Direct Evaluation of Live Uropathogenic Escherichia coli Adhesion

Article Cite This: Anal. Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/ac

Direct Evaluation of Live Uropathogenic Escherichia coli Adhesion and Efficiency of Antiadhesive Compounds Using a Simple Microarray Approach Ioanna Kalograiaki,†,‡,§,∇ Marta Abellán-Flos,∥,#,∇ Luis Á ngel Fernández,⊥ Margarita Menéndez,†,‡ Stéphane P. Vincent,*,∥ and Dolores Solís*,†,‡ †

Instituto de Química Física Rocasolano, CSIC, Serrano 119, 28006 Madrid, Spain CIBER de Enfermedades Respiratorias (CIBERES), Avda Monforte de Lemos 3-5, 28029 Madrid, Spain § Centro de Investigaciones Biológicas, CSIC, Ramiro de Maeztu 9, 28040 Madrid, Spain ∥ Département de Chimie, Laboratoire de Chimie Bio-Organique, University of Namur (UNamur), Rue de Bruxelles 61, B-5000 Namur, Belgium ⊥ Centro Nacional de Biotecnología (CNB), CSIC, Darwin 3, Campus UAM Cantoblanco, 28049 Madrid, Spain

Downloaded via KAOHSIUNG MEDICAL UNIV on October 7, 2018 at 06:15:34 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡

S Supporting Information *

ABSTRACT: Many pathogens use host glycans as docking points for adhesion. Therefore, the use of compounds blocking carbohydrate-binding adhesins is a promising strategy for fighting infections. In this work, we describe a simple and rapid microarray approach for assessing the bacterial adhesion and efficiency of antiadhesive compounds targeting uropathogenic Escherichia coli UTI89, which displays mannose-specific adhesin FimH at the tip of fimbriae. The approach consisted in direct detection of live fluorescently labeled bacteria bound to mannan printed onto microarray slides. The utility of the arrays for binding/inhibition assays was first validated by comparing array-derived results for the model mannose-binding lectin concanavalin A with data obtained by isothermal titration calorimetry. Growth phasedependent binding of UTI89 to the arrays was observed, proving the usefulness of the setup for detecting differences in FimH expression. Importantly, bacteria labeling and binding assays entailed minimal manipulation, helping to preserve the integrity of fimbriae. The efficiency of three different dodecamannosylated fullerenes as FimH-targeted antiadhesives was next evaluated in competition assays. The results revealed a superior activity of the mannofullerenes (5- to 18-fold per mannose residue) over methyl α-D-mannopyranoside. Moreover, differences in activity were detected for mannofullerenes differing in the structure/ length of the spacer used for grafting mannose onto the fullerene core, further demonstrating the sensitivity of the assay. Overall, the approach combines straightforward and time-saving protocols for microarray preparation, bacteria labeling, and binding assays, and it can be easily tailored to other bacteria bearing carbohydrate-binding adhesins.

G

specific adhesin of uropathogenic Escherichia coli, or Vibrio cholerae toxin, which binds to ganglioside GM1, are relevant examples of well-characterized lectins from important human pathogens. Intuitively, interfering with harmful host−pathogen glycan− lectin interactions could be the basis for the development of novel strategies to fight infections, which are increasingly needed due to the emergence and spreading of antimicrobial resistances and the ensuing appearance of the so-called “superbugs”.4−6 Identification of carbohydrate epitopes on

lycans coating cell surfaces serve as recognition signals for a variety of receptors, thereby playing key roles in cell physiology and mediating cell communication and adhesion processes.1 In particular, sensing by host receptors of nonself glycans exposed on pathogens’ surfaces triggers immune signaling and activation. Not surprisingly, some pathogens have developed different mechanisms for mimicking host glycans to evade defense responses. Moreover, recognition of such self-like glycans by host receptors may also be exploited by the pathogen for down-regulation of innate immunity or as mechanism for attachment and cell entry. Conversely, many pathogens, including viruses, bacteria, fungi, and parasites, use host glycans as docking points for adhesion or as targets for secreted toxins.2,3 The influenza virus hemagglutinin, which binds sialic acid-containing glycans, the mannose (Man)© XXXX American Chemical Society

