Selective targeting of dendritic cell-specific intercellular adhesion molecule-3-grabbing nonintegrin (Dc-Sign) with mannose-based glycomimetics: Synthesis and interaction studies of bis(benzylamide) derivatives of a pseudomannobioside

Article abstract of DOI:10.1002/chem.201202764

Dendritic cell-specific intercellular adhesion molecule-3-grabbing nonintegrin (DC-SIGN) and Langerin are C-type lectins of dendritic cells (DCs) that share a specificity for mannose and are involved in pathogen recognition. HIV is known to use DC-SIGN on DCs to facilitate transinfection of T-cells. Langerin, on the contrary, contributes to virus elimination; therefore, the inhibition of this latter receptor is undesired. Glycomimetic molecules targeting DC-SIGN have been reported as promising agents for the inhibition of viral infections and for the modulation of immune responses mediated by DC-SIGN. We show here for the first time that glycomimetics based on a mannose anchor can be tuned to selectively inhibit DC-SIGN over Langerin. Based on structural and binding studies of a mannobioside mimic previously described by us (2), a focused library of derivatives was designed. The optimized synthesis gave fast and efficient access to a group of bis(amides), decorated with an azide-terminated tether allowing further conjugation. SPR inhibition tests showed improvements over the parent pseudomannobioside by a factor of 3-4. A dimeric, macrocyclic structure (11) was also serendipitously obtained, which afforded a 30-fold gain over the starting compound (2). The same ligands were tested against Langerin and found to exhibit high selectivity towards DC-SIGN. Structural studies using saturation transfer difference NMR spectroscopy (STD-NMR) were performed to analyze the binding mode of one representative library member with DC-SIGN. Despite the overlap of some signals, it was established that the new ligand interacts with the protein in the same fashion as the parent pseudodisaccharide. The two aromatic amide moieties showed relatively high saturation in the STD spectrum, which suggests that the improved potency of the bis(amides) over the parent dimethyl ester can be attributed to lipophilic interactions between the aromatic groups of the ligand and the binding site of DC-SIGN. Receptor targeting: For the first time glycomimetics based on a mannose anchor have been tuned to selectively inhibit DC-SIGN (dendritic cell-specific intercellular adhesion molecule-3-grabbing nonintegrin) over Langerin. Based on structural and binding studies of a mannobioside mimic previously described, a focused library of derivatives was designed (see figure). Copyright

Full text of DOI:10.1002/chem.201202764

DOI: 10.1002/chem.201202764
Selective Targeting of Dendritic Cell-Specific Intercellular Adhesion
Molecule-3-Grabbing Nonintegrin (DC-SIGN) with Mannose-Based
Glycomimetics: Synthesis and Interaction Studies of Bis(benzylamide)
Derivatives of a Pseudomannobioside
Norbert Varga,[a] Ieva Sutkeviciute,[b, c, d] Cinzia Guzzi,[e] John McGeagh,[f]
Isabelle Petit-Haertlein,[b, c, d] Serena Gugliotta,[a] Jçrg Weiser,[f] Jesffls Angulo,[e]
Franck Fieschi,[b, c, g] and Anna Bernardi*[a]
Abstract: Dendritic cell-specific inter-
cellular adhesion molecule-3-grabbing
nonintegrin (DC-SIGN) and Langerin
are C-type lectins of dendritic cells
(DCs) that share a specificity for man-
nose and are involved in pathogen rec-
ognition. HIV is known to use DC-
SIGN on DCs to facilitate transinfec-
tion of T-cells. Langerin, on the contra-
ry, contributes to virus elimination;
therefore, the inhibition of this latter
receptor is undesired. Glycomimetic
molecules targeting DC-SIGN have
been reported as promising agents for
the inhibition of viral infections and for
the modulation of immune responses
mediated by DC-SIGN. We show here
for the first time that glycomimetics
based on a mannose anchor can be
tuned to selectively inhibit DC-SIGN
over Langerin. Based on structural and
binding studies of mannobioside
mimic previously described by us (2), a
focused library of derivatives was de-
signed. The optimized synthesis gave
fast and efficient access to a group of
tested against Langerin and found to
exhibit high selectivity towards DC-
SIGN. Structural studies using satura-
tion transfer difference NMR spectro-
scopy (STD-NMR) were performed to
analyze the binding mode of one repre-
sentative library member with DC-
SIGN. Despite the overlap of some sig-
nals, it was established that the new
ligand interacts with the protein in the
same fashion as the parent pseudodi-
saccharide. The two aromatic amide
moieties showed relatively high satura-
tion in the STD spectrum, which sug-
gests that the improved potency of the
a
bisACTHNUTRGNE(UNG amides), decorated with an azide-
terminated tether allowing further con-
jugation. SPR inhibition tests showed
improvements over the parent pseudo-
mannobioside by a factor of 3–4. A di-
meric, macrocyclic structure (11) was
also serendipitously obtained, which af-
forded a 30-fold gain over the starting
compound (2). The same ligands were
bisACTHNUTRGENUG(N amides) over the parent dimethyl
ester can be attributed to lipophilic in-
teractions between the aromatic groups
of the ligand and the binding site of
DC-SIGN.
Keywords: DC-SIGN
·
glycomi-
metics · HIV · NMR spectroscopy ·
proteins
Introduction
an interesting target for the design of antiviral agents.[1,2] Im-
mature dendritic cells in mucosal tissue use DC-SIGN to
recognize high-mannose glycans present on the viral enve-
lope glycoprotein gp120 of HIV. This recognition event ap-
pears to contribute to infection by promoting viral transmis-
sion in a manner dependent on the composition of gp120
The dendritic cell (DC) membrane C-type lectin receptor
dendritic cell-specific intercellular adhesion molecule-3-
grabbing nonintegrin (DC-SIGN) has been implicated in the
early stages of HIV infection and is, therefore, considered
[a] N. Varga,+ S. Gugliotta, A. Bernardi
[f] J. McGeagh, J. Weiser
Universitaꢀ degli Studi di Milano, Dipartimento di Chimica
via Golgi 19, 20133 Milano (Italy)
Anterio Consult&Research GmbH, Augustaanlage 23
68165 Mannheim (Germany)
[b] I. Sutkeviciute,+ I. Petit-Haertlein, F. Fieschi
Universitꢁ Grenoble I, Institut de Biologie Structurale
41 rue Jules Horowitz, Grenoble, 38027 (France)
[g] F. Fieschi
Institut Universitaire de France
103 boulevard Saint-Michel 75005 Paris (France)
[+] These authors contributed equally to this work.
[c] I. Sutkeviciute,+ I. Petit-Haertlein, F. Fieschi
CNRS, UMR 5075, Grenoble, 38000 (France)
[d] I. Sutkeviciute,+ I. Petit-Haertlein
CEA, DSV, Grenoble, 38000 (France)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202764.
[e] C. Guzzi, J. Angulo
Glycosystems Laboratory, Instituto de Investigaciones
Quꢂmicas (IIQ), CSIC - Universidad de Sevilla
Amꢁrico Vespucio 49, 41092 Sevilla (Spain)
4786
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 4786 – 4797

