Conformational Plasticity in Glycomimetics: Fluorocarbamethyl- L -idopyranosides Mimic the Intrinsic Dynamic Behaviour of Natural Idose Rings

Article abstract of DOI:10.1002/chem.201501249

Sugar function, structure and dynamics are intricately correlated. Ring flexibility is intrinsically related to biological activity; actually plasticity in L-iduronic rings modulates their interactions with biological receptors. However, the access to the

Full text of DOI:10.1002/chem.201501249

DOI: 10.1002/chem.201501249
Full Paper
&
Glycomimetics
Conformational Plasticity in Glycomimetics: Fluorocarbamethyl-l-
idopyranosides Mimic the Intrinsic Dynamic Behaviour of Natural
Idose Rings
Luca Unione,[a, c] Bixue Xu,[b, g] Dolores Díaz,[a] Sonsoles Martín-Santamaría,[a] Ana Poveda,[c]
Jo¼o Sardinha,[b, e] Amelia Pilar Rauter,[e] Yves BlØriot,[f] Yongmin Zhang,[b] F. Javier CaÇada,[a]
Matthieu Sollogoub,*[b] and Jesus JimØnez-Barbero*[a, c, d]
Abstract: Sugar function, structure and dynamics are intri-
cately correlated. Ring flexibility is intrinsically related to bio-
logical activity; actually plasticity in l-iduronic rings modu-
lates their interactions with biological receptors. However,
the access to the experimental values of the energy barriers
and free-energy difference for conformer interconversion in
water solution has been elusive. Here, a new generation of
fluorine-containing glycomimetics is presented. We have ap-
plied a combination of organic synthesis, NMR spectroscopy
and computational methods to investigate the conforma-
tional behaviour of idose- and glucose-like rings. We have
used low-temperature NMR spectroscopic experiments to
slow down the conformational exchange of the idose-like
rings. Under these conditions, the exchange rate becomes
slow in the 19F NMR spectroscopic chemical shift timescale
and allows shedding light on the thermodynamic and kinetic
features of the equilibrium. Despite the minimal structural
differences between these compounds, a remarkable differ-
ence in their dynamic behaviour indeed occurs. The impor-
tance of introducing fluorine atoms in these sugars mimics
is also highlighted. Only the use of 19F NMR spectroscopic
experiments has permitted the unveiling of key features of
the conformational equilibrium that would have otherwise
remained unobserved.
[a] L. Unione,+ Dr. D. Díaz, Dr. S. Martín-Santamaría, Prof. F. J. CaÇada,
Prof. J. JimØnez-Barbero
Introduction
Chemical and Physical Biology Department
Centro de Investigaciones Biológicas, CIB-CSIC
Ramiro de Maeztu 9, 28040 Madrid (Spain)
E-mail: jjbarbero@cicbiogune.es
The specific interaction between sugars and their natural re-
ceptors (e.g., lectins) is responsible for triggering many crucial
biological processes, cells–cell and cell–host dialogues,
immune response, etc.[1,2] Thus, from a biomedical viewpoint,
the knowledge about both the chemical and structural factors
that are decisive for establishing effective interactions are of
great interest for the development of new potential glycan-
based drugs.[3] Hence, the field of glycosciences has experi-
enced in the last years a blossom in research to reveal the
structure–activity relationships (SARs) that govern the effective-
ness of carbohydrate–protein interactions.[4,5] Thus, biophysical
techniques, such as X-ray crystallography or NMR spectrosco-
py,[6] among others, usually assisted by molecular modelling,
have been successfully employed to assess the structural, dy-
namic, and conformational requirements that favour the for-
mation of sugar–protein complexes.[7,8] However, natural
sugars are subject to degradation processes under natural con-
ditions in tissues and biofluids and therefore, a variety of
chemical analogues have been designed and synthesized to
overcome this problem. Thus, different glycomimetics have
been reported that compete with their natural counterparts
for the same receptors or enzymes acting, consequently, as
molecular probes or enzyme inhibitors.[9] Generally speaking,
two major groups of sugar mimics have been devised by sub-
stitution of either the endo- or exo-cyclic oxygen atom by an-
[b] Dr. B. Xu,+ Dr. J. Sardinha, Y. Zhang, Prof. M. Sollogoub
Sorbonne UniversitØs, UPMC Univ Paris 06, Institut Universitaire de France
UMR CNRS 8232, IPCM, 4, place Jussieu, 75005 Paris (France)
E-mail: matthieu.sollogoub@upmc.fr
[c] L. Unione,+ Dr. A. Poveda, Prof. J. JimØnez-Barbero
Infectious Disease Programme, CIC bioGUNE, 48160 Derio, Bizkaia (Spain)
[d] Prof. J. JimØnez-Barbero
IKERBASQUE, Basque Foundation for Science, 48011 Bilbao (Spain)
[e] Dr. J. Sardinha, Prof. A. P. Rauter
Centro de Química e Bioquímica
Faculdade de CiÞncias da Universidade de Lisboa, Edifício C8
1749-016 Lisboa (Portugal)
[f] Prof. Y. BlØriot
UniversitØ de Poitiers, UMR CNRS 7285, IC2MP, Equipe Synth›se organique,
Groupe Glycochimie, 4, avenue Michel Brunet, 86022 Poitiers Cedex
(France)
[g] Dr. B. Xu+
Present address: The Key Laboratory of Chemistry for Natural Products of
Guizhou Province and Chinese Academy of Sciences
202 Shachong South Road, Guiyang, 550002 (P. R. China)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201501249. It contains eighteen figures and
five tables including NMR spectra of the discussed compounds, NOEs and J
coupling analysis of the experimental data, structures, synthesis and com-
putational data.
