Discovery of two classes of potent glycomimetic inhibitors of pseudomonas aeruginosa LecB with distinct binding modes

Article abstract of DOI:10.1021/cb400371r

The treatment of infections due to the opportunistic pathogen Pseudomonas aeruginosa is often difficult, as a consequence of bacterial biofilm formation. Such a protective environment shields the bacterium from host defense and antibiotic treatment and secures its survival. One crucial factor for maintenance of the biofilm architecture is the carbohydrate-binding lectin LecB. Here, we report the identification of potent mannose-based LecB inhibitors from a screening of four series of mannosides in a novel competitive binding assay for LecB. Cinnamide and sulfonamide derivatives are inhibitors of bacterial adhesion with up to a 20-fold increase in affinity to LecB compared to the natural ligand methyl mannoside. Because many lectins of the host require terminal saccharides (e.g., fucosides), such capped structures as reported here may offer a beneficial selectivity profile for the pathogenic lectin. Both classes of compounds show distinct binding modes at the protein, offering the advantage of a simultaneous development of two new lead structures as anti-pseudomonadal drugs with an anti-virulence mode of action.

Full text of DOI:10.1021/cb400371r

Articles
pubs.acs.org/acschemicalbiology
Discovery of Two Classes of Potent Glycomimetic Inhibitors of
Pseudomonas aeruginosa LecB with Distinct Binding Modes
Dirk Hauck,†,‡,# Ines Joachim,†,# Benjamin Frommeyer,†,# Annabelle Varrot,§ Bodo Philipp,∥
Heiko M. Moller,† Anne Imberty,§ Thomas E. Exner,†,‡,⊥ and Alexander Titz*,†,‡
̈
‡
†Department of Chemistry and Zukunftskolleg, University of Konstanz, D-78457 Konstanz, Germany
§Centre de Recherche sur les Macromolec
Molec
∥Institute of Molecular Microbiology and Biotechnology, University of Munster, D-48149 Munster, Germany
́
ules Veg
́ ́ ́
etales (CERMAV)-CNRS, Universite Grenoble 1 and Institut de Chimie
́
ulaire de Grenoble BP53, F-38041 Grenoble cedex 9, France
̈
̈
S
* Supporting Information
ABSTRACT: The treatment of infections due to the oppor-
tunistic pathogen Pseudomonas aeruginosa is often difficult, as a
consequence of bacterial biofilm formation. Such a protective
environment shields the bacterium from host defense and
antibiotic treatment and secures its survival. One crucial factor
for maintenance of the biofilm architecture is the carbohydrate-
binding lectin LecB. Here, we report the identification of potent
mannose-based LecB inhibitors from a screening of four series
of mannosides in a novel competitive binding assay for LecB.
Cinnamide and sulfonamide derivatives are inhibitors of bac-
terial adhesion with up to a 20-fold increase in affinity to LecB
compared to the natural ligand methyl mannoside. Because
many lectins of the host require terminal saccharides (e.g.,
fucosides), such capped structures as reported here may offer a beneficial selectivity profile for the pathogenic lectin. Both classes
of compounds show distinct binding modes at the protein, offering the advantage of a simultaneous development of two new
lead structures as anti-pseudomonadal drugs with an anti-virulence mode of action.
Pseudomonas aeruginosa is a ubiquitous Gram-negative bacteri-
um accounting for a large number of nosocomial infections.1,2
Additionally, it frequently colonizes airways of cystic fibrosis
patients, ultimately leading to lung failure.3 As a result of the
bacterium’s high resistance toward antibiotics and its ability to
form biofilms, infections can turn chronic with a high mortality
rate among patients.4,5 In a biofilm, bacteria are embedded in
the extracellular matrix consisting of extracellular DNA, poly-
saccharides, and proteins.6 P. aeruginosa produces two soluble
lectins under quorum-sensing control: LecA (or PA-IL) and
LecB (or PA-IIL).7,8 Deletion of either lecA or lecB genes in P.
aeruginosa resulted in strains with weaker lung colonization and
systemic spread in a murine model of lung infections.9 Since
the two lectins are both virulence factors and necessary for
biofilm formation, their inhibition is considered a promising
approach for anti-pseudomonadal treatment.10−17 These tetra-
meric carbohydrate-binding proteins recognize specific mono-
saccharide residues in a calcium-dependent manner: LecA is
specific for D-galactose, whereas LecB binds L-fucosides and
D-mannosides.7 Williams and Reymond showed that treatment
of P. aeruginosa with multivalent carbohydrate-based inhibitors of
these lectins resulted in a reduced biofilm formation in vitro.13,18,19
Furthermore, inhalation of an aqueous galactose- and fucose-
containing aerosol resulted in a reduction of respiratory tract
infections with P. aeruginosa.20,21
LecB has an unusual high affinity for fucose residues, which
has been explained by the crystal structure of the complex: two
calcium ions in the binding site mediate the binding of one
saccharide ligand to the protein.22,23 This particularly high
affinity has prompted various groups to synthesize fucose-based
inhibitors of this lectin (reviewed in Imberty et al.12,15). The
Lewis blood group antigen Lewisa, a trisaccharide containing an
α-fucoside, is the best known monovalent ligand of LecB (Kd =
210 nM).24 The crystal structure of its complex revealed, in
addition to the recognition of fucose, an additional hydrogen bond
of the attached GlcNAc-6-OH with the receptor but no direct
interaction of the third saccharide moiety. Consequently, synthetic
inhibitors were simplified and based on the Fuc-α-1,4-GlcNAc
disaccharide or on fucose alone. This did, however, not lead to an
increase in affinity compared to the parent saccharide Lewisa, and
its saccharide character was maintained.15,25 In another example,
fucosylamides displayed a 3-fold lower affinity compared with that
of methyl α-L-fucoside, which has a Kd of 430 nM.26,27
These examples and all other synthetic fucose-derived inhib-
itors of LecB described to date bear terminal α-fucosides,12,15
Received: February 4, 2013
Accepted: May 29, 2013
Published: May 29, 2013
© 2013 American Chemical Society
1775
dx.doi.org/10.1021/cb400371r | ACS Chem. Biol. 2013, 8, 1775−1784

ACS Chemical Biology
Articles
Figure 2. (Top) Titration of fluorescein-labeled fucose-based reporter
ligand 6 with LecB and determination of the fluorescence polarization
reveals a strong and specific binding (EC50 = 697 nM, black triangles)
that could be completely inhibited by the addition of excess fucose
(1 mM, 16 μM LecB, black circle). Ligand 6 in the absence of LecB
showed fluorescence polarization values at background level (black
square). (bottom). Competitive inhibition of the binding of 6 to LecB
with inhibitors allowed the determination of IC50 values for fucose
(IC50 = 2.74 1.44 μM, black circles) and 1 (IC50 = 157.8 52.8 μM,
black squares); one representative titration of triplicates is shown. IC50
values for fucose and 1 correspond to 21 and 17 independent experi-
ments, respectively, and error bars show standard deviations.