FULL PAPER
glycans.[3] Ligands that can be used to explore the multiple
functions of DC-SIGN and to inhibit DC-SIGN-mediated
pathogen binding, are actively pursued by different
groups.[4,5,6] Selective inhibition of DC-SIGN, however, has
rarely been addressed, despite the presence in the immune
system of many other C-type lectins that share DC-SIGNs
ability to bind high-mannose oligosaccharides. Among these,
Langerin, which is expressed on the surface of a different
subset of antigens presenting cells known as Langerhans
cells (LCs), has been suggested to have protective effects
against HIV infection.[7] Selective DC-SIGN ligands, which
interact only weakly with Langerin, are potentially useful
therapeutic tools against sexually transmitted HIV infec-
tion.[6d,8]
Glycomimetic structures designed to inhibit DC-SIGN
have been reported.[4a–c,6] In our previous work, we focused
on the design and synthesis of monovalent N-fucosylami-
des[6b,d] or on mannose conjugates, such as 1–3 (Fig-
ure 1).[6a,c,g] These molecules have been conceived to be met-
abolically more stable than native oligosaccharides and can
be equipped with reactive moieties that enable a multimeric
presentation on different scaffolds. Indeed, the pseudotrisac-
charide 1, mimicking the linear Man-a-1,2-Man-a-1,6-tri-
mannoside, the D3 arm of the high mannose structure Man9,
was found to inhibit DC-SIGN binding to mannosylated
the amide R group, but most of them efficiently inhibited
DC-SIGN-mediated adhesion. The data showed that the ac-
tivity increased if the R group was lipophilic. However,
most compounds in this series possess a low solubility in
aqueous media, which limits their applicability. Preliminary
NMR spectroscopic studies of their interaction with DC-
SIGN also indicated that some nonspecific association oc-
curred through the allyl linker.
Recently, structural studies combining X-ray crystallogra-
phy and NMR spectroscopic analysis allowed us to charac-
terize the binding mode of 2 within the DC-SIGN binding
site and to demonstrate that it corresponds to a unique well-
defined orientation, making 2 a good lead compound for
chemical improvement of the binding affinity and specificity
towards DC-SIGN.[8] Analysis of the X-ray structure showed
that the methyl ester groups of 2 in the complex extend to-
wards the DC-SIGN surface in an area in which larger sub-
stituents can be accommodated. Larger groups in this posi-
tion can be expected to reach the protein surface and make
additional contacts. These observations validated the initial
strategy that had led to the synthesis of bisACTHNURTGNEU(GN amides) 3 and
suggested that their structures should be reconsidered with
the goal of increasing water solubility and removing the
allyl group, as the source of undesired nonspecific interac-
tions with the protein. In the group of compounds previous-
ly examined, the best balance of synthetic accessibility, se-
lectivity, and solubility was achieved for benzylamides 3a–b,
which showed an IC50 in the mm range in the adhesion assay.
These molecules provided interesting leads, which could be
improved by optimizing the nature of the benzylamide sub-
stituents and by replacing their allylether appendage with a
functional tether, which would also allow multimeric presen-
tation of the ligands.
Bovine Serum AlbuminACHTNURGTNE(UNG Man-BSA) with an IC50 of 130 mm
by surface-plasmon resonance (SPR).[6c] Moreover, when
presented on a tetravalent dendron, 1 was also found to in-
hibit HIV transinfection of CD4+T lymphocytes[6c] and of
cervical tissue explants[6f] at micromolar concentrations. We
have also recently reported a small library of mannose con-
jugates of general formula 3, which were tested by using a
dendritic cell adhesion assay to mannan-coated plates.[6 g]
The activity of these compounds was tuned by the nature of
With these goals in mind, a group of functionalized bis-
(benzylamides) of general for-
mula 4 (Scheme 1) was synthe-
sized. They contain an azidoe-
thanol linker, which can be ex-
ploited for conjugation, and dif-
ferent polar groups on the ben-
zylamide ring, to increase water
solubility and possibly improve
affinity for the receptor. These
molecules were tested for their
activity as DC-SIGN ligands by
using an SPR competition
assay. Selectivity against Lan-
gerin was also addressed. NMR
spectroscopic studies were per-
formed to analyze the interac-
tion of the ligands with DC-
SIGN and establish the epitope
region. Altogether these results
showed that bis(benzylamido)
pseudodimannosides bind selec-
tively to DC-SIGN and allowed
Figure 1. Structure of Man9 and of previously reported mannose-based DC-SIGN inhibitors 1–3.
the selection of 4h as the opti-
Chem. Eur. J. 2013, 19, 4786 – 4797
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4787

A. Bernardi et al.
tures. By only using p-xylylendi-
amine (10p, see Scheme S1 in
the Supporting Information)
one major product could be
chromatographically isolated in
modest yields (18%). MS and
13C NMR spectroscopic analysis
revealed that the dimeric mac-
rocyclic structure 11 was assem-
bled, as a 1:1 mixture of insepa-
rable diastereoisomers that was
tested as such in the following
assays. Finally, to analyze the
effect of the stereochemistry of
the cyclohexane scaffold, com-
pound 12, a stereoisomer of 4h,
Scheme 1. Synthesis of bis(benzylamides) 4.
mal ligand for the future con-
struction of multivalent systems.
In the course of this study a di-
valent macrocyclic ligand (11)
was also serendipitously ob-
tained, which showed a low mm
affinity for DC-SIGN in the
SPR assay used.
Results and Discussion
A
high-yielding and flexible
synthesis of the entire library
was achieved by small modifica-
tions of the previously reported
procedure.[6g] By starting from
Figure 2. The bis-benzylamides 4a–o, 11 and 12 synthesized and tested in this work.
enantiomerically pure (1S,2S)-
diacid 5,[9] the bis(p-nitropheny-
lester) was synthesized and the
double bond oxidized by using mCPBA (meta-chloroperben-
zoic acid), to afford epoxide 6. Copper-catalyzed epoxide
opening[10] with azidoethanol,[11] followed by mannosylation
of alcohol 7 with trichloroacetemidate 8[12] afforded the
pseudodisaccharide 9 in only four steps and with 42% over-
all yield. The two activated p-nitrophenyl esters, used as
acid-protecting groups throughout the sequence so far, were
transformed into amides by reaction with an excess of the
was synthesized by starting from the (1R,2R)-enantiomer of
5 (see Scheme S8 in the Supporting Information). Analysis
of the H NMR spectra confirmed that all compounds share
1
the same chair conformation of the cyclohexane ring ob-
served in 2[13] (see Figure S1 in the Supporting Information).
The activity of these compounds as inhibitors of DC-
SIGN binding to mannosylated BSA (Man-BSA) was tested
by SPR (Figure 3). In the assay, Man-BSA was immobilized
on a CM4 chip and the extracellular domain (ECD) of DC-
SIGN was injected over the Man-BSA surface in the ab-
sence or in the presence of increasing concentrations of the
inhibitors. The obtained IC50 values are shown in Table 1.
The parent pseudodisaccharide bis(methyl ester) 2 and the
pseudotrisaccharide 1 are included in the Figure 3 data as
reference compounds. Many of the diamides prepared
showed a remarkable increase in inhibitory activity com-
pared to diester 2 and some approached the affinity ob-
served for 1, a molecule of significantly higher structural
and synthetic complexity. Low solubility remains a problem
appropriate benzylamine (10), yielding bisACTHUNRTGNEUNG(amides) 4 after
Zemplꢁn deprotection of the sugar.
Docking studies[6g] had suggested that H-bond donors and
acceptors out of the plane of the aromatic ring could be
beneficial for interaction with the protein surface. This con-
sideration, together with water solubility and synthetic avail-
ability, drove the selection of benzylamines 10 and of the
group of 16 compounds 4a–o (Figure 2, Table 1) to be syn-
thesized, all bearing polar, oxygen-containing groups on the
aromatic ring. Bis(benzylamines) were also considered;
however, their reaction with 9 led mostly to complex mix-
4788
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 4786 – 4797

Targeting of DC-SIGN with Mannose-Based Glycomimetics
FULL PAPER
Table 1. Structure and activity of DC-SIGN ligands 4a–o, 11, and 12.
the aromatic residue around the
N-benzylic bond. Similarly,
compound 12, a diastereoisom-
er of 4h with the opposite con-
figuration (1R,2R) of the cyclo-
hexane ring, binds poorly to
DC-SIGN, which indicates that
the design of the pseudodisac-
charide scaffold is important to
optimize interaction with the
protein as suggested by the
structure of the 2/DC-SIGN
complex.[8] Remarkably, dimer
11 (1:1 mixture of isomers) with
an IC50 of 31 mm turned out to
be the most potent inhibitor of
the series, and one of the most
effective reported so far. The
dimension of the macrocycle in
[b]
[b]
Structure
Yield[a] IC50
Structure
Yield[a] IC50
[%]
(R=)
[%]
[mm]
(R=)
[mm]
4a
4b
67
296[c]
4j
65
367
398
73
75
441
814
4k
76
64
4c
4l
335
4d
4e
72
60
810[c] 4m
63
80
317
405
324
310
356
4n
4o
11
11
(OH-4–OH-4
distance:
22 ꢄ) is too short to span two
adjacent binding sites of the
DC-SIGN tetramer (which are
separated by no less than
35 ꢄ),[14] thus the potency of
this compound must derive
from proximity effects (statisti-
cal rebinding) or by aggregation
of the protein,[15] which, in turn,
could block the availability of
DC-SIGN binding sites and
thus effectively inhibit binding
4 f
4g
59
65
28
–
1461
31
dimer (Figure 2)
4h
4i
82
78
325
290
12
2
diastereoisomer of 4h (Figure 2) 65[d]
1421
986
to
immobilized
Man-BSA.
Since the effect is rather large,
protein oligomerization under
see Figure 1
–
[a] Unoptimized overall yield from 9. [b] In SPR competition test with immobilized Man-BSA. [c] Low solubil- the assay conditions appears to
ity in water. [d] Isolated as a byproduct from glycosylation of a batch of non enantiomerically pure 7 (see the
Supporting Information for details).
be the most likely explanation.
Protein aggregation by polyva-
lent carbohydrates is very well
documented;[16] however, given
for some of the structures (4a and 4d, shown in white in
Figure 3), but most compounds displayed good solubility in
water at the concentration required for the assay.
the heterogeneity of the compound and the low yield with
which it was synthesized, proving the mechanism underlying
the activity of 11 would be challenging. Furthermore, even if
an oligomerization mechanism could be demonstrated under
the conditions of the current assay (by using soluble DC-
SIGN extra-cellular domain (ECD)), it is very unlikely that
it will be operative in the biologically relevant conditions, in
which the protein would be membrane-bound. Thus the
issue was not pursued further.
In parallel, most of the library members were further
tested for inhibition of Langerin by using an SPR assay that
we have recently described.[6d] Selectivity for DC-SIGN
versus Langerin is particularly important to develop inhibi-
tors that would block sexually transmitted HIV infection.
As we have already remarked, while interaction with DC-
SIGN on mucosal DCs is used by the virus to invade the
host immune system, Langerin was shown to have protective
The group of diamides 4e–n, built upon the modeling sug-
gestion mentioned above, showed a remarkable activity, as
all the compounds were found to be only 2.5-fold less
potent than the pseudotrisaccharide 1 (Figure 1). Among
them, 4h, featuring a hydroxymethylene group in the para
position, was selected for structural modifications. The hy-
droxy group appears to play a role as a H-bond acceptor,
since the corresponding methyl ether shows the same inhibi-
tory potency (cf. 4h and 4i). Addition of fluorine atoms on
the ring (4k–l) or additional lipophylic groups in the prox-
imity of the acceptor (4j) did not significantly improve the
affinity. On the contrary, two methoxy substituents meta to
the hydroxymethylene group, as in compound 4o, had a neg-
ative effect, possibly as a result of a different orientation of
Chem. Eur. J. 2013, 19, 4786 – 4797
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4789