Chem. Eur. J. 2015, 21, 10513 – 10521
10513
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
other atom (e.g. carbon, sulfur,
nitrogen).[10,11] However, their
conformational preferences, as
well as their internal hydrogen-
bond patterns, might differ from
their natural analogues in a way
that could also influence their in-
teractions with the key residues
within the binding site of sugar
receptors.[12] For instance, for
carbasugars, for which the ring
oxygen has been replaced by
Figure 1. Sugar ring conformations of the a-IdoA unit. Structural representation of the sugar ring conformations
(chairs, C, and skew-boat, S) of the a-IdoA unit, observed in glycosaminoglycans, such as heparin and heparan sul-
fate. Carbon atoms are numbered according to their positions within the sugar ring.
a methylene group,[13] this substitution rules out the anomeric
effects, modifies the intramolecular hydrogen-bond pattern,
modulates the amphiphilicity of the sugar ring, and results in
changes in the flexibility and conformations population distri-
butions.[14]
biological activity, thus providing direct evidence on the recog-
nition of these conformers by AT-III.[26] Obviously, this dynamic
behaviour has key implications in the kinetics and thermody-
namics of the molecular recognition event. Nevertheless, the
access to idose mimics that retain conformational plasticity
and the quantification of the experimental values of the
energy barriers and free-energy differences for the chair–skew
boat interconversion processes in water solution has remained
elusive. Recent efforts using O-substituted Ido compounds
have provided energy values in organic solvents.[27,28] However,
given the intrinsic relative low energy barrier for this equilibri-
um, NMR spectroscopic experiments in water using hydroxylat-
ed natural compounds have failed to slow down the equilibri-
um to provide quantitative and non-ambiguous values.
Fluorine is not an endogenous nucleus, but medicinal
chemistry studies have demonstrated its useful introduction in
bioactive substances to improve their pharmacokinetics prop-
erties and to modulate its biological properties. From the NMR
perspective, the presence of 19F atoms may also deliver impor-
tant conformational and structural information, complementa-
1
ry to that provided by H and 13C nuclei.[15]
For all these reasons, fluorinated glycomimetics have re-
ceived careful attention.[16] Depending on the substituted posi-
tion, the fluorine substituent can have a remarkable effect
upon the physical and chemical properties of the molecule. It
could induce increase of lipophilicity, decrease in pKa values of
certain groups by OH–F electrostatic interaction, modulate the
hydrogen-bond acceptor/donator ability or foster the presence
of a particular ring conformation.[17]
On this basis, we herein present the synthesis and conforma-
tional analysis of different fluorine-containing Ido-mimetics.
Thus, we have investigated two gem-difluorocarbasugars 1a
and 2 both analogues to methyl-b-l-idopyranoside and pos-
sessing either a quaternary or a ternary C5 respectively. As
a model compound, we have also studied the corresponding
Glc analogue 1b (Figure 2), since Glc pyranose rings are usually
In this context, difluorinated carbasugars have been recently
synthesized.[18–20] Their suitability to act as improved sugar ana-
logues has been studied. Indeed, the presence of fluorine
atoms emulates, to a certain degree, the properties of the en-
docyclic oxygen[20] and, consequently, they are better candi-
dates to mimic natural glycans than their non-fluorinated ana-
logues. The possibility of the presence of intramolecular OHÀF
hydrogen bonds should also be explored, along with their
structural and conformational implications.[21] In principle, in
those cases for which the bound conformation resembles the
ground state of the natural saccharide, glycomimetics with
closer conformational behaviour would be desirable. One of
the paradigmatic cases of conformational dynamics in glyco-
sciences is that of the Iduronic acid (l-IdoA) moiety present in
heparin. The extent of conformational mobility of l-IdoA in
heparin oligosaccharides in both the free and bound states
has also been matter of debate,[22] especially focused on the
4C1-chair–skew boat-1C4 chair equilibrium of the pyranose rings
in the free and protein-bound states (Figure 1). Interestingly,
depending on the protein receptor, distinct conformations of
the l-IdoA ring are recognized.[23] Indeed, AT-III recognizes the
skew boat conformer,[24] while the IdoA rings of a heparin hex-
asaccharide maintain the chair–skew boat flexibility when
bound to FGF-1.[25] Fittingly, for AT-III case, Sinay¨ et al. prepared
skew-boat conformationally locked compounds that keep the
Figure 2. Schematic representation of glycomimetics discussed in this work.
1
conformationally stable. 19F and H homo- and heteronuclear
NMR spectroscopic methods have been applied in water and
dimethyl sulfoxide solutions to determine their intrinsic confor-
mational and structural properties. The experimental NMR
spectroscopic data have been supported by computational
methods to unambiguously unravel the structural and confor-
mational effects of the difluoromethylene function.
Synthesis of compounds 1a and 1b used a similar route to
the one we designed to make gem-difluorocarbadisacchar-
ides,[20] and involved a Pummerer reaction to make the difluori-
nated exo-glycal.[29] Sulfidation of alcohol 3[30] was achieved by
Chem. Eur. J. 2015, 21, 10513 – 10521
www.chemeurj.org
10514
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
treatment with bis(4-methoxyphenyl) disulfide and tributyl-
phosphine in DMF to afford sulfide 4 in 96% yield. Fluorination
of sulfide 4 was performed using a Pummerer reaction, sulfide
was first oxidized to the sulfoxides using m-chloroperbenzoic
acid (mCPBA), then treatment with (diethylamino)sulfur trifluor-
ide (DAST) gave monofluorosulfides 5 in 96% yield as a mixture
of two inseparable diastereomers. The second fluorination was
realized through reaction of 5 with Selectfluor in the presence
of DAST, followed by the addition of diisopropylethylamine
(DIPEA) to afford desired difluorosulfide 6 in 60% yield. Anoth-
er mCPBA oxidation of the difluorosulfide 6 gave sulfoxides 7
in 77% yield as an inseparable mixture of diastereomers. Ther-
molysis of difluorosulfoxides 7 was achieved using Bn3N in
Ph2O at 1908C under air for 120 h to give the difluoroalkene 8
in 36% yield. An alternative route to alkene 8 was also ex-
plored starting from known alkene 9 synthesized from 3. Hexo-
glucal 9[31] was converted into lactone 10 by ozonolysis,[32] and
10 was converted into 8 using a Wittig olefination. However,
this last step appeared to be rather delicate to implement.
Treatment of the difluorovinyl compound 8 with triisobutylalu-
minium (TIBAL) gave the desired carbacycle,[33–37] which was
oxidized into ketone 11 with Dess–Martin periodinane. Reac-
tion of freshly prepared Tamao’s reagent on 11 provided b-hy-
droxysilanes, which were subjected, without purification, to ox-
idative cleavage of the SiÀC bond by basic hydrogen peroxide
to give diols 12 and 13, which were separated. Final deprotec-
tions of both 12 and 13 through hydrogenolysis using Pd/C
gave the target gem-difluorocarbasugars 1a and 1b in quanti-
tative yields. (Scheme 1; for synthesis of compound 2, see the
Supporting Information).
