Article pubs.acs.org/ac

Carbohydrate Microarray for the Detection of Glycan−Protein Interactions Using Metal-Enhanced Fluorescence Jie Yang,*,† Anne Moraillon,† Aloysius Siriwardena,‡ Rabah Boukherroub,§ François Ozanam,† Anne Chantal Gouget-Laemmel,*,† and Sabine Szunerits*,§ †

Physique de la Matière Condensée, Ecole Polytechnique-CNRS, 91128 Palaiseau, France Laboratoire de Glycochimie des Antimicrobiens et des Agroressources (LG2A), (FRE 3517-CNRS), Université de Picardie Jules Verne, 33 Rue St Leu, 80039 Amiens, France § Institut d’Electronique, de Microélectronique et de Nanotechnologie (IEMN, CNRS-8520), Université Lille 1, Cité Scientifique, Avenue Poincaré B.P. 60069, 59652 Villeneuve d’Ascq, France ‡

S Supporting Information *

ABSTRACT: Carbohydrate arrays are potentially one of the most attractive tools to study carbohydrate-based interactions. This paper describes a new analytical platform that exploits metalenhanced fluorescence for the sensitive and selective screening of carbohydrate-lectin interactions. The chip consists of a glass slide covered with gold nanostructures, postcoated with a thin layer of amorphous silicon−carbon alloy (a-Si0.8C0.2:H). An immobilization strategy based on the formation of a covalent bond between propargyl-terminated glycans and surface-linked azide groups was used to attach various glycans at varying surface densities onto the interface and to fabricate a carbohydrate array via efficient local “click” chemistry strategy. The specific association of the new interface with fluorescently labeled lectins was assessed by fluorescence imaging and an excellent selectivity to specific proteins was achieved. Optimization of the surface architecture and the plasmonic transducer resulted in an enhancement of the fluorescence intensity by 1 order of magnitude, when compared to the corresponding substrate devoid of gold nanostructures. The limit of detection (LOD) of such microarrays is in the picomolar range, making it a promising system for development in pharmaceutical or biomedical applications.

A

(glass, ITO).34−36 We have shown that postcoating surfacebound gold nanostructures with dielectric thin films results not only in chemically and optically stable LSPR biosensors,37−41 but also allows modulation of its molecular fluorescence characteristics.42 The level of metal enhanced fluorescence (MEF) depends on a variety of different parameters such as the size, shape and resonance wavelength of the nanoparticles, the absorption/emission wavelength of the fluorescent molecules but also the distance between the fluorophores and the metallic nanoparticles.43−46 We have shown that a practical strategy for achieving optimal conditions for enhanced fluorescence on interfaces is to use a dielectric spacer based on hydrogenated amorphous silicon carbon alloy (a-Si1−xCx:H) above a metalnanoparticle layer.42 This amorphous spacer can be easily deposited as a thin-film by plasma-enhanced chemical vapor deposition (PECVD) in low-power regime and functionalized through robust Si−C covalent bonds.47,48 Furthermore, the optical properties of a-Si1−xCx:H can be adjusted by controlling the carbon content x and the thickness to minimize absorption in the visible range.49,50 For example, coating a random metal nanoparticles assembly with an amorphous silicon−carbon

dvances in biology continue to reveal that carbohydrate− protein interactions play an essential role in the development and maintenance of living systems as they are fundamental in many cellular processes such as cell adhesion, viral/bacterial infection, autoimmunity or tissue growth and engineering.1−3 Progress in oligosaccharide synthesis4,5 together with the advances in methods for engineering glycanmodified surfaces6−12 have accelerated the development of tools for the study of protein-carbohydrate interactions. The use of glycan arrays is of special interest as it allows the monitoring of binding events between surface-immobilized carbohydrates and protein in solution in a high-throughput manner.13,14 Despite the central place of fluorescence,15−17 and propagating surface plasmon resonance (SPR),18−22 as the major techniques for detecting carbohydrate−protein interactions, more effective alternatives have been sorted.23−26 For example, localized surface plasmon resonance (LSPR) has recently proven its utility for the characterization of carbohydrate−protein interactions.27,28 When surface plasmons are confined within nanostructures, localized optical modes are possible, and lead to highly localized electromagnetic fields around the nanostructures.29,30 Indeed much effort has been invested in finding the optimal metal nanostructure for the development of sensitive LSPR sensors.31−33 One of the common experimental schemes for LSPR sensing on surfaces is to immobilize metal nanostructures onto transparent interfaces © XXXX American Chemical Society

Received: November 10, 2014 Accepted: March 2, 2015

A

DOI: 10.1021/ac504262b Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 1. Functionalization scheme for the immobilization of glycans on metal-enhanced fluorescence structure, consisting of a glass slide covered with Au NSs postcoated with a-Si0.8C0.2:H layer.