Figure 1. (Top) The crystal structure of LecB in complex with
D-mannose reveals the hydrogen bond between O6 of the ligand and the
Ser23 side chain (yellow line). The loop from Val69 to Ser74 confines
a cleft adjacent to C6 of mannose (PDB code 1OUR22). (Bottom)
Methyl α-D-mannoside (1, Kd = 71 μM27) served as a lead structure
for the search for LecB inhibitors. Compound 1 was modified at its
primary hydroxyl group to target the adjacent cleft using different
linker chemistries. Libraries of triazoles 2, amines 3, amides 4, and
sulfonamides 5 were synthesized and analyzed for binding to LecB.
LecB also binds mannosides through coordination of the two
calcium ions to their secondary hydroxyl groups with the same
relative orientation as in fucose. Interestingly, the additional
6-OH of mannose forms a hydrogen bond to Ser23.22 Mannose,
however, lacks the equatorial methyl group of L-fucose, which
served as an explanation for the increased affinity of fucose with
respect to mannose derivatives (e.g., 1, Figure 1; Kd = 71 μM)
toward LecB.22,27 Adjacent to the primary 6-hydroxy group of
1,22 a cleft on the protein surface that is confined by the loop of
Val69 to Ser74 may be targeted by additional pharmacophores to
address affinity, specificity, and drug-like properties (Figure 1).
Here, we report on novel terminally modified mannosides as
ligands of LecB and inhibitors of the adhesion of P. aeruginosa
cells with a 20-fold increase in affinity compared to that of 1.
These small molecules may lead to more selective ligands for
LecB over other lectins of the host through the terminal cap-
ping of mannose.
which are also ubiquitous glycoconjugate epitopes in the host.
Such inhibitors therefore possess the potential to interfere with
other fucose-binding proteins of the host (e.g., DC-SIGN, MBL,
the selectins, etc.), which have among others important roles
in host defense and inflammatory response. The natural
LecB ligand Lewisa, for example, is also a potent ligand to
DC-SIGN28,29 and is therefore unsuitable as a drug. In general,
an inhibition with terminally unmodified saccharides may result
in unwanted off-target side effects. In nature, carbohydrate-
binding proteins generally recognize multivalently displayed
epitopes and thereby circumvent the promiscuity of individual
monosaccharides for a plethora of target receptors. Therefore,
the design of small molecule inhibitors of carbohydrate-binding
proteins with the potential for becoming a drug is a particular
challenge.14 Consequently, besides the poor pharmacokinetic
properties of structures such as Lewisa (Supplementary Table S3),
modifications of the natural saccharide ligands with the aim to
improve affinity and specificity toward the target of interest are
of crucial importance.
RESULTS AND DISCUSSION
■
To target the additional cleft on the protein surface, 1 was mod-
ified with a nitrogen at position 6, and a set of substituted triazoles 2,
amines 3, amides 4, and sulfonamides 5 were designed (Figure 1).
The different linking functional groups were chosen on the
1776
dx.doi.org/10.1021/cb400371r | ACS Chem. Biol. 2013, 8, 1775−1784

ACS Chemical Biology
Articles
a
Scheme 1. Synthesis of Fucose-Based Fluorescein-Labeled Reporter Ligand 6
a
Reagents and conditions: (i) La(OTf)3, 2-bromoethanol, 7.5 h, 70°C; (ii) Ac2O, pyridine, 12 h, rt, 46% (2 steps, α/β 4:1); (iii) NaN3, DMF, 3.25 h,
70°C, 70%; (iv) NaOMe, MeOH, 12 h, rt; (v) H2, Pd/C, EtOH, 20 h, rt, 66% (2 steps); (vi) FITC, NaHCO3, DMF, 6.5 h, rt, 96%.
a
Scheme 2. Synthesis of Mannose-Derived Triazoles 2, Amines 3, Amides 4, and Sulfonamides 5
a
Reagents and conditions: (i) TsCl, pyridine, rt, 24 h; (ii) NaN3, DMF, 70°C, 4 h; (iii) alkyne, CuSO4, Na ascorbate, rt, 2−48 h; (iv) H2,
Pd/C, EtOH, rt, 12 h; (v) acyl chlorides or carboxylic acids/EDC·HCl, TEA, DMF; (vi) RCHO, NaBH4, 12 h; (vii) sulfonyl chlorides,
TEA, DMF.
basis of facile access through selective reactions and to study
their influence on possible hydrogen bonding with Ser23 or salt
bridge formation with Asp96.