Results and Discussion
NMR spectroscopy
Compounds 1a and 1b only differ in the configuration of the
1
stereogenic center at C5 (Figure 2). Strikingly, the observed H,
and especially the 19F NMR spectra for both molecules, were
dramatically different (Figure 3). The assignment of the two 19F
Figure 3. 19F NMR (470 MHz) spectra of 1b (left) and 1a (right) in the solvent
D2O. Notice the broad shape of the axial fluorine resonance signal of 1a.
The integrals (1:1) are also shown.
resonances allowed us to determine that the broad signal ob-
served in 1a corresponds to the axially oriented fluorine atom.
As deduced from the visual inspection of the shape of the
equatorial fluorine in 1a, as well as of those of both fluorine
atoms in 1b, the behaviour of the axial fluorine of 1a was
rather unique. Given this particular feature, a detailed confor-
mational analysis by using NMR spectroscopic methods was
then performed. Tables 1 and 2
gather the chemical shift and J-
coupling data for the three
study molecules compared to
the data for natural methyl-l-
idopyranoside[38] and methyl-d-
glucopyranoside[39] used as refer-
ence compounds. The observa-
tions for the different molecules
are given below:
Compound 1b
Evidence for the major shape of
the six-membered ring of 1b
were extracted from the analysis
3
of the vicinal JHH coupling con-
stants (Table 1), measured with
and without 19F-decoupling con-
Scheme 1. Reagents and conditions: i) bis(4-methoxyphenyl) disulfide (1.5 equiv), nBu3P, DMF, RT, overnight, 96%;
ii) a) mCPBA (1.1 equiv), CH2Cl2, À208C to RT, overnight, 99%; b) DAST (5 equiv), CH2Cl2, 458C, 60 h, 96%; iii) Select-
fluor (1.25 equiv), DAST (0.18 equiv), CH3CN, RT, 45 min; then anhydrous DIPEA (1.5 equiv), RT, 45 min, 60%;
iv) mCPBA (1.05 equiv), CH2Cl2, À208C to RT, overnight, 77%; v) Bn3N (3 equiv) Ph2O, 1908C, 100 h, 36%; vi) a) I2,
PPh3, imidazole, b) NaH, DMF, RT, 12 h, 82%; vii) O3, Me2S, À788C; viii) CBr2F2, HMPT, À15 to 458C, 3 h; ix) TIBAL
(10 equiv), toluene, 508C, 30 min, 61%; x) Dess–Martin periodinane (3.7 equiv), CH2Cl2, RT, 2.5 h; xi) [(isopropoxy-di-
methylsilyl)methyl]magnesium chloride, 4 Š MS, THF, 08C, 1 h; xii) TBAF (4 equiv), KHCO3 (8 equiv), H2O2 (30 equiv),
THF/MeOH (1:1, v/v), RT, 45 min, 57% over three steps; xiii) H2, 10% Pd/C, MeOH, RT, 15 h, quant.
ditions (Figures S4 and S5 in the
Supporting Information). As ex-
3
pected, very large JH2H3 and
3JH3H4 values were observed in
water solution, thus demonstrat-
ing that it adopts a very major
Chem. Eur. J. 2015, 21, 10513 – 10521
www.chemeurj.org
10515
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
groups are favoured. No strong intramolecular hy-
drogen-bond between the hydroxyl moieties is
present in [D6]DMSO. Obviously, the presence of
competing water molecules in the D2O solution fur-
ther precludes this possibility. Through-space cou-
pling constants between the axial fluorine atom
with H2, H4 and H6 were also deduced. Indeed,
these couplings were also supported by heteronu-
3
n
Table 1. JHH and JHF coupling constants for 1a, 1b and 2 [Hz] in D2O solution at 300 K
and 500 MHz.[a]
3
3
3
3
2
3
3
4
2
Molecule JH1H2 JH2H3 JH3H4 JH4H5 JH6H6’ JH1Fax JH1Feq JH2Fax JH4Fax JH6Fax JH6’Feq JFeqFax
1a
1b
2
Glc
Ido
4.1
9.0 9.5
3.6 10.0 10.0
–
–
12.5 3.8[b] 7.3[b] 3.3 2.1 <1 2.8
270.0
278.0
270.0
–
12.5 1.7
3.7
9.0
–
2.5 3.4
2.8 2.8 n.a.
1.2 –
3.3
3.8
1.7
5.5 5.5 4.3 11.5 2.0
9.8 9.1 10.1 12.3
4.1 4.6 2.5
–
–
–
–
–
–
–
–
–
–
–
–
–
–
1
clear H–19F NOEs between the corresponding atom
[a] The observed values are in agreement with those expected for a very predominant
4C1 conformation for 1a and 1b isomers, while intermediate values for 2 are observed.
The experimental J coupling values for natural compounds are also provided. [b] Esti-
mated from the corresponding 1H NMR spectroscopic signals.
pairs in the HOESY spectra (Figure S6 in the Sup-
porting Information). No significant features were
observed in the 19F NMR spectra of 1b acquired at
low temperature (Figure S3 in the Supporting Infor-
mation). Therefore, all the NMR parameters and ob-
servations were in agreement with the existence of
4
Table 2. 1H and 19F NMR spectroscopic chemical shifts for 1a, 1b and 2
[Hz] in D2O solution at 300 K and 500 MHz.[a]
a C1 chair conformation, with no particular additional experi-
mental observations worth mentioning.
Molecule dH1 dH2 dH3 dH4 dH5 dH6 dH6’ dFax
dFeq
dOMe
1a
1b
2
Glc
Ido
3.82 3.64 3.60 3.50 –
3.75 3.54 3.72 3.42 –
3.89 3.96 À109.10 À119.40 3.48
3.66 3.76 À110.20 À119.40 3.47
Compound 1a (ido-like, difluoro, C5 is quaternary)
The 19F NMR spectrum of 1a was drastically different to that of
1b (Figure 3) at room temperature. Strikingly, the signal of the
axial fluorine sharpened in a noticeable manner upon decreas-
ing the temperature down to 238 K (using 20% of methanol,
Figure 4) or increasing it at 333 K. This fact suggests the exis-
3.80 3.90 3.92 3.55 2.58 4.00 3.96 À98.70 À110.0 3.56
4.79 3.54 3.65 3.38 3.63 3.85 3.74
4.69 3.53 3.73 3.75 4.09 3.79 3.82
–
–
–
–
3.50
3.45
[a] Experimental chemical shift data for methyl-b-l-idopyranoside and
methyl-a-d-glucopyranoside are also provided.