Received: September 17, 2018 Accepted: September 20, 2018

A

DOI: 10.1021/acs.analchem.8b04235 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Culture and Fluorescent Labeling of Bacteria. E. coli UTI89 wild type and ΔfimH strains were kindly provided by Dr. Julie Bouckaert (CNRS, Lille, France). Bacterial cultures were grown at 37 °C in enriched lysogeny broth (LB2YT) containing 8 g of tryptone, 5 g of yeast extract, and 5 g of NaCl per 500 mL. The concentration of bacteria was estimated by correlating turbidity measurements with the number of colonies obtained after plating serial dilutions on LB2YTagar. Fifteen milliliters of culture in the desired growth phase (OD600 = 0.3−0.5 in the exponential phase, corresponding to 0.88−1.13 × 108 colony-forming units (CFUs) per mL, and OD600 above 1 in the stationary phase, corresponding to 0.98− 1.16 × 109 CFUs) was collected and centrifuged at 4000g for 15 min at 4 °C. Pellets were washed with 10 mM HEPES, pH 7.4, containing 150 mM NaCl (HBS), and resuspended in 5 mL of the same buffer. For labeling, 10 μL of the redfluorescent nucleic acid stain SYTO 62 (Life Technologies) was added to the suspension at a final concentration of 10 μM. After 10 min in the dark, bacteria were centrifuged, washed with HBS containing 0.1% Tween 20 (HBS-T), and resuspended in 5 mL of the same buffer. Fimbriae Purification and Biotinylation. For permanent expression of type 1 pili34 with high biomass yield, UTI89 cells were cultivated at least two consecutive times in static LB2YT broth for 18 h at 37 °C. Fimbriae were then mechanically detached from the cells and selectively precipitated in the presence of 500 mM NaCl and 250 mM MgCl2, as described by Munera et al.35 Fimbria pellets were resuspended in 5 mM Tris-HCl, pH 8, overnight at 4 °C, and the samples were analyzed by SDS−PAGE, also as described.35 For determination of protein concentration, the buffer was exchanged to HBS, and the amino acid composition was determined quantitatively after gas phase acid hydrolysis (carried out at the Protein Chemistry Facility of the Biological Research Center, CSIC). For binding/inhibition assays, fimbria suspensions in HBS were biotinylated using a biotinamidocaproate ester derivative (GE Healthcare Life Sciences), following the manufacturer’s recommendations, and exhaustively dialyzed against HBS at 4 °C. Microarray Preparation. A mannan-poly-L-lysine conjugate was first prepared as follows. Forty microliters of glutaraldehyde (25% in water, grade I, Sigma-Aldrich) was added to 1 mL of 10 mg/mL mannan from Saccharomyces cerevisiae (Sigma-Aldrich) in 50 mM NaHCO3, pH 9.6. After incubation for 1 h at 20 °C, poly-L-lysine hydrobromide (30− 70 kDa, Sigma-Aldrich) was added at 200 μg/mL final concentration, and the mixture was incubated for 30 min at 20 °C with periodic shaking. Cross-linking was rendered irreversible by the addition of NaBH4 traces. The mannanpoly-L-lysine conjugate was then exhaustively dialyzed against Milli-Q water followed by 5 mM sodium phosphate buffer, pH 7.2, and 0.2 M NaCl (PBS), and its hexose concentration was determined by the phenol-sulfuric acid method,36 using pure mannan as reference. Conjugate dilutions from 6 to 0.1 mg/ mL in terms of mannan concentration, in PBS containing 47% glycerol and 0.06% Triton X100,7 were printed on a 16-pad layout on hydrogel-coated slides activated with N-hydroxysuccinimide ester (Nexterion Slide H, Schott AG), using a noncontact arrayer (Sprint, Arrayjet Ltd.). For concanavalin A (ConA, Sigma-Aldrich) competition assays, the glycoprotein ribonuclease B (Sigma-Aldrich) was similarly printed at 0.1−1 mg/mL. To enable postarray monitoring of the spots, the Cy3 fluorophore (GE Healthcare) was added to the mannan-poly-

the pathogen’s surface and characterization of the binding properties of microbial lectins is essential to this aim. Recently, we developed novel bacterial microarrays as efficient tools for profiling accessible carbohydrate structures and screening of their recognition by host receptors, also enabling the evaluation of potential inhibitors of these interactions.7−9 As a complementary approach, here we report on the potential of tailored carbohydrate microarrays for evaluation of the functionality of glycan-binding adhesins and testing of antiadhesive compounds using live bacteria, thereby preserving the natural arrangement, distribution, and density of adhesins on the pathogen’s surface. Furthermore, the clustered presentation of glycans in the arrays facilitates the establishment of multivalent interactions, as occurs in the bacteria−host cell interplay. This work focuses on uropathogenic Escherichia coli (UPEC), a bacterium that has developed resistance to common antibiotics (for example, fluoroquinolones are nowadays inefficient in more than 50% of patients in many parts of the world).10 One of the best characterized UPEC adhesins is FimH, which is displayed at the tip of type-1 fimbriae (a single FimH molecule per fimbria)11 and specifically binds to terminal mannosides of the highly mannosylated glycoprotein uroplakin 1a on bladder urothelial cells. FimH is also involved in infection by adherent-invasive E. coli in Crohn’s disease.12,13 Therefore, it is an attractive target for the development of antiadhesive strategies, and numerous mono- and multivalent mannosides have been designed as FimH antagonists and tested using different approaches.14−18 Since simultaneous interactions mediated by several fimbriae enhance adhesion, multivalent compounds potentially targeting several fimbrial adhesins have attracted great interest.19 A first difficulty is that the number of fimbriae and spacing between them can vary between cells. Moreover, expression of type 1 fimbriae is controlled by phase variation,20,21 which may be affected by growth conditions.22 Several proteins directly or indirectly regulate phase switching,21 e.g., the sigma factor RpoS, which is preferentially expressed in stationary phase and down-regulates type 1 fimbriae expression.23 Altogether, it seems evident that evaluation of FimH-targeted antiadhesive compounds should be performed using live bacteria grown under well-defined conditions. Classical hemagglutination- and microtiter plate-based assays have typically been used for detecting E. coli binding and screening of inhibitors.14,15,24,25 The applicability of glycan microarrays to this aim was also explored using a nonpathogenic surrogate of enterohemorrhagic E. coli as model bacterium.26−28 In this work, we developed a simple and rapid microarray approach for testing Man-targeted adhesion of live UPEC. The method proved to be suitable for evaluating the efficiency of FimH-mediated adhesion in exponential and stationary bacterial growth phases, and for testing the antiadhesive potential of dodecamannosylated fullerenes, a class of compounds previously found to serve as multivalent ligands for FimH.14,29

■

EXPERIMENTAL SECTION Glycofullerenes Synthesis. Glycofullerenes were synthesized by adapting the synthetic procedures previously described for the preparation of dodecavalent fullerenes.30−33 Reagents and conditions, detailed procedures, and analytical data of the novel compounds are gathered in the Supporting Information. B


Article

Analytical Chemistry and ribonuclease B solutions at 1 μg/mL final concentration. Printed slides were kept overnight under the environmental control system of the arrayer (JetMosphere, Arrayjet Ltd.) for maintaining appropriate temperature (25 °C) and relative humidity (∼50%), and next scanned to locate and grid Cy3 signals (excitation at 532 nm, green laser) using a GenePix Autoloader 4200AL (Axon Instruments) microarray scanner. Then, slides were washed with Milli-Q water and HBS, and any potentially remaining active group was deactivated by incubation for 90 min with 50 mM ethanolamine in 0.1 M boric acid, pH 8.5. Following thorough washing with Milli-Q water, deactivated slides were directly used for binding assays or dried under a nitrogen stream and stored at −20 °C with desiccant. Microarray Binding and Inhibition Assays. Mannanfunctionalized microarray slides were incubated with SYTO 62-stained bacterial suspensions in HBS-T at OD600 = 1 (100 μL per pad) for 1 h at room temperature, in the absence or presence of different concentrations of the tested compounds. After a short wash with HBS and Milli-Q water, slides were dried under a nitrogen stream and scanned for SYTO 62 signals (excitation at 635 nm, red laser) using the maximum laser power and a resolution of 10 μm. For ConA binding/ inhibition assays, lectin was biotinylated by incubation with a biotinamidocaproate ester derivative (GE Healthcare Life Sciences) as described previously.8 Slides were washed with HBS-T and incubated for 75 min at room temperature in the same buffer with 0.1−0.4 μM biotinylated ConA, in the absence or presence of inhibitors. Slides were then washed twice with HBS-T and incubated for 35 min with 1 μg/mL AlexaFluor 647 (AF647)-labeled streptavidin (Invitrogen) in HBS-T. After washing with HBS and Milli-Q water, slides were dried under a nitrogen stream and scanned for AF647 signals (excitation at 635 nm, maximum laser power, and 10 μm resolution). The same procedure was followed for testing the binding of biotinylated fimbriae at a concentration of 75 μg/ mL. In each case, the gain of the photomultiplier tube was adjusted depending on the intensity of the signals and the specific aim of the assay. SYTO 62 and AF647 fluorescence signals were quantified with GenePix Pro 6.1 software (Molecular Devices). For spot segmentation, a fixed 100-μm circle and the grid of Cy3 signals as reference for spot coordinates were employed. Median spot intensities minus median local background were used for final quantitation. The inhibitory activity (IC50 values) for the tested compounds was estimated by semilogarithmic plotting of the percentage of bound bacteria/lectin versus inhibitor concentration, taking as 100% the binding in the absence of inhibitors, and sigmoidal fit to experimental data using Origin 7.0 software (OriginLab). Isothermal Titration Calorimetry. ITC analyses of ConA−ligand interactions were carried out at 22 °C either in a MicroCal VP-ITC calorimeter (GE Healthcare) or in a Malvern Microcal PEAQ-ITC instrument. Lectin was exhaustively dialyzed against HBS, pH 7.4, and ligand solutions (at 11.7−15 mM for monovalent ligands and 0.2 mM for mannofullerenes) were prepared using the last dialysis buffer. Protein concentration was determined from the absorbance at 280 nm, using an extinction coefficient E1 cm0.1% of 1.24. Titrations were carried out by stepwise injection of the ligand into the reaction cell loaded with the ConA solution at a monomer concentration of 79−239 μM. Titration data were analyzed with Microcal-ITC Origin software. The concen-

tration of functional ConA determined by titration with methyl α-D-mannopyranoside (MeαMan) was used as input for ligand concentration when fitting titrations with dodecamannosylated fullerenes using the “ligand-in-cell” option.

L-lysine

■

RESULTS AND DISCUSSION Establishment of a Microarray Setup for Evaluation of Man-Targeted UPEC Adhesion. A strong preference of FimH toward oligomannosides exposing Manα(1−3)Man has been reported.24,37 Therefore, mannan from Saccharomyces cerevisiae, which displays multiple terminal α(1,3)-linked Man residues, was selected as the probe for immobilization in the arrays. Indeed, mannan has been shown to be an efficient ligand for two different FimH-displaying E. coli laboratory strains.25,27 Regarding the selection of the microarray surface, preliminary studies evidenced a poor performance in UPEC binding assays of mannan-coated nitrocellulose slides, a surface previously used for the development of bacterial microarrays.7−9 As an alternative, we used commercial hydrogelcoated glass surfaces displaying carboxylic acids activated with N-hydroxysuccinimide ester. These esters react with primary amines releasing N-hydroxysuccinimide and yielding stable amide bonds, thereby enabling covalent immobilization of amine-containing biomolecules, e.g., amino-derivatized glycans.38 Therefore, taking advantage of the mannoprotein bound to the polymannose backbone, mannan was conjugated to poly-L-lysine using glutaraldehyde,39 likely through a reaction of the dialdehyde with the ε-amino group of lysine residues from the mannoprotein and poly-L-lysine.40 The rationale behind mannan−poly-L-lysine conjugation was that the profuse presence of amino groups in the poly-L-lysine moiety should facilitate the covalent immobilization on the succinimide-activated slides. To assess the efficiency of mannan coupling and immobilization in the arrays, serial dilutions of the conjugate were used for printing, and the binding of the Man-specific lectin ConA was examined. Strong binding signals showing the expected mannan concentrationdependent trend were detected (Figure 1, upper panel), thereby validating the approach used for microarray preparation. Of note, comparable ConA binding to slides freshly prepared and after storage for 6 or 12 months at −20 °C was observed (Figure 1, lower panel), demonstrating the stability of the arrays. Having proved the robustness of the mannan arrays for lectin-binding assays, we evaluated the binding of UPEC strain UTI89, originally isolated from a patient with acute bladder infection.41 Different methods have been reported for monitoring binding, e.g., indirect detection using biotinylated bacteria and fluorescent streptavidin, or direct detection of fluorescent bacteria using transformed strains expressing green fluorescent protein15,25 or bacteria labeled with a fluorescent dye.27,28 Pursuing the development of a rapid and simple microarray approach suitable for testing adhesion of pathogenic bacterial isolates, fluorescent labeling was the method of choice. Thus, live UTI89 was labeled by 10 min of incubation with the red-fluorescent cell-permeable nucleic acid dye SYTO 62 and directly used for binding assays. The behavior of bacteria in exponential and stationary growth phases was compared. As shown in Figure 2A,D, UTI89 in the exponential phase strongly bound to the mannan arrays in a probe concentration-dependent manner, and the binding was successfully inhibited in the presence of MeαMan (Figure 2B,D), evidencing carbohydrate-mediated adhesion. In conC