For a rapid biochemical evaluation of the LecB antagonists,
we established a competitive binding assay based on fluores-
cence polarization. Such assays were shown useful for inhibitor
screening of other lectins, e.g., galectins30 and FimH.31 A fluorescein-
labeled, fucose-based reporter ligand 6 was synthesized (Scheme 1)
starting from a lanthanum triflate catalyzed fucosylation of
2-bromoethanol following the procedure of Dasgupta et al.32 in
an α/β anomeric ratio of 4:1. Acetylation of 2-bromoethyl
fucoside, nucleophilic replacement of the bromide by sodium
1777
dx.doi.org/10.1021/cb400371r | ACS Chem. Biol. 2013, 8, 1775−1784

ACS Chemical Biology
Articles
a
Table 1. Mannose-Derived Libraries of Triazoles 2, Amines 3, Amides 4, and Sulfonamides 5 as LecB Inhibitors
triazoles
amines
R
IC50 [μM]
578.1
264.1
372.6
279.4
293.8
ns
SD [μM]
76.5
R
IC50 [μM]
226.8
SD [μM]
85.0
2a
2b
2c
2d
2e
2f
CH2OH
CH2CH2Ph
Ph
3a
3b
3c
3d
3e
CH2Ph
113.6
90.1
CH2(4-NO2)Ph
CH2(4-CN)Ph
CH2(4-Cl)Ph
H
116.9
65.2
114.9
19.4
3-MePh
4-MePh
4-tBuPh
4-FPh
120.2
211.6
88.5
353.5
29.4
2g
240.0
149.9
amides
sulfonamides
R
Me
IC50 [μM]
SD [μM]
R
IC50 [μM]
15.9
49.7
20.9
36.1
3.4
SD [μM]
3.3
4a
4b
4c
4d
4e
4f
281.3
110.2
75.8
123.8
28.1
15.8
39.2
13.2
9.8
5a
5b
5c
5d
5e
5f
Ph
Ph
4-MePh
36.9
4.6
4-NO2Ph
3,5-diNO2Ph
CH2Ph
4-BrPh
121.0
73.8
4-NO2Ph
2,4,6-Me3-Ph
2,4,6-iPr3-Ph
(E)-CHCHPh
16.0
0.4
CHPh2
161.2
37.4
ns
4g
(E)-CHCHPh
5.0
5g
86.9
11.1
a
IC50 values were determined with a competitive fluorescence polarization assay. A minimum of 3 independent experiments were performed, and
standard deviations are given. ns = not soluble at 10 mM in aqueous buffer with 10% DMSO. SD = standard deviation.
azide, and deacetylation under Zemplen
́
conditions gave 9 in
6-modified mannosides showed binding to LecB. However, a
striking difference in potency was detected between the different
linker groups. 1,2,3-Triazoles as in 2 are motifs resulting from
the so-called ‘click chemistry’ cycloaddition, which proved to
be a versatile method of diversification.36 The use of triazoles as
linkers in 2 led, however, to a strong reduction in binding
affinity compared to 1 (IC50 = 157.8 μM) with IC50 values
ranging from 240 to 578 μM (Table 1). Amines 3 were
designed to form a putative salt bridge with Asp96 of LecB:
Adam et al.10 analyzed a mutant form of LecB, where Ser22 was
changed to Ala22 (S22A mutant), in complex with 1 by X-ray
crystallography. In this mutant, the primary hydroxyl group of
1 no longer forms a hydrogen bond with Ser23 but establishes
a hydrogen bond with Asp96. However, the existence of such a
salt bridge between the protonated amines with Asp96 is un-
likely according to the high IC50 values for 3a−e (115−353 μM,
Table 1). We also studied the protonated form of 3e in com-
plex with LecB by molecular dynamics simulation (Supple-
mentary Figures S1, S2). Consistent with the experimental data,
the amino group in 3e does not establish an interaction with
Asp96 but forms a stable hydrogen bond to Ser23 as observed
for 1 (Supplementary Figure S1). Even an enforced interaction
of the protonated amino group in 3e with Asp96 in the start
conformation of the simulation resulted in a conformational
change to establish the hydrogen bond with Ser23 after a few
picoseconds (Supplementary Figure S2). This may explain the
weaker affinities observed for amines 3a−e compared to 1.
Surprisingly, amides 4 were generally more potent inhibitors
of LecB with IC50 values from 37 to 161 μM (Table 1). Within
the amide series 4, only acetamide 4a (IC50 = 281 μM) was a
weaker ligand than 1. When lipophilic substituents were added
(4b−g), an increase in binding affinity to LecB was observed
and (E)-cinnamide 4g (IC50 = 37 μM) was identified as the
best ligand among this series. 4g was subsequently studied in
LecB-mediated hemagglutination of sheep erythrocytes18 as
well as agglutination of baker’s yeast.37 In both assays, 4g showed
good yield. After reduction, 2-aminoethyl fucoside (10) was
subsequently treated with fluorescein-5-isothiocyanate (FITC)
and base to yield the reporter ligand 6 in 96% yield. Upon
incubation of 6 with a dilution series of LecB, a dose-dependent
increase of fluorescence polarization was observed as a result of
the binding to LecB (Figure 2, top). The observed binding of
6 to LecB (EC50 = 697 nM) is in good agreement with the
known dissociation constant of methyl α-L-fucoside.27 The
carbohydrate specificity of the binding of 6 was demonstrated
by addition of excess fucose, resulting in complete displacement
of 6 from the protein. For the screening of LecB ligands, con-
stant concentrations of 6 and LecB were titrated with serial
dilutions of inhibitors. The assay was evaluated using fucose
and 1 as ligands with known Kd values (Figure 2, bottom): IC50
values of 2.74 1.44 μM (N = 21) for fucose and 157.8
52.8 μM (N = 17) for 1 are in good agreement with the
corresponding dissociation constant of fucose (Kd = 2.9 μM24)
and 1 (Kd = 71 μM27).