4C1 chair conformation. No significant variations of chemical
shifts and coupling constants were appreciated upon decreas-
ing the temperature (Figure S7 in the Supporting Information)
from 298 down to 248 K (adding 20% of deuterated metha-
nol). Similar couplings were observed for 1b in [D6]DMSO solu-
3
tion. Intermediate values for the JH,OH coupling constants were
observed, ranging between 5.0 and 6.6 Hz (Table 3). Tempera-
ture coefficient factors were also of medium size for all the hy-
droxyl groups, between 4.9 and 7.0 ppb8À1 (Table 4). These
facts suggest that no particular orientations of the hydroxyl
Figure 4. From bottom to top. 19F NMR (470 MHz) spectra of 1a at 238, 273
and 298 K (left), in D2O in presence of 20% of methanol and at 283, 323 and
333 K (right) in D2O.
3
Table 3. Observed JH,OH coupling constants [Hz] in DMSO solution at
300 K and 500 MHz.
Coupling constants [Hz]
3JH4,OH4
3JH2,OH2
3JH3,OH3
3JH6’,OH6
3JH6,OH6
tence of a dynamic process, which especially affects the trans-
verse relaxation features of Fax.
1a
1b
5.5
6.1
4.1
5.0
5.6
6.6
4.6
5.9
7.2
5.9
3
3
Curiously, the JH2H3 and JH3H4 coupling constant values
(Table 1) were also relatively large in water solution, which sug-
gests that a major 4C1 chair conformation indeed exists at
room temperature (see below, in the Discussion). As for 1b,
similar couplings were observed in [D6]DMSO solution, togeth-
er with medium-sized values (between 4.1 and 7.2 Hz) for the
3JH,OH couplings constants, (Table 3), and a narrow range of
temperature coefficients (between 4.6 and 6.9 ppb8À1) for the
hydroxyl groups, (Table 4) which suggests the presence of con-
formational averaging around the corresponding CÀO bonds.
Again, there is no strong intramolecular hydrogen bond be-
tween the hydroxyl moieties in the employed solvents, in con-
Table 4. Temperature coefficients measured for the different hydroxyl
groups of 1a and 1b, from the analysis of the 1H NMR spectra recorded
in DMSO between 298 and 343 K.[a]
Dd/DT [ppbKÀ1
]
Fax
8.4
Feq
OH4
OH2
OH3
OH6
OH5
1a
1b
17.2
20
6.9
6.2
6.5
7.0
6.5
7.0
5.1
5.7
4.6
4.9
18
[a] The temperature coefficients of the fluorine atoms were deduced in
D2O using the same temperature range.
Chem. Eur. J. 2015, 21, 10513 – 10521
www.chemeurj.org
10516
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
trast with the observations for protected fluorine-containing
carbohydrates in non-polar solvents.[40] Long-range coupling
constants could also be detected, from the inspection of the
1H NMR spectrum, between the axial fluorine with H2
(medium), H4 (small) and H6’ (2.8 Hz). Again, the correspond-
ing heteronuclear 1H-19F NOEs were observed for the 1H/19F
pairs in the HOESY spectrum (Figures S10 and S11 in the Sup-
porting Information). Indeed, additional information on the ge-
ometry of 1a was obtained through the careful analysis of the
1H/1H homo- and 1H/19F heteronuclear NOE experiments
(Figure 5 and Table S2 in the Supporting Information). Two key
cross-peaks were observed for 1a. There is a medium-size H3/
H6 NOE, and a weak H3/H6’ one. Moreover, a clear heteronu-
clear H6/Feq NOE was also observed, while the corresponding
H6’/Feq NOE was rather weak. The vicinal couplings between
the two H6 protons in 1a and the corresponding OH6 were
somehow different (2.6 Hz of difference). These findings are
only compatible with the existence of conformational averag-
ing in solution. The HOESY spectrum in DMSO (Figure S13 in
the Supporting Information) showed the Fax/OH5 NOE, be-
sides those observed in D2O solution. No NOEs with other hy-
droxyl groups were observed for the two 19F atoms.
The observed data suggests that the differences in the
19F NMR spectra between 1b and 1a derive from a different
dynamic behaviour of both molecules. Nevertheless, still for
3
3
1a, the relatively large JH2H3 and JH3H4 coupling constant
4
values point out the existence of a major C1 conformation for
the six-membered ring. Moreover, the simultaneous spatial
proximity between the H3/H6 and Feq/H6 pairs, together with
the through-space coupling constant between Fax and H6’
strongly suggests the existence of two orientations around
C5ÀC6, as schematized in Figure 5. Moreover, in order to justify
2
the weak H6’/Feq NOE, a minor contribution of the S5a confor-
mer has to be also considered (see also below).
Low-temperature NMR experiments in water solution (with
20% methanol) provided additional information on the nature
of the conformational equilibrium. It was observed that H1
shifted downfield upon decreasing temperature, while H2, H3
and H4 were shifted upfield (Figure S14 in the Supporting In-
formation). Strikingly, H6 and H6’ interchanged chemical shifts
Figure 5. Top panel, schematic representation of the three major conformations present for 1a. The arrows highlight the existence of NOE cross-peaks. No
unique geometry is able to completely explain the NOE data, but a combination of all the represented structures does it. w is defined by (O6-C6-C5-C5a) and
4
(O6-C6-C5-C4) torsion angles. The C1 gt conformer also justifies the H6’/Fax long-range coupling constant given its relative W-like arrangement between the
two coupled nuclei. Left, 2D NOESY spectra (700 ms mixing time) of compound 1a at 600 MHz and 298 K in D2O. Key NOEs are highlighted in the strip taken
1
at H3 frequency. No other NOEs are observed above the noise level. Right, H-19F HOESY spectrum (470/500 MHz Bruker spectrometer) of 1a in D2O in the
presence of 20% methanol (800 ms mixing time). The strips taken at frequency of equatorial fluorine are highlighted.
Chem. Eur. J. 2015, 21, 10513 – 10521
www.chemeurj.org
10517
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
during the cooling process. As mentioned above, the axial different possible conformers for the molecules discussed here
19F NMR signal became sharper at low temperature and no ad- (Figure 7).
ditional 19F signals were evident.
