Glutathione Immunosensing Platform Based on Total Internal

Apr 25, 2014 - Antonio García Marín , María Jesús Hernández , Eduardo Ruiz , Jose María Abad , Encarnación Lorenzo , Juan Piqueras , Jose Luis ...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/ac

Glutathione Immunosensing Platform Based on Total Internal Reflection Ellipsometry Enhanced by Functionalized Gold Nanoparticles Antonio García-Marín,†,∥ José M. Abad,‡,§,∥ Eduardo Ruiz,† Encarnación Lorenzo,*,‡,§ Juan Piqueras,† and José L. Pau*,† †

Grupo de Electrónica y Semiconductores, Departamento de Física Aplicada, Universidad Autónoma de Madrid, Cantoblanco, 28049 Madrid, Spain ‡ Departamento de Química Analítica y Análisis Instrumental, Universidad Autónoma de Madrid, Cantoblanco, 28049 Madrid, Spain § Instituto Madrileño de Estudios Avanzados en Nanociencia (IMDEA-Nanociencia), Fáraday, 9, Campus UAM, Cantoblanco, 28049 Madrid, Spain S Supporting Information *

ABSTRACT: An immunosensor to detect small molecules, such as glutathione (GSH), has been developed by combination of ellipsometry and Kretschmann surface plasmon resonance (SPR). The Au thin film used for surface plasmon polariton (SPP) excitation is functionalized with anti-GSH to specifically bind GSH. At low concentrations, the small refractive index changes caused by the low molecular weight of GSH induced only negligible shifts in the plasmon resonant energy during GSH binding. To improve sensitivity, gold nanoparticles (AuNPs) are functionalized with glutathione acting as amplifiers of the antigen−antibody interaction. Changes induced by the AuNP adsorption are monitored using Ψ and Δ ellipsometric functions. After performing competitive assays using solutions containing different concentrations of free GSH and a constant amount of functionalized AuNPs, it was concluded that the resonant energy linearly shifts as the relative concentration of free GSH increases. A detection limit for free GSH in the nanomolar range is found, demonstrating the effectiveness of AuNPs to enhance the sensitivity to immunoreactions in total internal reflection ellipsometry.

S

polariton (SPP) excitation in Krestchmann configuration, in a technique known as total internal reflection ellipsometry (TIRE). This is especially relevant in the Δ case, in which larger magnitude changes are found in comparison to Ψ. Due to the strong variations of this function around the resonance energy, total internal reflection ellipsometry is claimed to have 10 times larger sensitivity than regular SPR in response to refractive index changes in the vicinity of the metal surface.7 However, the binding of a low-molecular-weight molecule only gives rise to small changes in the refractive index, making its detection difficult by SPR or TIRE techniques. To enhance sensitivity to small molecules, gold nanoparticles (AuNPs) can be used as signal amplifiers.8 These particles exhibit a strong absorption band around 520 nm that arises from the excitation of localized surface plasmon resonance (LSPR) by incident light. Their surfaces can be modified in many ways to incorporate biomolecules from their well-known attachment

urface plasmon resonance (SPR) sensor technology is a powerful technique to study biomolecular interactions whose role for the development of new generation sensing tools and protocols has notoriously grown in the past decade.1 One of the main advantages is that SPR does not require fluorescent labels, which are often toxic and can interfere with the biological processes under analysis. The technique has been widely investigated in immunosensing technologies, including applications such as medical diagnosis of illnesses,2 food safety,3 environmental monitoring,4 or detection of explosives.5 Although SPR enables the investigation of biomolecular interactions by probing refractive index changes on the sensor chip surface, its coupling to ellipsometry constitutes a more sensitive and accurate method.6 Common SPR methods measure changes in the intensity of a p-polarized reflected beam as a function of angle. In ellipsometry, the tangent of the psi (Ψ) function is defined as the amplitude ratio of the pcomponent of the reflected beam over the s-component, and delta (Δ) function is defined as the difference between the phase shift for both components. It has been demonstrated that both functions present strong evidence of surface plasmon © 2014 American Chemical Society

Received: January 27, 2014 Accepted: April 25, 2014 Published: April 25, 2014 4969

dx.doi.org/10.1021/ac5005212 | Anal. Chem. 2014, 86, 4969−4976

Analytical Chemistry

Article

Figure 1. (a) Scheme of the ellipsometric setup; the green wave inside the prism represents the surface plasmon polariton; typical shapes of Ψ and Δ functions around the surface plasmon (SP) resonance; resonance energy (EΨmin) and amplitude (Φsp) are defined as the position of the Ψ function minimum and the peak-to-peak angle shift in Δ function, respectively. (b) Scheme of the competitive immunoassay between free GSH and GSHAuNPs solution performed on the functionalized sensor chip.