A. Bernardi et al.
Figure 3. SPR results representing inhibition of DC-SIGN ECD binding
to Man-BSA surface by mannose-based compounds: inhibition curves (a)
and extracted corresponding IC50 values of mannose-based ligands (b). In
light grey: IC50 bars of the reference compounds 1 and 2. In white: com-
pounds that displayed low solubility under the assay conditions.
Figure 4. Inhibition of Langerin ECD binding to the Man-BSA/dextran
SPR surface by mannose-based compounds: inhibition curves (a), DC-
SIGN and langerin Inhibition level obtained for 0.5 mm of the com-
pounds (b).
bis(methylester) 2 is endowed with a modest selectivity for
effects against HIV infection.[7] To get an insight into the se-
lectivity of ligands 1–4, they were tested in the same concen-
tration range as inhibitors of DC-SIGN (20 mm) and of Lan-
gerin (15.5 mm) binding to dextran/Man-BSA surface. As can
be seen from the inhibition curves (Figure 4a), none of the
tested compounds were able to fully inhibit Langerin bind-
ing to the surface, despite the lower concentration of Lan-
gerin used in the assay relative to DC-SIGN ECD concen-
tration (due to limited availability of the protein). The most
potent inhibitors were found to be dimer 11 and the referen-
ces 1 and 2. Because most of the compounds at the highest
concentration used did not reach 50% inhibition of Langer-
in, it was not possible to calculate their IC50 values. Never-
theless, to get a comparison of DC-SIGN versus Langerin
inhibition, the inhibition levels (%) achieved at three com-
pound concentrations, 0.5, 1.0, and/or 1.5 mm, were deter-
mined. Data obtained at 0.5 mm are plotted in Figure 4b
(see Figure S2 in the Supporting Information for DC-SIGN
and Langerin inhibition levels at 1 and 1.5 mm ligand con-
centration). The data show that the pseudo-disaccharide
DC-SIGN as previously reported.[8] The corresponding pseu-
dotrisaccharide 1 and many of the bisACTHNUTRGNEUNG(amides) 4, that is, 4a,
4 f, 4h, and 4i, were found to inhibit DC-SIGN with an im-
proved selectivity relative to Langerin, a very desirable fea-
ture for further development as antiviral agents. Fucose-
based ligands, both natural[17] and unnatural[6d] have been re-
ported to recognize DC-SIGN selectively. However, man-
nose-containing oligosaccharides appear to bind both lec-
tins[17,18] and it has long been believed that selectivity could
not be achieved with mannose-based ligands. Our current
results show that this goal can be reached.
With this information in hand, bisACHTUNTRGNEUNG(amide) 4h was selected
for future development in multivalent constructs, since this
ligand appears to combine good affinity and selectivity for
DC-SIGN with good solubility, synthetic accessibility, and
atom efficiency. This ligand was therefore chosen for struc-
tural analysis by NMR spectroscopy to eventually identify
ligand-receptor contacts that are essential for complex for-
mation and compare its binding mode to that observed for 2
in the X-ray structure of its complex with DC-SIGN.[8] Ini-
4790
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 4786 – 4797

Targeting of DC-SIGN with Mannose-Based Glycomimetics
tial attempts were frustrated because the high-field (500 and
FULL PAPER
1
600 MHz) H NMR spectrum of 4h showed some overlap-
ping of signals in areas crucial for our analysis. Previous
NMR work on the parent bis(methylester) 2 had shown that
its triazole derivative 14 (Scheme 2) is a better substrate for
spectroscopic analysis. The triazole residue in the linker of
Figure 5. STD NMR spectroscopic study of the interaction of ligand 13
with DC-SIGN in buffered D2O (150 mm NaCl, 4 mm CaCl2, 25 mm d-
Tris, pD 8). a) 1H NMR reference spectrum (off-resonance frequency
40 ppm) and b) STD spectrum (on-resonance frequency d=0 ppm) of a
sample containing 13 (1 mm) and DC-SIGN ECD (19 mm) at 258C
(500 MHz) and a saturation time of 0.5 s. Key proton signals are labeled.
tified. The STD spectrum of 13 in the presence of DC-SIGN
ECD (50:1 ligand/protein ratio) at room temperature is re-
ported in Figure 5b.
To obtain information on the binding epitope, the initial
slope approximation (STD0)[20] was used, to reduce interfer-
ences from differences in relaxation properties of each
proton of the ligand and from rebinding processes.[21] The
experimental STD growth curves are collected in Figure 6.
Despite the signal overlap problems, a detailed compari-
son of the STD data obtained for 13 and 14 allows us to
qualitatively establish that the two molecules share the same
binding mode. For 14, quantitative analysis of the STD re-
sults had shown a very good match with the three-dimen-
sional structure of the DC-SIGN/2 complex determined by
X-ray crystallography (Figure 7a): binding occurs through
coordination of the calcium ion by OH-3 and OH-4 of the
mannose residue in an orientation which allows the cyclo-
hexane ring to establish a van der Waals contact with the
V351 side chain of the receptor; this results in important
STD signals for the mannose moiety and for all cyclohexane
Scheme 2. Triazole derivatives used for NMR spectroscopic interaction
studies.
14 did not affect binding to DC-SIGN ECD relative to 2,
neither in terms of affinity nor in terms of protein-ligand
contacts, but it significantly improved the dispersion of sig-
1
nals in the H NMR spectrum and allowed a full characteri-
zation of the binding epitope.[8] Thus, based on this experi-
ence, 4h was transformed into the corresponding triazole 13
(Scheme 2). For comparison, and to examine the role played
by the scaffold configuration in pseudomannobioside
mimics, the diastereomeric bisACTHNUGTRNEUNG(amide) 12 was also trans-
formed into the corresponding triazole 15 and its interaction
with DC-SIGN was studied by NMR spectroscopy.
The spectrum of 13 is indeed more resolved than the spec-
trum of 4h, but not fully resolved. In particular the 6ax
proton of the cyclohexane ring (H6ax(C)) and the 3 proton
of mannose (H3(M)), both belonging to the epitope of 14,
overlap with protons 6eq (H6eq(C)) and 3eq (H3eq(C)) (m,
at d=1.89–1.82 ppm) and with proton 2 (H2(C)) (m, at d=
3.84 ppm) of the cyclohexane ring, respectively (Figure 5a,
see Scheme 2 for numbering of the cyclohexane ring).
Nonetheless saturation transfer difference (STD) NMR
spectroscopic studies could be performed and the data were
compared with those obtained for the reference ligand 14.
STD-NMR spectra is a useful and robust technique to
obtain structural information on molecular recognition proc-
esses within the range of fast exchange kinetics (KD from
the 10 nm to 10 mm range).[19] Attempts to resolve signal
overlapping by varying the temperature were ineffective
and some of the most interesting signals could not be quan-
Figure 6. Experimental STD growth curves of 13 as a function of satura-
tion time.
Chem. Eur. J. 2013, 19, 4786 – 4797
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4791