For the Glc-configurated 1b, the theoretical calculations well
4
predicted the C1 as unique possible conformer. The other two
1
1
local minima, C4 and S3, displayed much larger relative free
energy, about 11.0 and 8.0 kcalmolÀ1, respectively. For this
system, the transition-state structures were also deduced to
build the potential energy surface. The activation energies
were estimated as approximately
Compound 2
The 19F NMR spectrum of 2 displayed two broad 19F NMR spec-
troscopic signals (Figure 6) at room temperature. Interestingly,
15.0 kcalmolÀ1 for the 3E enve-
lope and 13.5 kcalmolÀ1 for the
5aH5 half-chair shapes.
1
For 2, the alternative C4 chair
was predicted as the global min-
4
imum, with the C1 chair destabi-
lized in approximately 1.7 kcal
molÀ1. In this case, the equatorial
orientation of the two bulky
groups is probably the impetus
for the calculated energy values.
2
In this case, the S5a conformer
displayed the less favourable
energy value.
For 1a, the 1C4 form was
strongly destabilized with re-
1
Figure 6. Variable temperature H NMR (left) and decoupled 19F-{1H} NMR spectrum (right) of 2 in D2O (with 20%
spect to the global minimum,
methanol) at 500 (1H) or 470 (19F) MHz. From top to bottom: A) 318, B) 298 K, C) 283 K, D) 273 K, E) 263 K, F) 253 K.
Notice that the chemical shifts of the H NMR spectroscopic signals do not show major shifts with temperature.
4
the C1 conformer. The presence
1
of several 1,3-diaxially oriented
the signals sharpened in a noticeable manner upon decreasing
the temperature and two new clear 19F NMR spectroscopic sig-
nals appeared at 253 K (using a 20% of methanol, Figure 6).
The conformational equilibrium is slow in the 19F NMR chemi-
cal shift scale at this temperature. Using the observed coales-
cence temperature between 283 and 298 K and the estimated
chemical shift difference of the two set of fluorine signals at
253 K, the energy barrier was calculated to be approximately
11.8Æ0.4 kcalmolÀ1 [41]
.
The relative populations of the two sig-
nals were 70:30, which indicated that the free-energy differ-
ence was of only approximately 0.4 kcalmolÀ1
.
Because of the severe proton overlapping, the J coupling
constants were extracted from spectral simulation (Figure S16
3
3
in the Supporting Information). The JH2H3 and JH3H4 (Table 1)
values were intermediate (ca. 5.5 Hz for both) in solution,
which suggests the presence of a conformational equilibrium,
as shown by the two sets of 19F signals at 253 K. In contrast to
the observations for 1a, in this case, it was observed that none
1
of the H NMR spectroscopic signals were significantly shifted
at low temperature. As will be described below in the discus-
sion, this behaviour also contains key conformational informa-
tion.
Computational methods
Figure 7. Ball and stick representation and relative steric energy values of
the major conformers of 1a, 1b and 2, according to DFT calculations (see
also the Supporting Information).
The molecular modelling protocol described in the Experimen-
tal Section was adopted for structure assessment, providing
Chem. Eur. J. 2015, 21, 10513 – 10521
www.chemeurj.org
10518
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
4
groups provides the structural basis for these predictions. The
2S5a skew boat displays most of the bulky groups in pseudo-
equatorial orientations, resulting in a relative low destabilizing
energy value (4.0 kcalmolÀ1).
similar, as expected for the two alternative C1 and 1C4-chair
conformers. The free-energy difference between the two forms
is approximately 0.4 kcalmolÀ1, favouring the C4 chair in the
1
same trend as the energy differences estimated by the calcula-
tions. The energy barrier for interconversion is relatively low
(ca. 11.8 kcalmolÀ1), extremely difficult to access by variable
The conformational behaviour of idose rings and derivatives
thereof has been a matter of investigation for years.[22,42–45] It is
well known that for l-idopyranoses, the theoretically more fa-
vourable 1C4 chair displays three axially oriented hydroxyl
groups, with the corresponding steric consequences. The alter-
1
temperature H NMR spectroscopic experiments in water solu-
tion. The use of 19F NMR spectroscopy has permitted access to
this value, due to the wide accessible chemical shift range.
Indeed, computational chemistry calculations using ab initio
methods (DFT (B3LYP) in vacuum with the different basis sets:
4
native C1 chair places the bulky hydroxymethyl group at the
axial orientation, with the corresponding collapse. Therefore,
alternative skew-boat conformers are also present in the con-
formational equilibrium, depending on the hydroxyl substitu-
tion, chemical environment, and solvent.[46–48] Therefore, the
description of the conformational flexibility of these rings in
terms of thermodynamic parameters as activation free energy,
and entropic and enthalpy contributions to the Gibbs free
energy difference represents a challenge to the experimental
study. Such a description implies access to a detailed and accu-
rate proton-proton coupling constant analysis, which is ham-
pered by signal broadening and or overlapping. In fact, the sit-
uation becomes even more arduous, and most of the times in-
accessible, in cases of fast and medium conformational equili-
bria in the NMR chemical shift timescale, especially when it in-
volves entropically favoured isoforms, such as skew boat
conformers, which is often the case for idose ring derivatives.
As mentioned above, for ido-like sugars, flexibility is intrinsi-
cally related to biological activity. The sulfated l-iduronic rings
represent the paradigmatic example of how plasticity modu-
lates the interaction with biological receptors. From the chemi-
cal perspective, these molecular recognition processes are the
consequence of the balance between enthalpy and entropy
factors and solvation/desolvation effects. In this context, since
conformational entropy becomes an issue, it is essential to
consider that the conformational entropy of chair and skew
boat conformers is intrinsically different. Chair conformers are
defined in well-characterized potential energy wells, while the
conformational entropy of skew boat conformers is larger, due
to the low-energy cost geometry interconversions that con-
duct to basically the same conformer. The energy well for skew
boat conformers is much wider than that for the chairs.
4
1
6-31+ +G, for the C1 and C4 geometries; 6-31+ +G +freq,
for S5a; and 6-31+ +GTS +freq, for 3E and 5aH5), provided
2
energy values fairly similar to those experimentally detected
3
(12.5 kcalmolÀ1 for E and 13.0 for 5aH5).