conjugated Alexa Fluor 647 dye were purchased from Molecular Probes Invitrogen. All other chemicals of the highest available quality were purchased from Sigma-Aldrich. Ultrapure water (Milli-Q, 18 MΩ cm) was used for the preparation of the solutions and for all rinses. The synthesis of α-propargyl mannoside is described in reference.51 The synthesis and NMR characterization of β-propargyl lactoside is described in the Supporting Information S1. Preparation of Glass Substrates Covered with Gold Nanostructures (Au NSs/Glass). Glass slides (75 × 25 × 1 mm3) were first cleaned in ethanol, rinsed in Milli-Q water then cleaned in piranha solution (H2SO4/H2O2 = 3/1 v/v) at room temperature, rinsed copiously with Milli-Q water and dried under a stream of nitrogen. The clean substrates were then transferred into an evaporation chamber. Gold nanostructures (Au NSs) deposition was carried out by thermal evaporation of a 4 nm thick gold film using a MEB 550 S evaporation machine (Plassys, France), followed by thermal annealing. Postdeposition annealing of the gold-covered slides was carried out at 500 °C for 1 min under nitrogen atmosphere using a rapid thermal annealer (Jipelec Jet First 100). The reproducibility of the metal evaporation was evaluated by measuring the LSPR signals of a batch of 8 samples. The standard deviation in the wavelength (λmax) and maximum absorption (Imax) is typically 2 nm and 0.02 abs units, respectively. Coating with Amorphous Silicon−Carbon Alloy. Amorphous silicon−carbon alloy (a-Si0.8C0.2:H) layers were deposited onto Au NSs/glass, bare glass or onto a crystalline silicon prism using plasma-enhanced chemical vapor deposition (PECVD) in a “low-power” regime (plasma = 13.56 MHz and 100 mW cm−2, substrate temperature = 250 °C, pressure = 35 mTorr). The gas flow rates were 6.3 sccm SiH4 and 27 sccm CH4. The layer thickness was adjusted by controlling the deposition time (1.2 μm h−1). Surface Functionalization. Formation of CarboxydecylTerminated Surfaces. The a-Si0.8C0.2:H-coated Au NSs/glass

alloy containing 20% of carbon (a-Si0.8C0.2:H) of about 5 nm in thickness followed by covalent attachment of oligonucleotides resulted in an ultrasensitive MEF-LSPR response, allowing for designing DNA biosensor with a detection limit of DNA hybridization in the low femtomolar range.42 Motivated by these results, we investigate in this paper the utility of metal nanoparticles enhanced fluorescence (MEF) in the characterization of carbohydrate-lectin interactions in an array format. To illustrate our approach, the specific recognition between two fluorescently labeled lectins, concanavalin A (ConA) and peanut agglutinin (PNA) with mannose- and lactose-modified LSPR interface was selected for study. The glycan array was formed on amorphous silicon− carbon (a-Si0.8C0.2:H) coated gold nanostructures (Au NSs) through a multistep reaction as shown in Figure 1. Acid functional groups are introduced onto the surface by photochemical hydrosilylation of undecylenic acid onto hydrogenated a-Si0.8C0.2:H thin films, followed by amidation with glutamic acid (Glu) to effectively double the density of acid functions. Further derivatization with oligo(ethylene glycol) (OEG) chains carrying terminal azido functions serves to limit nonspecific protein adsorption and also allows efficient bioconjugation of propargyl-terminated glycan moieties via copper(I) catalyzed “click” reaction.51 Optimization of the MEF sensitivity was possible through varying the thickness of the a-Si0.8C0.2:H coating. The practical potential of the glycan microarray thus fabricated was demonstrated in a thermodynamic study.

■

EXPERIMENTAL SECTION Materials. All cleaning reagents (H2O2, 30%; H2SO4, 96%), and etching (HF, 50%) reagents were of VLSI grade and supplied by Carlo Erba. Undecylenic acid (99%) was purchased from Acros Organics and 10X PBS (phosphate buffer saline) buffer pH 7.4 from Ambion. Alexa Fluor 647 labeled concanavalin A (ConA), peanut agglutinin (PNA), and NHSB

DOI: 10.1021/ac504262b Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 2. (A) AFM images of a glass slide covered with Au NSs formed through thermal annealing of a 4 nm-thick gold film before (a) and after coating with 3 (b), 5 (c), and 8 nm (d) of a-Si0.8C0.2:H layer; (B) UV−vis absorption spectra of Au NSs/glass coated with a-Si0.8C0.2:H of various thicknesses.

buffer. After spotting, the slides were stored in a desiccator overnight at controlled humidity (∼75%), then rinsed in 1× PBS/0.1% SDS for 10 min and quickly in Milli-Q water and dried in a stream of N2. Interaction with Lectins. The glycan array was incubated in Alexa Fluor 647-labeled PNA or ConA solutions at various concentrations in 1× PBS containing 0.005% Tween 20 for 1 h in a locked hybridization chamber. The array was thereafter washed with 1× PBS/0.1% SDS (10 min), 0.2× PBS (2 min), and 0.1× PBS (2 min) and finally with deionized water before being dried under a stream of argon prior to analysis. For the isotherm construction, the spotting of ten replicate zones was performed on the same slide. Each replicate zone contains labeled control spots as an internal reference to eliminate local inhomogeneities on the slide. A first incubation of ConA was done simultaneously at 10 concentrations, followed by a second incubation of PNA at the same concentrations. Instrumentation. Contact Mode-Atomic Force Microscopy (CM-AFM). AFM images were obtained using a Pico SPM microscope (Molecular Imaging, Phoenix, AZ) in contact mode with silicon nitride cantilevers (Nanoprobe, spring constant = 0.12 N m−1) under a N2 atmosphere. UV−vis Spectrometry. Absorption spectra were recorded with a Cary 50 Scan UV/vis spectrophotometer. The wavelength range was 350−800 nm. IR-ATR. The IR spectra were recorded with s and ppolarization over the 900−400 cm−1 spectral range using a Bruker Equinox FTIR spectrometer coupled to a homemade, nitrogen-gas purged external ATR compartment (100 scans, 4 cm−1 resolution). The ATR element is a crystalline silicon prism prepared from a double-side polished float-zone purified, 30−40 Ω cm, n-type (111)-oriented silicon crystal (Siltronix). A typical prism has a dimension of 15 × 15 × 0.5 mm3, the two opposite sides being 50° beveled, providing ∼22 reflections. Spotter. The deposition of a few nanoliters of a “click” mixture containing glycan analogs was performed using a pin spotter from Biorobotics Microgrid II in a controlled environment (T ≈ 20 °C and humidity ≈ 50%).