chemistry.9 AuNPs binding to thin Au films leads to large changes in the polariton resonance through optical coupling between the evanescent wave in the metal surface of the film and the resonant scattering in the NPs.10 This property can be used to amplify low molecular weight molecule responses, improving their detection limit. As a case of study, we have chosen glutathione (GSH) due to the importance of this biological molecule. GSH is the most abundant intracellular nonprotein low-molecular-weight thiol in almost all living organisms predominantly in eukaryotic cells,11 playing a critical role in numerous processes.12 These include protection against oxidative stress in cells maintaining the intracellular redox balance13 and reducing hydrogen peroxide through glutathione reductase, detoxification of xenobiotics, modulation of the immune response,14 protein folding,15 and transport of organic sulfur.16 It is known that many dangerous diseases in humans are related to low GSH levels, including Parkinson, Alzheimer,17 cancer,18 cystic fibrosis, cardiovascular, inflammatory,19 immune,20 and metabolic diseases.21 Since GSH constitutes an important biomarker, in the last years there has been an increasing interest in its selective and quantitative detection. To date, various methods for GSH detection have been published. Typically, GSH can be quantified by spectrofluorimetry, using commercial kits which comprise Ellman’s reagents or bimane drivatives such as monobromobimane, becoming fluorescent after binding to GSH.22 These methods do not need to achieve low detection

limits since intracellular GSH concentrations usually range from 0.1 to 10 mM.23 However, for other applications where a higher sensitivity is required, other techniques such as electrochemistry,24 high-performance liquid chromatography25 and gas chromatography,26 or quartz crystal microbalance27 have been employed, enabling the detection at lower levels. All these methods have advantages but also drawbacks, such as the lower selectivity against other amino acids or of being timeconsuming, laborious, and complicated. So far, very few GSH detection methods have been published, combining a high sensitivity and specificity.24a,28 In this paper, we present an inmunosensor for GSH quantification based on TIRE and enhanced by the use of colloidal AuNPs, taking the advantage of sensitivity of the ellipsometry technique besides selectivity of the biorecognition antibody−antigen. Schematic representation of the ellipsometric setup and the functionalized sensor are shown in Figure 1. The sensor is composed of a Au thin film on a quartz substrate functionalized by a GSH specific antibody on its surface, covalently attached through a self-assembled monolayer (SAM) of 3,3′-dithiodipropionic acid di(N-hydroxysuccinimide ester) (DTSP). Competitive immunoassay is performed between GSHfunctionalized gold nanoparticles (GSH-AuNPs) and free GSH in solution. The energy shift caused by the GSHAuNPs coupling in TIRE experiments can be correlated to the amount of GSH present in the sample, since an increasing 4970

dx.doi.org/10.1021/ac5005212 | Anal. Chem. 2014, 86, 4969−4976

Analytical Chemistry

Article

range. The light spot has an elliptical shape with a 5.0 × 3.0 mm2 area. The SPR resonance in the TIRE spectra presents a characteristic reduction of the Ψ function due to the extinction of the p-polarized light component, and a strong angle shift of the Δ function. The resonance energy is determined by the position of the Ψ function minimum (EΨmin) and the resonance amplitude by the peak-to-peak angle shift in Δ function (Φsp) (inset of Figure 1a). First, actual thicknesses of the AZO/Au bilayer are obtained from the fitting of the output characteristics in the as-deposited samples. An external incident angle of 47° is chosen to maximize the amplitude of the plasmon resonance in the setup. Due to refraction effects in the air− prism interface, the effective incident angle in the AZO/Au bilayer is almost perpendicular to the input face of the prism. In order to find out the NP concentration that produces a largest SPR shift after the interaction between the antibody layer and the GSH-AuNPs, samples are exposed to several solutions with NP concentrations ranging from 10 to 80 nM. Finally, the competitive immunoassay is performed, adjusting the GSH-AuNPs concentration to this value and varying the relative concentration of free GSH. A scheme of the sensor concept is shown in Figure 1b. To estimate reaction times on the Au surface, real-time ellipsometry is performed at an external incident angle of 85°, while the Au surface is in direct contact with different solutions. The prism hosting the sample underneath is directly put on a homemade flow cell, sealed with a rubber O-ring, and connected to a peristaltic pump to inject the as-prepared solutions.

concentration of GSH reduces the available binding sites for GSH-AuNPs and their final surface density.



