Author’s Accepted Manuscript An ultrasensitive sandwich type electrochemiluminescence immunosensor for triiodothyronine detection using silver nanoparticledecorated graphene oxide as a nanocarrier Hung-Tao Chou, Chien-Yu Fu, Chi-Young Lee, Nyan-Hwa Tai, Hwan-You Chang www.elsevier.com/locate/bios

PII: DOI: Reference:

S0956-5663(15)30063-4 http://dx.doi.org/10.1016/j.bios.2015.04.060 BIOS7628

To appear in: Biosensors and Bioelectronic Received date: 4 February 2015 Revised date: 19 April 2015 Accepted date: 20 April 2015 Cite this article as: Hung-Tao Chou, Chien-Yu Fu, Chi-Young Lee, Nyan-Hwa Tai and Hwan-You Chang, An ultrasensitive sandwich type electrochemiluminescence immunosensor for triiodothyronine detection using silver nanoparticle-decorated graphene oxide as a nanocarrier, Biosensors and Bioelectronic, http://dx.doi.org/10.1016/j.bios.2015.04.060 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

An ultrasensitive sandwich type electrochemiluminescence immunosensor for triiodothyronine detection using silver nanoparticle-decorated graphene oxide as a nanocarrier Hung-Tao Choua, Chien-Yu Fub, Chi-Young Leea, Nyan-Hwa Taia,*, Hwan-You Changb,† a

b

Department of Materials Science and Engineering,

Department of Medical Science, National Tsing Hua University, Hsin Chu, 300 Taiwan *Corresponding author. Tel.: +886 3 5742568 †

Corresponding author. Tel.: +886 3 5742910

E-mail addresses: [email protected] (N.-H. Tai) [email protected] (H.-Y. Chang) Abstract An ultrasensitive electrochemiluminescence (ECL) immunosensor was constructed to detect 3,3’,5–triiodothyronine (T3). The system employed T3-conjugated, silver nanoparticle-decorated carboxylic graphene oxide (Ag@fGO-T3) as a carrier and anti-T3 antibody-tris(2,2’-bipyridyl) ruthenium(II) (Ru(bpy)32+) as a probe. The Ag@fGO-T3 and Ru(bpy)32+ complex could be

mobilized rapidly to the anode in the reaction chamber through electrophoresis. The fGO is reduced electrochemically at the electrode, and the electrons could transfer from an anode to the Ru(bpy)32+. The complex is excited at the electrode and an ECL signal is produced upon reacting with tripropylamine (TPrA). Because of its large surface area and excellent conductivity, Ag@fGO could enhance ECL signal significantly in the system. Quantitative measurement of T3 could be achieved in the range from 0.1 pg/mL to 0.8 ng/mLwith a detection limit of 0.05 pg/mL. In addition, the novel immunosensor showed good specificity in the presence of serum, indicating its high potential in clinical use. Keywords Electrochemiluminescence; Sandwich immunosensor; Graphene oxide; Ru(bpy)32+/TPrA; Silver nanoparticle Introduction Electrochemiluminescence (ECL) analysis, combining the advantages of both electrochemical and chemiluminescence analyses, has received much attention due to its high sensitivity, low background signals, simple optical setup, ease of control, low reagent cost, and versatility for a variety of analytes. ECL technology has been applied in several areas, including medical diagnosis, environmental surveillance, food and water hygiene monitoring, as well as toxic agent detection.

A typical ECL reaction is consist of two components: chemical complex includes a transitional metal such as ruthenium, osmium, platinum, and palladium for light production and a co-reactant that recycles the transitional metal complex. The commercially available tris(2,2’-bipyridyl) ruthenium(II) (Ru(bpy)32+)/tripropylamine (TPrA) system and its derivatives are the favorite system for ECL assay due to their chemical stability, long excited state lifetimes, and high luminescence efficiency in a wide range of pH values. The working mechanism of ECL analysis involves the production of reactive intermediates (Ru(bpy)33+) from the stable metal complex (Ru(bpy)32+) used as precursors at the surface of the anode. The excited-state intermediate then transfers the energy to co-reactant (TPrA) and return to the ground state, accompanied with light emission. By integrating highly specific antibodies in the system, ECL immunoassays have become a powerful clinical diagnosis tool for early stage disease detection (Miao and Bard, 2004; Jin et al., 2013). The advancement of nanotechnology has further improved the sensitivity of ECL assays. Because of their large specific surface area, nanomaterials are frequently adopted as carriers for loading a large quantity of sensing molecules to increase electrocatalytic activity or amplify the electrochemical signals (Yu et al., 2013; Qian et al., 2010; Song et al., 2013; Xu et al., 2011; Bottini et al., 2006). In addition, carbon nanomaterials, such as nanotubes, spheres, and graphene display excellent electrical