Article


core was found to have an effect on the inhibitory potential of the mannofullerene.14 On the basis of these precedents, the efficiency of these compounds as E. coli antiadhesives was evaluated using the established microarray setup. Three fullerene hexakis-adducts (1−3, Figure 3), bearing 12 Man residues grafted onto the C60 core by different spacers, were synthesized by 12 simultaneous CuAAC “click” reactions between synthetic monovalent azido-sugar derivatives and the trimethylsilyl-protected polyalkyne fullerene scaffold developed by Nierengarten and collaborators.48 Glycofullerene 4, bearing 12 galactose (Gal) moieties, was prepared in the same way to be used as the negative control. The structure of all the compounds was confirmed by infrared, 1H and 13C NMR spectroscopies, and mass spectrometry (detailed in the Supporting Information). As a proof of concept, mannofullerenes 1 and 2 were first tested as inhibitors of ConA and their relative inhibitory potency compared with the thermodynamic parameters of ConA−fullerene interactions determined by ITC (Table 1). The monovalent R1 Man-derivative bearing a terminal azide (compound 6 in Scheme S-1, here designated R1−azide) or triazole group (designated R1−triazole), as present in the glycofullerenes (Figure 3), was also tested to assess the influence of the aglycon chain and triazole ring in the recognition. It is important to mention that ConA is a tetrameric lectin potentially able to establish simultaneous interactions with different fullerene molecules, which is also possible for UPEC cells (but not for isolated FimH molecules, which are monomeric). In comparison to the mannandisplayed Manα(1−3)Man epitope, ConA shows superior affinity for the Manα(1−3)[Manα(1−6)]Man core of glycoprotein N-glycans, with the α(1−6)Man occupying the primary binding subsite, and the affinity is further enhanced for high-mannose chains bearing α(1−2)-linked Man residues.49 Therefore, ribonuclease B (RNaseB), a glycoprotein containing high-mannose Man5−9 glycoforms strongly bound by ConA,50 was used as array-printed ligand to evaluate the efficiency of mannofullerenes as ConA inhibitors. A dosedependent inhibition of ConA binding to RNaseB by the two mannofullerenes was observed, as illustrated in Figure S-6A for mannofullerene 1. In contrast, galactofullerene 4, tested in parallel, was not inhibitory, proving that the activity of mannofullerenes was entirely due to Man-specific recognition. IC50 values were estimated from the plot of % inhibition of binding versus fullerene concentration (Figure S-6B) and compared with that obtained for MeαMan, used as reference, and for the monovalent Man-derivatives R1−azide and R1− triazole (Table 1). A seeming multivalency effect was observed for fullerenes 1 and 2, which exhibited a 60- to 90-fold higher inhibitory potency than MeαMan when normalized per Man moiety (Table 1). Of note, the presence of the aglycon chain and triazole ring in the R1 derivatives tested did not improve their inhibitory potential over MeαMan. Calorimetric titration of ConA with the three glycofullerenes (Supporting Information, Figure S-7) confirmed the lack of binding to galactofullerene 4 and yielded the thermodynamic parameters for mannofullerenes 1 and 2 compiled in Table 1. For both compounds, the average number of binding sites for ConA per dodecafullerene molecule was around 7, indicating that simultaneous binding of ConA to the 12 mannose residues is not possible, likely due to steric hindrance (discussed in the Supporting Information). It is worth mentioning that samples recovered from the calorimeter at the end of the experiment

Figure 1. Binding of ConA to mannan−poly-L-lysine immobilized on hydrogel-coated microarray slides. Upper panel: mannan−poly-Llysine was printed as sextuplicates at decreasing concentrations, and the binding of 0.4 μM biotin-labeled ConA was monitored as described in the Experimental Section. White spots indicate detector saturation and correspond to intensity values above 40 000 relative fluorescence units (rfu). Lower panel: binding of 0.1 μM biotinylated ConA to freshly prepared slides (white columns) and to slides stored for 6 (light gray) and 12 months (gray columns) at −20 °C. All experiments were carried out in HBS, pH 7.4.

trast, bacteria in stationary phase only gave weak binding signals at the highest mannan concentrations (Figure 2C). Moreover, no significant binding was detected for the mutant strain UTI89Δf imH, while isolated UTI89 fimbriae gave strong MeαMan-inhibitable signals (Figure 2D), altogether supporting that UTI89 binding to the arrays was mediated by FimH. Analogous results were obtained for a different E. coli strain also displaying FimH-terminated type 1 fimbriae (K-12, substrain MG1655) 4 2 and its afimbriated mutant (MG1655Δf im)43 (Figure 2D). Thus, the established microarray setup proved to be efficient for assessing Man-targeted adhesion of UTI89 and MG1655, discriminating between bacterial cells differing in FimH expression. Therefore, we next examined the applicability of the setup for the evaluation of potential inhibitors of FimH-mediated bacterial adhesion. Microarray-Assisted Evaluation of Antiadhesive Compounds. As mentioned above, multivalent compounds are attractive candidates as E. coli antiadhesives. Therefore, different FimH-targeted structures with varying glycan densities and spatial arrangements, ranging from mediumsized scaffolds to nanoparticles, have been designed and tested for bioactivity.18 Although simultaneous interactions with more than one fimbria of the same bacterial cell might only be possible for large compounds, medium-sized glycoclusters can aggregate E. coli cells through binding to different bacterial cells, and binding enhancements due to statistically favored rebinding or formation of cross-linked lattices could also be expected.44−47 This was the case for a dodecamannosylated fullerene tested as inhibitor of UTI89-mediated hemagglutination, which exhibited an inhibition titer 30-fold lower (2.5fold per Man residue) than the monovalent Man derivative grafted onto the fullerene core.14 Interestingly, the structure and length of the spacer between the sugar unit and the C60 D


Article


Figure 2. Binding of E. coli cells and isolated fimbriae to mannan microarrays. Mannan−poly-L-lysine was printed as sextuplicates at decreasing concentrations, and the binding of SYTO 62-stained bacteria was tested. (A−C) Fluorescence signals detected for UTI89 in exponential (A,B) and stationary (C) growth phase, in the absence (A,C) or presence of 0.1 M MeαMan (B). (D) Fluorescence intensity of the binding signals obtained for SYTO 62-labeled UTI89, UTI89Δf imH, MG1655, and MG1655Δf im cells in exponential growth phase and for biotin-labeled fimbriae isolated from UTI89, in the absence (white) and presence (gray) of 5 mM MeαMan. Please note that since the detection system used for cells and fimbriae is different, their respective fluorescence intensities are not directly comparable. (E,F) Representative scans showing the inhibition of the binding of UTI89 in exponential growth phase by increasing concentrations of mannofullerenes 1 (E) and 3 (F). (G) Inhibitory activity of mannofullerenes 1 (□), 2 (●), and 3 (▲) on the binding to mannan−poly-L-lysine printed at 6 mg/mL. Data shown are the mean of at least four different experiments, and error bars indicate standard deviations. Solid lines correspond to sigmoidal fits to experimental data. All experiments were carried out in HBS, pH 7.4. For scanning, the gain of the photomultiplier tube was adjusted to 500, except for panel D in which a gain of 700 was used.

Figure 3. Structures of mannofullerenes 1−3 and galactofullerene 4. Twelve Man (R1−3) or Gal (R4) residues were grafted onto the C60 core via CuAAC, displaying polyethylene glycol (R1,2,4) or heptyl (R3) linkers, as described in the Experimental Section of the Supporting Information.

42-fold lower than that for MeαMan, analyzed in parallel, demonstrating a clear multivalency effect. Moreover, direct comparison of mannofullerene 1 and the monovalent

were visibly cloudy, evidencing cross-linking of mannofullerenes by ConA. The dissociation constant (Kd) obtained for each individual ConA−mannoside binding event was 16- to E


Article


Table 1. Microarray- and ITC-Derived Parameters for the Binding of ConA to Mannofullerenes 1 and 2 and Monovalent Mannosides mannofullerene parameter Microarray Competition Assay IC50 (μM) relative potency per moleculea per Man residueb ITC Analysis Nc Kd (μM)e ΔH (kcal mol−1) ΔS (cal mol−1 degree−1)

monovalent mannoside

1

2

R1−azide

R1−triazole

MeαMan

0.64 ± 0.01

0.42 ± 0.01

830 ± 50

780 ± 80

460 ± 50

713 59.4

1088 90.7

0.55 0.55

0.56 0.56

1 1

6.8 ± 0.7 6.85 ± 0.3 −10.1 ± 0.1 −10.6 ± 0.4

7±1 2.5 ± 0.7 −10 ± 1 −8 ± 4

0.78 ± 0.01 70 ± 1 −9.5 ± 0.1 −13.0 ± 0.4

0.86 ± 0.01 106.0 ± 0.6 −7.71 ± 0.07 −8.0 ± 0.2

1.5 ± 0.1 125 ± 25 −4.8 ± 0. 2 2±1

1d 125 ± 25d −7.3 ± 0.3d −7 ± 2d

a Relative potency taking the IC50 value of MeαMan as a unit. bRelative potency normalized per mannose unit. cFor 1 and 2, N is the number of mannose units occupied by ConA per dodecamannosylated fullerene. For MeαMan, N is the relative number of functional ConA monomers. d Parameters recalculated per mole of ConA monomer (see the Supporting Information). eCalculated per ConA monomer. Data correspond to the mean of at least three independent experiments.

derivatives R1−azide and R1−triazole indicated that the multivalent display in the mannofullerene resulted in 10- to 18-fold enhancement in binding affinity (Table 1). In addition, the Kd obtained for mannofullerene 2 was found to be 2.9-fold lower than that for mannofullerene 1, due to a smaller entropic penalty, in line with the greater inhibitory potency observed for the former in the microarray competition assays. Overall, ITC data were in agreement with the binding trends observed in the microarray competition experiments, thus validating the arrayderived results. Therefore, we next examined the potential of mannofullerenes as E. coli antiadhesives using the established microarray setup. The binding to the mannan arrays of UTI89 in exponential growth phase was examined in the absence and presence of increasing concentrations of glycofullerenes 1−4. As representatively shown in Figure 2E,F for mannofullerenes 1 and 3, inhibition of the binding was already observed at mannofullerene concentrations of 0.3 μM. In contrast, galactofullerene 4 was not inhibitory, proving that the antiadhesive activity of mannofullerenes was entirely due to Man-mediated interactions. IC50 values (Table 2), estimated from the plot of % inhibition of binding versus mannofullerene concentration (Figure 2G), evidenced a clearly enhanced inhibitory activity for the three mannofullerenes compared to that of MeαMan, simultaneously tested, with 5- to 18-fold higher relative potency per Man residue (Table 2). However, when testing mannofullerene 1 as inhibitor of the binding of isolated fimbriae, only a 2-fold enhancement in inhibitory potency per Man residue was observed. Furthermore, IC50 values obtained for the monovalent Man-derivatives R1−azide and R1−triazole (Table 2), tested in parallel as inhibitors of UTI89 cells and fimbriae, were comparable to those obtained for MeαMan, indicating that the aglycon chain and triazole ring are not directly involved in the recognition and confirming a clear multivalency effect for mannofullerene 1 at the level of cells. Interestingly, mannofullerene 3, whose respective monovalent derivative was previously reported to be an excellent ligand for the recombinant FimH lectin domain,51 was 2.6- to 3.6-fold less inhibitory than mannofullerenes 1 and 2, which contain more flexible polyethylene glycol spacers, a rather surprising result considering the superiority of heptyl-mannose over pegylated mannosides as FimH ligand.46 This result nicely illustrates that the behavior of isolated adhesins may not