slide was exposed to HF vapor for 15 s before placing it into a Schlenk tube containing previously degassed neat undecylenic acid and irradiated at 312 nm (6 mW cm−2) for 3 h. After a final rinse in hot acetic acid (75 °C) for 40 min,52 the sample was dried under nitrogen flow. Formation of NHS Ester-Terminated Surfaces. The acidterminated slide was immersed in 60 mL of an aqueous solution of N-hydroxysuccinimide, NHS (5 mM), and N-(3(dimethylamino)propyl)-N′-carbodiimide, EDC (5 mM) and allowed to react for 90 min at 15 °C.53 The resulting surface was copiously rinsed with Milli-Q water and dried under a stream of nitrogen. Formation of Glutamic Acid-Terminated Surfaces. The NHS ester-activated slide was immersed in 20 mM of glutamic acid (Glu) in 1× PBS (V = 20 mL) at pH ∼8 for 3 h at room temperature. The resulting surface was rinsed with 1× PBS/ 0.1% SDS for 10 min and copiously with Milli-Q water. The Glu-terminated surface was dried under a stream of nitrogen. Formation of Azido-Terminated Surfaces. The Gluterminated surface was activated again for the generation of NHS-ester function according to the above-described procedure. Then, the activated slide was immersed in 20 mM of NH2−C2H4−EG8−N3 in 1× PBS (V = 20 mL) at pH ∼8 for 4 h at room temperature. The resulting surface was copiously rinsed with 1X PBS, followed by a surfactinated rinse (1× PBS/ 0.1% SDS for 15 min; 0.2× PBS for 5 min; 0.1× PBS for 5 min) and finally with Milli-Q water.54 The azido-OEG surface was dried under a stream of argon. Formation of Glycan-Modified Surfaces. Propargyl glycans were “spotted” onto the azide-terminated slide from mixture of respective propargyl-glycan (3 mM), sodium ascorbate (15 mol %), and copper sulfate (CuSO4) (5 mol %) in the spotting buffer composed of 0.3 M phosphate (Na2HPO4/NaH2PO4)/ 0.005% Tween 20/0.001% Sarkosyl. Different mixtures of propargyl glycans diluted in propargyl alcohol at 30, 10, and 3 mol % were spotted in the same conditions, the total concentration of propargyl functions being kept at 3 mM. The spotting of the propargyl-conjugated control molecules was achieved using propargyl-conjugated Alexa 647 (7.7 mM), sodium ascorbate (101 μM), CuSO4 (4 μM) in the spotting C

DOI: 10.1021/ac504262b Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry Fluorescence measurements. For end-point measurements, the fluorescence was measured at different exposure times with a Diagarray scanner (Genewave, France) using a CCD camera and a 16-bit. The values of fluorescence intensity in the histogram correspond to the difference of the measured intensity averaged over the spot and that of the fluorescence background recorded in the vicinity of the spot.

■

RESULTS AND DISCUSSION Assembly of Carbohydrates on LSPR Interfaces. Figure 1 shows a schematic presentation of the formation of the sensor architecture used in this work. It consists of a random assembly of gold nanostructures (Au NSs) coated with a-Si0.8C0.2:H film of different thicknesses (0−12 nm). The AFM image in Figure 2Aa reveals the height of these Au NSs to be ∼13 nm and their average diameter to be ∼30 nm before coating. Coating the Au NSs with a-Si0.8C0.2:H of various thicknesses preserves the topography of the Au NSs (Figure 2Ab-c). On the contrary, the thickness of a-Si0.8C0.2:H strongly affects the LSPR response of the Au NSs. The absorbance spectra of the glass/Au NSs before and after coating with a-Si0.8C0.2:H thin films are displayed in Figure 2B. Coating the Au NSs with a-Si0.8C0.2:H results in a significant red shift of the plasmon band from 538 nm (Au NSs only) to 598 nm with an increase in the peak full width at halfmaximum (fwhm) from 70 nm (Au NSs only) to 124 nm due to the change of the refractive index surrounding the nanostructures, which becomes larger as the thickness increases (see Table 1).