In contrast, compound 1a shows a unique conformational
behaviour. The chemical shift and coupling constant values
drastically changed upon temperature variation (Figure S14 in
the Supporting Information) indicating that its conformational
distributions depend on the temperature. This observation
strongly suggests that the conformational entropy of the con-
tributing geometries is different. At low temperature, the en-
thalpy-favoured conformer should be predominant, since the
entropy contribution to free energy will be largely attenuated
(DG=DHÀTDS). Fittingly, H2, H3, and H4 shift upfield more
than d=0.1 ppm upon decreasing temperature. Concomitant-
ly, H1 shifted downfield. This fact provides evidence that, at
low temperature, the predominant conformer of 1a displays
H2, H3 and H4 in axial orientation, while H1 shows an equato-
rial arrangement. Therefore, the major and enthalpy-favoured
4
conformer is the C1 chair. The other participating conformer
should display a skew boat geometry since its contribution to
the conformational equilibrium strongly decreases at low tem-
2
perature. Computational chemistry calculations found the S5a
conformer as the most stable skew boat form, with a relative
energy of about 4.0 kcalmolÀ1 with respect to the 4C1 chair.
1
The C4 form was strongly destabilized, by more than 10 kcal
molÀ1. In fact, for the S5a conformer, H2, H3 and H4 display
2
3
3
a quasi-axial orientation, providing large JH2H3 and JH3H4 cou-
plings.
3
3
Since these observed JH2H3 and JH3H4 couplings at room
temperature for both molecules were already rather large
(above 9 Hz), the contribution of the skew boat conformers
would have been clearly neglected from the inspection of the
1H and 13C NMR spectra, unless the 19F NMR spectra would not
have shown dramatically broad signals.
Interestingly, the conformational behaviour of these 19F-con-
taining glycomimetics remarkably resembles the intrinsic flexi-
bility of the natural Ido-configurated sugars (Table S5 in the
Supporting Information). Although this fact might not com-
pletely surprising, as a matter of fact, regular Ido-like carbasu-
gars,[49] with a CH2 group mimicking the endocyclic oxygen
atom, did not show any conformational plasticity (Table S5 in
the Supporting Information). In contrast, the molecules pre-
sented herein, with CF2 moieties resembling the endocyclic
oxygen atom show important conformational plasticity. The
dynamic process has been quantified in terms of energy barri-
Under these premises, the obtained results can now be ac-
counted for in a satisfactory manner. Compound 1b (Glc-like)
4
displays a unique C1 chair conformer with a very well defined
geometry, as in the natural compound. Similarly to the natural
Ido-like molecule, compound 2, displays significant conforma-
tional distortions. There is a clear-cut behaviour for compound
2. 19F-based variable temperature experiments demonstrate
the existence of a conformational equilibrium between two
forms with an approximately 70:30 population distribution.
The observed coupling constants are in fact in agreement with
4
1
a 70:30 distribution between the canonical C1 (minor) and C4
(major) chair forms (Table S4 in the Supporting Information).
No changes in either chemical shifts or coupling constants are
observed with temperature, which indicates that the confor-
mational distribution is temperature-independent. Therefore,
the conformational entropy of the contributing geometries is
Chem. Eur. J. 2015, 21, 10513 – 10521
www.chemeurj.org
10519
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the J coupling constants were extracted from spectral simulation
using MestreNova software. The method of determining activation
energy parameters is through the estimation of the coalescence
temperature and chemical shift difference for each of the fluorine
signals measured in the spectrum at lower temperature (253 K)
giving 2900 Hz for low-field signals and 4010 Hz for higher field
signals.[41] The exchange rate for both fluorine signals can be now
estimated by applying Equation (1):
ers and free-energy differences. The destabilizing energies for
the C1 conformer in 2 is 0.4 kcalmolÀ1 above that for the C4
chair. However, when C5 is modified (as in 1a, with one addi-
tional OH substituent), there is a participation of skew boat
conformers, while the proportion of the 1C4 conformer is
strongly diminished. The OH3 group displays an equatorial ori-
4
1
4
2
entation in the C1 and S5a geometries, minimizing the influ-
ence of steric conflicts with the additional OH5. However, it
1
would adopt an axial disposition in the C4 form, provoking
pDv
p
kc ¼
ð1Þ
important additional steric clashes. For the Glc-like molecules,
the conformational behaviour of the CF2 analogue also mimics
that of natural glucopyranosides. Therefore, these glycomimet-
ics can behave as conformational bioisosteres.
2
After observing the fluorine NMR spectra at different temperatures
(Figure 6), it was possible to enclose the activation energy barrier
between the limiting values calculated for 283 and 298 K by apply-
ing the Eyring equation [see Eq. (2)]:
Conclusions
ꢀ
ꢁ
ꢂ
ꢃ
KBTc
h
DG° ¼ RTc ln
ÀlnðkcÞ
ð2Þ
A new generation of fluorine-containing glycomimetics is pre-
sented. The importance of introducing fluorine atoms in these
glycomimetics is also highlighted. First, only the use of
19F NMR spectroscopic experiments has permitted the detec-
tion of a dynamic process of paramount significance that
would have been otherwise remained unobserved. Additional-
ly, only in the presence of fluorine atoms at C5a, the Ido-like
six-membered ring recovers its required flexibility, absent in
regular CH2-Ido-carbasugars,[38] while the presence of a bulky
substituent at position C5 strongly reduces the ring flexibility
and introduces important steric clashes. A recent report from
our group[20] has shown that the CF2 moiety partially recovers
the exo-anomeric effect typical of oligosaccharides, thus resem-
bling the conformational behaviour of glycans around the gly-
cosidic linkage. Herein, we also demonstrate that the presence
of the fluorine in the ring additionally restores the plasticity of
Ido-like six-membered rings.
in which R is the gas constant, Tc the coalescence temperature, KB
the Boltzmann constant, h the Planck constant, and kc the deter-
mined exchange rate. With this above equation, the activation
energy barrier relative to the low-field signal lies between the limit-
ing values of 11.6 kcalmolÀ1 at 283 K and 12.2 kcalmolÀ1 at 298 K,
giving an average value of 11.9 kcalmolÀ1 at the estimated coales-
cence temperature of 290 K, while for the high-field signal it lies
between 11.4 and 12.0 kcalmolÀ1 with an average value of 11.7 kcal
molÀ1 at 290 K. The final estimation for the activation energy is
DG° =11.8 kcalmolÀ1 Æ0.4.