The common precursor for the designed LecB ligands,
methyl 6-azido 6-deoxy α-D-mannopyranoside (11), was ob-
tained following the published protocol of Ferguson and co-
workers,33 omitting the acetylation/deacetylation sequence
(Scheme 2). A library of substituted triazoles 2 was obtained
through copper-catalyzed Huisgen 1,3-dipolar cycloaddi-
tion34,35 with alkynes. For the sulfonamide (5), amide (4),
and amine (3) derived LecB ligands, the azido group in 11 was
reduced by hydrogenation. The resulting primary amine 3e was
then reacted with either sulfonyl chlorides, acyl chlorides, or
carbodiimide-activated carboxylic acids to give the correspond-
ing 6-sulfonamido 6-deoxy mannosides 5 or 6-amido 6-deoxy
mannosides 4, respectively. For the amine derived library 3,
aldehydes were reductively aminated using amine 3e and sodium
borohydride.
We then analyzed triazoles 2, amines 3, amides 4, and
sulfonamides 5 in the competitive binding assay (Table 1). All
1778
dx.doi.org/10.1021/cb400371r | ACS Chem. Biol. 2013, 8, 1775−1784

ACS Chemical Biology
Articles
a stronger activity in inhibiting agglutination than 1 (Supple-
mentary Figure S9). Microcalorimetric titrations of LecB with 4g
revealed a Kd of 18.5 μM (Table 2, Supplementary Figure S10),
replacements38 for amides with distinct differences to carbox-
amides in conformational and hydrogen bonding properties. A
set of aryl sulfonamides was synthesized and tested in the
competitive binding assay (5a−f, Table 1). All members of this
subset were equal or stronger ligands than cinnamide 4g, with 5e
being the most potent inhibitor (IC50 = 3.4 μM, Kd = 3.3 μM) of
all identified 6-modified mannosides corresponding to a 20-fold
increase in binding affinity compared to 1 (Kd = 71 μM) as
determined by isothermal microcalorimetry (Table 2).
To get insight into the binding modes of the potent members
of both amides and sulfonamides, we performed co-
crystallization of both ligands (4g, 5e) with LecB. Crystals
diffracting to 1.41 Å resolution were obtained for the complex
of 5e. The asymmetric unit contains one tetramer of LecB, with
very similar orientation of the ligand 5e in each binding site
(Figure 3). The mannoside residue of 5e coordinates to the
two protein-bound calcium ions, in an identical orientation as
mannose22 is recognized by LecB. In addition, the sulfonamide
nitrogen forms a hydrogen bond with Asp96 and the phenyl ring
has lipophilic contacts with the protein surface. This hydrogen bond
likely accounts for the enthalpy-driven binding of 5e (Table 2).
Table 2. Comparison of Binding Thermodynamics of 4g and
a
5e with LecB Compared to 127 by Isothermal Calorimetry
Ka [M−1
]
Kd [μM] ΔH [kcal/mol] −TΔS [kcal/mol]
N
4g
5e
1
53 950
302 000
14 200
18.5
3.3
−4.3
−7.9
−4.3
−2.3
0.4
0.94
0.71
0.94
71.0
−1.4
a
Four and two independent titrations were performed for 4g and 5e,
and average values are given. Compound aggregation was not observed
for all ITC titrations and was further verified by UV spectroscopy
(Supplementary Figure S19).
which is in good agreement with its IC50 value of 37.4 μM ob-
tained from the competitive binding assay, as well as the increase
in potency observed in both agglutination assays. Encouraged by
the results for the terminally amide capped mannoside 4g, we
investigated sulfonamides 5 as linking groups and common
Figure 3. (A) General structure of the LecB tetramer complexed with sulfonamide 5e. (B) Binding site of chain A with representation of hydrogen
bond network between protein amino acids, ligand, and conserved water molecules. (C) Superimposition of the four binding sites.
Figure 4. (Left) The crystal structure of sulfonamide 5e in complex with LecB shows an extended binding of the lipophilic substituent to the surface
of the protein; the sulfonamide nitrogen forms a hydrogen bond with the carboxylate of Asp96. (Right) One representative pose of one 40-ns
molecular dynamics simulation of 4g with LecB indicates an opening of the β-sandwich lectin fold to accommodate the lipophilic cinnamide moiety.
The saccharide moiety of 4g binds in the same orientation as in sulfonamide 5e in the crystal structure; calcium ions are shown in pink, and the
surface is colored by lipophilicity ranging from blue (hydrophilic) to orange (lipophilic).
1779
dx.doi.org/10.1021/cb400371r | ACS Chem. Biol. 2013, 8, 1775−1784

ACS Chemical Biology
Articles
Figure 5. (A) Binding of 4g to LecB was detected by 1H NMR. The spectra were referenced to equal intensity on the 13C satellite peak of Tris at
3.53 ppm. A strong reduction of signal intensity of 4g (in buffer, black trace) in the presence of 2 equiv of LecB (red trace) indicates binding to the
protein. The region containing the water and main Tris signal do not contain indicative signals for 4g (3.6 to 6.4 ppm) and was removed from the
spectrum for clarity reasons. (B) 15N-labeled LecB was analyzed by 1H,15N-TROSY-HSQC NMR experiments in the absence (blue) and in presence
of 0.5 equiv 4g (red). Reduction in intensities of numerous peaks were observed, indicative for a large area of interaction and/or significant structural
changes upon ligand binding that occur on an intermediate time scale. The signals below 102 ppm are folded.
Since crystals obtained for cinnamide 4g in complex with
LecB did not diffract, we performed extensive molecular
dynamics (MD) simulations: starting structures were generated
by aligning the saccharide moiety of 4g onto 1 in the crystal
structure22 and the cinnamido substituent manually oriented in
12 different poses, and 40-ns MD simulations were performed
for each of these starting poses. All simulations resulted in only
two distinctly different poses (Supplementary Figure S5). In
the first one, the saccharide part of 4g is coordinated to both
Ca2+ ions and the cinnamide moiety opens a cleft and binds
inside the formerly closed β-sandwich (Figure 4, right). In a super-
position with the crystal structure, the lowest energy structure of
4g bound to LecB indicated nearly unchanged β-sheets of the
β-sandwich fold despite the intercalating ligand, whereas slight
changes in the loops were observed (Supplementary Figure S6).