Table 2. Inhibitory Potential of Mannofullerenes and Monovalent Mannosides on the Binding of UTI89 Cells and Isolated Fimbriae to Mannan Microarrays relative potency

UTI89 cells

fimbriae

per Man residuec

compound

IC50a (μM)

per moleculeb

mannofullerene 1

0.20 ± 0.01

154

13

mannofullerene 2 mannofullerene 3 R1−azide R1−triazole MeαMan mannofullerene 1 R1−azide R1−triazole MeαMan

0.14 ± 0.01 0.52 ± 0.02 32 ± 6 33 ± 6 31 ± 2 38 ± 5 81 ± 4 82 ± 4 80 ± 4

215 59 0.97 0.94 1 2.1 0.99 0.98 1

18 5 0.97 0.94 1 0.17 0.99 0.98 1

a

Concentration required for 50% inhibition of the binding to mannan−poly-L-lysine printed at 6 mg/mL. bRelative potency taking the IC50 value of MeαMan as a unit. cRelative potency normalized per mannose unit.

exactly reflect the behavior of the entire bacteria. Indeed, at the tip of type-1 fimbriae, the lectin domain of FimH is stabilized in a low-affinity state by the C-terminal pilin domain, which is anchored to the subsequent FimG subunit. However, when the domains separate, e.g., under shear stress originating from urine flow, a high-affinity state of the lectin domain is induced.52 Consequently, small or multimeric ligands may display distinct behaviors when assayed on the isolated FimH lectin or in the context of a whole-cell competition assay. Thus, the results point out an advantage of the technology developed in this study since the ultimate objective is to discover or characterize antiadhesive molecules, i.e., molecules that act as inhibitors of host−pathogen interactions at the cell level, which may differ from the discovery of ligands of individual soluble proteins. Moreover, mannofullerene 2 was found to be slightly more efficient than mannofullerene 1, which bears a shorter spacer. Overall, the microarray competition assays were sensitive to small changes in the inhibitory potential of mannofullerenes arising from variations in the flexibility and length of the spacer. F


Article


■

Author Contributions

CONCLUSIONS The FimH-tailored microarray setup proved to be efficient for assessing Man-targeted adhesion of the uropathogenic E. coli strain UTI89, discriminating between bacterial cells with different levels of FimH expression, and for comparative evaluation of antiadhesive compounds. A first advantage of the approach, compared to other sophisticated methods for preparation of carbohydrate microarrays,53,54 is the use of commercially available slide surfaces that enable a straightforward covalent immobilization of amino-derivatized glycans. Notably, the stability of the arrays generated allows the utilization of a set of simultaneously printed slides for comparative analyses for at least 1 year. Another advantage is the use of live bacteria readily labeled with a fluorescent cellpermeable nucleic acid dye, not harming interactions taking place at the surface. Labeled bacteria are directly used for binding/inhibition assays involving a single incubation step, so the functionality of the bacterial adhesin is evaluated very rapidly and with minimal manipulation, thereby reducing contamination risks and helping to prevent the breakage of the FimH-containing fimbrial tip.11 Moreover, the efficiency of antiadhesive compounds can be simultaneously examined, the assay requiring small quantities of the compounds thanks to the miniaturization characteristic of the microarray format. The sensitivity of the competition assay was demonstrated by the differences in activity detected for mannofullerenes differing only in the structure/length of the spacer between mannose and the fullerene core. Altogether, the established approach benefits from the simplicity of all steps, from microarray preparation to bacteria labeling and binding assays, therefore representing a valuable improvement over more laborious and time-consuming protocols used previously. Importantly, the setup can be easily tailored to other bacteria bearing carbohydrate-binding adhesins by immobilizing a suitable ligand on the microarrays, thus setting the grounds for the development of simple platforms of use in microbiology, biomedicine, or clinical fields.

■

∇

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We gratefully acknowledge financial support from the Spanish Ministry of Economy and Competitiveness (Grant BFU201570052-R), the Spanish Ministry of Science, Innovation and Universities (Grant BIO2017-89081-R, AEI/FEDER, UE), the CIBER of Respiratory Diseases (CIBERES), an initiative from the Spanish Institute of Health Carlos III (ISCIII), and the Marie Curie Initial Training Networks DYNANO (Grant PITN-GA-2011-289033) and GLYCOPHARM (Grant PITNGA-2012-317297). I.K. and M.A.F were funded by Marie Curie contracts from the European Commission. We thank the Protein Chemistry Facility of the Biological Research Center (CSIC) for quantitative amino acid analysis, as well as Professor Ernesto Garciá and Professor José Manuel Andreu for letting us use the BSL-2 laboratory for bacterial culture and binding assays and the PEAQ-ITC, respectively.