Computational methods
A conformational search on these molecules 1a, 1b and 2 was
performed by using Macromodel at the Maestro suite of programs,
with the MM3* force field. Different local minima conformers were
chosen within a conservative 20 kcalmolÀ1 threshold from the
global minimum. Their expected coupling constant values and
proton–proton distances (related to NOEs) were estimated from
the corresponding structures using Maestro. The transition-state
geometries for the interconversion process were chosen based on
the well known Cremer–Pople sphere conformational routes. Then,
the MM3*-optimized structures were used as starting conforma-
tions for additional ab initio calculations. Thus, DFT geometry opti-
mizations were performed with the Gaussian 03 program using the
hybrid B3LYP functional and the 6–31+ +G(d,p) basis set followed
by vibrational frequency analysis. This protocol allowed assessing
whether the optimized structures were true energy minima, transi-
tion states, or saddle points. For the transition-states structures,
the final geometry optimization was achieved by applying the TS
Berny algorithm. In all cases, the presence of solvent was account-
ed for by the integral equation formalism polarizable continuum
model (IEFPCM). The NMR isotropic shielding constants were calcu-
lated using the standard Gauge Independent Atomic Orbital
(GIAO) approach. The experimental and calculated NMR coupling
constants were then compared.
Thus, the combination of NMR spectroscopic experiments
and computational methods has permitted the demonstration
that these idose-like analogues resemble the conformational
plasticity of the natural parent molecules that is anticipated to
be required for the key molecular recognition process and ulti-
mately for biological activity.
Experimental Section
NMR Spectroscopy
19F NMR spectroscopic experiments were performed 470 MHz with
a Bruker AVANCE spectrometer equipped with the proper fluorine
probe SEF, at 298 K unless otherwise stated while low temperature
experiments were done with Bruker DRX 500 MHz equipped with
BBOF plus probe. 1H NMR spectroscopic experiments were per-
formed at 600 and 700 MHz with a Bruker AVANCE spectrometer
equipped with TXI probe. Experiments were performed in D2O,
[D6]DMSO and in D2O in the presence of 20% methanol for low
temperature analysis. The concentration employed was 2 mm for
all the discussed molecules. In addition to standard 1D 1H NMR
spectra, COSY, TOCSY, NOESY and HOESY (800 ms mixing time) and
1H/13C HSQC experiments based on the standard BRUKER sequen-
ces were also acquired, in order to assign the resonance of all NMR
spectroscopic signals. Because of the severe proton overlapping,
Acknowledgements
We thank the Ministry of Economy and Competitiveness of
Spain for financial support (Grants CTQ2012-32025 and
Chem. Eur. J. 2015, 21, 10513 – 10521
www.chemeurj.org
10520
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[23] J. Turnbull, A. Powell, S. Guimond, Trends Cell Biol. 2001, 11, 75–82.
[24] W. Li, D. J. Johnson, C. T. Esmon, J. A. Huntington, Nat. Struct. Mol. Biol.
2004, 11, 857–862.
[25] A. Canales, R. Lozano, B. López-MØndez, J. Angulo, R. Ojeda, P. M. Nieto,
M. Martín-Lomas, G. GimØnez-Gallego, J. JimØnez-Barbero, FEBS J. 2006,
273, 4716–4727.
CTQ2011-22724). We also thank EU for the ITN/ETN Marie Curie
Actions Glycopharm, Dynano, and Tollerant as well as the
China Scholarship Council for a PhD fellowship to B.X., and
FundaÅao para a CiÞncia e Tecnologia for a research grant
(J.S.).
[26] S. K. Das, J.-M. Mallet, J. Esnault, P.-A. Driguez, P. Duchaussoy, P. Sizun,
J.-P. HØrault, J.-M. Herbert, M. Petitou, P. Sinay¨, Angew. Chem. Int. Ed.
2001, 40, 1670–1673; Angew. Chem. 2001, 113, 1723–1726.
[27] J. Rçnnols, S. Manner, A. Siegbahn, U. Ellervik, G. Widmalm, Org. Biomol.
Chem. 2013, 11, 5465–5472.
Keywords: carbasugars
fluorine · l-idose · NMR spectroscopy
·
conformational equilibrium
·
[28] J. Rçnnols, S. Manner, U. Ellervik, G. Widmalm, Org. Biomol. Chem. 2014,
12, 8031–8035.
[1] D. Solís, N. V. Bovin, A. P. Davis, J. JimØnez-Barbero, A. Romero, R. Roy, K.
Smetana, Jr., H.-J. Gabius, Biochim. Biophys. Acta 2015, 1850, 186–235.
[2] H.-J. Gabius, S. AndrØ, J. JimØnez-Barbero, A. Romero, D. Solís, Trends
Biochem. Sci. 2011, 36, 298–313.
[3] S. A. W. Gruner, E. Locardi, E. Lohof, H. Kessler, Chem. Rev. 2002, 102,
491–514.
[4] Q. H. Fan, K. A. Claunch, S. Striegler, J. Med. Chem. 2014, 57, 8999–9009.
[5] A. Ardµ, P. Blasco, D. Varón Silva, V. Schubert, S. AndrØ, M. Bruix, F. J.
CaÇada, Jr., H.-J. Gabius, C. Unverzagt J. JimØnez-Barbero, J. Am. Chem.
Soc. 2013, 135, 2667–2675.
[6] L. Unione, S. Galante, D. Díaz, F. J. CaÇada, J. JimØnez-Barbero, Med.
Chem. Commun. 2014, 5, 1280–1289.
[7] M. C. Fernµndez-Alonso, D. Díaz, M. A. Berbis, F. Marcelo, F. J. CaÇada, J.
JimØnez-Barbero, Curr. Protein Pept. Sci. 2012, 13, 816–830.
[8] A. Canales, A. Mallagaray, M. A. Berbís, A. Navarro-Vµzquez, G. Domí-
nguez, F. J. CaÇada, S. AndrØ, Jr. , H.-J. Gabius, J. PØrez-Castells, J. JimØ-
nez-Barbero, J. Am. Chem. Soc. 2014, 136, 8011–8017.
[9] A. García-Herrero, E. Montero, J. L. MuÇoz, J. F. Espinosa, A. Viµn, J. L.
García, J. L. Asensio, F. J. CaÇada, J. JimØnez-Barbero, J. Am. Chem. Soc.
2002, 124, 4804–4810.
[10] P. Vidal, V. Roldós, M. C. Fernµndez-Alonso, B. Vauzeilles, Y. BlØriot, F. J.
CaÇada, S. AndrØ, Jr. , H.-J. Gabius, J. JimØnez-Barbero, J. F. Espinosa, S.
Martín-Santamaría, Chem. Eur. J. 2013, 19, 14581–14590.