Figure 3. FTIR-ATR spectra in p-polarization of the acid- (a), NHS ester- (b), Glu- (c), Glu-NHS ester- (d), N3- (e), and mannoseterminated surfaces (f) on a-Si0.8C0.2:H coated silicon wafer. The reference spectrum is the hydrogen-terminated surface.

analysis of the carbonyl band, we obtain a density of acid chains of 7.2 × 1013 cm−2, corresponding to about ∼1/3 of that grafted on a crystalline silicon surface.51 The lower functionalization yield is correlated with the presence of methyl groups on the a-Si0.8C0.2:H surface, as observed previously.38,55,56 In order to increase the surface density of OEG segments and improve the resistance against nonspecific adsorption, we increased the surface density of the acid chains by a two-step amidation process using glutamic acid (Glu). The carboxylic acid terminated a-Si0.8C0.2:H overcoating was activated with carbodiimide (EDC) in the presence of N-hydroxysuccinimide (NHS) to form NHS-ester function. The IR spectrum clearly shows the three peaks at 1734, 1784, and 1811 cm−1 related to the stretching modes of the NHS-ester (Figure 3b). The aminolysis reaction with Glu results in the disappearance of the triplet band and the appearance of the amide I vibration band at 1650 cm−1 together with a broad carbonyl band related to the acid function of Glu (Figure 3c). The quantitative analysis of the activated and amidated surfaces leads to a surface concentration of 5 × 1013 and of 1 × 1014 cm−2, respectively, indicating that the density of acid functional groups further increased by ∼50% over the initial grafting step. The Gluterminated monolayer was then modified with the amino-OEG precursor, NH2−C2H4-EG8-N3 via EDC/NHS coupling to form azide-terminated monolayer. From the IR-ATR spectrum shown in Figure 3e, the amidation was evidenced by the increase of the amide I and II vibration bands. By comparing the peak of initial Glu moiety with that of unreacted acid function, the amidation yield can be assessed, with a value of 60% corresponding to a density of OEG chains of 6 × 1013 cm−2. The final step was “clicking” propargyl-glycan, in this case propargyl-mannoside, in the presence of Cu(I) catalyst generated in situ from sodium ascorbate and CuSO4. The presence of mannosyl units is revealed in Figure 3f by the increase in the absorption bands related to the νC−O and νC−C at 1030−1130 cm−1. Since the azide density on a-Si0.8C0.2:H is much lower than that on crystalline silicon surface,51 there is less steric hindrance among mannoside molecules favoring a “click” yield approaching unity. Therefore, we estimate the

Table 1. λmax and FWHM of the Absorbance Spectra of Au NSs/Glass Coated with a-Si0.8C0.2:H of Various Thicknesses thickness of a-Si0.8C0.2:H (nm)

λmax (nm)

fwhm (nm)

0 3 5 8 12

538 561 572 595 598

70 89 99 120 124

The LSPR transducer was functionalized with glycan ligands using the Cu(I)-catalyzed “click” reaction as schematically outlined in Figure 1. This strategy has been recently established on a crystalline silicon surface.51 The long and densely packed surface intercalated oligo(ethylene glycol) chains with a density of 1.6 × 1014 cm−2 has proven to limit undesired adsorption of nonspecific lectins, and the “click” strategy allowed the anchoring of glycans with a “click” yield of 75% limited probably by steric hindrance. To demonstrate the reliability of this functionalization strategy on a-Si0.8C0.2:H, quantitative FTIR-ATR measurements were performed on silicon prisms coated with a 40 nm-thick a-Si0.8C0.2:H layer (Figure 3). It means that absorbance data presented in Figure 3 are computed using a control blank amorphous silicon carbon coated Au NSs substrate. The reference spectrum is the hydrogenated a-Si0.8C0.2:H surface, which accounts for the large negative band at ∼2080 cm−1 corresponding to the stretching modes of the Si−H vibration on surface and on the bulk. The first step of the modification is the hydrosilylation reaction of undecylenic acid with hydrogenated a-Si0.8C0.2:H. The IR-ATR spectrum (Figure 3a) confirms the success of the grafting of carboxydecyl chains via Si−C covalent bonds by the presence of the νCO of the acid function at 1716 cm−1 and the two νCH2 of the alkyl chains at 2857 and 2930 cm−1. From a quantitative D

DOI: 10.1021/ac504262b Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 4. (A) Fluorescence images of the spotted array after interaction with ConA (0.9 mg/mL) (a) and PNA (0.9 mg/mL) (b), followed by a rinse with 1× PBS/0.1% SDS. The substrate is N3-terminated Au NSs/glass slide coated with 5 nm a-Si0.8C0.2:H; (B) Histogram of fluorescence intensity for diluted glycan spots.

Figure 5. (a) Histogram of fluorescence intensity of spotted 100% mannose and 100% lactose arrays after interaction with ConA and PNA (90 μg/ mL) followed by a rinse with 1× PBS/0.1% SDS on different substrates. The fluorescence exposure time was 2 s. The inset fluorescence images and spotting schemes are presented accordingly. (b) Histogram of fluorescence intensity of spotted 100% mannose and 100% lactose on a 3 nm thick aSi0.8C0.2:H/Au NSs array after interaction with ConA and PNA at various concentrations followed by PBS rinse. The fluorescence exposure time was 5 s. The inset fluorescence images and spotting schemes are presented accordingly.

density of “clicked” mannoside on a-Si0.8C0.2:H to be higher than 5 × 1013 cm−2. Generation of Glycan Arrays and Detection of Glycan−Lectin Interactions. In the case of a-Si0.8C0.2:H coated LSPR transducer interfaces, the immobilization of alkynyl-terminated glycans was performed locally by spotting of the respective propargyl-modified glycans in the “click” solution. The most striking advantage of this approach is that a glycan array can be built up in an easy manner where not only different glycans can be immobilized, but mixed glycan layers can be formed and their density can be controlled at will. As with other array-based technologies, the effect of the density of attached glycans, and thus glycan accessibility, on microarray performance is of utter importance in determining the strength and selectivity of its interaction with proteins.18,57,58 In our case, two propargyl glycans, mannoside and lactoside, were spotted at varying dilutions (100, 30, 10, 3 mol %) using propargyl alcohol as an inert diluting partner. Figure 4A shows the fluorescence images of the glycan array after incubation

with respectively, Alexa647-labeled ConA (specific to mannose) and PNA (specific to lactose) in two distinct areas (a and b). As expected, interaction of the glycans with specific lectins results in a clear fluorescence signal, whereas no fluorescent background was observed with nonbinding controls, indicating the excellent specificity of the glycan array and its marked antifouling capacity. We also notice that the size of the spots decreases as the sugar concentration is reduced. This is plausibly due to the hydrophilicity or the potential surfactant effect of the polyhydroxyl structured mannose and lactose. A higher concentration of sugar might be expected to lower the evaporation rate of buffer, thus leading to a larger spot size. Figure 4B depicts the obtained fluorescence intensity using different glycan densities. For both lectins, the binding intensity is optimum for the 100 mol % glycan surfaces. However, the binding efficiency is affected by the surface glycan density. The binding response of PNA is less sensitive from 30 to 3 mol % lactose concentration with an equivalent efficiency level, whereas the binding response of ConA decreases with the E

DOI: 10.1021/ac504262b Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

molecule, whereas ConA features 3 dye units per molecule. The picomolar LOD of ConA-mannose binding is currently one of the lowest reported, and comparable with the picomolar levels observed for fluorescence microarrays (see Supporting Information Table S1).64 For comparison, the reported limits of detection of ConA-mannose binding include those using a quartz crystal microbalance (∼1.3 nM),24,65,66 SPR (∼1.5 nM)9,21,67 cantilever deflection (∼9.6 nM),68 and electrochemical techniques, such as voltammetry (7 nM)69 and impedance spectroscopy (5 nM).70 The high sensitivity of the MEF-based carbohydrate array described here is thus seen to hold great promise. To determine the interaction strength between lectins and surface-linked glycans, adsorption isotherms for both test lectins were measured. The binding assay was performed on one slide comprised of ten replicate zones, which allows the simultaneous readout of the intensity values of either ConA or PNA solutions at 10 different concentrations. The data can be fitted to Langmuir isotherms as shown in Figure 6 for the 100%

mannose density. This phenomenon is probably related to the increase expected from multivalent binding.21,18,57 On crystalline silicon, we demonstrate that the optimum conditions for multivalent interactions is a compromise between the surface concentration of glycan ligands in terms of density and spacing.59 This effect gives rise to a maximum efficiency in the concentration range 3−9 × 1013 cm−2 for bivalent interaction of mannose with Lens culinaris lectin. Such an order of magnitude is consistent with the behavior recorded here for ConA, which exhibits a similar dimeric structure to that of Lens culinaris lectin. Optimization of the Metal Nanostructures-Enhanced Fluorescence Response. The enhancement of the fluorescence signal of a given fluorophore (in our case Alexa 647 with λexcitation= 650 nm and λemission = 665 nm) by the LSPR transducer is strongly determined by the distance of metalfluorophore as well as by the position of the plasmonic band. To optimize this response, a-Si0.8C0.2:H coatings of varying thicknesses were deposited on Au NSs/glass substrates and the enhancement factor was evaluated by comparing with bare slide without Au NSs coated with a-Si0.8C0.2:H. The bare slide with an a-Si0.8C0.2:H optimized thickness of 124 nm shows optical properties equivalent to conventional glass slide while keeping the same surface chemistry modification based on Si−C bonds.42,60 Figure 5a shows the resulting fluorescence images on different interfaces. The highest fluorescence signal occurs with a 3 nm-thick a-Si0.8C0.2:H layer, which is 1 order of magnitude higher than that recorded on a-Si0.8C0.2:H interfaces (124 nm) for both interactions. Considering that the thickness of carboxydecyl chain is ∼1.2 nm,61 the length of OEG chain is 2.5 nm (2.74 Å for each ethylene glycol unit)62 and the distance from the triazole to the terminal hydroxyl group of mannose is ∼0.8 nm, the total thickness of the functionalized monolayer is estimated roughly to be 4.5 nm. By adding the 3 nm of the amorphous layer, the actual distance from the fluorophore to the Au NSs is ∼7.5 nm, depending on the position of dye molecule labeled on the protein. From the dimension of ConA (7 × 7 × 5 nm) and PNA (6.5 × 6.5 × 3.7 nm) (protein databank), we are able to estimate that this distance is around 10 nm, which is in accordance with the theoretical optimal coupling,63 even though the LSPR position of the 3 nm coating (561 nm) is somehow far from the excitation wavelength of the fluorophore (650 nm). The best overlap between the excitation wavelength of the dye and the LSPR interface is that at 12 nm. However, the large fwhm and the large distance of fluorophoreAu NSs which seems to be out of the MEF enhancement window could explain its less efficient LSPR coupling effect. The sensitivity of this protein biosensor can be alternatively expressed by the limit of detection (LOD). Figure 5b shows the diagram of fluorescence intensity of the microarray upon interaction with various concentrations of ConA and PNA. The lowest distinguished level is 4.5 ng/mL (37.5 pM) for ConA (MW = 120 kDa) and 90 ng/mL (870 pM) for PNA (MW = 103 kDa), indicating that the LOD is in the picomolar range. The fluorescence response as a function of the analyte concentration (Supporting Information Figure S1) is linear at least up to one-fourth of the dynamic range of the detection system in the conditions used in the present study. We found that the linear range is 37.5 pM−75 nM (4.5 ng/mL−9 μg/ mL) for ConA and 870 pM−870 nM (90 ng/mL−90 μg/mL) for PNA. The lower sensitivity of PNA might be related to the fact that the PNA employed carries only 2 dye moieties per

Figure 6. Binding isotherm of 100% mannose-ConA and 100% lactose-PNA measured by the optimized MEF-based microarray. The data is fitted to the Langmuir model. The fluorescence exposure time was 5 s. The inset table is the summary of the binding KD values for diluted surface glycan concentrations.

mannose-ConA and 100% lactose-PNA. Dissociation constants KD in the order of micromolar for the two interactions are determined. The KD values for the diluted glycan surfaces are also summarized in the inset table (Figure 6). In all cases, the order of KD at the level of μM indicates that these interactions are plausibly multivalent as the affinity of a monovalent binding event is reported to be 2 orders of magnitude lower,23,71 even when the surface glycan concentration is as low as 1 mol %. As both ConA and PNA are tetrameric proteins, the binding of either with appropriate surface glycans could be at least a bivalent interaction.

■

CONCLUSION We take advantage of the coupling between LSPR (Au NSs) and fluorescence to fabricate sensitive carbohydrate microarrays F

DOI: 10.1021/ac504262b Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

(14) Song, E.-H.; Pohl, N. L. B. Curr. Opin. Chem. Biol. 2009, 13, 626−632. (15) Liang, P.-H.; Wang, S.-K.; Wong, C.-H. J. Am. Chem. Soc. 2007, 129, 11177−11184. (16) Park, S.; Lee, M.-R.; Pyo, S.-J.; Shin, I. J. Am. Chem. Soc. 2004, 126, 4812−4819. (17) Wang, X.; Matei, E.; Deng, L.; Koharudin, L.; Gronenborn, A. M.; Ramström, O.; Yan, M. Biosens. Bioelectron. 2013, 47, 258−264. (18) Dhayal, M.; Ratner, D. M. Langmuir 2009, 25, 2181−2187. (19) Penezic, A.; Deokar, G.; Vignaud, D.; Pichonat, E.; Happy, H.; Subramanian, P.; Gasparović, B.; Boukherroub, R.; Szunerits, S. Plasmonics 2014, 9, 677−683. (20) Safina, G. Anal. Chim. Acta 2012, 712, 9−29. (21) Smith, E. A.; Thomas, W. D.; Kiessling, L. L.; Corn, R. M. J. Am. Chem. Soc. 2003, 125, 6140−6148. (22) Tyagi, A.; Wang, X.; Deng, L.; Ramström, O.; Yan, M. Biosens. Bioelectron. 2010, 26, 344−350. (23) Kussrow, A.; Kaltgrad, E.; Wolfenden, M. L.; Cloninger, M. J.; Finn, M. G.; Bornhop, D. J. Anal. Chem. 2009, 81, 4889−4897. (24) Lyu, Y.-K.; Lim, K.-R.; Lee, B. Y.; Kim, K. S.; Lee, W.-Y. Chem. Commun. 2008, 4771−4773. (25) Min, I.-H.; Choi, L.; Ahn, K.-S.; Kim, B. K.; Lee, B. Y.; Kim, K. S.; Choi, H. N.; Lee, W.-Y. Biosens. Bioelectron. 2010, 26, 1326−1331. (26) Zhang, G.-J.; Huang, M. J.; Ang, J. A. J.; Yao, Q.; Ning, Y. Anal. Chem. 2013, 85, 4392−4397. (27) Bellapadrona, G.; Tesler, A. B.; Grünstein, D.; Hossain, L. H.; Kikkeri, R.; Seeberger, P. H.; Vaskevich, A.; Rubinstein, I. Anal. Chem. 2011, 84, 232−240. (28) Yonzon, C. R.; Jeoung, E.; Zou, S.; Schatz, G. C.; Mrksich, M.; Van Duyne, R. P. J. Am. Chem. Soc. 2004, 126, 12669−12676. (29) Mayer, K. M.; Hafner, J. H. Chem. Rev. 2011, 111, 3828−3857. (30) Stewart, M. E.; Anderton, C. R.; Thompson, L. B.; Maria, J.; Gray, S. K.; Rogers, J. A.; Nuzzo, R. G. Chem. Rev. 2008, 108, 494− 521. (31) Dahlin, A.; Zäch, M.; Rindzevicius, T.; Käll, M.; Sutherland, D. S.; Höök, F. J. Am. Chem. Soc. 2005, 127, 5043−5048. (32) Guo, L.; Kim, D.-H. Biosens. Bioelectron. 2012, 31, 567−570. (33) Jain, P. K.; Huang, X.; El-Sayed, I.; El-Sayed, M. Plasmonics 2007, 2, 107−118. (34) Bendikov, T. A.; Rabinkov, A.; Karakouz, T.; Vaskevich, A.; Rubinstein, I. Anal. Chem. 2008, 80, 7487−7498. (35) Szunerits, S.; Praig, V. G; Manesse, M.; Boukherroub, R. Nanotechnology 2008, 19, 195712. (36) Hall, W. P.; Ngatia, S. N.; Van Duyne, R. P. J. Phys. Chem. C 2011, 115, 1410−1414. (37) Galopin, E.; Noual, A.; Niedziółka-Jö nsson, J.; JönssonNiedziółka, M.; Akjouj, A.; Pennec, Y.; Djafari-Rouhani, B.; Boukherroub, R.; Szunerits, S. J. Phys. Chem. C 2010, 113, 15921− 15927. (38) Galopin, E.; Touahir, L.; Niedziółka-Jönsson, J.; Boukherroub, R.; Gouget-Laemmel, A. C.; Chazalviel, J.-N.; Ozanam, F.; Szunerits, S. Biosens. Bioelectron. 2010, 25, 1199−1203. (39) Niedziółka-Jönsson, J.; Barka, F.; Castel, X.; Pisarek, M.; Bezzi, N.; Boukherroub, R.; Szunerits, S. Langmuir 2010, 26, 4266−4273. (40) Saison-Francioso, O.; Lévêque, G.; Akjouj, A.; Pennec, Y.; Djafari-Rouhani, B.; Szunerits, S.; Boukherroub, R. J. Phys. Chem. C 2012, 116, 17819−17827. (41) Szunerits, S.; Shalabney, A.; Boukherroub, R.; Abdulhalim, I. Rev. Anal. Chem. 2012, 31, 15. (42) Touahir, L.; Galopin, E.; Boukherroub, R.; Gouget-Laemmel, A. C.; Chazalviel, J.-N.; Ozanam, F.; Szunerits, S. Biosens. Bioelectron. 2010, 25, 2579−2585. (43) Chen, Y.; Munechika, K.; Ginger, D. S. Nano Lett. 2007, 7, 690−696. (44) Lakowicz, J. R.; Shen, Y.; D’Auria, S.; Malicka, J.; Fang, J.; Gryczynski, Z.; Gryczynski, I. Anal. Biochem. 2002, 301, 261−277. (45) Pompa, P. P.; Martiradonna, L.; Torre, A. D.; Sala, F. D.; Manna, L.; De Vittorio, M.; Calabi, F.; Cingolani, R.; Rinaldi, R. Nat. Nanotechnol. 2006, 1, 126−130.