This model is consistent with the large entropic contribution to
binding of 4g to LecB (Table 2). In contrast, the other pose of
the simulation positioned the cinnamide moiety into the solvent
and thus contradicts the thermodynamic fingerprint of 4g in
comparison to 1 and can therefore be ruled out.
1780
dx.doi.org/10.1021/cb400371r | ACS Chem. Biol. 2013, 8, 1775−1784

ACS Chemical Biology
Articles
1
Figure 6. Binding of 5e to LecB was analyzed by H,15N-TROSY-HSQC NMR experiments in the absence (blue) and presence of 5e (red).
(A) Loss of signal intensity of defined signals of 15N-labeled LecB was observed upon addition of 5e to 2 equiv of LecB, and new signals
appeared. (B) For Sulfonamide 5e, the 1:1 complex with LecB could be analyzed, and chemical shift perturbation was observed for ca. 15
signals. In contrast to intermediate exchange of 4g, ligand binding of 5e to LecB occurs on a slow time scale, as deduced from the appearance
of distinct sets of signals.
To gain further insight into the molecular recognition of
ligands by LecB, we monitored the interaction of 1 and 4g by
NMR spectroscopy (Figure 5, Supplementary Figure S12).
Interestingly, 1H,15N-TROSY-HSQC experiments of
15N-labeled LecB with 1 did not show significant chemical
shift perturbation but led to strong line broadening of distinct
amide signals. Nevertheless, binding of 1 was clearly evident
also in our NMR experiments. At ligand-to-protein ratios of less
1781
dx.doi.org/10.1021/cb400371r | ACS Chem. Biol. 2013, 8, 1775−1784

ACS Chemical Biology
Articles
than or equal to 1:1 the resonances of 1 showed extensive line
broadening leading to vanishing signal intensities. At higher
excess of 1 line widths characteristic of the free sugar were
observed. This indicates that even this monosaccharide ligand
is in slow exchange on the NMR time scale and that the
resonances of the bound state are broadened beyond detection.
In contrast, cinnamide 4g causes massive changes of the 1H,15N
correlation of the 15N-labeled LecB. Addition of 4g showed
extensive global line broadening of protein resonances and
vanishing of LecB as a whole (Figure 5). This indicates that the
interaction of 4g with its binding site on LecB causes chemical
shift changes not only locally at the binding site but also ex-
tending to the whole lectin. Apparently, the binding is taking
place on an intermediate time scale leading to extensive exchange
broadening. This behavior is consistent with the proposed
binding mode of 4g, where the cinnamide moiety penetrates
into the β-sandwich causing a slight but global rearrangement
of the whole lectin domain (Figure 4, Supplementary Figure
S6). To further compare the binding between LecB and 4g and
5e, respectively, we analyzed sulfonamide 5e in complex with
LecB by 1H,15N-TROSY-HSQC NMR spectroscopy. In
contrast to the global effect due to binding of the cinnamide
4g (Figure 5B), line broadening of distinct signals of LecB was
observed upon binding of 5e (Figure 6A). The appearance of
signals in addition to the signals of the unligated protein indicates
a slow exchange between bound and unbound state on the NMR
time scale (Figure 6A). For the fully ligated protein (Figure 6B),
ca. 15 signals differ in chemical shift and/or signal intensity. This
number is in good agreement with the approximate area of the
binding site of LecB for 5e as determined by X-ray crystallography.
To assess the potential of 4g and 5e as possible lead struc-
tures for the search of inhibitors of P. aeruginosa host cell
binding and biofilm formation, we finally developed a bacterial
adhesion assay. In analogy to a plasmid-based system developed
for fimbrial adhesion of E. coli bacteria by Lindhorst et al.,39 we
generated genomically CFP-tagged wild type P. aeruginosa and
analyzed the adhesion to fucose-coated microtiter plates
(Figure 7). The expression of both lectins, LecA and LecB, is
induced in the stationary phase.40 P. aeruginosa cultures grown
for 48 h showed LecB-dependent and fucose-specific binding to
the surface, and this adhesion could be inhibited by addition of
LecB-specific ligands, i.e., fucose and 1. The controls glucose or
methyl α-D-glucoside showed only a negligible effect, which was
also observed by Lindhorst et al.41 for E. coli adhesion. All
inhibitors of LecB showed only a partial reduction of bacterial
adhesion. The fact that LecA has a low affinity for fucose42 is
likely to account for the observed residual binding to the coated
fucose, which increases its affinity due to its polyvalent nature.
The binding of a P. aeruginosa lecA knockout strain to a surface
coated with ligands of LecB could be completely inhibited
with inhibitors of LecB, supporting this hypothesis.43 For
a therapeutic application, this partial redundancy needs to
be addressed by development and co-application of a suitable
LecA inhibitor. Here, both, cinnamide 4g and sulfonamide 5e
proved to be even slightly more potent than fucose for
the inhibition of LecB-mediated adhesion of P. aeruginosa and
thus will serve as lead structures for further development.
Conclusion. In summary, we developed a new competitive
binding assay for LecB that allows in situ detection of the binding.
Because it is run in a 384-well format, this assay is amenable to
high-throughput screening of small molecule libraries. Current
enzyme linked lectin assays (ELLA) for LecB19,25 are based on
the ELISA protocol, offering the advantage of evaluation of
inhibitors over a broad affinity range. However, this format
requires numerous handling steps and suffers from low
reproducibility between plates,44 as a result of the low affinity
nature of carbohydrate−protein interactions as compared to
antibody−antigen binding in ELISA. Our new assay allowed the
screening of four novel series of ligands for LecB inhibition.