EXPERIMENTAL SECTION Materials. L-reduced GSH, 3,3′-dithiodipropionic acid di(N-hydroxysuccinimide ester) (DTSP), acetone, and dimethyl sulfoxide (DMSO) are purchased from Sigma-Aldrich (St. Louis, MO). GSH polyclonal antibody (anti-GSH) is acquired from Abcam (U.K.). Matching index fluid is purchased from Thorlabs (Germany, ref. G608N). All chemicals are used as received without further purification. All solutions are prepared with ultrapure water (18 MΩ cm) obtained from a Direct-Q 3 UV Millipore purification system. Sensor Fabrication and Immobilization of Anti-GSH. A 50 nm thick Al-doped ZnO (AZO) layer is synthesized on 2.0 × 2.0 cm2 quartz substrates at 300 °C by rf-magnetron sputtering, using a 99.999% pure ZnO w/2% Al2O3 target and Ar plasma. Afterward, a 45 nm Au layer is thermally evaporated on the AZO thin film. The AZO layer improves the adherence of the Au layer to the substrate and presents an optical transmittance which exceeds 75% at the SPP resonant energy, avoiding absorption losses found in conventional Cr or Ti adherence layers.29 The Au surface is further functionalized, immersing the samples in a 4 mM DTSP in DMSO solution and left overnight. DTSP adsorbs onto Au surfaces, through the disulfide group,30 giving rise to the chemical modification of the surface with N-succinimidyl-3-thiopropionate, which can react with amino groups present in the antibody enabling covalent bonding through the formation of amide bonds.31 Then, the surface is rinsed a few times with acetone and blown with dry nitrogen. Afterward, anti-GSH is covalently immobilized for 2 h on the functionalized Au surface by incubation in a 150 nM antibody physiological solution. Unreacted succinimidyl moieties on the chip surface are deactivated by further soaking the chip into physiological solution for 1 h. Finally, the samples are rinsed a few times with water, blown with dry nitrogen, and stored in the fridge for further use. AuNPs Synthesis and Functionalization. Aqueous sodium citrate-stabilized gold nanoparticles (Cit-AuNPs) are synthesized by following the Turkevich’s method.32 The ratio between chloroauric acid and sodium citrate is set to obtain NPs with an average diameter of 14 nm. To functionalize the NP surface, 200 μL of a free GSH aqueous solution (0.2 M) is added to 4 mL of Cit-AuNPs solution and left overnight for stabilization. Afterward, a purification process is carried out in order to remove GSH excess and reaction byproducts by using several cycles of ultrafiltration through Amicon Ultra Centrifugal Filters with ultracel-100 K membrane at 4000 rpm, employing water as a diluent. The free GSH concentration in the NP solution after purification is lower than 5 pM. The final concentration of GSH-AuNPs is adjusted to 40 nM for the competitive assay. Total Internal Reflection Ellipsometry (TIRE). The samples are measured by SPR-TIRE to find out optimal experimental conditions for SPR excitation. Following the Kretschmann configuration, samples are attached to a rightangle fused-silica prism, using a matching index fluid on the backside of the substrate in order to minimize reflections at the prism/substrate interface. The experimental setup is represented in Figure 1a. Ψ and Δ functions are recorded using a Jobin Yvon UVISEL spectroscopic phase-modulated ellipsometer. The measurements are recorded in the 1.5−2.3 eV spectral



RESULTS AND DISCUSSION Synthesis and Characterization of Glutathione-Capped AuNPs. Au nanoparticles are first prepared by the aqueous citrate reduction method. Transmission electron microscopy (TEM) of the nanoparticles (Figure S1 of the Supporting Information) shows a narrow particle size distribution with an average diameter of 14 ± 1 nm. UV/ visible absorbance spectrum of these Cit-AuNPs exhibits, as evidence of LSPR excitation, the characteristic surface plasmon band with a peak wavelength (λp) of 518 nm (Figure S2a of the Supporting Information, black curve). Fourier-transform infrared spectroscopy (FT-IR, Figure S2, panels b and c, of the Supporting Information, black curve) shows the presence of the citrate stabilizer in the AuNPs with absorbances at 2955, 2867, 1609, 1512, and 1456 cm−1, corresponding to the vibrational modes νasy(−CH2), νsym(−CH2), νasy(−CO2¯), νsym(−CH2), and νsym(−CO2¯), respectively. The band at approximately 3380 cm−1 is attributed to the ν(−OH) vibration, and the bands at 1257 and 1184 cm−1 are correlated with ν(−CO). The AuNPs are capped with glutathione by addition of an aqueous solution of GSH immediately after preparation. The sulfhydryl group in GSH strongly chemisorbs to the Au surface displacing the adsorbed citrate ions. As a result, λp is red shifted about 1.6 nm (Figure S2a of the Supporting Information, red curve) due to the change in the dielectric permittivity of the NP surface. This shift is in agreement with that expected from the theoretical model proposed by Templeton et al. for AuNPs capped with an organic layer, which predicts a λp red shift of around 1.0−2.0 nm for thicknesses between 0.8 and 2.0 nm.33 It is also noticeable that there is a damping effect on the resonance after capping. This behavior is characteristic of thiolcontaining compounds chemisorbed to metallic NPs, which 4971