[11] M. Risseeuw, M. Overhand, G. W. Fleet, M. I. Simone, Amino Acids 2013,
45, 613–689.
[12] J. JimØnez-Barbero, J. F. Espinosa, J. L. Asensio, F. J. CaÇada, A. Poveda,
Adv. Carbohydr. Chem. Biochem. 2000, 56, 235–284.
[13] E. Montero, A. García-Herrero, J. L. Asensio, K. Hirai, S. Ogawa, F. San-
toyo-Gonzµlez, F. J. CaÇada, J. JimØnez-Barbero, Eur. J. Org. Chem. 2000,
1945–1952.
[14] B. López-MØndez, C. Jia, Y. Zhang, P. Sinay¨, L.-H. Zhang, J. JimØnez-Bar-
bero, M. Sollogoub, Chem. Asian J. 2008, 3, 51–58.
[15] A. Vulpetti, U. Hommel, G. Landrum, R. Lewis, C. Dalvit, J. Am. Chem.
Soc. 2009, 131, 12949–12959.
[16] E. Leclerc, X. Pannecoucke, M. Eth›ve-Quelquejeu, M. Sollogoub, Chem.
Soc. Rev. 2013, 42, 4270–4283.
[17] E. Prchalovµ, O. Stepµnek, S. Smrcek, M. Kotora, Future Med. Chem.
2014, 6, 1201–1229.
[18] A. Deleuze, C. Menozzi, M. Sollogoub, P. Sinay¨, Angew. Chem. Int. Ed.
2004, 43, 6680–6683; Angew. Chem. 2004, 116, 6848–6851.
[19] J. Sardinha, S. Guieu, A. Deleuze, M. C. Fernµndez-Alonso, A. P. Rauter, P.
Sinay¨, J. Marrot, J. JimØnez-Barbero, M. Sollogoub, Carbohydr. Res. 2007,
342, 1689–1703.
[29] E. T. Jarvi, J. R. McCarthy, S. Mehdi, D. P. Matthews, M. L. Edwards, N. J.
Prakash, T. L. Bowlin, P. S. Sunkara, J. Med. Chem. 1991, 34, 647–656.
[30] H. Paulsen, M. Paal, Carbohydr. Res. 1983, 113, 203–218.
[31] D. Semeria, M. Philippe, J.-M. Delaumeny, A.-M. Sepulchre, S. D. Gero,
Synthesis 1983, 710–713.
[32] G.-N. Wang, G. Reinkensmeier, S.-W. Zhang, J. Zhou, L.-R. Zhang, L.-H.
Zhang, T. D. Butters, X.-S. Ye, J. Med. Chem. 2009, 52, 3146–3149.
[33] S. K. Das, J.-M. Mallet, P. Sinay¨, Angew. Chem. Int. Ed. Engl. 1997, 36,
493–496; Angew. Chem. 1997, 109, 513–516.
[34] M. Sollogoub, J.-M. Mallet, P. Sinay¨, Tetrahedron Lett. 1998, 39, 3471–
3472.
[35] M. Sollogoub, A. J. Pearce, A. HØrault, P. Sinay¨, Tetrahedron: Asymmetry
2000, 11, 283–294.
[36] A. J. Pearce, M. Sollogoub, J.-M. Mallet, P. Sinay¨, Eur. J. Org. Chem. 1999,
2103–2117.
[37] M. Sollogoub, J.-M. Mallet, P. Sinay¨, Angew. Chem. Int. Ed. 2000, 39, 362–
364; Angew. Chem. 2000, 112, 370–372.
[38] B. M. Sattelle, B. Bose-Basu, M. Tessier, R. J. Woods, A. S. Serianni, A.
Almond, J. Phys. Chem. B 2012, 116, 6380–6386.
[39] M. U. Roslund, P. Tähtinen, M. Niemitz, R. Sjçholm, Carbohydr. Res. 2008,
343, 101–112.
[40] G. T. Giuffredi, V. Gouverneur, B. Bernet, Angew. Chem. Int. Ed. 2013, 52,
10524–10528; Angew. Chem. 2013, 125, 10718–10722.
[41] J. Von Sandstrçm, Dynamic NMR Spectroscopy, Academic Press, London
1982.
[42] J. Rçnnols, S. Manner, U. Ellervik, G. Widmalm, Org. Biomol. Chem. 2014,
12, 8031–8035.
[43] J. R. Snyder, A. S. Serianni, J. Org. Chem. 1986, 51, 2694–2702.
[44] M. B. Sattelle, U. S. Hansen, J. Gardiner, A. Almond, J. Am. Chem. Soc.
2010, 132, 13132–13134.
[45] A. Canales, J. Angulo, R. Ojeda, M. Bruix, R. Lozano, G. GimØnez-Gallego,
M. Martín-Lomas, P. M. Nieto, J. JimØnez-Barbero, J. Am. Chem. Soc.
2005, 127, 5778–5779.
[46] A. Ard›vol, X. BiarnØs, A. Planas, C. Rovira, J. Am. Chem. Soc. 2010, 132,
16058–16065.
[47] J. Angulo, P. M. Nieto, M. Martín-Lomas, Chem. Commun. (Camb) 2003,
1512–1513.
[48] X. BiarnØs, A. Ard›vol, A. Planas, C. Rovira, A. Laio, M. Parrinello, J. Am.
Chem. Soc. 2007, 129, 10686–10693.
[49] E. SäwØn, M. U. Roslund, I. Cumpstey, G. Widmalm, Carbohydr. Res.
2010, 345, 984–993.
[20] B. Xu, L. Unione, J. Sardinha, S. Wu, M. Eth›ve-Quelquejeu, A. P. Rauter,
Y. BlØriot, Y. Zhang, S. Martín-Santamaría, D. Díaz, J. JimØnez-Barbero, M.
Sollogoub, Angew. Chem. Int. Ed. 2014, 53, 9597–9602; Angew. Chem.
2014, 126, 9751–9756.
[21] C. Dalvit, A. Vulpetti, ChemMedChem 2012, 7, 262–272.
[22] B. Casu, J. Choay, D. R. Ferro, G. Gatti, J.-C. Jacquinet, M. Petitou, A. Pro-
vasoli, M. Ragazzi, P. Sinay¨, G. Torri, Nature 1986, 322, 215–216.
Received: March 30, 2015
Revised: May 8, 2015
Published online on June 10, 2015
Chem. Eur. J. 2015, 21, 10513 – 10521
www.chemeurj.org
10521
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim