Functionalization of Azide-Terminated Silicon Surfaces with Glycans

Publication History ...... This material is available free of charge via the Internet at http://pubs.acs.org. pdf. jp309866d_si_001.pdf (326.24 kb) ...
2 downloads 0 Views 985KB Size
Article pubs.acs.org/JPCC

Functionalization of Azide-Terminated Silicon Surfaces with Glycans Using Click Chemistry: XPS and FTIR Study A. C. Gouget-Laemmel,*,† J. Yang,† M. A. Lodhi,‡ A. Siriwardena,*,‡ D. Aureau,§ R. Boukherroub,∥ J.-N. Chazalviel,† F. Ozanam,† and S. Szunerits*,∥ †

Physique de la Matière Condensée, Ecole Polytechnique-CNRS, France Laboratoire des Glucides, Université de Picardie Jules Verne, 33 rue saint Leu, 80039 Amiens, France § Institut Lavoisier, UVSQ-CNRS UMR 8180, 78035 Versailles, France ∥ Institut de Recherche Interdisciplinaire (IRI, USR 3078), Université Lille 1 France, Parc de la Haute Borne, 50 Avenue de Halley, BP 70478, 59658 Villeneuve d’Ascq, France ‡

S Supporting Information *

ABSTRACT: Efficient functionalization of silicon substrates is important for the development of silicon-based sensors. Organic monolayers directly bonded to hydrogen-terminated silicon substrates via Si−C bonds display enhanced stability toward hydrolytic cleavage. Here, we show that monolayers presenting a high density of terminal azide groups are amenable to bioconjugation with alkynyl-derivatized glycans via a copper-catalyzed azide−alkyne 1,3-dipolar cycloaddition. The prerequisite azide-functionalized silicon surface is fabricated via hydrosilylation of undecylenic acid with hydrogen-terminated silicon substrate followed by reaction of the thus formed monolayer of acid groups with short, bifunctional oligoethylene oxide chains carrying an amine function at one terminus and an azido group at the other. The possibility to functionalize these azido-surfaces with alkynyl-derivatized glycans such as propargyl mannose through a click protocol is demonstrated and evidenced using X-ray photoelectron spectroscopy and Fourier-transform infrared spectroscopy. In addition, the interaction of these mannose-adorned silicon substrates with glycan binding proteins such as Lens culinaris lectin is investigated. The data establishes clearly the specificity of the interaction of this newly fabricated silicon surface for mannose-selective proteins as well as its reusability, thereby demonstrating its potential as a sensor.

1. INTRODUCTION Interactions of glycans with glycan binding proteins are key in the regulation of a plethora of physiological and pathological events.1 The study of protein−carbohydrate interactions has been challenged by the structural complexity and heterogeneity of cell surface glycans, and the typically weak affinities of their binding with glycan binding proteins (GBP or lectins). The study of the roles of glycans in cellular function (glycomics) using microarray platforms has consequently lagged behind that of genes and proteins (genomics and proteomics). The choice of chemistry for linking glycans to any surface is crucial to its ultimate utility as a sensor. Physisorption approaches are generally convenient but, in the case of small sugars, suffer from limitations due to the weakness of the van der Waals interaction forces with the surface.2,3 Self-assembled monolayers of thiolfunctionalized carbohydrates have found wide use to decorate gold interfaces.4−6 However, the packing of such molecules to form structurally well-defined monolayers is not guaranteed.7 Alternatively, Mercey et al. used pyrrole-derivatized oligosaccharides formed by electrochemical copolymerization to link glycans to SPR subtrates.8 The Cu(I)-catalyzed Huisgen 1,3dipolar cycloaddition between azide and alkyne functions to form the corresponding triazole has also generated enormous © 2012 American Chemical Society

interest as a surface modification strategy as it allows for coupling of functional molecules to surfaces with high selectivity and high yields under mild conditions.9,10 This reaction has been widely applied for the fabrication of glycanterminated surfaces.11−14 An important consideration in the construction of a glycanfunctionalized sensor surface is the choice of the substrate. Silicon-based substrates offer the advantage over organic assemblies on metal surfaces in that a well-established technology for the bulk manufacturing of silicon material exists. The dominant strategy for functionalizing silicon surfaces is based on common siloxane chemistries.15−17 However, the moisture sensitivity of silanization and the instability typical of bound silanes present drawbacks to the development of siliconbased conjugation methods.18 In addition, the low surface coverage, the hydrolytic instability of silane layers together with the formation of multilayer silane networks further limit exploitation of such layers for sensing.19 An alternative approach based on a more robust surface functionalization Received: October 5, 2012 Revised: November 30, 2012 Published: December 11, 2012 368