dx.doi.org/10.1021/ac5005212 | Anal. Chem. 2014, 86, 4969−4976

Analytical Chemistry

Article

typically reduce the oscillator strength and broaden the LSPR band due to charge localization near the AuNP surface.34 The FT-IR spectrum of the GSH-AuNPs (Figure S2, panels b and c, of the Supporting Information, red curve) is different to that obtained for Cit-AuNPs. It exhibits characteristic bands associated with the vibrational modes of glutathione, in agreement with the assignments reported in the literature.35 The large band at the 1660−1600 cm−1 range is attributed to amide I modes (CO symmetric stretching from amide bonds and CN symmetric stretching, νs(CO) and νs(CN), and NH3+ antisymmetric bending band δ(NH3+), which arises at 1602 cm−1). The presence of amide II modes are observed in the spectral region between 1550 and 1530 cm−1, associated with NH in-plane bend and CN symmetric stretching vibrations [δ(NH) and νs(CN)]. The peak at 1396 cm−1 corresponds to the COO− symmetric stretching [νs(COO−)], and the band at 3200 cm−1 is assigned to the contribution of several overlapped bands, NH3+ antisymmetric and symmetric stretching modes [νas(NH3+) and νs(NH3+)]. Additionally, CH2 antisymmetric and symmetric stretching modes [νas(CH2) and νs(CH2)] are found at 2922 and 2854 cm−1, respectively. All these results demonstrate the effective functionalization of AuNPs by glutathione. Development and Characterization of the Sensor Chip. SPR-TIRE sensor chips are fabricated for operation in a Krestchmann configuration setup as described in Experimental Section. Instead of Cr or Ti intermediate layers, commonly employed in SPR chips, an Al-doped ZnO (AZO) layer is deposited between the Au thin film and the glass substrate in order to avoid absorption losses produced by those materials.36 This layer, besides improving the adherence of Au to the glass substrate, helps to enhance the energy transfer from the incident radiation to the evanescent wave, thanks to the high transparency of the AZO at the resonant energy under the chosen experimental conditions. In air, the numerical simulation of the ellipsometric functions using Fresnel equations shows that the introduction of the AZO layer only produces a slight red-shift in the Δ function in comparison to glass chips with plain Au. On the other hand, the introduction of Cr for thicknesses in the usual nanometer range produces larger red shifts and strong attenuations which can hinder the SPR analysis (Figure S3 of the Supporting Information). In the TIRE experimental setup, the resonance for the AZO/ Au bilayer in air is found at energies of about 2.0 eV under an external incident angle of 47° (Figure 2). The position of the resonance is fitted using numerical simulations and a four-layer model under the fused silica prism (model A, Table 1). From top-to-bottom, the model includes a 50 nm thick AZO layer whose optical constants are found by numerical inversion of ellipsometric data, independently taken on different AZO thin films on Si substrates.29,37 This layer is followed by a 42 nm thick Au film whose optical constants are taken from the literature.38 The thicknesses for both layers are experimentally determined by stylus profilometry. Next, the fitting works require the introduction of an effective medium approximation (EMA) layer formed by Au and air to account for the surface roughness of the Au layer. The material compositions and thickness of this layer are varied to fit the energy of the final resonance, providing best fitting values of 60% for Au and 40% for air and a 5.2 nm total thickness. Within this model, an accurate fitting of the resonance energy is obtained. Au surface is coated with a DTSP SAM by dipping the sensor chip in a 4 mM DTSP/DMSO solution, and the ellipsometric

Figure 2. TIRE-SPR spectra measured before (black □) and after (red ○) antibody immobilization for both ellipsometric functions (Ψ and Δ).

functions in TIRE mode are measured in air. A small resonant energy shift is observed in the spectrum being within the experimental error, probably due to differences in Au surface roughness and the low refractive index change introduced by the thin SAM in the vicinity of the metal surface. In order to provide evidence of the DTSP monolayer formation, electrochemical reductive desorption of the attached layer is carried out using cyclic voltammetry. Cyclic voltammogram (Figure S4 of the Supporting Information) shows a characteristic peak at −1.10 V associated with the reductive desorption of the thiol. A surface coverage of 9.5 × 10−10 mol/ cm2 is estimated from the charge under the reductive desorption peak, assuming a value of 1 electron per sulfur atom. This value is in agreement with the typical surface density of molecules reported for the maximum thiol coverage on Au surfaces,39 which confirms the surface modification. The surface, now bearing an “activated ester” termination, is subsequently made to react with the anti-GSH. The immobilization is expected to take place by nucleophilic attack of primary amino groups of the antibody to the terminal Nsuccinimidyl esters in the monolayer.30 After antibody immobilization, the TIRE spectrum is newly measured in air. As shown in Figure 2, a significant red shift of the resonant energy is obtained due to the strong change in the refractive index caused by the adlayer. In addition, the resonance strength slightly attenuates because of the screening effect of the charges in the antibody layer, which reduces the amplitude of the polariton in the metal surface. In order to fit the resonance energy after antibody immobilization, the four-layer model (model A) in Table 1 is modified as follows: the Au/air EMA layer is replaced by an Au/bioorganic EMA layer with the same thickness, below which a uniform bioorganic layer is introduced. 4972

dx.doi.org/10.1021/ac5005212 | Anal. Chem. 2014, 86, 4969−4976

Analytical Chemistry

Article

Table 1. Layer Models for the Sensor Chip That Best Fit the Resonance Energy Position Before (left) and After (right) Antibody Immobilizationa model A ambient

EMA layer substrate

composition fused silica AZO Au 60% Au 40% air air

thickness (nm)

model B ambient

50 42 5.2

EMA layer

substrate a

composition

thickness (nm)

fused silica AZO Au 60% Au 40% antibody antibody air

50 42 5.2 3.8

The layers are ordered from bottom to top in the setup. Ambient fused silica layer accounts for the prism optical characteristics.