A. Bernardi et al.
and H2(M)), but showing only a single contact between the
protein and the cyclohexane ring, through the H4 proton
(H4(C); see Table S1 and Figure S3 in the Supporting Infor-
mation). These data are in agreement with the lower affinity
measured for 15 in the SPR experiments and further support
the notion that appropriate design of the cyclohexane ring
configuration is crucial to the activity of the mannobioside
mimics in this series.
Analysis of the NOESY spectra of 13 revealed the clear
existence of the same diagnostic cross-peaks critically defin-
ing inter-residue contacts (highlighted in Figure 8) in both
NOESY (Figure 8a, free-ligand state) and TR-NOESY (Fig-
Figure 7. a) X-ray structure of the DC-SIGN/2 complex (from ref. [8]).
b) Docked pose of 13 that qualitatively accounts for the STD data.
protons, particularly H6ax(C) (see Figure S3 in the Support-
ing Information). Although no accurate integration of some
of the key protons was possible for 13, overall the same pat-
tern of STD intensities emerged (for a comparison of the
STD0 values, see the Table S1 in the Supporting Informa-
tion; for a comparison of the full STD spectra see Figure S3
in the Supporting Information) leaving little doubt that 13
has the same binding mode as 14. Additionally, the STD0
NMR spectroscopic signals obtained for 13 are higher than
those for the corresponding protons of 14 under the same
conditions (see Table S1 in the Supporting Information). If
both binding modes are the same, this is also indicative of a
higher affinity of 13 relative to the parent compound.
Indeed, the bis(benzylamido) moieties of 13 appear to es-
tablish further contacts with the protein (H(Ar), Figures 5
and 6), which can explain the significant increase in affinity
for 13 compared to 14. Docking of mimic 13 in the 1L4 crys-
tal structure[22] was carried out by using the quantum polar-
ized ligand docking (QPLD) workflow[23] based on the dock-
ing algorithm Glide.[24] This procedure identified poses con-
sistent with the crystal structure of the DC-SIGN/2 com-
plex,[8] and also displayed additional interactions between
the bis(benzylamido) moieties and the protein surface. A
docking pose that qualitatively accounts for the STD data is
shown in Figure 7b.
Figure 8. Expansions of NOESY experiments at 258C (500 MHz) of 13.
a) Free state, mixing time 500 ms. b) In the presence of DC-SIGN ECD,
mixing time 500 ms. Labels indicate some key NOE peaks.
ure 8b, bound ligand state). This suggests that there is no
major variation in the conformation of 13 upon binding to
DC-SIGN, or, in other words, that the lectin recognizes the
main conformation(s) existing in solution.
The STD spectra of the diastereomeric ligand 15 obtained
under the same conditions suggested a very different bind-
ing mode involving the mannose moiety (H3(M), H4(M),
4792
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 4786 – 4797

Targeting of DC-SIGN with Mannose-Based Glycomimetics
FULL PAPER
2 were activated with 50 mL of an 0.2m 1-ethyl-3-(3-dimethylaminopro-
pyl)carbodiimide) (EDC)/0.05m N-hydroxysuccinimide (NHS) mixture,
then Fc2 was functionalized with mannosylated bovine serum albumin
(Man-BSA) by injecting protein (60 mgmLꢀ1) prepared in sodium acetate
(10 mm, pH 4). Finally, the remaining activated groups of both flow cells
were blocked with ethanolamine (30 mL, 1m). After blocking, the three
flow cells were treated with HCl (10 mL, 10 mm) to remove unspecific
bound protein and then EDTA (20 mL, 50 mm) to expose surface to re-
generation protocol. On average, 2000 RU of Man-BSA were immobi-
lized after these steps.
Conclusion
We have shown here that benzylic bisACTHNUTRGNEUNG(amides) of general
formula 4 bind to DC-SIGN with affinities approaching that
of the synthetically more demanding pseudotrisaccharide 1.
Replacement of an allylic linker and inclusion of substitu-
ents designed to increase water solubility afforded a class of
molecules that outperform the group of amides of general
formula 3[6 g] previously found to bind somewhat nonspecifi-
cally to the protein and to display low solubility. Indeed,
NMR interaction studies showed no sign of nonspecific asso-
ciation between DC-SIGN and compound 13, derived from
4h. These studies also allowed us to determine that 13 binds
to DC-SIGN with the same binding mode of the parent
pseudodisaccharide bis(methylester) 2, that was recently de-
termined by high-resolution X-ray analysis.[8] Among the
group of molecules synthesized, the dimeric structure 11 was
found to display the best affinity for DC-SIGN (31 mm).
However, this ligand was not pursued further, since its activ-
ity is likely to depend on its ability to cross-link the DC-
SIGN ECD tetramer, a feature that can contribute to in-
creased selectivity in the format of the assay performed (in-
hibition of DC-SIGN ECD binding to immobilized Man-
BSA), but might be not relevant in the real physiological sit-
uation (DC-SIGN “immobilized” on DC surface). Among
the molecules prepared here, ligand 4h was selected for fur-
ther studies, due to a combination of good solubility, syn-
thetic accessibility, activity, and selectivity. Polyvalent pre-
sentations of this ligand on dendrimeric scaffolds are pres-
ently being examined and will be reported in due course.
Probably the most important characteristic that we could
determine for compounds 4a–o is their selectivity for DC-
SIGN against a second mannose-binding lectin, Langerin,
also expressed by a (different) subset of dendritic cells and
endowed with protective properties against HIV infection.
Selective inhibition of DC-SIGN versus Langerin is consid-
ered an important requirement for the development of anti-
infective agents able to block the first stages of infection
during sexual transmission of the virus. The current observa-
tion that amides 4 display enhanced selectivity that remains
significant up to 100% DC-SIGN inhibition shows that the
goal of selective mannose-based DC-SIGN ligands can be
achieved. More generally, these results strongly suggest that
the activity and selectivity of glycomimetic molecules based
on a monosaccharide anchor can be tuned by structural
modifications of secondary elements in the molecule and
bode well for the development of selective low molecular
weight lectin ligands.
For inhibition studies, mixtures of each lectin (DC-SIGN ECD (20 mm)
or Langerin (15.5 mm)) with increasing concentrations of inhibiting com-
pounds were prepared in a running buffer composed of Tris (25 mm,
pH 8), NaCl (150 mm), CaCl2 (4 mm), and P20 surfactant (0.005%), and
each sample (13 mL) was injected onto the surfaces at a 5 mLminꢀ1 flow
rate. Concentrations of inhibiting compounds ranged from 0.7 to
4583 mm. The resulting sensorgrams were reference surface corrected,
except for the case of Langerin; because this lectin displayed high affinity
to the dextran matrix, a dextran/Man-BSA surface was considered as a
combined ligand of Langerin, as previously described.[6d] The lectin-bind-
ing responses were extracted from the sensorgrams, converted to percent
residual activity values (y) with respect to lectin-alone binding, and then
plotted against corresponding compound concentrations. The 4-parame-
ter logistic model [Eq. (1)] was fitted to the plots, and the IC50 values
were calculated by using the values of fitted parameters (Rhi, Rlo, A1, and
A2) and Equation (2).
RhiꢀRlo
y ¼ Rhiꢀ
ꢀ
ꢁ
A2
ð1Þ
ð2Þ
c
1 þ
A1
ꢀ
ꢁ
1
A2
RhiꢀRlo
Rhiꢀ50
IC50 ¼ A1
ꢀ1
NMR spectroscopy experiments: NMR spectroscopic experiments were
recorded on a Bruker Avance DRX 500 MHz instrument equipped with
a 5 mm inverse triple-resonance probe head. NMR samples were pre-
pared in D2O (550 mL, 99.9%) and, for the experiments in the presence
of the receptor, in the same amount of buffer D2O (NaCl (150 mm),
CaCl2 (4 mm), d-Tris (25 mm, pD 8) and DC-SIGN ECD (19 mm). The
concentration of the ligand was 5 mm for assignment experiments (1H,
COSY, TOCSY, NOESY, HSQC) and 1 mm for both STD NMR and
transferred NOESY spectroscopic experiments.
STD NMR spectroscopic experiments were carried out at 10, 25, and
358C by using a train of Gaussian shaped pulses of 49 ms (field strength
of ca. 80 Hz), an interpulse delay of 1 ms[19] and 15 ms spin-lock pulse
(field strength of 3.7 kHz) prior to acquisition. The on-resonance fre-
quency was set to d=0 ppm and the off-resonance frequency was
40 ppm. Appropriate blank experiments were performed to assure the
lack of direct saturation of the ligand protons. Saturation times of 0.5, 1,
1.5, 2, 3, 4, and 5 s were used to obtain the STD build-up curves.
The binding epitope was characterized by the analysis of initial slopes of
the STD intensities[26]: the experimental (I0ꢀIsat/I0) curves were fitted to
an exponential function described by the equation: STD (tsat)=STDmax
(1ꢀe ꢀksat·tsat), which allows us to calculate STD at zero saturation time
[20]
(initial slopes) by the product of resulting parameters STDmax and ksat
.
NOESY experiments were carried out at 0.1, 0.2, 0.3, and 0.5 s, by using
a phase-sensitive pulse program with gradient pulses in the mixing time
and with presaturation.[27,28]
Docking studies
Experimental Section
Protein setup: By starting from the crystal structure (resolution=1.55 ꢄ)
of human DC-SIGN in a complex with Man4 (PDB code 1L4),[22] a mo-
lecular model of the protein was constructed. By using the ꢅProtein Prep-
aration Wizardꢀ within the Maestro graphical interface,[29] the crystal
structure was modified by deleting all crystal waters, assigning bond
orders and adding hydrogen atoms. Protonation states of basic and acidic
residues were assigned by optimization of the hydrogen-bonding net-
work. An Impref minimization was carried out on the final protein struc-
DC-SIGN ECD protein (residues 66–404) were overexpressed and puri-
fied as described previously.[14] Langerin ECD construct, comprising resi-
dues 68–328, were overexpressed as inclusion body, then refolded and pu-
rified to homogeneity in a functional form as already described.[25] Sur-
face plasmon resonance experiments were performed on a Biacore 3000
by using a CM4 chip, functionalized at 5 mLminꢀ1. Flow cells (Fc) 1 and
Chem. Eur. J. 2013, 19, 4786 – 4797
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4793