dx.doi.org/10.1021/jp309866d | J. Phys. Chem. C 2013, 117, 368−375

The Journal of Physical Chemistry C

Article

mL), acetic anhydride (15 mL) was added dropwise at 0 °C under argon. After 4 h, at room temperature, the mixture was quenched by adding aqueous HCl (1 M, ca. 500 mL). The mixture was extracted with dichloromethane, the organic phase dried over Na2SO4 and filtered, and the solvent evaporated under vacuum to give the crude product (5.42 g, 83.4%). The practically pure per-O-acetyl mannose (2g, 5.13 mmol) was placed in a two-necked round-bottom flask under argon and dissolved in dry dichloromethane (32 mL), and then, propargyl alcohol (1.21 mL, 20.51 mmol) was added followed by the dropwise addition of BF3OEt2 (3.39 mL, 26.8 mmol) at 0 °C. After 22 h, the reaction mixture was quenched by slow addition of a saturated NaHCO3 solution. The mixture was extracted carefully (generation of CO2) with dichloromethane, and the organic phase was dried over Na2SO4 and concentrated in vacuo. The title compound was obtained pure by flash chromatography using cyclohexane/ethyl acetate (7/3) (26%). 1H NMR (300 MHz, CDCl3) δ = 1.98 (s, 3H, CH3); 2.03 (s, 3H, CH3); 2.10 (s, 3H, CH3); 2.16 (s, 3H, CH3); 2.47 (t, 1H, CH); 4.06 (m, 2H, CH2−C); 4.26 (d, 2H, CH2− OAc); 4.28 (dt, H); 5.02 (d, H, anomeric C−H); 5.25−5.4 (m, 3H). 13C NMR δ = 20.50−20.90 (4C, CH3); 54.93 (1C, CH2− C); 62.30 (1C, CH2−OAc); 66.02 (1C, CH); 68.91 (1C, CH); 68.96 (1C, CH); 69.34 (1C, CH); 75.50 (1C, −C≡); 77.84 (1C, ≡CH); 96.22 (1C, anomeric C); 169.65 (1C, C O); 169.79 (1C, CO); 169.91 (1C, CO); 170.59 (1C, CO). 2.2.2. α-Propargyl Mannoside. Per-O-acetyl propargyl-αmannoside (200 mg, 0.518 mmol) was dissolved in dry MeOH (12 mL), and sodium methoxide was added (18.9 mg 0.349 mmol). The reaction mixture was stirred vigorously for 2 h, then treated with Amberlite resin (acid form) until neutral pH is obtained, filtered, and concentrated in vacuo to yield the title compound (110 mg, 97.4%). 1H NMR (300 MHz, CD3OD) δ = 2.85 (t, H, CH); 3.45−3.59 (m, 4H); 3.80 (m, 2H, CH2OH); 4.27 (d, 2H, CH2−C); 4.96 (d, H, anomeric C− H). 13C NMR δ = 54.87 (1C, CH2−C); 62.82 (1C, CH2− OH); 68.49 (1C, CH); 72.05 (1C, CH); 72.52 (1C, CH); 75.10 (1C, CH); 76.01 (1C, −C≡); 80.69 (1C, ≡CH); 99.87 (1C, anomeric C). 2.3. Preparation of Carbohydrate-Derivatized Surfaces. 2.3.1. Acid-Terminated Surface. The silicon platelet was first cleaned in a 1/3 H2O2/H2SO4 piranha solution at 100 °C and rinsed with Milli-Q water. It was subsequently etched in a 50% HF solution for 5 s and rinsed with Milli-Q water. The hydrogen-terminated Si surface was placed at room temperature in a Schlenk tube containing previously deoxygenated neat undecylenic acid solution and irradiated at 312 nm (6 mW cm−2) for 3 h.30 The excess of unreacted and physisorbed reagent was removed by a final rinse in hot acetic acid for 30 min.31 Then, the sample was dried under nitrogen flow. 2.3.2. NHS-Functionalized Surface. The conversion of the acid functions to the corresponding succinimidyl ester was accomplished as follows: the acid-functionalized surface was immersed in 10 mL of an aqueous solution of NHS (5 mM) and EDC (5 mM) and allowed to react for 90 min at 15 °C.32 The resulting surface was copiously rinsed with Milli-Q water and dried under a stream of argon. 2.3.3. Azido-OEO Surface. The NHS-terminated surface was reacted with 10 mM of NH2−C2H4−OEO−N3 in 1× PBS at pH ≈ 8 overnight at room temperature. The resulting surface was copiously rinsed with 1× PBS, followed by a surfactinated rinse (1× PBS/0.1% SDS for 15 min; 0.2× PBS for 5 min; 0.1×

strategy relies on the modification of hydrogen-terminated silicon surfaces via hydrosilylation chemistry using functional alkenes and gives highly stable functional monolayers due to the low polarity of the thus formed silicon−carbon linkage.20−24 However, direct attachment of azide-functionalized alkenes using this latter approach fails, as the azido group readily decomposes via the formation of a highly reactive nitrene intermediate during photochemically- or thermallyactivated hydrosilylation.13 Nevertheless, the required azidefunctionalized surface can be formed by the hydrosilylation reaction of hydrogenated silicon with bromo-terminated alkenes followed by reaction with sodium azide.25 An alternate solution where alkynyl-functionalized monolayers bound to silicon via Si−C bonds can be clicked with azido-derivatives has been reported.26−28 Gooding and co-workers used commercially available 1,8-nonadiyne to introduce alkynyl groups onto the silicon surface in a thermally activated hydrosilylation reaction where subsequent grafting of azide-terminated oligo(ethylene oxide) was achieved via a copper-catalyzed click reaction with a modest yield of 42−51%.26,28 Decreasing the density of the alkynyl chains by codeposition with alkyl chains increased the yield of the click reaction to 90% but did not provide sufficient density of oligo(ethylene oxide) chains on the silicon substrates to limit their nonspecific adsorption of proteins. This same issue was recently revisited by Cai and coworkers13 who linked trimethylgermanyl protected enyne groups to silicon via a photoactivated hydrosilylation strategy. Subsequent removal of the protection group and click reactions proceeded in a single step in good yield. Herein, we report on a hydrosilylation reaction of undecylenic acid with hydrogen-terminated silicon to form acid-terminated silicon, followed by reaction of the terminal acid groups with bifunctional short oligoethylene oxide (OEO) chains carrying an amine (−NH2) function at one terminus and an azido group (−N3) at the other. The −NH2 group can be attached through an amide-linkage to carboxylic acid partners directly via a EDC/NHS coupling strategy. The advantages of surface-linked N3 groups over surface-linked alkyne moieties have been highlighted previously.29 The possibility of these azido-terminated surfaces being clicked with alkynyl-modified glycans such as mannose is demonstrated by using X-ray photoelectron spectroscopy and quantitative FTIR spectroscopy in the attenuated total reflection (ATR) geometry. Finally, the interaction of mannose with glycan binding proteins (GBP’s) such as lectins is investigated.

2. EXPERIMENTAL SECTION 2.1. Materials. All chemicals were of reagent grade or higher and were used as received without further purification. All cleaning reagents (H2O2, 30%; H2SO4, 96%, absolute EtOH anhydrous) and hydrofluoric acid (HF, 50%) were of RSE grade and supplied by Carlo Erba. The undecylenic acid (99%) was supplied by Acros organics. All other chemicals were purchased from Sigma-Aldrich. Lectins from Lens culinaris (LENS) and from Arachis hypogaea (PNA) were obtained from Aldrich and were prepared in phosphate buffer saline (PBS) 1×. Ultrapure water (Milli-Q, 18 MΩcm) was used for the preparation of the solutions and for all rinses. 2.2. Synthesis of Alkynyl-Terminated Acetylated Mannose (Alkynyl-ManOAc) and Alkynyl-Terminated Mannose (Alkynyl-Man). 2.2.1. Per-O-acetyl-α-propargyl Mannoside. In a two-round-bottom flask containing Dmannose (3 g, 16.65 mmol) and anhydrous pyridine (51 369

dx.doi.org/10.1021/jp309866d | J. Phys. Chem. C 2013, 117, 368−375

The Journal of Physical Chemistry C

Article

Figure 1. Chemically modified crystalline silicon surfaces featuring decyl chains with terminal carboxy functions (a). Introduction of OEO spacers bearing terminal azido moieties via conjugation with the biorthogonal spacer, H2N−C2H4-EO8−N3 (b). Click reaction with an propargyl-derivatized mannose analoueg (c). Interaction with a specific GBP (d) and with a nonspecific GBP (e).

PBS for 5 min)33 and finally with Milli-Q water. The azidoOEO surface was dried under a stream of argon. 2.3.4. Clicking of Alkynyl-Terminated-Molecules. The azido−OEO surface was immersed in degassed solutions of either 3 mM per-O-acetylated in DMSO/EtOH/H2O (1/8/5 v/v/v) or free propargyl mannoside in MeOH/H2O (1/1 v/v) containing 5 mol % CuSO4·5H2O, 20 mol % sodium ascorbate. After 24 h, the sample was washed twice with EtOH, twice with EDTA solution (0.1 M), and finally rinsed thoroughly with deionized water. The clicked surface was then dried under a stream of argon. 2.3.5. Interaction with Lectins. The clicked surface is placed in direct contact with a solution of either PNA (1 mg/mL in 1× PBS) or LENS (at concentrations varying from 0.1 to 3 mg/mL in 1× PBS) and left for 1 h in a hybridization chamber. The cover slide was removed with 1× PBS, and the surface was washed with a surfactinated rinse (1× PBS/0.1% SDS for 5 min; 0.2× PBS for 2 min; 0.1× PBS for 2 min) and finally with deionized water. The lectin-treated surface was dried under a stream of argon prior to analysis. 2.4. Safety Considerations. The mixture H2SO4/H2O2 (piranha) solution is a strong oxidant. It reacts violently with organic materials. It can cause severe skin burns. It must be handled with extreme care in a well-ventilated fume hood, while wearing appropriate chemical safety protection. HF is a hazardous acid, which can result in serious tissue damage if burns are not appropriately treated. Etching of silicon should be performed in a well-ventilated fume hood with appropriate safety considerations: face shield and double layered nitrile gloves. 2.5. Surface Characterization. 2.5.1. X-ray Photoelectron Spectroscopy (XPS). XPS experiments were performed on

Thermo-VG Escalab 220iXL or Thermo K-Alpha spectrometers at the “ILV_CEFS2” center. A monochromatic Al Kα Xray line was used for the excitation. The samples are kept under nitrogen before the introduction inside the preparation chamber of the XPS analyzer. The detection was performed perpendicularly to the sample surface, using a constant energy analyzer mode (pass energy 20 eV). 2.5.2. Attenuated Total Reflection Fourier Transform Infrared (ATR-FTIR). For infrared characterization of the functionalized layer, Fourier-transform infrared spectroscopy in the attenuated total reflection geometry (ATR-FTIR) was chosen owing to its high sensitivity. For that purpose, homemade crystalline (111) silicon prisms were used as the ATR element. The dimensions of the prism (typically 15 × 15 × 0.5 mm3) limited the infrared path length in silicon, providing access to observable vibrations as low as 1000 cm−1. The spectra were recorded on a Bruker Equinox FTIR spectrometer coupled to a homemade, nitrogen-gas purged external ATR compartment. The spectra were collected with 100 scans in s- and p-polarization over the 950−4000 cm−1 spectral range with a 4 cm−1 resolution. They are displayed as absorbance per reflection (computed using natural logarithm) by using a reference spectrum recorded prior to surface modification and then normalizing the spectrum to display the actual number of reflections N determined by measuring the exact prism dimensions and angle (for the results presented here, N ≈ 22 with a bevel angle of 48°). 2.5.3. UV−vis Spectroscopy. The UV−vis spectra were recorded in the range 400−700 nm with a Cary 300 spectrophotometer and a quartz cell Hellma 6040-UV type (10 mm light path for a maximum 1400 μL volume). To make the calibration curve, 60 μL of phenol (5%), 60 μL of D370

dx.doi.org/10.1021/jp309866d | J. Phys. Chem. C 2013, 117, 368−375

The Journal of Physical Chemistry C

Article

O stretching mode at 1715 cm−1, the C−OH in plane mode at 1410 cm−1, and the two CH2 stretching modes at 2855 and 2930 cm−1 confirm that the grafting takes place without oxide formation since no Si−O−Si band is seen around 1050 cm−1 (Figure 2a). By making a quantitative analysis of the infrared data, as described in detail by Faucheux et al.,31 integration of the peak area of the CO band for s- and p-polarization enables to determine the molecular density of linked carboxydecyl groups, which is found to be N ≈ 2.0 ± 0.2 × 1014 mol cm−2, being slightly lower than that reported on atomically flat (111) silicon surfaces (N = 2.5 ± 0.2 × 1014 mol cm−2).31,34 The carboxylic acid functional group can be easily modified under mild conditions by a two-step procedure for covalently attaching amine-terminated probes. The acid function was first converted into an activated ester group (Figure 2b). The near disappearance of the acid peak at 1715 cm−1 and the appearance of new peaks at 1744, 1788, and 1816 cm−1, typical of stretching modes of the carbonyl functions of the activated ester, are consistent with the formation of the succinimidyl ester. The relative amount of activated ester formed on the surface was determined from a quantitative analysis of the above three peaks, as developed by Moraillon et al.34 The concentration is found to be N ≈ 1.6 × 1014 mol cm−2 corresponding to an activation yield of 80%. Aminolysis of the activated-ester moieties with H2N−C2H4−EO8−N3 results in the surface attachment through amide bond formation. In Figure 2c, two bands at 1642 and 1548 cm−1 characteristic of carbonyl and CNH vibrations of the amide group (amide I and II, respectively) are clearly observable. The vibrational bands at 1105 and ∼2820−2960 cm−1 are attributed to the C−O−C and OCH2 stretching modes, respectively, of the ethylene oxide chain. Moreover, bands characteristic of the stretching mode of the azido group are present at 2109 cm−1 and superimposed on the negative band of the stretching modes of SiHx. This band is more apparent when the spectrum is compared with that of the oxidized surface SiOx as a reference (inset in Figure 2c). The aminolysis yield was determined by fitting the peak at 1734 cm−1 corresponding to the remaining acid groups and was found to be 90 ± 10% (Figure 2c). The incorporation of the N3-function is clearly evidenced from XPS data. In the highresolution XPS N1s spectrum (Figure 3a), the band at 405.4 eV

mannose solution (a series of concentrations, 5.1, 8.5, 17, 25.5, and 34 μM) and 900 μL of sulphuric acid (96%) were mixed at 15 °C for 20 min before measuring the UV−vis spectra. Quantification of mannose clicked on the silicon prism was possible by dipping the prism into a stirred mixture of 60 μL of phenol (5%), 60 μL of water, and 900 μL of sulphuric acid at 15 °C for 20 min, decanting the solution into the quartz cell for UV−vis and comparing the measurement relative to the calibration curve. The reference sample is an acid-terminated silicon prism.

3. RESULTS AND DISCUSSION 3.1. Formation of Azide-Terminated Silicon Surfaces. Figure 1 depicts schematically the stepwise assembly of the carbohydrate-modified silicon substrates investigated in this work. Our approach for the preparation of azide-terminated clickable monolayers on silicon substrates is based on the selective hydrosilylation of undecylenic acid on hydrogenterminated silicon surfaces followed by activation and aminolysis of the acid functions with a short bifunctional oligoethylene oxide linker (H2N−C2H4−EO8−N3) carrying an amine function at one terminus and an azido group at the other (where EO stands for ethylene oxide, OCH2CH2, unit). FTIR spectroscopy in ATR geometry and XPS measurements were used to analyze the chemical composition and the nature of the chemical bonding on the silicon surface after each chemical modification step. Figure 2 represents the FTIR-ATR spectra of the various grafted monolayers, the reference spectra being the hydrogenated surface SiHx, which accounts for the negative bands at ∼2100 cm−1. After hydrosilylation of undecylenic acid by photochemical irradiation, the vibrational bands of the C

Figure 3. XPS narrow scans of the N1s region of the azidefunctionalized surface (a) and the mannose-functionalized surface (b).

is attributable to the azide function and arises from the central electron-deficient nitrogen (NNN) and another at 401.8 eV corresponding to the two lateral nitrogen atoms. These two bands appear with a ratio 2:1, in accordance with the incorporation of an N3 group.35 An additional band at 400.7 eV is attributed to the nitrogen atom of the amide function. In addition, high-resolution XPS spectra of Si2p and C1s are consistent with the FTIR data (see Supporting Information,

Figure 2. FTIR-ATR spectra in p-polarization of acid-terminated silicon surface before (a) and after activation with EDC/NHS (b), amidation with H2N−C2H4−EO8−N3 (c), and click reaction with acetylated-mannose, alkynyl-ManOAc, followed by the interaction with PNA (d). The reference spectra are of the hydrogenated silicon surface, SiHx. The fits of the carbonyl and amide group vibration peaks are represented in blue. The inset shows the peak related to the stretching mode of N3 of the azide surface, referred to the SiOx surface. 371

dx.doi.org/10.1021/jp309866d | J. Phys. Chem. C 2013, 117, 368−375

The Journal of Physical Chemistry C

Article

S1): the data supports that no oxidation has occurred (absence of emission peaks around 103 eV) and that amide-linked ethylene oxide chains are present (a peak at 287.2 eV ascribed to the ether C−O linkages of the ethylene oxide moieties and one at 288.8 eV ascribed to the carbonyl function of the amide). 3.2. Clicking of Acetylated Mannose. The per-Oacetylated propargyl-mannose analogues, featuring 4 acetyl functions, was clicked first to the azide-functionalized interface. The presence of the acetyl groups allows the reaction process to be followed more easily by FTIR than with unprotected sugars. The per-O-acetylated analogue was clicked on the azideterminated surface in a DMSO/H2O/EtOH mixture in the presence of Cu(I) catalyst, generated in situ from CuSO4 and sodium ascorbate. The success of the reaction is confirmed by the presence of the band at 1754 cm−1 assigned to the CO stretch of the acetyl groups and by the absence of the band at 2109 cm−1 corresponding to the stretching mode of the azide group in the FTIR-ATR spectrum of the clicked surface (Figure 2d). The broad band centered at 1050 cm−1 is ascribed to the stretching modes of C−O−C of the mannose cycle and also to the partial oxidation of the silicon surface occurring in the presence of the Cu(I) catalyst. The amount of clicked acetylated mannose was determined through integration of the absorbance band at 1754 cm−1. From a calibration curve of different concentrations of acetylated mannose in acetonitrile (see Supporting Information, S2), the IR cross-section of the carbonyl stretch vibration of acetylated mannose can be quantitatively determined. By properly taking into account the intensity of the IR electric field at the interface, this crosssection can be used to determine the surface concentration of acetylated mannose clicked at the surface from the area of the carbonyl modes in the IR spectrum of Figure 2d. This number is actually obtained from two contributions, N∥ and N⊥. These contributions correspond to the equivalent number of vibrators associated with the values of the components of the dynamic dipole parallel and perpendicular to the surface, which experience distinct values of the IR electric field at the surface. We obtain N∥ = 5.7 × 1015 As molecules/cm−2 and N⊥ = 5.5 × 1015 (1.96 Ap − 1.78As) molecules/cm−2 where As,p stands for the integrated absorbance per reflection of the νCO mode (expressed in cm−1) measured in s- and p-polarization. For the area analysis of the 1750 cm−1 band, note that the νCO mode of the remaining acid is also superimposed to that of the acetyl mode. The amount of clicked acetylated mannose was estimated to be N ≈ 1.2 × 1014 molecules cm−2, which corresponds to a yield of ∼75% for the click reaction. The less than optimal value suggests that degradation of the azido surface takes place under the click conditions. Figure 4 shows the change with time of an azido-terminated surface in contact with the solution used for the click reaction in the absence any alkyne precursor, as revealed by FTIR spectroscopy. Quantitative analysis shows that, after only 4 h of immersion in the click medium, ∼90% of the azide functions have already reacted. The data suggests that competition between the click reaction and the Cu(I)-catalyzed reduction of the azide functions is a limitation of the coupling strategy. However, the surface area occupied by one clicked per-O-acetylated mannose unit may be estimated using molecular modeling (Alchemy 2000). The result depends on the relative orientation of the triazole unit with respect to the surface. When estimated with reference to a surface concentration of 1.2 × 1014 cm−2 determined by FTIR, the values indicate close packing of per-

Figure 4. Stability of the azide-functionalized surface after immersion in a click solution (without propargyl-mannose) in MeOH/PBS 1× mixture as a function of time. (a) FTIR-ATR spectra, in p-polarization, in the range 1950−2250 cm−1, of the activated-carboxyl surface, SiNHS, and the azide-functionalized surface, Si−N3, before immersion. The reference spectra are of the oxidized silicon surface, SiOx. The fits of the vSiHx and vN3 peaks are represented in blue. The symbol ■ refers to the integrated absorbance of Si-NHS and ● to that of Si−N3. (b) Integrated absorbance of the vSiHx and vN3 stretching vibrations of the azide-terminated surface before and after 4 h, 16 h, and 4 days of immersion.

O-mannose units occurs at the surface. The data suggests that steric hindrance effects are the most plausible origin of the low yield of the click reaction. When the clicked nonacetylated mannose surface is considered instead, a surface concentration of 1.2 × 1014 cm−2 yields a surface coverage varying in the range from 0.57 to 0.88. To establish their potential for development as sensors, the absence of nonspecific protein interactions needed to be confirmed. Thus, the surface clicked per-O-acetylated mannose was exposed to either PNA or LENS lectin solutions (1 mg mL−1) in 1× PBS for 1 h. As the lectins are well-known to display no affinity for acetylated mannose, the surfaces clicked with this analogue served to establish the ability of the oligoethylene oxide linker to prevent nonspecific protein adsorption. When the OEO chain contains only three ethylene oxide units, we observe the irreversible adsorption of both PNA and LENS lectins to the surface (as evidenced by the presence of the characteristic amide I and II bands of the peptide bonds of the lectins in the corresponding FTIR spectra). In contrast, nonspecific adsorption of lectins was completely absent when alternate eight-unit ethylene oxide chains were used as linkers. In that case, no increase of the amide bands is observed in the FTIR-ATR spectrum as shown in Figure 2d. Recently, we demonstrated that low-density grafted poly(ethylene glycol) monolayers on (111) Si exhibit efficient protein-repellent performance but only when OEO chains are of sufficient length to be able to adopt an entangled arrangement (≥16).33 In the present case, however, the protein-repellent performance is achieved even for surfaces containing ethylene oxide chains of medium length (≥8) probably because of the higher density of the grafted monolayers obtained here (∼1.6 × 1014 cm−2). 3.3. Clicking of Alkynyl-Terminated Mannose and Interaction with Lectins. The click reaction of propargyl 372

dx.doi.org/10.1021/jp309866d | J. Phys. Chem. C 2013, 117, 368−375

The Journal of Physical Chemistry C

Article

mannose with the azide-terminated silicon surface was performed under the optimized conditions in a MeOH/H2O solution. The FTIR-ATR spectrum of the corresponding mannose-terminated silicon surface shows the disappearance of the νN3 band, which is indicative of the completion of the reaction (Figure 4a). XPS allowed for a more detailed analysis of the mannose-decorated surface (Figure 3b). Comparison between the N1s high-resolution XPS spectrum before (Figure 3a) and after clicking mannose to the surface allows the following conclusions. The peak at high binding energies (405.4 eV) characteristic of the central nitrogen in the azido function is absent, and two distinct peaks at 402.7 (N−C) and 401 eV (NN) characteristic of the formation of a triazole moiety are instead observed.35,36 A peak at 401 eV suggests an additional contribution of the nitrogen atom from the amide function. A less intense peak apparent at 400 eV is attributed to an additional contribution from an −NH2 moiety and might derive from the reduction of an azide function as also revealed by FTIR-ATR. The ratio of the N1s components 0.75/2.5/0.25 at 402.7/401/400 eV is in accordance with the theoretically expected values, taking into account that the yield of the click reaction is ∼75%. Overlap between the various contributions of the N1s spectrum precludes unambiguous determination of the yield of the click reaction, whose value is more accurately deduced from an FTIR analysis. An alternative colorimetric method based on the phenol-sulfuric acid assay (PSA) was used as an independent method for calculating this value.37,38 This method is based on the measured peak absorbance at ∼480 nm of a colored aromatic conjugate product formed between phenol and a reducing carbohydrate. The amount of sugar present on the surface, effectively cleaved off during treatment with concentrated sulphuric acid and transferred into the solution, is determined with reference to a UV−vis calibration curve generated using varying concentrations of D-mannose as the standard reference sugar (see Supporting Information, S3). The concentration of mannose derived using the PSA method allows the yield of the click reaction to be estimated as 70%, which is in line with the value determined by FTIR for the quantification of surface clicked per-O-acetylated mannose. The efficiency and specificity of the GBP interaction of the newly fabricated surface was evaluated by incubating it first with PNA and then with LENS lectin at 1 mg/mL in 1× PBS solution for periods of up to 1 h. Only LENS lectin selectively recognizes mannose residues. The IR-ATR spectrum reveals no adsorption of PNA to the mannose-clicked surface (Figure 5b, whereas adsorption of LENS lectin is revealed by the significant increase in bands ascribed to amide I and II vibrations by a factor of 2.1−2.2 (in Figure 5c). The reversibility of the observed mannose−LENS lectin interaction was established by disrupting it by washing with a concentrated D-mannose solution. After 1 h immersion in 20 mM mannose aqueous solution and a final surfactinated rinse, analysis of the amide bands by FTIR-ATR (data not shown) demonstrated that the regenerated surface contained less than 2% of LENS, establishing its reusability as a sensor. The same surface was again incubated with a PNA solution at 1 mg/mL for 1 h. After a surfactinated rinse, there is almost no PNA (less than 5%) present on the surface as revealed by FTIR-ATR (data not shown) confirming that the surface interaction is indeed specific for mannose selective lectins. Figure 6 represents the integrated IR absorbance of the amide I band of LENS lectin as a function of lectin concentration and, as expected, a curve that tends to saturate above 1.5 mg/mL is generated. This curve can

Figure 5. FTIR-ATR spectra in p-polarization of mannose-functionalized silicon surface before (a) and after interaction with PNA (b) or LENS lectins (c). The fits of the carbonyl and amide vibrational peaks are represented in blue. The reference spectra are the hydrogenated silicon surface SiHx. Notice that spectrum a is similar to that of Figure 2d, except for the presence of the carbonyl vibrations of acetylated substituents in the latter case.