solution is injected in the flow cell. Immediately, a strong variation in Δ is observed associated with the binding process of the NPs to antibody molecules immobilized onto the surface. The signal reaches a steady level after around 35 min. From these studies, 1 h is set for further immunoassays to ensure that the binding process reaches equilibrium. The interaction between GSH-AuNPs and the antibody layer is further verified by field emission-scanning electron microscopy (FE-SEM). The FE-SEM images corresponding to the antiglutathione modified sensor before (Figure S5a of the Supporting Information) and after (Figure S5b of the Supporting Information) the immunological reaction, show the presence of the GSH-AuNPs as tiny bright dots on the antibody layer. These can be more clearly observed in the image taken at higher magnification (inset of Figure S5b of the Supporting Information). The distribution is fairly uniform, with no signs of NP aggregation, confirming their binding to the anti-GSH layer. Given that NP and the antibody have an almost similar size, it is fair to assume that each NP binds to a single antibody molecule, giving rise to a random NP pattern over the antibody layer. The average particle density obtained from a raster scan over an area of 330 × 330 nm2 is estimated to be 158 ± 13 NP/μm2. In order to determine the optimal value for the immunosensor response, the GSH-AuNPs concentration effect is investigated. For this purpose, the antibody-modified sensors are exposed to GSH-AuNPs solutions of different concentrations for 1 h by drop-casting on the surface of the sensor chips. After washing with water and blowing the surface gently with dry nitrogen, the sensor chip is measured again in the TIRE mode. Ψ and Δ functions are recorded, and the spectra represented in Figure 4 (panels a and b). ΔEΨmin in Figure 4c represents the shift of the resonance energy after the GSH-AuNPs binding. As can be observed, this shift shows a linear increase as the concentration increases up to a 40 nM concentration and is related to the strong variation of the refractive index as GSH-AuNPs attach to the surface and the rearrangement of the antibody layer. By contrast, the highest concentration employed (80 nM) presented a strong attenuation and broadening of the resonance, as can be observed in Figure 4 (panels a and b). This behavior is likely related to a reduction in the polariton amplitude caused by an inhomogeneous optical coupling with the AuNP resonance. Mock et al. have also observed the attenuation of the SPR band in Au films in the presence of AuNPs placed at a nanometric distance by inserting a separation dielectric layer.41 From a practical viewpoint, a 40 nM concentration is chosen as optimal for carrying out the immunosensing reaction, since it provides the largest shift of the resonance energy without signal attenuation.

For the sake of simplicity, the organic layer is represented by a medium with a constant refractive index of 1.5 and a null extinction coefficient.4a It is worth noticing that the DTSPSAM effect is neglected, since its refractive index barely modifies the spectral position of the resonance. In the new fivelayer model (labeled as model B in Table 1), the thickness of the uniform bioorganic layer is varied to fit the resonance energy in the TIRE spectrum, yielding a best value of 3.8 ± 0.2 nm. If the thickness of the EMA Au/antibody layer is also considered by weighting the contribution of the antibody, a total layer thickness of 5.9 nm is obtained. This value is within the range of 4−6 nm, reported in atomic force microscopy studies for immobilized IgG molecules40 and is adequate to provide effective optical coupling between the AuNPs LSPR and the surface polariton during antibody−antigen recognition. The antibody immobilization process is also studied in real time using a flow-cell to pump the anti-GSH solution, as described in Experimental Section. Since the refractive index of water is higher, the resonance is found at larger angles and lower energies than in air. Thus, the kinetic study is performed at 1.5 eV and an external angle of 85°. The formation of the antibody layer is monitored by studying the Δ variation over time. The use of this function is motivated by its higher sensitivity in comparison to Ψ.7b As the antibody is immobilized on the surface, the Δ value increases monotonously up to reaching a steady level after 25 min (Figure 3). This reaction time confirms the strong binding affinity of the antibody to the DTSP-functionalized Au surface. Immunosensing Response. The kinetic of GSH-AuNPs binding process to the anti-GSH functionalized sensor is also investigated in real time (Figure 3). Once the sensor is functionalized by anti-GSH and any unbound antibody is removed by flowing a physiological solution, a GSH-AuNPs

Figure 3. Increment of Δ function (δΔ) measured in situ during the immobilization of the antibody and GSH-AuNPs, recorded at 1.5 eV. 4973

dx.doi.org/10.1021/ac5005212 | Anal. Chem. 2014, 86, 4969−4976

Analytical Chemistry

Article

Figure 4. (a and b) TIRE-SPR spectra collected in air for Ψ and Δ functions, respectively. Curves represent the spectrum obtained for the immobilized antibody (black □) and for the immobilized GSH-AuNPs on this antibody layer at different concentrations, namely 10 nM (red ○), 20 nM (green △), 40 nM (blue ▽) and 80 nM (purple ◇). Energy shift (ΔEΨmin) caused by (c) immobilization of GSH-AuNPs at different concentrations. (d) Calibration curve of the SPR shifts as a function of the free GSH concentration in mixed solutions of GSH-AuNPs and free GSH; relative SPR shift is defined as the ratio between the energy shift caused by the mixed solution and the shift caused by a 40 nM GSH-AuNPs control solution; error bars represent the standard deviation of three measurements.

In order to check the sensitivity of the TIRE resonance technique to the detection of GSH, the sensor chip is also exposed to a GSH solution without NPs. In this case, as expected for low molecular weight molecules, the resonance energy does not give any change (data not shown), demonstrating the effective amplification of the interaction between the GSH and the antibody layer by the use of GSHAuNPs in the ellipsometric setup. On the basis of the above results, an immunoassay for glutathione detection is developed based on the competition between free GSH and GSH-AuNPs for the active sites of the antibody. The immunosensor chip is exposed for 1 h to solutions containing a constant concentration of GSH-AuNPs mixed with increasing concentrations of free GSH from 10 to 100 nM. Once the surface is washed with water and blown with nitrogen, TIRE measurements are carried out. Relative SPR shift is defined as the ratio between the energy shift caused by the mixed solution and the shift caused by a 40 nM GSHAuNPs control solution. Since the free GSH competes with the GSH-AuNPs for the antibody sites, the resonance shift decreases as the GSH concentration increases. Figure 4d shows the calibration curve obtained. Each point is the average of three measurements. Numerical results are also summarized in Table S1 of the Supporting Information. Linear fitting of the experimental data yield a linear correlation in the range from 6 to 100 nM (r = 0.993). Detection (LD = 3 × SD/|slope|) and determination limits (LQ = 10 × SD/|slope|) are calculated to be 6 and 18 nM, respectively, for a standard deviation (SD) determined from three measurements of the blank (40 nM GSH-AuNP sample). In terms of interferences, these could arise from AuNP desorption from the immunosensing surface, causing a signal too large for the true amount of free GSH in the sample. However, it is expected that those are neglected during the analysis time due to the strength of the binding between anti-

GSH and GSH-AuNPs, besides to the high stability of the glutathione monolayer-protected gold clusters.42 The reproducibility of immunosensors is evaluated from the relative standard deviation (RSD) of the response obtained for 30 nM GSH concentration with 3 different sensors prepared in the same manner. A RSD value of 5% is obtained. Therefore, the present immunosensor can successfully detect the GSH with higher or similar sensitivity in comparison with other methods reported in the literature. In addition, this ellipsometric method could be extended to the detection of other low molecular weight molecules by using the suitable antibody and analyte-functionalized AuNPs.



CONCLUSIONS The feasibility of GSH immunosensing based on TIRE detection enhanced by using AuNPs modified with GSH has been demonstrated. Sensor chips are fabricated from AZO/Au bilayers on transparent substrates and Au layer is functionalized with a SAM of antibodies against GSH. Since free GSH causes a negligible effect on the resonance, a competitive immunoassay is conceived based on the use of different concentrations of free GSH mixed with a fixed amount of GSH-AuNPs. A linear relationship between the free GSH concentration and the energy shift of the resonance is found, allowing high-sensitivity detection of GSH at concentrations in the nanomolar range.



ASSOCIATED CONTENT

S Supporting Information *

Additional experimental procedures; TEM micrograph of citrate-stabilized gold nanoparticles (Figure S1); absorbance and FT-IR spectra of the as-synthesized Cit-AuNPs and GSHAuNPs (Figure S2); comparison of the simulated Δ function around the resonance energy for different intermediate layers in the sensor (Figure S3); voltammogram of reductive thiol desorption of the Au layer modified with a DTSP SAM (Figure 4974

dx.doi.org/10.1021/ac5005212 | Anal. Chem. 2014, 86, 4969−4976

Analytical Chemistry

Article

J.; Williams, D. E. Biosens. Bioelectron. 2012, 36, 250−256. (c) Nabok, A. V.; Tsargorodskaya, A.; Hassan, A. K.; Starodub, N. F. Appl. Surf. Sci. 2005, 246, 381−386. (7) (a) Arwin, H.; Poksinski, M.; Johansen, K. Appl. Opt. 2004, 3028−3036. (b) Moirangthem, R. S.; Chang, Y.-C.; Hsu, S.-H.; Wei, P.-K. Biosens. Bioelectron. 2010, 25, 2633−2638. (c) Ü stündağ, Z.; Cağlayan, M. O.; Güzel, R.; Pişkin, E.; Solak, A. O. Analyst 2011, 136, 1464−1471. (8) (a) Zamborini, F. P.; Bao, L.; Dasari, R. Anal. Chem. 2012, 84, 541−576. (b) Lyon, L. A.; Musick, M. D.; Natan, M. J. Anal. Chem. 1998, 70, 5177−5183. (9) (a) Pingarrón, J. M.; Yáñez-Sedeño, P.; González-Cortés, A. Electrochim. Acta 2008, 53, 5848−5866. (b) Daniel, M.-C.; Astruc, D. Chem. Rev. 2004, 104, 294−346. (c) Zhang, J.; Xuab, X.; Yang, X. Analyst 2012, 137, 1556−1558. (10) (a) Honga, X.; Hall, E. A. H. Analyst 2012, 137, 4712−4719. (b) Cao, X.; Ye, Y.; Liu, S. Anal. Biochem. 2011, 417, 1−16. (c) Moirangthem, R. S.; Chang, Y.-C.; Wei, P.-K. Biomed. Opt. Express 2011, 2, 2569−2576. (11) (a) Meister, A. Science 1983, 220, 472−477. (b) Corazza, A.; H, I.; Sadler, P. J. Eur. J. Biochem. 1996, 236, 697−705. (c) Hedley, D. W.; Chow, S. Cytometry 1994, 15, 349−358. (12) (a) Sies, H. Free Radical Biol. Med. 1999, 27, 916−921. (b) Meister, A.; Anderson, M. E. Annu. Rev. Biochem. 1983, 52, 711− 760. (13) Gutscher, M.; Pauleau, A. L.; Marty, L.; Brach, T.; Wabnitz, G. H.; Samstag, Y.; Meyer, A. J.; Dick, T. P. Nat. Methods 2008, 5, 553− 559. (14) (a) Buttke, T. M.; Sandstrom, P. A. Immunol. Today 1994, 15, 7−10. (b) MacMicking, J.; Xie, Q. W.; Nathan, C. Annu. Rev. Immunol. 1997, 15, 323−350. (15) Appenzeller-Herzog, C. J. Cell. Sci. 2011, 124, 847−855. (16) May, M. J.; Vernoux, T.; Leaver, T.; Van Montagu, M.; Inzé, D. J. Exp. Bot. 1998, 49, 649−667. (17) Pocernich, C. B.; Butterfield, D. A. Biochim. Biophys. Acta 2012, 1822, 625−630. (18) Lusini, L.; Tripodi, S. A.; Rossi, R.; Giannerini, F.; Giustarini, D.; Vecchio, M. T. d.; Barbanti, G.; Cintorino, M.; Tosi, P.; Simplicio, P. D. Int. J. Cancer 2001, 91, 55−59. (19) (a) Ruan, E. A.; Rao, S.; Bun-lick, J. S.; Stryker, S. J.; Telford, G. L.; Otterson, M. F.; Opara, E. C.; Koch, T. R. Nutr. Res. (N.Y.) 1997, 17, 463−473. (b) Rahman, I.; Biswas, S. K.; Jimenez, L. A.; Torres, M.; Forman, H. J. Antioxid. Redox Signal. 2005, 7, 42−59. (20) Galera, R. M. L.; Giménez, J. C. J.; Ronsano, J. B. M.; Cardona, R. M. S.; Via, M. A. A.; Roca, C. A.; Puigbert, J. M. T. Clin. Chim. Acta 1996, 254, 63−72. (21) (a) Mayatepek, E.; Jaeken, J. Inborn Metabolic Diseases 2012, 423−430. (b) Meister, A. J. Biol. Chem. 1988, 263, 17205−17208. (22) Meyer, A. J.; May, M. J.; Fricker, M. Plant J. 2001, 27, 67−78. (23) (a) Hwang, C.; Sinskey, A. J.; Lodish, H. F. Science 1992, 257, 1496−1502. (b) Rabenstein, D. L.; Millis, K. K.; Weaver, K. H. J. Org. Chem. 1993, 58, 4144−4146. (24) (a) Miao, P.; Liu, L.; Nie, Y.; Li, G. Biosens. Bioelectron. 2009, 24, 3347−3351. (b) White, P. C.; Lawrence, N. S.; Davis, J.; Compton, R. G. Electroanalysis 2002, 14, 89−98. (c) Ndamanisha, J. C.; Bai, J.; Qi, B.; Guo, L. Anal. Chem. 2009, 386, 79−84. (d) Stobiecka, M.; Deeb, J.; Hepel, M. Electrochemical Society Transactions 2009, 19, 15−32. (25) (a) Lu, C.; Zu, Y.; Yam, V. W. W. J. Chromatogr., A 2007, 1163, 328−332. (b) Zhang, W.; Wan, F.; Zhu, W.; Xu, H.; Ye, X.; Cheng, R.; Jin, L.-T. J. Chromatogr., B 2005, 818, 227−232. (c) Vacek, J.; Klejdus, B.; Petrlová, J.; Lojková, L.; Kubá, V. Analyst 2006, 131, 1167−1174. (26) Kataoka, H.; Takagi, K.; Makita, M. Biomed. Chromatogr. 1995, 9, 85−89. (27) Gerdon, A. E.; Wright, D. W.; Cliffel, D. E. Anal. Chem. 2005, 77, 304−310. (28) Xu, H.; Hepel, M. Anal. Chem. 2011, 83, 813−819. (29) Pau, J. L.; Abad, J. M.; Hernández, M. J.; Cervera, M.; Ruiz, E.; Nuñez, C. G.; Lorenzo, E.; Piqueras, J. Investigation of Surface Plasmon Resonance in Au Nanoparticles Deposited on ZnO:Al Thin

S4); SEM images of GSH-AuNPs binding on the anti-GSHfunctionalized sensor (Figure S5); and resonance energy shifts obtained in the competitive immunoassay (Table S1). This material is available free of charge via the Internet at http:// pubs.acs.org.



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Tel. +34914974488. Fax: +34914974931. *E-mail: [email protected]. Tel. +34914978607. Fax: +34914974895. Author Contributions ∥

A. G.-M. and J. M. A. contributed equally. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work has been supported by DGUI of Comunidad Autónoma de Madrid (CAM), UAM (project N° CCG10UAM/MAT-5731), CAM (project N° S2009/PPQ-1642, AVANSENS), TEC2010-20796, BIO2012-34835, and UAMBanco Santander projects. J.M.A. and J.L.P. acknowledge research funding from the Spanish Ministry of Economy and Competitiveness through the “Ramón y Cajal” program. The authors thank M. J. Hernández and M. Cervera for their fruitful discussions. We are also grateful to Prof. V. M. Fernández, Dr. M. Pita, and Dr. A. L. de Lacey (ICP-CSIC) for providing FTIR measurement facilities.