A. Bernardi et al.
ture by using the OPLS2005 force field,[30] converging all heavy atoms to
a root-mean-squared deviation (RMSD) of 0.30 ꢄ of the crystallographic
positions.
4,5-Epoxy-cyclohexan-1,2-dicarboxylic
acid
bis(4-nitrophenyl)ester
(1S,2S) (6): EDC·HCl (394 mg, 2.05 mmol, 3.5 equiv) was added to a sol-
ution of the diacid 5[10] (100 mg, 0.59 mmol, 1 equiv) in dry THF (5.8 mL)
with stirring and under a nitrogen atmosphere. After 10 min, p-nitrophe-
nol (245 mg, 1.76 mmol, 3 equiv) was added. The solution was stirred at
room temperature for 2 h. After completion of the reaction, the solvent
was evaporated under reduced pres-
sure, the residue was taken up in Et2O,
washed with 1m HCl, saturated
Na2CO3 (3ꢆ), and water; the mixture
was then dried over sodium sulfate.
Ligand conformational search: A conformational search was carried out
on each ligand by using a mixed torsional/low-mode sampling method.[31]
The conformational sampling was performed by using the OPLS2005
force field,[30] an implicit water model[32] with constant dielectric, and
with the van der Waals, electrostatic and hydrogen bond cutoffs set to 8,
20, and 4 ꢄ, respectively. Extended torsion sampling was used with the
maximum number of steps set to 1000. The energy window for saving
structures was set to 42 kJmolꢀ1 (10.04 kcalmolꢀ1). The resulting confor-
mations were clustered by using an average linkage clustering technique,
with representatives closest to the centroid of each cluster selected for
docking.
The solvent was evaporated under re-
duced pressure to yield 74% of pure
(1S,2S)-4-cyclohexen-1,2-dicarboxylic
acid bis(p-nitrophenyl) (PNP) ester as
Docking: The protein structure described above was used to generate a
grid to be employed in the docking study. An ꢅenclosing boxꢀ with dimen-
sions of 36 ꢄ, centered on the ligand in the crystal structure, was set;
nonplanar conformations of amide bonds were penalized, and the partial
charge cutoff was set to 0.25. No constraints were used so as to allow li-
gands to freely explore the grid.
a pale-yellow solid.
[a]2D0 = +129.6 (c=1 in CHCl3); 1H NMR (400 MHz, CDCl3): d=8.28–
8.22 (m, 4H; H10), 7.27–7.22 (m, 4H; H9), 5.83 (appd, J=2.8 Hz, 2H; H4,
H5), 3.27–3.19 (m, 2H; H1, H2), 2.78–2.68 (m, 2H; H3ps–eq, H6ps–eq), 2.48–
2.37 ppm (m, 2H; H3ps–ax, H6ps–ax); 13C NMR (100 MHz, CDCl3): d=172.8
(C7); 155.4 (C11); 145.7 (C8); 125.5 (C10); 124.9 (C5, C4); 122.5 (C9); 41.5
(C1, C2); 28.0 ppm (C3, C6).
Docking studies were carried out by using the QPLD[23] workflow based
on Glide.[24] This technique allows for charge polarization of the ligands
induced by the protein environment, which is essential within highly
charged active sites, such as DC-SIGN. QPLD improves the partial
charges on the ligand atoms in Glide docking run by replacing them with
charges generated from a quantum-mechanical (QM) calculation on the
ligand in the field of the receptor. By using the previously described grid,
along with the input structures obtained from conformational clustering,
QPLD was performed. Initial charges for each ligand were generated
with a semi-empirical method and were rigidly docked in Glide Extra
Precision (XP) mode[33] with van der Waals radii scaled by 0.8. The QM
charges were calculated by using density functional theory at the B3LYP/
3-21g level of theory by using the Jaguar program.[34] The generated
poses from the initial docking stage were subsequently re-docked rigidly
in XP mode by using the calculated QM charges for the ligands.
mCPBA (77%, 891 mg, 3.98 mmol, 1.2 equiv) was added to a solution of
the PNP ester prepared in the previous step (1367 mg, 3.32 mmol,
1 equiv) in dry CH2Cl2 (11 mL) under stirring. The reaction was stirred
under nitrogen at room temperature
for 16 h. The solvent was evaporated
at reduced pressure and then the reac-
tion mixture was diluted with EtOAc
and washed with saturated NaHCO3
(3ꢆ) and with water. The organic
phase was dried over sodium sulfate
and the solvent evaporated under re-
duced pressure to yield 91% of pure
product 6 as a white solid.
[a]2D0 = +82.2 (c=1.1 in CHCl3); 1H NMR (400 MHz, CDCl3): d=8.28–
8.22 (m, 4H; H10), 7.27–7.21 (m, 4H; H9), 3.43–3.36 (m, 1H; H4 or H5),
3.35–3.31 (m, 1H; H4 or H5), 3.32–3.24 (m, 1H; H1 or H2), 3.08–3.00 (m,
1H; H1 or H2), 2.79–2.71 (m, 1H, H3eq or H6eq), 2.63–2.54 (m, 1H; H3eq or
Redundant binding modes, in which no calcium coordination was ob-
served, were removed by filtering docked poses by the OH-3–Ca2+ dis-
tance. Distances greater than 4 ꢄ were discarded and the remaining
poses were minimized, and binding energies estimated, with the Prime
MM-GBSA program.[35]
H
H
6eq), 2.42–2.33 (m, 1H; H3ax or H6ax), 2.21–2.12 ppm (m, 1H; H3ax or
6ax); 13C NMR (100 MHz, CDCl3): d=172.6, 171.5 (C7); 155.4, 155.3
Synthesis
(C8); 149.9 (C11); 125.5 (C10); 122.5, 122.6 (C9); 51.8, 50.3 (C4, C5); 40.1,
38.1 (C1, C2); 26.8, 26.3 ppm (C3, C6); MS (ESI): m/z (%): for
[C20H16N2O9Na]+: 451.3; found: 452.0; HRMS: m/z calcd for
[C20H16N2O9Na]+: 451.07480; found: 451.07525
General: Dichloromethane, methanol, N,N-diisopropylethylamine
(DIPEA), and triethylamine (TEA) were dried over calcium hydride;
tetrahydrofuran (THF) was distilled over sodium, and N,N-dimethylace-
tamide (DMA) was dried over activated molecular sieves. Reactions re-
1,2-Cyclohexanedicarboxylic acid, (1S,2S,4S,5S)-5-(2-azidoethoxy)-4-hy-
droxy-1,2-bis-p-nitrophenyl ester (7): Epoxide
1
quiring anhydrous conditions were performed under nitrogen. H and 13C
6 (50 mg, 0.12 mmol,
spectra were recorded at 400 MHz on a Bruker AVANCE-400 instru-
ment. Chemical shifts (d) for 1H and 13C spectra are expressed in ppm
relative to the internal standard (CDCl3: d=7.24 ppm for 1H and d=
77.23 ppm for 13C; CD3OD: d=3.31 ppm for 1H and d=49.15 ppm for
13C). Signals were abbreviated as s, singlet; brs, broad singlet; d, doublet;
t, triplet; q, quartet; m, multiplet. Sugar signals were numbered as cus-
tomary; cyclohexane protons are indicated with the letter C followed by
numbers. The unusual numbering of the pseudodisaccharide derivatives
was adopted to facilitate the comparison with the native disaccharide.
Mass spectra were obtained with a ThermoFisherLCQapparatus (ESI
ionization), or iontrap ESI Esquire 6000 from Bruker, or a Microflex ap-
paratus (MALDI ionization) from Bruker, or Apex II ICR FTMS (ESI
ionization–HRMS). Specific optical rotation values were measured at
589 nm by using a Perkin–Elmer 241 with a 1 dm cell. TLC analysis was
carried out with pre-coated Merck F254 silica gel plates. Flash chroma-
tography (FC) was carried out with Macherey–Nagel silica gel 60 (230–
400 mesh). Amines 10a–e and 10p are commercially available. The syn-
thesis of amine 10 f–o is reported in the Supporting Information file. The
synthesis of 4h, 11, 13, and 15 is described below. The synthesis of all
other amides (4a–g, 4i–o, and 12) are reported in the Supporting Infor-
mation.
1 equiv) and Cu(OTf)2 (4 mg, 0.01 mmol, 0.1 equiv) were added to a solu-
AHCTUNGTRENNUNG
tion of 2-azidoethanol[11] (ca. 600 mg, 7 mmol, 58 equiv) in dichlorome-
thane (3 mL) and the mixture was stirred under nitrogen at room temper-
ature. After completion (16 h), the solvents were removed under reduced
pressure and the crude was purified by flash chromatography (hexane
with gradient of ethyl acetate from 20 to 50%) to yield 70% of pure
product 7 as a colorless wax.
4794
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 4786 – 4797