conductivity that facilitates electron transfer at the electrode interface, an essential step of ECL. Furthermore, grafting functional groups on carbon nanomaterials is relatively straightforward that improves the versatility of the materials. As a result, carbon nanomaterials have drawn much attention in biomedical studies. For example, Xu and colleagues achieved enhanced sensitivity using a carbon nanodot@Ag hybrid material in an ECL immunoassay for cancer cell detection (Song et al., 2013). Bottini’s group greatly improved the sensitivity of an ECL immunosensor using a single-wall carbon nanotube modified electrode coupled with silica nanoparticles (NP) loaded with Ru(bpy)32+ (Bottini et al., 2006). Among various carbon nanomaterials, graphene oxide (GO), a starting material for the preparation of graphene sheets, is of particular interest. The material contains abundant oxygen-containing functional groups, such as epoxide, hydroxyl, and carboxylic groups that have been conveniently employed for modification and used in many biomedical applications. It has been shown previously that GO could have a protein loading ratio more than twice of that of other nanocarriers (Zhang et al., 2010; Li et al., 2013) . In addition, GO has high chemical stability, a property essential for the use in immunoassay systems (Razmi and Mohammad-Rezaei, 2013; Shiddiky et al., 2012; Srivastava et al., 2013; Xu et al., 2011; Lai et al., 2013). Nevertheless, currently GO is primarily used in electrochemical immunoassays. Very few studies

have taken the advantages of GO and used it in ECL systems. In this work, we describe a novel competitive sandwich type immunoassay using GO as a sensing molecule carrier to detect 3,3’,5-triiodothyronine (T3), the most widely used marker of thyroid function (Liao et al., 2013). Three major improvements over the conventional ECL assay were made in the currently reported system. First, the functionalized GO provides high loadings of anti-T3 antibody. Second, the functionalized GO were decorated with silver nanoparticle to improve the conductivity of the nanomaterial. Third, electrophoresis was used to rapidly concentrate the sensing complex at the anode while coupling with the ECL reaction (An et al., 2010; Chavez-Valdez et al., 2013; Dong et al., 2011). As a result, this novel ECL system demonstrates excellent properties in biosensing. 1. Experimental Suppliers of the raw materials and chemicals used in this study are shown in the Supplementary section. 1.1 Preparation of silver nanoparticle decorated GO Preparation of GO functionalized with carboxylic groups was performed as described previously (Zhang et al., 2010; Li et al., 2013) and the procedure is shown in the Supplementary section. A room temperature one-pot process to decorate silver nanoparticles on GO, developed previously in our laboratory (Y. Li et al., 2013), was

adopted in this study. Briefly, functionalized GO was dispersed in ethanol and was subjected to sonication for 1 h, mixed with 10.0 mM silver nitrate/ethanol solution, followed by sonication again for 1 h. After centrifugation and washing to remove impurities, the product (Ag@fGO) was kept in ethanol to prevent oxidation and was kept at 4˚C. The topology of Ag@fGO was examined using an atomic force microscope (AFM, Veeco Nanoscope 3100, Japan) and a field-emission scanning electron microscope (SEM, JSM-6500, JEOL, Japan). Energy dispersive X-ray analysis (EDX, model 7418, UK) was also carried out to evaluate the Ag nanoparticle loading on GO. Functional groups on GO and fGO were analyzed by Fourier Transform Infrared Spectroscopy (FTIR, Spectrum RX, Perkin-Elmer). 1.2 Preparation of Ag@fGO-T3 To produce a GO-based nanocarrier for T3, this study first conjugated streptavidin (SA) to Ag@fGO following the method described previously (Shiddiky et al., 2012). In short, 1.0 mg Ag@fGO was dispersed in 10 mL of 0.1 M 2-(N-morpholino) ethanesulfonic acid (MES) buffer, then 10 mg of N-hydroxysuccinimide (NHS) and 2.0 mg of N-(3-Dimethylaminopropyl) -N’-ethylcarbodiimide HCl (EDC) were added sequentially into the solution. The intermediate product Ag@fGO-O-acylisourea was collected by centrifugation, washed with MES buffer, mixed with the SA solution (0.04 g/mL in PBS) at 25˚C for 2 h, followed by stirring for overnight in a