Using the crystal structure of D-mannose with LecB, we
synthesized various 6-modified mannosides targeting a cleft on
the protein surface. Mannose-derived amides were identified as
potent inhibitors with cinnamide 4g as the best identified
inhibitor among this series. In addition, a set of sulfonamide-
derived ligands were screened, and 5e was identified as a
mannose-derived, terminally modified, high affinity inhibitor of
LecB (Kd = 3.3 μM). The crystal structure of sulfonamide 5e in
complex with LecB revealed an additional hydrogen bond of
the sulfonamide nitrogen with Asp96 and a binding of the
aromatic ring to the surface of LecB. Because we could not
obtain diffracting crystals of LecB with 4g, we characterized the
binding in detail using a combination of biochemical assays,
biomolecular NMR spectroscopy, molecular dynamics simu-
lation, and bacterial adhesion. Our results suggest that the
cinnamide moiety in 4g opens a hydrophobic cavity within the
β-sandwich of LecB and binds distinctly differently than all
previously crystallized ligands of LecB, including 5e. Note-
worthy, 5g, the sulfonamide homologue of 4g carrying the same
substituent, is a 2-fold weaker ligand. This supports the two
different binding modes, which are likely to result from the
conformational difference of carboxamides (planar, trans) and
sulfonamides (staggered, gauche). Finally, it was shown for the
first time that such glycomimetic mannose-derived inhibitors
are able to inhibit the adhesion of whole P. aeruginosa bacteria.
In contrast to the natural ligand Lewisa, which is unsuitable
as potential drug (see earlier), both 4g and 5e comply with
Lipinski’s rule of five and its extensions for oral availability
(Supplementary Table S3). Because many lectins in the host
recognize terminal fucosides, inhibitors based on this epitope
are likely to interfere with important receptors involved in
inflammation and immunity. We have identified and optimized
Figure 7. Cinnamide 4g and sulfonamide 5e can inhibit the adhesion
of P. aeruginosa to a fucose-coated surface (black bars). The inhibition
of bacterial adhesion was tested in the presence of fucose (Fuc), 1, 4g,
5e, methyl α-D-glucoside (MeGlc), and D-glucose (Glc) all at 200 μM
or in the absence of ligands (w/o). Furthermore, control experiments
with a glucose-coated surface were conducted (white bars). Three
independent experiments were performed. One representative graph is
shown, and error bars represent the standard deviation of triplicates of
one experiment.
1782
dx.doi.org/10.1021/cb400371r | ACS Chem. Biol. 2013, 8, 1775−1784

ACS Chemical Biology
Articles
and PA-IIL are controlled by quorum sensing and by RpoS. J. Bacteriol.
182, 6401−6411.
terminally capped mannosides as inhibitors of LecB. Such
capping modifications are expected to yield selective inhibitors
for the pathogenic lectin over lectins of the host. Studies
addressing the selectivity and in vivo efficacy of the reported
compounds are ongoing. Knowledge gained from this work will
further improve LecB inhibitors, which may lead to a novel
class of anti-virulence drugs. Due to such a mode of action, a
strongly reduced potential of developing resistant strains can be
anticipated,45 and follow-up compounds may constitute a
suitable treatment of nosocomial infections and respiratory
tract infections of cystic fibrosis patients.
(9) Chemani, C., Imberty, A., de Bentzmann, S., Pierre, M.,
́
Wimmerova, M., Guery, B. P., and Faure, K. (2009) Role of LecA
and LecB lectins in Pseudomonas aeruginosa-induced lung injury and
effect of carbohydrate ligands. Infect. Immun. 77, 2065−2075.
́
(10) Adam, J., Pokorna, M., Sabin, C., Mitchell, E. P., Imberty, A.,
́
and Wimmerova, M. (2007) Engineering of PA-IIL lectin from
Pseudomonas aeruginosa - Unravelling the role of the specificity loop
for sugar preference. BMC Struct. Biol. 7, 36.
(11) Bajolet-Laudinat, O., Girod-de Bentzmann, S., Tournier, J. M.,
Madoulet, C., Plotkowski, M. C., Chippaux, C., and Puchelle, E.
(1994) Cytotoxicity of Pseudomonas aeruginosa internal lectin PA-I to
respiratory epithelial cells in primary culture. Infect. Immun. 62, 4481−
4487.
(12) Bernardi, A., Jimenez-Barbero, J., Casnati, A., De Castro, C.,
́
Darbre, T., Fieschi, F., Finne, J., Funken, H., Jaeger, K.-E., Lahmann,
M., Lindhorst, T. K., Marradi, M., Messner, P., Molinaro, A., Murphy,
ASSOCIATED CONTENT
* Supporting Information
Details of molecular modeling studies and all experimental
■
S
1
procedures, as well as H NMR and 13C NMR spectra. This
material is available free of charge via the Internet at http://
pubs.acs.org.
P. V., Nativi, C., Oscarson, S., Penades
Renaudet, O., Reymond, J.-L., Richichi, B., Rojo, J., Sansone, F.,
́
, S., Peri, F., Pieters, R. J.,
Schaffer, C., Turnbull, W. B., Velasco-Torrijos, T., Vidal, S., Vincent,
̈
S., Wennekes, T., Zuilhof, H., and Imberty, A. (2012) Multivalent
glycoconjugates as anti-pathogenic agents. Chem. Soc. Rev. 42, 4709−
4727.
AUTHOR INFORMATION
Corresponding Author
*E-mail: alexander.titz@uni-konstanz.de.
■
́
(13) Diggle, S. P., Stacey, R. E., Dodd, C., Camara, M., Williams, P.,
and Winzer, K. (2006) The galactophilic lectin, LecA, contributes to
biofilm development in Pseudomonas aeruginosa. Environ. Microbiol. 8,
1095−1104.
Present Address
⊥Theoretical Medicinal Chemistry and Biophysics, Institute of
Pharmacy, University of Tubingen, D-72076 Tubingen,
̈
̈
(14) Ernst, B., and Magnani, J. L. (2009) From carbohydrate leads to
glycomimetic drugs. Nat. Rev. Drug Discovery 8, 661−677.