Targeting of DC-SIGN with Mannose-Based Glycomimetics
FULL PAPER
[a]2D0 = +33.8 (c=1.1 in CHCl3); 1H NMR (400 MHz, CDCl3): d=8.27–
8.17 (m, 4H; H12), 7.31–7.23 (m, 4H; H11), 4.19–4.14 (m, 1H; H2), 3.88–
3.82 (m, 1H; H7a), 3.73–3.63 (m, 2H; H7b, H1), 3.49–3.29 (m, 4H; H4, H5,
H
8a,b), 2.40–2.10 ppm (m, 4H; H3, H6); 13C NMR (100 MHz, CDCl3): d=
172.7, 172.6 (C9); 155.5 (C10); 145.8, 145.7 (C13); 125.5 (C12); 122.6, 122.6
(C11); 76.5 (C1); 68.7 (C7); 66.3 (C2); 51.1 (C8); 39.4, 39.1 (C4, C5); 30.5,
27.0 ppm (C3, C6); MS (ESI): m/z calcd for [C22H21N3O10Na]+: 512.4
[MꢀN2+Na]+ ; found: 512.0.
1,2-Cyclohexanedicarboxylic acid, (1S,2S,4S,5S)-4-(2-azidoethoxy)-5-
[(2,3,4,6-tetra-O-benzoyl-a-d-mannopyranosyl)oxy]-1,2-bis(p-nitrophenyl
ester) (9): A mixture of the acceptor 7 (37 mg, 0.071 mmol, 1 equiv) and
the donor 8[12] (65 mg, 0.086 mmol, 1.2 equiv) was co-evaporated from
toluene three times. Powdered and activated 4 ꢄ molecular sieves (acid
washed) were added; and the mixture was kept under vacuum for a few
hours and then dissolved with dry CH2Cl2 (1 mL). After cooling at
ꢀ208C, TMSOTf (3 mL, 0.014 mmol, 0.2 equiv) was added to the mixture
whilst stirring. The reaction was stirred at ꢀ208C for 1 h and was then
quenched with Et3N. The mixture was warmed to room temperature and
filtered over a Celite pad. The filtrate was evaporated at reduced pres-
sure and the crude product purified by flash chromatography (toluene
with a gradient of ethyl acetate from 0 to 10%) to yield 85% of pure
product 9 as a white foam.
Yield=82%; [a]2D0 = +12.1 (c=0.81 in CH3OH); 1H NMR (400 MHz,
CD3OD): d=7.29 (d, 4H; H12, J12ꢀ13 =8 Hz), 7.23 (d, 4H; H13, J13ꢀ12
8 Hz), 4.96 (d, 1H; H1, J1ꢀ2 =1.6 Hz), 4.58 (s, 4H; H15a,b), 4.31 (s, 4H;
10a,b), 4.08–4.03 (m, 1H; C2), 3.93–3.89 (m, 1H; H2), 3.89–3.84 (m, 1H;
=
H
H6a), 3.84–3.65 (m, 5H; C1, H6b, H7a,b, H3,), 3.64–3.54 (m, 2H; H4, H5),
3.47–3.35 (m, 2H; H8a,b), 3.02–2.85 (m, 2H; C4,C5), 2.06–1.86 ppm (m,
4H; C3, C6); 13C NMR (100 MHz, CD3OD): d=177.2, 177.0 (C9); 141.7
(C14); 139.2 (C11); 128.6, 128.7 (C12); 128.3 (C13); 100.4 (C1); 76.6 (C3);
75.7 (C5); 72.7 (CC1); 72.5 (C2); 72.4 ( CC2); 69.3 (C7); 68.9 (C4); 65.1
(C15); 63.2 (C6); 52.1 (C8); 43.8 (C10); 42.1, 41.9 (CC4, CC5); 29.9, 29.0 ppm
(CC3, CC6); HRMS: m/z calcd for: [C32H43N5O11Na]+: 696.28568; found:
696.28423.
Synthesis of macrocycle 11:
A solution of p-xylylbenzylamine 10p
(3.8 mg, 0.028 mmol, 0.5 equiv) in MeCN (0.55 mL) was added dropwise
under nitrogen to a flask charged with scaffold 9 (60 mg, 0.055 mmol,
1 equiv). The reaction was stirred for 5 h. TLC (dichloromethane/MeOH
9:1, dichloromethane/MeOH 9:1+1% TEA, hexane/AcOEt 6:4) indicat-
ed the presence of scaffold 9, whereas amine 10p was not observed.
Therefore, another portion of amine 10p (0.5 equiv) was added and the
reaction mixture was stirred for an additional 16 h after which TLC anal-
ysis indicated almost no 9 and a complex mixture of products. The sol-
vent was removed under reduced pressure and the crude was purified by
flash chromatography (silica, hexane with gradient of EtOAc from 30 to
70%) to afford 18.6 mg of benzoyl-protected intermediate (TLC, hexane/
EtOAc 2:8, Rf =0.33); MS (ESI): m/z calcd for [C104H98N10O26Na]+:
1926.9; found: 1926.6.
[a]2D0 =ꢀ18.0 (c=0.5 in CHCl3); 1H NMR (400 MHz, CDCl3): d=8.30–
8.20 (m, 4H; H12), 8.09–8.00 (m, 4H; HBz), 7.99–7.94 (m, 2H; HBz,), 7.83–
7.79 (m, 3H; HBz), 7.63–7.48 (m, 3H; HBz), 7.46–7.20 (m, 12H; HBz, H11),
A solution of sodium methoxide in MeOH (1m, 100 mL, 0.037 mmol,
4 equiv) was added to the solution of the product obtained in the previ-
ous reaction (18.6 mg, 0.0097 mmol, 0.18 equiv) in dry MeOH (0.7 mL).
After 45 min, the reaction mixture was diluted with methanol and neu-
tralized with prewashed Amberlite IRA 120-H+. The resin was filtered
off and the filtrate was concentrated under reduced pressure. The crude
product was purified by flash chromatography (CHCl3 with a gradient of
MeOH from 0 to 20% with 10% water in methanol) to afford 8.7 mg of
product.
6.07 (t, 1H; H4, J4ꢀ3 =J4ꢀ5 =10.0 Hz), 5.90 (dd, 1H; H3, J3ꢀ4 =10.0, J3ꢀ2
3.3 Hz), 5.75 (dd, 1H; H2, J2ꢀ1 =1.7, J2ꢀ3 =3.3 Hz), 5.34 (d, 1H; H1, J1ꢀ2
=
=
1.7 Hz), 4.73 (dd, 1H; H6b, J6bꢀ5 =2.9, J6aꢀ6b =12.0 Hz), 4.53 (dd, 1H; H6a,
6aꢀ5 =5.3, J6aꢀ6b =12 Hz), 4.48–4.40 (m, 1H; H5), 4.23–4.18 (m, 1H; C2),
3.89–3.84 (m, 1H; C1), 3.74–3.67 (m, 1H; H7a), 3.60–3.52 (m, 1H; H7b),
3.44–3.34 (m, 2H; C4, C5), 3.35–3.23 (m, 2H; H8), 2.52–2.41 (m, 2H; C3eq
6eq), 2.25–2.09 ppm (m, 2H; C3ax, C6ax); 13C NMR (100 MHz, CDCl3):
J
,
C
d=172.3, 172.2 (C9); 166.3, 165.8, 165.7 (COBZ); 155.5, 155.4 (C10); 145.8
(C13); 133.9, 133.8, 133.6, 133.5 (CHBZ); 130.1, 130.1, 129.9, 129.9, 129.9
(CHBZ); 129.3, 129.2, 129.1, 129.0 (CquatBZ); 128.9, 128.7, 128.9, 128.4
(CHBZ); 125.5, 125.4 (C12); 122.7, 122.6 (C11); 96.8 (C1); 74.9 (CC1); 72.3
(CC2); 70.9 (C2); 70.1 (C5); 69.9 (C3); 68.7 (C7); 67.4 (C4); 63.5 (C6); 51.0
(C8); 39.4, 39.2 (CC4,CC5); 27.8 (CC3); 27.0 ppm (CC6); MS (ESI): m/z
calcd for [C56H47N5O19Na]+: 1117.0; found 1116.1; HRMS: m/z calcd for
[C56H47N5O19Na]+: 1116.27575; found: 1116.27734.
Synthesis of N1,N2-bis[4-(hydroxymethylene)benzyl]amide (4h): The
amine 10h (38 mg, 0.275 mmol, 3 equiv) was added to a 0.1m solution of
9 (100 mg, 0.091 mmol, 1 equiv) in dry acetonitrile (0.9 mL) whilst stir-
ring and under a nitrogen atmosphere at room temperature. After com-
pletion (3 h; TLC, hexane/EtOAc 1:1 and 2:8) the solvent was evaporat-
ed under reduced pressure. The crude product was dissolved in dry meth-
anol (c=0.1m), under nitrogen at room temperature and a 1m solution of
sodium methoxide in MeOH (2 equiv) was added. After reaction comple-
tion (1 h), the reaction mixture was diluted with MeOH and neutralized
with prewashed Amberlite IRA 120-H+. The resin was filtered off and
the filtrate was concentrated under reduced pressure. The crude was pu-
rified by flash chromatography (CHCl3 with gradient of methanol from 0
to 20%).
Yield=15%; 1H NMR (400 MHz, CD3OD): d=7.11 (s, 8H; H12), 4.96
(d, 2H; H1, J2ꢀ1 =1.7 Hz), 4.72–4.65 (m, 4H; H10a), 4.05–4.02 (m, 2H;
C2), 3.91 (dd, 2H; H2, J2ꢀ1 =1.7, J2ꢀ3 =3.2 Hz), 3.88–3.81 (m, 2H; H6b),
3.80–3.62 (m, 14H; H10b, H3, H7, C1, H6b), 3.61–3.52 (m, 2H: H4, H5),
3.45–3.33 (m, 2H; H8), 2.98–2.81 (m, 2H; C4, C5), 2.00–1.86 ppm (m, 4H;
C3, C6); 13C NMR (100 MHz, CD3OD): d=177.1, 177.0, 176.8, 176.7 (C9);
138.9, 138.9 (C11); 128.9 (C12); 100.3 (C1); 76.7 (CC1); 75.7 (C5); 72.7 (C3);
72.6 (C2); 72.5 (CC2); 69.3 (C7); 68.9 (C4); 63.2 (C6); 52.2 (C8); 43.8 (C10);
Chem. Eur. J. 2013, 19, 4786 – 4797
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4795