■

REFERENCES

(1) Gabius, H.-J.; André, S.; Jiménez-Barbero, J.; Romero, A.; Solís, D. Trends Biochem. Sci. 2011, 36, 298−313. (2) Kulkarni, A. A.; Weiss, A. A.; Iyer, S. S. Med. Res. Rev. 2010, 30, 327−393. (3) Nizet, V.; Varki, A.; Aebi, M. In Essentials of Glycobiology, 3rd ed.; Varki, A., Cummings, R. D., Esko, J. D., Stanley, P., Hart, G. W., Aebi, M., Darvill, A. G., Kinoshita, T., Packer, N. H., Prestegard, J. H., Schnaar, R. L., Seeberger, P. H., Eds.; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 2017. (4) Davies, J.; Davies, D. Microbiol. Mol. Biol. Rev. 2010, 74, 417− 433. (5) Khan, S. N.; Khan, A. U. Breaking the spell: Combating multidrug resistant ‘superbugs’. Front. Microbiol. 2016, 7. DOI: 10.3389/fmicb.2016.00174 (6) Rappuoli, R.; Bloom, D.; Black, S. Nature 2017, 552, 165−167. (7) Campanero-Rhodes, M. A.; Llobet, E.; Bengoechea, J. A.; Solis, D. RSC Adv. 2015, 5, 7173−7181. (8) Kalograiaki, I.; Euba, B.; Proverbio, D.; Campanero-Rhodes, M. A.; Aastrup, T.; Garmendia, J.; Solís, D. Anal. Chem. 2016, 88, 5950− 5957. (9) Kalograiaki, I.; Campanero-Rhodes, M. A.; Proverbio, D.; Euba, B.; Garmendia, J.; Aastrup, T.; Solís, D. Chapter Two - Bacterial Surface Glycans: Microarray and QCM Strategies for Glycophenotyping and Exploration of Recognition by Host Receptors. In Methods in Enzymology; Imperiali, B., Ed.; Academic Press, 2018; Vol. 598, pp 37−70. (10) World Health Organisation. http://www.who.int/mediacentre/ factsheets/fs194/en/ (accessed Jan, 2018). (11) Knight, S. D.; Bouckaert, J. Structure, Function, and Assembly of Type 1 Fimbriae. In Glycoscience and Microbial Adhesion; Lindhorst, T. K., Oscarson, S., Eds.; Springer: Berlin, Germany, 2009; pp 67− 107. (12) Barnich, N.; Carvalho, F. A.; Glasser, A.-L.; Darcha, C.; Jantscheff, P.; Allez, M.; Peeters, H.; Bommelaer, G.; Desreumaux, P.; Colombel, J.-F.; Darfeuille-Michaud, A. J. Clin. Invest. 2007, 117, 1566−1574. (13) Sivignon, A.; Bouckaert, J.; Bernard, J.; Gouin, S. G.; Barnich, N. Expert Opin. Ther. Targets 2017, 21, 837−847. (14) Durka, M.; Buffet, K.; Iehl, J.; Holler, M.; Nierengarten, J.-F.; Taganna, J.; Bouckaert, J.; Vincent, S. P. Chem. Commun. 2011, 47, 1321−1323.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.8b04235.

■

I.K. and M.A.-F. contributed equally as cofirst authors.

Description of the synthesis and characterization of monomeric glycoside derivatives and glycofullerenes, including NMR and mass spectra of new compounds; details on the microarray evaluation of mannofullerenes as inhibitors of ConA; and ITC analysis of ConA− mannofullerene interactions (PDF)

AUTHOR INFORMATION

Corresponding Authors

*(S.P.V.) E-mail: [email protected]. *(D.S.) E-mail: [email protected]. ORCID

Dolores Solís: 0000-0002-8148-1875 Present Address

M.A.-F.: Matière Molle et Chimie, É cole Supérieure de Physique et de Chimie Industrielles de la Ville de Paris (ESPCI)−CNRS, UMR-7167, Paris Sciences et Lettres (PSL) Research University, 10 Rue Vauquelin, 75005 Paris, France.

