Y. Huang, M. M. L. Zulueta, J.-Y. Liu, W.
Chang, S.-C. Hung, Nat. Chem. 2011, 3, 557–563.
[23] E. A. Yates, F. Santini, M. Guerrini, A. Naggi, G. Torri, B. Casu,
Carbohydr. Res. 1996, 294, 15–27.
[24] S. Macura, B. T. Farmer, L. R. Brown, J. Magn. Reson. 1986, 70,
493–499.
[25] K. Stott, J. Keeler, Q. N. Van, A. J. Shaka, J. Magn. Reson. 1997,
125, 302–324.
[26] H. Hu, J. Magn. Reson. 2006, 182, 173–177.
[27] S. Macura, R. R. Ernst, Mol. Phys. 1980, 41, 95–117.
[28] J. Angulo, P. M. Nieto, M. Martin-Lomas, Chem. Commun. 2003,
1512–1513.
[35] S. Pꢇrez, C. Meyer, A. Imberty, Mol. Eng. 1995, 5, 271–300
[36] M. Clark, R. D. Cramer, N. Vanopdenbosch, J. Comput. Chem. 1989,
10, 982–1012.
[37] D. M. York, T. A. Darden, L. G. Pedersen, J. Chem. Phys. 1993, 99,
8345–8348.
[38] R. J. Woods, R. A. Dwek, C. J. Edge, B. Fraser-Reid, J. Phys. Chem.
1995, 99, 3832–3846.
[39] C. J. M. Huige, C. Altona, J. Comput. Chem. 1995, 16, 56–79.
[40] H. J. C. Berendsen, J. P. M. Postma, W. F. Vangunsteren, A. Dinola,
J. R. Haak, J. Chem. Phys. 1984, 81, 3684–3690.
[41] H. G. Petersen, J. Chem. Phys. 1995, 103, 3668–3679.
[42] C. A. G. Haasnoot, F. A. A. M. de Leeuw, C. Altona, Tetrahedron
1980, 36, 2783–2792.
[29] D. R. Ferro, A. Provasoli, M. Ragazzi, B. Casu, G. Torri, V. Bossen-
nec, B. Perly, P. Sinay, M. Petitou, J. Choay, Carbohydr. Res. 1990,
195, 157–167.
Received: August 1, 2012
Published online: November 9, 2012
Chem. Eur. J. 2012, 18, 16319 – 16331
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16331