A. Bernardi et al.
41.9, 41.8 (CC4, CC5); 30.0, 29.1 ppm (CC3, CC6); HRMS: m/z calcd for
5.0 Hz), 4.56 (s, 2H; H15), 4.56 (s, 2H; H15), 4.29 (s, 2H; H10), 4.27 (s, 2H;
H10), 3.99–3.92 (m, 1H; H7a), 3.90–3.84 (m, 2H; H7b, C2), 3.84–3.80 (m,
1H; H6a), 3.78–3.71 (m, 2H; H2, H4), 3.71–3.59 (m, 2H; H5, H6b), 3.59–
3.49 (m, 2H; C1, H3), 2.94–2.84 (m, 1H; C4), 2.70–2.59 (m, 1H; C5), 1.98–
1.71 ppm (m, 4H; C3, C6); 13C NMR (100 MHz, CD3OD): d=177.0 (C9);
141.7, 141.6 (C14); 139.2, 139.2 (C11); 128.6 128.5 (C12); 128.3, 128.3 (C13);
125.7 (C16); 101.8 (C1); 75.8 (C5); 75.2 (D1); 73.9 (CC2); 72.6, 72.5 (C2,
C4); 69.0 (C3); 68.3 (C7); 65.1 (C15); 63.3 (C6); 56.7 (C18); 51.9 (C8); 43.8,
43.7 (C10); 42.3, 41.7 (CC4, CC5); 31.0, 29.2 ppm (CC3, CC6); HRMS: m/z
calcd for [C35H47N5O12Na]+: 752.31189; found: 752.31123
[C48H66N10O18Na]+: 1093.44543; found: 1093.44341.
Synthesis of 13 (triazol derivative of 4h): Bisamide 4h (20 mg,
0.030 mmol, 1 equiv), propargyl alcohol (5 mg, 0.090 mmol, 3 equiv), cop-
per(II) sulfate pentahydrate (0.7 mg, 0.003 mmol, 0.1 equiv), sodium as-
corbate (2.3 mg, 0.012 mmol, 0.4 equiv), and tris[(1-benzyl-1H-1,2,3-tria-
zol-4-yl)methyl]amine (TBTA, 3.1 mg, 0.006 mmol, 0.2 equiv) were dis-
solved in THF/H2O (1 mL, 1:1). After reaction completion (4 h; TLC,
CHCl3/MeOH 8:2) the solvent was removed under reduced pressure and
the resulting crude was purified by flash chromatography (silica, chloro-
form with gradient of MeOH from 10 to 30%) to afford 14.8 mg of pure
product.
Acknowledgements
The project was supported under the EU ITN Marie-Curie program
(CARMUSYS Grant number 213592).
[1] Y. van Kooyk, T. B. H. Geijtenbeek, Nat. Rev. Immunol. 2003, 3,
697–709.
[2] M. Anderluh, G. Jug, U. Svajger, N. Obermajer, Curr Med Chem.
2012, 19, 992–1007.
[3] T. van Montfort, D. Eggink, M. Boot, M. Tuen, C. E. Hioe, B.
Berkhout, R. W. Sanders, J. Immunol. 2011, 187, 4676–4685.
[4] a) M. J. Borrok, L. L. Kiessling, J. Am. Chem. Soc. 2007, 129,
12780–12785; b) K. C. A. Garber, K. Wangkanont, E. E. Carlson,
L. L. Kiessling, Chem. Commun. 2010, 46, 6747–6749; c) S. L. Man-
gold, L. R. Prost, L. L. Kiessling, Chem. Sci. 2012, 3, 772–777;
d) T. H. Tran, R. El Baz, A. Cuconati, J. Arthos, P. Jain, Z. K. Khan,
J. Antivir. Antiretrovir. 2011, 3, 49–54.
[5] a) F. Lasala, E. Arce, J. R. Otero, J. Rojo, R. Delgado, Antimicrob.
Agents Chemother. 2003, 47, 3970–3972; b) J. Rojo, R. Delgado, J.
Antimicrob. Chemother. 2004, 54, 579–581; c) J. J. Reina, O. S. Mal-
donado, G. Tabarani, F. Fieschi, J. Rojo, Bioconjugate Chem. 2007,
18, 963–969.
Yield=70%; [a]2D0 = +11.7 (c=0.30 in CH3OH); 1H NMR (400 MHz,
CD3OD): d=8.00 (s, 1H; H16), 7.31–7.17 (m, 8H; H12, H13), 4.88 (brs,
1H; H1), 4.65 (s, 2H; H18), 4.59 (t, 2H; H8, J7ꢀ8 =5.0 Hz), 4.56 (s, 4H;
H15), 4.32–4.23 (m, 4H; H10a,b), 3.98–3.86 (m, 3H; H7, C2), 3.88–3.81 (m,
2H; H2, H6a), 3.72–3.61 (m, 3H; C1, H6b, H3,), 3.58–3.48 (m, 2H; H4, H5),
2.88–2.77 (m, 1H; C4), 2.72–2.60 (m, 1H; C5), 1.94–1.64 ppm (m, 4H; C3,
C6); 13C NMR (100 MHz, CD3OD): d=177.0, 176.9 (C9); 141.7 (C14);
139.2 (C11); 128.5 (C12); 128.3 (C13); 125.5 (C16); 100.4 (C1); 76.1 (C3);
75.7 (C5); 72.7 (CC1); 72.5 (C2); 72.2 (D2); 69.0 (C4); 68.3 (C7); 65.1 (C15);
63.2 (C6); 56.7 (C18); 51.8 (C8); 43.8 (C10); 41.9, 41.8 (CC4, CC5); 29.6,
28.9 ppm (CC3, CC6); HRMS: m/z calcd for [C35H47N5O12Na]+: 752.31189;
found: 752.31099.
[6] a) J. J. Reina, S. Sattin, D. Invernizzi, S. Mari, L. Martinez-Prats, G.
Tabarani, F. Fieschi, R. Delgado, P. M. Nieto, J. Rojo, A. Bernardi,
Chemmedchem 2007, 2, 1030–1036; b) G. Timpano, G. Tabarani, M.
Anderluh, D. Invernizzi, F. Vasile, D. Potenza, P. M. Nieto, J. Rojo,
F. Fieschi, A. Bernardi, Chembiochem 2008, 9, 1921–1930; c) S.
Sattin, A. Daghetti, M. Thepaut, A. Berzi, M. Sanchez-Navarro, G.
Tabarani, J. Rojo, F. Fieschi, M. Clerici, A. Bernardi, ACS Chem.
Biol. 2010, 5, 301–312; d) M. Andreini, D. Doknic, I. Sutkeviciute,
J. J. Reina, J. Duan, E. Chabrol, M. Thepaut, E. Moroni, F. Doro, L.
Belvisi, J. Weiser, J. Rojo, F. Fieschi, A. Bernardi, Org. Biomol.
Chem. 2011, 9, 5778–5786; e) J. Luczkowiak, S. Sattin, I. Sutkevi-
ciute, J. J. Reina, M. Sanchez-Navarro, M. Thepaut, L. Martinez-
Prats, A. Daghetti, F. Fieschi, R. Delgado, A. Bernardi, J. Rojo, Bio-
conjugate Chem. 2011, 22, 1354–1365; f) A. Berzi, J. J. Reina, R.
Ottria, I. Sutkeviciute, P. Antonazzo, M. Sanchez-Navarro, E. Chab-
rol, M. Biasin, D. Trabattoni, I. Cetin, J. Rojo, F. Fieschi, A. Bernar-
di, M. Clerici, Aids 2012, 26, 127–137; g) N. Obermajer, S. Sattin, C.
Colombo, M. Bruno, U. Svajger, M. Anderluh, A. Bernardi, Mole-
cules Molecular Diversity 2011, 15, 347–360.
Synthesis of 15 (triazol derivative of 12): Bisamide 12 (see Scheme S8 in
the Supporting Information, 11 mg, 0.016 mmol, 1 equiv), propargyl alco-
hol (4.6 mg, 0.081 mmol, 5 equiv), copper(II) sulfate pentahydrate
(0.4 mg, 0.001 mmol, 0.1 equiv), sodium ascorbate (1.3 mg, 0.006 mmol,
0.4 equiv), and TBTA (1.72 mg, 0.003 mmol, 0.2 equiv) were dissolved in
THF/H2O (1 mL, 1:1). After reaction completion (12 h; TLC, CHCl3/
MeOH 8:2), the solvent was removed under reduced pressure and the re-
sulting crude was purified by flash chromatography (silica, chloroform
with gradient of methanol from 10 to 30%) to afford 4.2 mg of pure
product.
[7] L. de Witte, A. Nabatov, M. Pion, D. Fluitsma, M. A. de Jong, T. de
Gruijl, V. Piguet, Y. van Kooyk, T. B. Geijtenbeek, Nat. Med. 2007,
13, 367–371.
[8] C. G. M. Thꢁpaut, I. Sutkeviciute, S. Sattin, R. Ribeiro-Viana, N.
Varga, J. R. Eric Chabrol, J. Angulo, A. Bernardi, P. M. Nieto, F. Fi-
eschi, J. Am. Chem. Soc. 2013, DOI: 10.1021/ja3053305.
[9] A. Bernardi, D. Arosio, D. Dellavecchia, F. Micheli, Tetrahedron:
Asymmetry 1999, 10, 3403–3407.
[10] A. Bernardi, D. Arosio, L. Manzoni, F. Micheli, A. Pasquarello, P.
Seneci, J. Org. Chem. 2001, 66, 6209–6216.
[11] Y. H. Dong, X. M. Liang, H. Z. Yuan, S. H. Qi, F. H. Chen, D. Q.
Wang, Green Chem. 2008, 10, 990–994.
Yield=35%; [a]2D0 = +30.1 (c=0.20 in CH3OH); 1H NMR (400 MHz,
CD3OD): d=8.00 (s, 1H; H16), 7.31–7.14 (m, 8H; H12, H13), 4.84 (brs,
1H; H1), 4.88 (brs, 1H; H1), 4.65 (s, 2H; H18), 4.59 (t, 2H; H8, J7ꢀ8
=
4796
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 4786 – 4797