#

G


Article

Analytical Chemistry (15) Hartmann, M.; Papavlassopoulos, H.; Chandrasekaran, V.; Grabosch, C.; Beiroth, F.; Lindhorst, T. K.; Röhl, C. FEBS Lett. 2012, 586, 1459−1465. (16) de Ruyck, J.; Lensink, M. F.; Bouckaert, J. IUCrJ 2016, 3, 163− 167. (17) Mydock-McGrane, L. K.; Hannan, T. J.; Janetka, J. W. Expert Opin. Drug Discovery 2017, 12, 711−731. (18) Sattin, S.; Bernardi, A. Trends Biotechnol. 2016, 34, 483−495. (19) 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, P. V.; Nativi, C.; Oscarson, S.; Penades, S.; Peri, F.; Pieters, R. J.; et al. Chem. Soc. Rev. 2013, 42, 4709−4727. (20) Abraham, J. M.; Freitag, C. S.; Clements, J. R.; Eisenstein, B. I. Proc. Natl. Acad. Sci. U. S. A. 1985, 82, 5724−5727. (21) Schwan, W. R. World J. Clin. Infect. Dis. 2011, 1, 17−25. (22) Gally, D. L.; Bogan, J. A.; Eisenstein, B. I.; Blomfield, I. C. J. Bacteriol. 1993, 175, 6186−6193. (23) Dove, S. L.; Smith, S. G. J.; Dorman, C. J. Mol. Gen. Genet. 1997, 254, 13−20. (24) Bouckaert, J.; Mackenzie, J.; de Paz, J. L.; Chipwaza, B.; Choudhury, D.; Zavialov, A.; Mannerstedt, K.; Anderson, J.; Piérard, D.; Wyns, L.; Seeberger, P. H.; Oscarson, S.; De Greve, H.; Knight, S. D. Mol. Microbiol. 2006, 61, 1556−1568. (25) Hartmann, M.; Horst, A. K.; Klemm, P.; Lindhorst, T. K. Chem. Commun. 2010, 46, 330−332. (26) Park, S.; Lee, M.-R.; Shin, I. Bioconjugate Chem. 2009, 20, 155− 162. (27) Lee, M.-r.; Shin, I. Org. Lett. 2005, 7, 4269−4272. (28) Disney, M. D.; Seeberger, P. H. Chem. Biol. 2004, 11, 1701− 1707. (29) Beaussart, A.; Abellán-Flos, M.; El-Kirat-Chatel, S.; Vincent, S. P.; Dufrêne, Y. F. Nano Lett. 2016, 16, 1299−1307. (30) Abellán Flos, M.; García Moreno, M. I.; Ortiz Mellet, C.; García Fernández, J. M.; Nierengarten, J.-F.; Vincent, S. P. Chem. Eur. J. 2016, 22, 11450−11460. (31) Abellan-Flos, M.; Tanc, M.; Supuran, C. T.; Vincent, S. P. Org. Biomol. Chem. 2015, 13, 7445−7451. (32) Buffet, K.; Gillon, E.; Holler, M.; Nierengarten, J.-F.; Imberty, A.; Vincent, S. P. Org. Biomol. Chem. 2015, 13, 6482−6492. (33) Nierengarten, J.-F.; Iehl, J.; Oerthel, V.; Holler, M.; Illescas, B. M.; Munoz, A.; Martin, N.; Rojo, J.; Sanchez-Navarro, M.; Cecioni, S.; Vidal, S.; Buffet, K.; Durka, M.; Vincent, S. P. Chem. Commun. 2010, 46, 3860−3862. (34) Hultgren, S. J.; Schwan, W. R.; Schaeffer, A. J.; Duncan, J. L. Infect. Immun. 1986, 54, 613−620. (35) Munera, D.; Hultgren, S.; Fernández, L. Á . Mol. Microbiol. 2007, 64, 333−346. (36) Dubois, M.; Gilles, K.; Hamilton, J. K.; Rebers, P. A.; Smith, F. Nature 1951, 168, 167. (37) Cavallone, D.; Malagolini, N.; Monti, A.; Wu, X.-R.; SerafiniCessi, F. J. Biol. Chem. 2004, 279, 216−222. (38) Alvarez, R. A.; Blixt, O. Identification of Ligand Specificities for Glycan-Binding Proteins Using Glycan Arrays. In Methods in Enzymology; Academic Press, 2006; Vol. 415, pp 292−310. (39) Manzano, A. I.; Javier Cañada, F.; Cases, B.; Sirvent, S.; Soria, I.; Palomares, O.; Fernández-Caldas, E.; Casanovas, M.; JiménezBarbero, J.; Subiza, J. L. Glycoconjugate J. 2016, 33, 93−101. (40) Migneault, I.; Dartiguenave, C.; Bertrand, M. J.; Waldron, K. C. BioTechniques 2004, 37, 790−802. (41) Mulvey, M. A.; Schilling, J. D.; Hultgren, S. J. Infect. Immun. 2001, 69, 4572−4579. (42) Guyer, M. S.; Reed, R. R.; Steitz, J. A.; Low, K. B. Cold Spring Harbor Symp. Quant. Biol. 1981, 45, 135−140. (43) Blomfjeld, I. C.; McClain, M. S.; Eisenstein, B. I. Mol. Microbiol. 1991, 5, 1439−1445. (44) Lundquist, J. J.; Toone, E. J. Chem. Rev. 2002, 102, 555−578. (45) Jayaraman, N. Chem. Soc. Rev. 2009, 38, 3463−3483.

(46) Almant, M.; Moreau, V.; Kovensky, J.; Bouckaert, J.; Gouin, S. G. Chem. - Eur. J. 2011, 17, 10029−10038. (47) Bouckaert, J.; Li, Z.; Xavier, C.; Almant, M.; Caveliers, V.; Lahoutte, T.; Weeks, S. D.; Kovensky, J.; Gouin, S. G. Chem. - Eur. J. 2013, 19, 7847−7855. (48) Iehl, J.; Nierengarten, J.-F. Chem. - Eur. J. 2009, 15, 7306−7309. (49) Gupta, D.; Oscarson, S.; Raju, T. S.; Stanley, P.; Toone, E. J.; Brewer, C. F. Eur. J. Biochem. 1996, 242, 320−326. (50) González, L.; Bruix, M.; Díaz-Mauriño, T.; Feizi, T.; Rico, M.; Solís, D.; Jiménez-Barbero, J. Arch. Biochem. Biophys. 2000, 383, 17− 27. (51) Bouckaert, J.; Berglund, J.; Schembri, M.; De Genst, E.; Cools, L.; Wuhrer, M.; Hung, C.-S.; Pinkner, J.; Slättegård, R.; Zavialov, A.; Choudhury, D.; Langermann, S.; Hultgren, S. J.; Wyns, L.; Klemm, P.; Oscarson, S.; Knight, S. D.; De Greve, H. Mol. Microbiol. 2005, 55, 441−455. (52) Le Trong, I.; Aprikian, P.; Kidd, B. A.; Forero-Shelton, M.; Tchesnokova, V.; Rajagopal, P.; Rodriguez, V.; Interlandi, G.; Klevit, R.; Vogel, V.; Stenkamp, R. E.; Sokurenko, E. V.; Thomas, W. E. Cell 2010, 141, 645−655. (53) Gerland, B.; Goudot, A.; Ligeour, C.; Pourceau, G.; Meyer, A.; Vidal, S.; Gehin, T.; Vidal, O.; Souteyrand, E.; Vasseur, J.-J.; Chevolot, Y.; Morvan, F. Bioconjugate Chem. 2014, 25, 379−392. (54) Hoang, A.; Laigre, E.; Goyard, D.; Defrancq, E.; Vinet, F.; Dumy, P.; Renaudet, O. Org. Biomol. Chem. 2017, 15, 5135−5139.

H


Direct Evaluation of Live Uropathogenic Escherichia coli Adhesion

Recommend Documents