ABBREVIATIONS GSH, reduced glutathione; TIRE, total internal reflection ellipsometry; AuNPs, gold nanoparticles; Cit-AuNPs, citratestabilized gold nanoparticles; AZO, Al-doped ZnO; EMA, effective medium approximation; DTSP, 3,3′-Dithiodipropionic acid di(N-hydroxysuccinimide ester); SAM, Self-assembled monolayer; SPP, surface plasmon polariton; FE-SEM, field emission scanning electron microscopy; LSPR, localized surface plasmon resonance



REFERENCES

(1) (a) Hoa, X. D.; Kirk, A. G.; Tabrizian, M. Biosens. Bioelectron. 2007, 23, 151−160. (b) Homola, J. Anal. Bioanal. Chem. 2003, 377, 528−539. (c) Jain, P. K.; Huang, X.; El-Sayed, I. H.; El-Sayed, M. A. Plasmonics 2007, 2, 107−118. (2) (a) Lee, J.-H.; Kim, B.-C.; Oh, B.-K.; Choi, J.-W. Nanomedicine 2013, 9, 1018−1026. (b) Souto, D. E. P.; Silva, J. V.; Martins, H. R.; Reis, A. B.; Luz, R. C. S.; Kubota, L. T.; Damos, F. S. Biosens. Bioelectron. 2013, 46, 22−29. (3) (a) Ashley, J.; Li, S. F. Y. Biosens. Bioelectron. 2013, 48, 126−131. (b) Nanduri, V.; Bhunia, A. K.; Tu, S.-I.; Paoli, G. C.; Brewster, J. D. Biosens. Bioelectron. 2007, 23, 248−252. (c) Estevez, M.-C.; Belenguer, J.; Gomez-Montes, S.; Miralles, J.; Escuela, A. M.; Montoyad, A.; Lechuga, L. M. Analyst 2012, 137, 5659−5665. (4) (a) Nabok, A.; Tsargorodskaya, A.; Mustafa, M. K.; Székács, I.; Starodub, N. F.; Székács, A. Sens. Actuators., B 2011, 154, 232−237. (b) Mauriz, E.; Calle, A.; Montoya, A.; Lechuga, L. M. Talanta 2006, 69, 359−364. (5) Tanaka, Y.; Yatabe, R.; Nagatomo, K.; Onodera, T.; Matsumoto, K.; Toko, K. IEEE Sensors J. 2013, 13, 4452−4458. (6) (a) Poksinski, M.; Arwin, H. Thin Solid Films 2004, 455−456, 716−721. (b) Le, N. C. H.; Gubala, V.; Clancy, E.; Barry, T.; Smith, T. 4975

dx.doi.org/10.1021/ac5005212 | Anal. Chem. 2014, 86, 4969−4976

Analytical Chemistry

Article

Films. In Proceedings of the 8th Spanish Conference on Electron Devices, CDE’2011, Palma de Mallorca, Illes Balears, Feb 8−11, 2011. (30) Darder, M.; Takada, K.; Pariente, F.; Lorenzo, E.; Abruña, H. D. Anal. Chem. 1999, 71, 5530−5537. (31) Wagner, P.; Hegner, M.; Kernen, P.; Zaugg, F.; Semenza, G. Biophys. J. 1996, 70, 2052−2066. (32) Turkevich, J.; Stevenson, P. C.; Hillier, J. Discuss. Faraday Soc. 1951, 11, 55−75. (33) Templeton, A. C.; Pietron, J. J.; Murray, R. W.; Mulvaney, P. J. Phys. Chem. B 2000, 104, 564−570. (34) (a) Mulvaney, P. Langmuir 1996, 12, 788−800. (b) Garcia, M. A.; Venta, J. d. l.; Crespo, P.; LLopis, J.; Penadés, S.; Fernández, A.; Hernando, A. Phys. Rev. B 2005, 72, 241403. (35) (a) Qian, W.; Krimm, S. Biopolymers 1994, 34, 1377−1394. (b) Picquart, M.; Grajcar, L.; Baron, M. H.; Abedinzadeh, Z. Biospectroscopy 1999, 5, 328−337. (36) Chang, C.-C.; Chiu, N.-F.; Lin, D. S.; Chu-Su, Y.; Liang, Y.-H.; Lin, C.-W. Anal. Chem. 2010, 82, 1207−1212. (37) Hernandez, M. J.; Garrido, J.; Piqueras, J. J. Vac. Sci. Technol., B 1994, 12, 581−584. (38) Aspnes, D. E.; Kinsbron, E.; Bacon, D. D. Phys. Rev. B 1980, 21, 3290−3299. (39) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Chem. Rev. 2005, 105, 1103−1169. (40) Browning-Kelley, M. E.; Wadu-Mesthrige, K.; Hari, V.; Liu, G. Y. Langmuir 1997, 13, 343−350. (41) Mock, J. J.; Hill, R. T.; Degiron, A.; Zauscher, S.; Chilkoti, A.; Smith, D. R. Nano Lett. 2008, 8, 2245−2252. (42) Schaaff, T. G.; Knight, G.; Shafigullin, M. N.; Borkman, R. F.; Whetten, R. L. J. Phys. Chem. B 1998, 102, 10643−10646.

4976

dx.doi.org/10.1021/ac5005212 | Anal. Chem. 2014, 86, 4969−4976