Targeting of DC-SIGN with Mannose-Based Glycomimetics
FULL PAPER
[12] D. J. Lee, R. Kowalczyk, V. J. Muir, P. M. Rendle, M. A. Brimble,
Carbohydr. Res. 2007, 342, 2628–2634.
D. E. Shaw, P. Francis, P. S. Shenkin, J. Med. Chem. 2004, 47, 1739–
1749.
[25] M. Thꢁpaut, J. Valladeau, A. Nurisso, R. Kahn, B. Arnou, C. Vives,
S. Saeland, C. Ebel, C. Monnier, C. Dezutter-Dambuyant, A. Imber-
ty, F. Fieschi, Biochemistry 2009, 48, 2684–2698.
[26] M. Mayer, B. Meyer, J. Am. Chem. Soc. 2001, 123, 6108–6117.
[27] J. Jeener, B. H. Meier, P. Bachmann, R. R. Ernst, J. Chem. Phys.
1979, 71, 4546–4553.
[28] R. Wagner, S. Berger, J. Magn. Reson., Ser. A 1996, 123, 119–121.
[29] Schrçdinger Suite 2011 Protein Preparation Wizard: Epik ver-
sion 2.2, Schrçdinger, LLC, New York, NY, 2011; Impact version 5.7,
Schrçdinger, LLC, New York, NY, 2011; Prime version 3.0, Schrç-
dinger, New York, NY, 2011.
[30] J. L. Banks, H. S. Beard, Y. Cao, A. E. Cho, W. Damm, R. Farid,
A. K. Felts, T. A. Halgren, D. T. Mainz, J. R. Maple, R. Murphy,
D. M. Philipp, M. P. Repasky, L. Y. Zhang, B. J. Berne, R. A. Fries-
ner, E. Gallicchio, R. M. Levy, J. Comput. Chem. 2005, 26, 1752–
1780.
[31] I. Kolossvꢇry, W. C. Guida, J. Am. Chem. Soc. 1996, 118, 5011–5019.
[32] a) W. C. Still, A. Tempczyk, R. C. Hawley, T. Hendrickson, J. Am.
Chem. Soc. 1990, 112, 6127–6129; b) J. Weiser, P. S. Shenkin, W. C.
Still, J. Comput. Chem. 1999, 20, 217–230.
[33] R. A. Friesner, R. B. Murphy, M. P. Repasky, L. L. Frye, J. R. Green-
wood, T. A. Halgren, P. C. Sanschagrin, D. T. Mainz, J. Med. Chem.
2006, 49, 6177–6196.
[13] S. Mari, H. Posteri, G. Marcou, D. Potenza, F. Micheli, F. Canada, J.
Jimenez-Barbero, A. Bernardi, Eur. J. Org. Chem. 2004, 5119–5125.
[14] G. Tabarani, M. Thepaut, D. Stroebel, C. Ebel, C. Vives, P. Vachette,
D. Durand, F. Fieschi, J. Biol. Chem. 2009, 284, 21229–21240.
[15] R. J. Pieters, Org. Biomol. Chem. 2009, 7, 2013–2025.
[16] K. H. Schlick, C. K. Lange, G. D. Gillispie, M. J. Cloninger, J. Am.
Chem. Soc. 2009, 131, 16608–16609, and references therein; C. Sisu,
A. J. Baron, H. M. Baron, Branderhorst, S. D. Connell, C. A. Weij-
ers, R. de Vries, E. D. Hayes, A. V. Pukin, M. Gilbert, R. J. Pieters,
H. Zuilhof, G. M. Visser, W. B. Turnbull, Chembiochem 2009, 10,
329–337, and references therein.
[17] C. Galustian, C. G. Park, W. Chai, M. Kiso, S. A. Bruening, Y. S.
Kang, R. M. Steinman, T. Feizi, International Immunology 2004, 16,
853–866.
[18] H. Feinberg, M. E. Taylor, N. Razi, R. McBride, Y. A. Knirel, S. A.
Graham, K. Drickamer, W. I. Weis, J. Mol. Biol. 2011, 405, 1027–
1039.
[19] M. Mayer, B. Meyer, Angew. Chem. Int. Ed. 1999, 38, 1784–1788.
[20] M. Mayer, T. L. James, J. Am. Chem. Soc. 2004, 126, 4453–4460.
[21] J. Angulo, P. M. Enriquez-Navas, P. M. Nieto, Chem. Eur. J. 2010,
16, 7803–7812.
[22] Y. Guo, H. Feinberg, E. Conroy, D. A. Mitchell, R. Alvarez, O.
Blixt, M. E. Taylor, W. I. Weis, K. Drickamer, Nat. Struct. Mol. Biol.
2004, 11, 591–598.
[23] A. E. Cho, V. Guallar, B. Berne, R. A. Friesner, J. Comput. Chem.
2005, 26, 915–931.
[24] R. A. Friesner, J. L. Banks, R. B. Murphy, T. A. Halgren, J. J. Klicic,
D. T. Mainz, M. P. Repasky, E. H. Knoll, M. Shelley, J. K. Perry,
[34] Jaguar, version 7.8, Schrçdinger, LLC, New York, NY, 2011.
[35] Prime, version 3.0, Schrçdinger, LLC, New York, NY, 2011.
Received: August 1, 2012
Revised: December 17, 2012
Published online: February 18, 2013
Chem. Eur. J. 2013, 19, 4786 – 4797
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4797