for the detection of protein-glycan interactions. The deposition of a dielectric amorphous silicon−carbon alloy thin layer on a glass substrate coated with Au NSs allows covalent immobilization of a glycan monolayer via Si−C bonds in a multistep modification process. The recognition specificity on such interfaces is shown to be excellent and features limited nonspecific adsorption. The striking level of metal enhancedfluorescence was achieved with a 3 nm-thick amorphous silicon−carbon coating where the distance between the fluorophore and metal nanostructures in the layer is expected to be optimal at around 10 nm distance. The sensitivity of the, structured carbohydrate microarray described here is estimated to be enhanced by an order of magnitude with respect to the corresponding nonmetal embedded bare glass slide. A detection limit in the picomolar range is determined, which is hitherto one of the lowest values among the reported carbohydrate sensors.

■

ASSOCIATED CONTENT

S Supporting Information *

The synthesis of β-propargyl lactoside, and Figure S1 and Table S1 of LOD. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS J.Y. thanks the Ecole Polytechnique for Ph.D. financial support (EDX grant). We gratefully acknowledge financial support from the Centre National de Recherche Scientifique (CNRS), the Université Lille 1, the Région Nord Pas de Calais and the Institut Universitaire de France (IUF).

■

REFERENCES

(1) Ambrosi, M.; Cameron, N. R.; Davis, B. G. Org. Biomol. Chem. 2005, 3, 1593−1608. (2) Rudd, P. M.; Wormald, M. R.; Dwek, R. A. Trends Biotechnol. 2004, 22, 524−530. (3) Karlsson, K. A. Curr. Opin. Struct. Biol. 1995, 5, 622−635. (4) Wu, C.-Y.; Wong, C.-H. Chem. Commun. 2011, 47, 6201−6207. (5) Bertozzi, C. R.; Kiessling, L. L. Science 2001, 291, 2357−2364. (6) Barras, A.; Martin, F. A.; Bande, O.; Baumann, J.-S.; Ghigo, J.-M.; Boukherroub, R.; Beloin, C.; Siriwardena, A.; Szunerits, S. Nanoscale 2013, 5, 2307−2316. (7) Linman, M. J.; Taylor, J. D.; Yu, H.; Chen, X.; Cheng, Q. Anal. Chem. 2008, 80, 4007−4013. (8) Mann, D. A.; Kanai, M.; Maly, D. J.; Kiessling, L. L. J. Am. Chem. Soc. 1998, 120, 10575−10582. (9) Mercey, E.; Sadir, R.; Maillart, E.; Roget, A.; Baleux, F.; LortatJacob, H.; Livache, T. Anal. Chem. 2008, 80, 3476−3482. (10) Norberg, O.; Lee, I. H.; Aastrup, T.; Yan, M.; Ramström, O. Biosens. Bioelectron. 2012, 34, 51−56. (11) Sun, X.-L.; Stabler, C. L.; Cazalis, C. S.; Chaikof, E. L. Bioconjugate Chem. 2006, 17, 52−57. (12) Wang, H.; Zhang, Y.; Yuan, X.; Chen, Y.; Yan, M. Bioconjugate Chem. 2011, 22, 26−32. (13) Houseman, B. T.; Mrksich, M. Chem. Biol. 2002, 9, 443−454. G

DOI: 10.1021/ac504262b Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry (46) Ray, K.; Badugu, R.; Lakowicz, J. R. Langmuir 2006, 22, 8374− 8378. (47) Lehner, A.; Steinhoff, G.; Brandt, M. S.; Eickhoff, M.; Stutzmann, M. J. Appl. Phys. 2003, 94, 2289−2294. (48) Touahir, L.; Moraillon, A.; Allongue, P.; Chazalviel, J.-N.; Henry de Villeneuve, C.; Ozanam, F.; Solomon, I.; Gouget-Laemmel, A. C. Biosens. Bioelectron. 2009, 25, 952−955. (49) Touahir, L.; Jenkins, A. T. A.; Boukherroub, R.; GougetLaemmel, A. C.; Chazalviel, J. N.; Peretti, J.; Ozanam, F.; Szunerits, S. J. Phys. Chem. C 2010, 114, 22582−22589. (50) Tessler, L. R.; Solomon, I. Phys. Rev. B 1995, 52, 10962−10971. (51) Gouget-Laemmel, A. C.; Yang, J.; Lodhi, M. A.; Siriwardena, A.; Aureau, D.; Boukherroub, R.; Chazalviel, J.-N.; Ozanam, F.; Szunerits, S. J. Phys. Chem. C 2013, 117, 368−375. (52) Faucheux, A.; Gouget-Laemmel, A. C.; Henry de Villeneuve, C.; Boukherroub, R.; Ozanam, F.; Allongue, P.; Chazalviel, J.-N. Langmuir 2006, 22, 153−162. (53) Sam, S.; Touahir, L.; Salvador Andresa, J.; Allongue, P.; Chazalviel, J. N.; Gouget-Laemmel, A. C.; Henry de Villeneuve, C.; Moraillon, A.; Ozanam, F.; Gabouze, N.; Djebbar, S. Langmuir 2009, 26, 809−814. (54) Perez, E.; Lahlil, K.; Rougeau, C.; Moraillon, A.; Chazalviel, J.N.; Ozanam, F.; Gouget-Laemmel, A. C. Langmuir 2012, 28, 14654− 14664. (55) Touahir, L.; Niedziółka-Jönsson, J.; Galopin, E.; Boukherroub, R.; Gouget-Laemmel, A. C.; Solomon, I.; Petukhov, M.; Chazalviel, J.N.; Ozanam, F.; Szunerits, S. Langmuir 2010, 26, 6058−6065. (56) Touahir, L.; Jenkins, A. T. A.; Boukherroub, R.; GougetLaemmel, A. C.; Chazalviel, J.-N.; Peretti, J.; Ozanam, F.; Szunerits, S. J. Phys. Chem. C 2010, 114, 22582−22589. (57) Sato, Y.; Yoshioka, K.; Murakami, T.; Yoshimoto, S.; Niwa, O. Langmuir 2011, 28, 1846−1851. (58) Disney, M. D.; Seeberger, P. H. Chem. Biol. 2004, 11, 1701− 1707. (59) Yang, J.; Chazalviel, J.-N.; Siriwardena, A.; Boukherroub, R.; Ozanam, F.; Szunerits, S.; Gouget-Laemmel, A. C. Anal. Chem. 2014, 86, 10340−10349. (60) Touahir, L.; Moraillon, A.; Allongue, P.; Chazalviel, J. N.; Henry de Villeneuve, C.; Ozanam, F.; Solomon, I.; Gouget-Laemmel, A. C. Biosens. Bioelectron. 2009, 25, 952−955. (61) Gorostiza, P.; Henry de Villeneuve, C.; Sun, Q. Y.; Sanz, F.; Wallart, X.; Boukherroub, R.; Allongue, P. J. Phys. Chem. B 2006, 110, 5576−5585. (62) Harder, P.; Grunze, M.; Dahint, R.; Whitesides, G. M.; Laibinis, P. E. J. Phys. Chem. B 1998, 102, 426−436. (63) Malicka, J.; Gryczynski, I.; Gryczynski, Z.; Lakowicz, J. R. Anal. Biochem. 2003, 315, 57−66. (64) Zhou, X.; Zhou, J. Biosens. Bioelectron. 2006, 21, 1451−1458. (65) Zhang, Y.; Luo, S.; Tang, Y.; Yu, L.; Hou, K.-Y.; Cheng, J.-P.; Zeng, X.; Wang, P. G. Anal. Chem. 2006, 78, 2001−2008. (66) Zhang, Y.; Telyatnikov, V.; Sathe, M.; Zeng, X.; Wang, P. G. J. Am. Chem. Soc. 2003, 125, 9292−9293. (67) Gondran, C.; Dubois, M.-P.; Fort, S.; Cosnier, S.; Szunerits, S. Analyst 2008, 133, 206−212. (68) Gruber, K.; Horlacher, T.; Castelli, R.; Mader, A.; Seeberger, P. H.; Hermann, B. A. ACS Nano 2011, 5, 3670−3678. (69) Sugawara, K.; Shirotori, T.; Hirabayashi, G.; Kuramitz, H.; Tanaka, S. J. Electroanal. Chem. 2004, 568, 7−12. (70) Szunerits, S.; Niedziǒłka-Jönsson, J.; Boukherroub, R.; Woisel, P.; Baumann, J.-S.; Siriwardena, A. Anal. Chem. 2010, 82, 8203−8210. (71) Dam, T. K.; Oscarson, S.; Roy, R.; Das, S. K.; Pagé, D.; Macaluso, F.; Brewer, C. F. J. Biol. Chem. 2005, 280, 8640−8646.

H

DOI: 10.1021/ac504262b Anal. Chem. XXXX, XXX, XXX−XXX