Figure 6. Integrated absorbance of the amide I vibrations as a function of the concentration of LENS lectin.

be fitted according to the Langmuir adsorption isotherm model. We consider Aeq as the integrated absorbance of adsorbed LENS at the surface, which is equal to Aeq =

A max KAC LENS 1 + KAC LENS

(1)

where Amax corresponds to the maximum adsorbed amount of LENS lectin with increasing lectin concentration. The affinity constant (KA) for the interaction can be determined using Figure 6 by fitting the data to the Langmuir equation. We find a value of 2.37 mg−1 mL for KA, i.e., 1.09 × 105 M−1, which is of the magnitude expected for typical mannose/GBP interactions but in the upper range of values reported in the literature.39,8,11,12 Such a large value for KA observed with the 373

dx.doi.org/10.1021/jp309866d | J. Phys. Chem. C 2013, 117, 368−375

The Journal of Physical Chemistry C

Article

(7) Houseman, B. T.; Mrksich, M. Chem. Biol. 2002, 9, 443−454. (8) Mercey, E.; Sadir, R.; Maillart, E.; Roget, A.; Baleux, F.; LortatJacob, H.; Livache, T. Anal. Chem. 2008, 80, 3476−3482. (9) Moses, J. E.; Moorhouse, A. D. Chem. Soc. Rev. 2007, 36, 1249− 1262. (10) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. Angew. Chem., Int. Ed. 2002, 41, 2596−2599. (11) Szunerits, S.; Niedziolka-Jonsson, J.; Boukherroub, R.; Woisel, P.; Baumann, J. S.; Siriwardena, A. Anal. Chem. 2010, 82, 8203−8210. (12) Zhang, Y.; Luo, S. Z.; Tang, Y. J.; Yu, L.; Hou, K. Y.; Cheng, J. P.; Zeng, X. Q.; Wang, P. G. Anal. Chem. 2006, 78, 2001−2008. (13) Qin, G. T.; Santos, C.; Zhang, W.; Li, Y.; Kumar, A.; Erasquin, U. J.; Liu, K.; Muradov, P.; Trautner, B. W.; Cai, C. Z. J. Am. Chem. Soc. 2010, 132, 16432−16441. (14) Santoyo-Gonzalez, F.; Hernandez-Mateo, F. Chem. Soc. Rev. 2009, 38, 3449−3462. (15) Washburn, A. L.; Gunn, L. C.; Bailey, R. C. Anal. Chem. 2009, 81, 9499−9506. (16) Jo, S.; Park, K. Biomaterials 2000, 21, 605−616. (17) Harbers, G. M.; Emoto, K.; Greef, C.; Metzger, S. W.; Woodward, H. N.; Mascali, J. J.; Grainger, D. W.; Lochhead, M. J. Chem. Mater. 2007, 19, 4405−4414. (18) Cattani-Scholz, A.; Pedone, D.; Dubey, M.; Neppl, S.; Nickel, B.; Feulner, P.; Schwartz, J.; Abstreiter, G.; Tornow, M. ACS Nano 2008, 2, 1653−1660. (19) Lapin, N. A.; Chabal, Y. J. J. Phys. Chem. B 2009, 113, 8776− 8783. (20) Ciampi, S.; Harper, J. B.; Gooding, J. J. Chem. Soc. Rev. 2010, 39, 2158−2183. (21) Linford, M. R.; Chidsey, C. E. D. J. Am. Chem. Soc. 1993, 115, 12631−12632. (22) Boukherroub, R.; Morin, S.; Bensebaa, F.; Wayner, D. D. M. Langmuir 1999, 15, 3831−3835. (23) Boukherroub, R.; Wayner, D. D. M. J. Am. Chem. Soc. 1999, 121, 11513−11515. (24) Aureau, D.; Varin, Y.; Roodenko, K.; Seitz, O.; Pluchery, O.; Chabal, Y. J. J. Phys. Chem. C 2010, 114, 14180−14186. (25) Marrani, A. G.; Dalchiele, E. A.; Zanoni, R.; Decker, F.; Cattaruzza, F.; Bonifazi, D.; Pratoc, M. Electrochim. Acta 2008, 53, 3903−3909. (26) Ciampi, S.; Bocking, T.; Kilian, K. A.; James, M.; Harper, J. B.; Gooding, J. J. Langmuir 2007, 23, 9320−9329. (27) Hurley, P. T.; Nemanick, E. J.; Brunschwig, B. S.; Lewis, N. S. J. Am. Chem. Soc. 2006, 128, 9990−9991. (28) Ciampi, S.; Eggers, P. K.; Le Saux, G.; James, M.; Harper, J. B.; Gooding, J. J. Langmuir 2009, 25, 2530−2539. (29) Devaraj, N. K.; Collman, J. P. QSAR Comb. Sci. 2007, 26, 1253− 1260. (30) Voicu, R.; Boukherroub, R.; Bartzoka, V.; Ward, T.; Wojtyk, J. T. C.; Wayner, D. D. M. Langmuir 2004, 20, 11713−11720. (31) Faucheux, A.; Gouget-Laemmel, A. C.; Henry de Villeneuve, C.; Boukherroub, R.; Ozanam, F.; Allongue, P.; Chazalviel, J.-N. Langmuir 2006, 22, 153−162. (32) 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 2010, 26, 809−814. (33) Perez, E.; Lahlil, K.; Rougeau, C.; Moraillon, A.; Chazalviel, J. N.; Ozanam, F.; Gouget-Laemmel, A. C. Langmuir 2012, 28, 14654− 14664. (34) Moraillon, A.; Gouget-Laemmel, A. C.; Ozanam, F.; Chazalviel, J.-N. J. Phys. Chem. C 2008, 112, 7158−7167. (35) Collman, J. P.; Devaraj, N. K.; Eberspacher, T. P. A.; Chidsey, C. E. D. Langmuir 2006, 22, 2457−2464. (36) Li, Y.; Wang, J.; Cai, C. Z. Langmuir 2011, 27, 2437−2445. (37) Dubois, M.; Gilles, K. A.; Hamilton, J. K.; Rebers, P. A.; Smith, F. Anal. Chem. 1956, 28, 350−356. (38) Panagiotopoulos, C.; Sempere, R. Limnol. Oceanogr.: Methods 2005, 3, 419−454.