(15) Imberty, A., Chabre, Y. M., and Roy, R. (2008) Glycomimetics
and glycodendrimers as high affinity microbial anti-adhesins. Chem.
Eur. J. 14, 7490−7499.
Germany.
Author Contributions
#These authors contributed equally to this work.
Notes
(16) Sharon, N. (2006) Carbohydrates as future anti-adhesion drugs
for infectious diseases. Biochim. Biophys. Acta 1760, 527−537.
(17) Tielker, D., Hacker, S., Loris, R., Strathmann, M., Wingender, J.,
Wilhelm, S., Rosenau, F., and Jaeger, K.-E. (2005) Pseudomonas
aeruginosa lectin LecB is located in the outer membrane and is
involved in biofilm formation. Microbiology 151, 1313−1323.
(18) Kadam, R. U., Bergmann, M., Hurley, M., Garg, D., Cacciarini,
M., Swiderska, M. A., Nativi, C., Sattler, M., Smyth, A. R., Williams, P.,
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Generous funding was received from the Zukunftskolleg, the
Deutsche Forschungsgemeinschaft, and the Fonds der Chem-
ischen Industrie. The authors acknowledge fruitful discussions
with R. Sommer, D. Strasser, S. Wagner, and V. Wittmann. J.
Hartig is kindly acknowledged for generous support. We are
grateful to B. Furones and E. Gillon for technical assistance, A.-
L. Steck is acknowledged for the HRMS measurements. We
acknowledge the use of beamline Proxima 1 at the Soleil
synchrotron (France).
́
Camara, M., Stocker, A., Darbre, T., and Reymond, J.-L. (2011) A
glycopeptide dendrimer inhibitor of the galactose-specific lectin LecA
and of Pseudomonas aeruginosa biofilms. Angew. Chem., Int. Ed. 50,
10631−10635.
(19) Johansson, E. M. V., Crusz, S. A., Kolomiets, E., Buts, L., Kadam,
́
R. U., Cacciarini, M., Bartels, K.-M., Diggle, S. P., Camara, M.,
Williams, P., Loris, R., Nativi, C., Rosenau, F., Jaeger, K.-E., Darbre, T.,
and Reymond, J.-L. (2008) Inhibition and dispersion of Pseudomonas
aeruginosa biofilms by glycopeptide dendrimers targeting the fucose-
specific lectin LecB. Chem. Biol. 15, 1249−1257.
(20) von Bismarck, P., Schneppenheim, R., and Schumacher, U.
(2001) Successful treatment of Pseudomonas aeruginosa respiratory
tract infection with a sugar solution–a case report on a lectin based
therapeutic principle. Klin. Padiatr. 213, 285−287.
(21) Hauber, H.-P., Schulz, M., Pforte, A., Mack, D., Zabel, P., and
Schumacher, U. (2008) Inhalation with fucose and galactose for
treatment of Pseudomonas aeruginosa in cystic fibrosis patients. Int. J.
Med. Sci. 5, 371−376.
REFERENCES
■
(1) Cross, A., Allen, J. R., Burke, J., Ducel, G., Harris, A., John, J.,
Johnson, D., Lew, M., MacMillan, B., and Meers, P. (1983)
Nosocomial infections due to Pseudomonas aeruginosa: review of
recent trends. Clin. Infect. Dis. 5, S837−845.
(2) Peleg, A. Y., and Hooper, D. C. (2010) Hospital-acquired
infections due to GRAM-negative bacteria. New Engl. J. Med 362,
1804−1813.
(3) Tummler, B., and Kiewitz, C. (1999) Cystic fibrosis: an inherited
̈
susceptibility to bacterial respiratory infections. Mol. Med. Today 5,
351−358.
(22) Loris, R., Tielker, D., Jaeger, K.-E., and Wyns, L. (2003)
Structural basis of carbohydrate recognition by the lectin LecB from
Pseudomonas aeruginosa. J. Mol. Biol. 331, 861−870.
(4) Poole, K. (2011) Pseudomonas aeruginosa: resistance to the max.
Front. Microbiol. 2, 65.
(5) Davies, D. (2003) Understanding biofilm resistance to
antibacterial agents. Nat. Rev. Drug Discovery 2, 114−122.
(6) Flemming, H.-C., and Wingender, J. (2010) The biofilm matrix.
Nat. Rev. Microbiol. 8, 623−633.
(7) Gilboa-Garber, N. (1982) Pseudomonas aeruginosa lectins.
Methods Enzymol. 83, 378−385.
(8) Winzer, K., Falconer, C., Garber, N. C., Diggle, S. P., Camara, M.,
and Williams, P. (2000) The Pseudomonas aeruginosa lectins PA-IL
(23) Mitchell, E., Houles, C., Sudakevitz, D., Wimmerova, M.,
́
Gautier, C., Perez, S., Wu, A. M., Gilboa-Garber, N., and Imberty, A.
(2002) Structural basis for oligosaccharide-mediated adhesion of
Pseudomonas aeruginosa in the lungs of cystic fibrosis patients. Nat.
Struct. Biol. 9, 918−921.
́
(24) Perret, S., Sabin, C., Dumon, C., Pokorna, M., Gautier, C., Galanina,
O., Ilia, S., Bovin, N., Nicaise, M., Desmadril, M., Gilboa-Garber, N.,
1783
dx.doi.org/10.1021/cb400371r | ACS Chem. Biol. 2013, 8, 1775−1784

ACS Chemical Biology
Articles
Wimmerova,
́
M., Mitchell, E. P., and Imberty, A. (2005) Structural basis for
(42) Garber, N., Guempel, U., Belz, A., Gilboa-Garber, N., and Doyle,
R. J. (1992) On the specificity of the D-galactose-binding lectin (PA-I)
of Pseudomonas aeruginosa and its strong binding to hydrophobic
derivatives of D-galactose and thiogalactose. Biochim. Biophys. Acta
1116, 331−333.
the interaction between human milk oligosaccharides and the bacterial
lectin PA-IIL of Pseudomonas aeruginosa. Biochem. J. 389, 325−332.