present sensor might well be due to its binding of LENS lectin through multiple surface-attached mannose moieties via some multivalent mechanism. Indeed, multivalency is a very commonly encountered phenomena underpinning various carbohydrate−GBP interactions that we and others have studied.40−42 Further studies will be necessary to draw clearcut conclusions relating to the molecular basis of these observations, and these are presently underway.

4. CONCLUSIONS We have developed a versatile platform featuring azideterminated oligo(ethylene oxide) groups presented on silicon surfaces that is amenable to direct tethering of alkynylfunctionalized glycans via Cu(I) catalyzed click protocol. The presence of eight ethylene oxide unit tethering spacers prevents nonspecific binding of proteins to the modified silicon surface and thus addresses successfully a key issue in sensor development. Quantitative FTIR spectroscopy, XPS, and a colorimetric UV−vis assay were explored to establish the amount of clicked mannose present on the silicon substrate and was determined to be about 70%. The resulting mannoseterminated surface was shown to selectively capture mannoserecognizing lectins and provides a new carbohydrate sensor with potential for the probing glycan/protein binding events.



ASSOCIATED CONTENT

S Supporting Information *

Additional information as described in the text. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*(A.C.G.) E-mail: [email protected]. Phone: +33 1 69 33 46 80. Fax: +33 1 69 33 47 99. (A.S.) E-mail: [email protected]. Phone: +33 3 22 82 76 73. Fax: +33 3 22 82 75 60. (S.S.) E-mail: sabine.szunerits@ iri.univ-lille1.fr. Phone: +33 3 20 19 79 87. Fax: +33 3 20 19 78 84. 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). R.B. and S.S. wish to thank the Université Lille 1 and the région Nord Pas de Calais for financial support. A.S. is grateful to the Higher Education Commission, Pakistan, for financial support and a postdoctoral fellowship to M.A.L.



REFERENCES

(1) Rudd, P. M.; Wormald, M. R.; Dwek, R. A. Trends Biotechnol. 2004, 22, 524−530. (2) Wang, D. N.; Liu, S. Y.; Trummer, B. J.; Deng, C.; Wang, A. L. Nat. Biotechnol. 2002, 20, 275−281. (3) Willats, W. G. T.; Rasmussen, S. E.; Kristensen, T.; Mikkelsen, J. D.; Knox, J. P. Proteomics 2002, 2, 1666−1671. (4) Grant, C. F.; Kanda, V.; Yu, H.; Bundle, D. R.; McDermott, M. T. Langmuir 2008, 24, 14125−14132. (5) Mann, D. A.; Kanai, M.; Maly, D. J.; Kiessling, L. L. J. Am. Chem. Soc. 1998, 120, 10575−10582. (6) Sato, Y.; Yoshioka, K.; Murakami, T.; Yoshimoto, S.; Niwa, O. Langmuir 2012, 28, 1846−1851. 374

dx.doi.org/10.1021/jp309866d | J. Phys. Chem. C 2013, 117, 368−375

The Journal of Physical Chemistry C

Article

(39) Lis, H.; Sharon, N. Chem. Rev. 1998, 98, 637−674. (40) Arranz-Plaza, E.; Tracy, A. S.; Siriwardena, A.; Pierce, J. M.; Boons, G. J. J. Am. Chem. Soc. 2002, 124, 13035−13046. (41) Siriwardena, A.; Jorgensen, M. R.; Wolfert, M. A.; Vandenplas, M. L.; Moore, J. N.; Boons, G. J. J. Am. Chem. Soc. 2001, 123, 8145− 8146. (42) Thobhani, S.; Ember, B.; Siriwardena, A.; Boons, G. J. J. Am. Chem. Soc. 2003, 125, 7154−7155.



NOTE ADDED AFTER ISSUE PUBLICATION In the manuscript published in volume 117, issue 1, pages 368−375, two corresponding authors were missing. The authors Sabine Szunerits and Aloysius Siriwardena were added as corresponding authors, along with their contact information. The corrected version was reposted to the Web on February 21, 2013.

375

dx.doi.org/10.1021/jp309866d | J. Phys. Chem. C 2013, 117, 368−375