́ ́
(25) Marotte, K., Sabin, C., Preville, C., Moume-Pymbock, M.,
́
Wimmerova, M., Mitchell, E. P., Imberty, A., and Roy, R. (2007) X-ray
(43) Joachim, I., Wagner, S., Philipp, B., and Titz, A. Unpublished
work.
structures and thermodynamics of the interaction of PA-IIL from
Pseudomonas aeruginosa with disaccharide derivatives. ChemMed-
Chem 2, 1328−1338.
(44) Maierhofer, C., Rohmer, K., and Wittmann, V. (2007) Probing
multivalent carbohydrate-lectin interactions by an enzyme-linked lectin
assay employing covalently immobilized carbohydrates. Bioorg. Med.
Chem. 15, 7661−7676.
(45) Clatworthy, A. E., Pierson, E., and Hung, D. T. (2007)
Targeting virulence: a new paradigm for antimicrobial therapy. Nat.
Chem. Biol. 3, 541−548.
(26) Andreini, M., Anderluh, M., Audfray, A., Bernardi, A., and
Imberty, A. (2010) Monovalent and bivalent N-fucosyl amides as high
affinity ligands for Pseudomonas aeruginosa PA-IIL lectin. Carbohydr.
Res. 345, 1400−1407.
(27) Sabin, C., Mitchell, E. P., Pokorna,
Wimmerova, M., and Imberty, A. (2006) Binding of different
́
M., Gautier, C., Utille, J.-P.,
́
monosaccharides by lectin PA-IIL from Pseudomonas aeruginosa:
thermodynamics data correlated with X-ray structures. FEBS Lett. 580,
982−987.
(28) Appelmelk, B. J., van Die, I., van Vliet, S. J., Vandenbroucke-
Grauls, C. M. J. E., Geijtenbeek, T. B. H., and van Kooyk, Y. (2003)
Cutting edge: carbohydrate profiling identifies new pathogens that
interact with dendritic cell-specific ICAM-3-grabbing nonintegrin on
dendritic cells. J. Immunol. 170, 1635−1639.
(29) Guo, Y., Feinberg, H., Conroy, E., Mitchell, D. A., Alvarez, R.,
Blixt, O., Taylor, M. E., Weis, W. I., and Drickamer, K. (2004)
Structural basis for distinct ligand-binding and targeting properties of
the receptors DC-SIGN and DC-SIGNR. Nat. Struct. Mol. Biol. 11,
591−598.
(30) Sorme, P., Kahl-Knutsson, B., Huflejt, M., Nilsson, U. J., and
̈
Leffler, H. (2004) Fluorescence polarization as an analytical tool to
evaluate galectin-ligand interactions. Anal. Biochem. 334, 36−47.
(31) Han, Z., Pinkner, J. S., Ford, B., Obermann, R., Nolan, W.,
Wildman, S. A., Hobbs, D., Ellenberger, T., Cusumano, C. K.,
Hultgren, S. J., and Janetka, J. W. (2010) Structure-based drug design
and optimization of mannoside bacterial FimH antagonists. J. Med.
Chem. 53, 4779−4792.
(32) Dasgupta, S., Rajput, V. K., Roy, B., and Mukhopadhyay, B.
(2007) Lanthanum trifluoromethanesulfonate-catalyzed facile syn-
thesis of per-O-acetylated sugars and their one-pot conversion to S-
aryl and O-alkyl/aryl glycosides. J. Carbohydr. Chem. 26, 91−106.
(33) Cottaz, S., Brimacombe, J. S., and Ferguson, M. A. (1993) An
imino-linked carba-disaccharide alpha-D-mannosidase inhibitor. Car-
bohydr. Res. 247, 341−345.
(34) Rostovtsev, V. V., Green, L. G., Fokin, V. V., and Sharpless, K. B.
(2002) A stepwise huisgen cycloaddition process: copper(I)-catalyzed
regioselective ″ligation″ of azides and terminal alkynes. Angew. Chem.,
Int. Ed. 41, 2596−2599.
(35) Tornøe, C. W., Christensen, C., and Meldal, M. (2002)
Peptidotriazoles on solid phase: [1,2,3]-triazoles by regiospecific
copper(I)-catalyzed 1,3-dipolar cycloadditions of terminal alkynes to
azides. J. Org. Chem. 67, 3057−3064.
(36) Kolb, H. C., Finn, M. G., and Sharpless, K. B. (2001) Click
Chemistry: Diverse chemical function from a few good reactions.
Angew. Chem., Int. Ed. 40, 2004−2021.
(37) Eshdat, Y., and Sharon, N. (1982) Escherichia coli surface
lectins. Methods. Enzymol. 83, 386−391.
(38) Patani, G. A., and LaVoie, E. J. (1996) Bioisosterism: A rational
approach in drug design. Chem. Rev. 96, 3147−3176.
(39) Hartmann, M., Horst, A. K., Klemm, P., and Lindhorst, T. K.
(2010) A kit for the investigation of live Escherichia coli cell adhesion
to glycosylated surfaces. Chem. Commun. (Camb) 46, 330−332.
(40) Dotsch, A., Eckweiler, D., Schniederjans, M., Zimmermann, A.,
̈
Jensen, V., Scharfe, M., Geffers, R., and Haussler, S. (2012) The
̈
Pseudomonas aeruginosa transcriptome in planktonic cultures and
static biofilms using RNA sequencing. PLoS One 7, e31092.
(41) Grabosch, C., Hartmann, M., Schmidt-Lassen, J., and Lindhorst,
T. K. (2011) Squaric acid monoamide mannosides as ligands for the
bacterial lectin FimH: covalent inhibition or not? ChemBioChem 12,
1066−1074.
1784
dx.doi.org/10.1021/cb400371r | ACS Chem. Biol. 2013, 8, 1775−1784