refrigerator. The modified Ag@fGO was collected by centrifugation, and the successful binding of SA on the GO was verified using biotin (5-fluorescein) as a probe and quantified with a fluorometer (Wallace Victor2 1420 Multilabel Counter, Perkin-Elmer). Preparation of T3 conjugated Ag@fGO was carried out by incubating 250 g Ag@fGO-SA with biotinylated T3 (0.276 ng/mL) under stirring for overnight at 4˚C. The reaction product, Ag@fGO-T3, was collected by centrifugation at 12,000 rpm for 5 min, re-dispersed in 1.0 mL PBS, and stored at 4˚C prior to use. 1.3 Detection of T3 using Ag@fGO-T3 This study designed a competitive sandwich immunoassay for the detection of T3. The T3 sample to be tested was first incubated with anti-T3 antibody-Ru(bpy)32+ solutions (75 ng/mL) at room temperature for 20 min. Then, Ag@fGO-T3 was added into the T3/ anti-T3 antibody- Ru(bpy)32+ mixture and incubated for 1 h. The Ag@fGO-T3 was collected by centrifugation and washed thoroughly with PBS. The precipitate was resuspended in PBS containing 5.0 mM TPrA, then transferred to a transparent cell customized for ECL assay. The ECL cell was equipped a three-electrode system with an Ag/AgCl reference electrode and two platinum electrodes (Pt electrode, Ф = 0.8 mm), served as working and counter electrode, respectively. The electrodes were connected to an Autolab PGSTAT30 & FRA2

electrochemistry workstation which provides triangular potential scan for ECL measurement. The applied potential was in the range of 0.2 - 1.6 V performing with a scan rate of 100 mV/s in both electrophoresis and ECL detection. The low operation potential can prevent Ag@fGO-T3/T3/ anti-T3 antibody-Ru(bpy)32+ complex from damage or separating, providing stable conditions for ECL detection. The cell was placed in a luminescence analyzer (TD-20/20, Turner Designs) for ECL measurement. In this study thyroxine (T4), bovine serum albumin (BSA,10 pg/mL), and fetal bovine serum (FBS, 1%) were individually included in the ECL assay to determine the specificity of the assay. 2. Results and Discussion 2.1 Acting mechanism of the Ag@fGO-based ECL immunosensor The strategy of the ECL biosensor developed in this study is shown in Fig. 1. First, a GO-based nanocarrier for T3 is produced by modifying GO with carboxylic groups, streptavidin, and T3 sequentially. This nanocarrier competes with the T3 in testing sample for anti-T3 antibody-Ru(bpy)32+. Then, the GO nanocarrier/T3/ anti-T3 antibody-Ru(bpy)32+ complex is mobilized to the anode and produced ECL signals under a DC electric field. The assay is based on a competitive test principle: the higher concentration of T3 in the testing sample, less GO nanocarrier/T3/ anti-T3 antibody-Ru(bpy)32+ complex could be formed, and thus lower ECL signals.

The ECL process could be expressed as following:

Ag @ fGO  T3 / T3 / anti  T3 antibody  Ru  bpy 3 - e   2

Ag @ fGO  T3 / T3 / anti  T3 antibody  Ru  bpy 3

3

TPrA - e  TPrA 



 TPrA  + H 

Ag @ fGO  T3 / T3 / anti  T3 antibody  Ru  bpy 3 +TPrA   3

Ag @ fGO  T3 / T3 / anti  T3 antibody  Ru  bpy 3  + products 2

Ag @ fGO  T3 / T3 / anti  T3 antibody  Ru  bpy 3   2

Ag @ fGO  T3 / T3 / anti  T3 antibody  Ru  bpy 3

2

+ h

Fig. 1. Schematic diagram of the ECL immunosensor developed in this study. (a) Functionalization of graphene oxide (GO) nanosheets with carboxylic groups. (b) Decoration of silver nanoparticles on functionalized GO (fGO). (c) Conjugation of streptavidin on Ag@fGO using the EDC/NHS strategy, then immobilization of

biotinylated T3 on Ag@fGO through binding on streptavidin. (d) Addition in the assay mixture containing anti-T3 antibody- Ru(bpy)32+, which binds free T3 in serum or Ag@fGO-T3 in a competitive manner. (e) Mobilization of Ag@fGO-T3 nanoprobes/anti-T3 antibody- Ru(bpy)32+ complex to the anode through electrophoresis by applying a voltage. (f) Oxidation of Ag@fGO-T3 nanoprobes/anti-T3 antibody- Ru(bpy)32+ to Ag@fGO-T3 nanoprobes/anti-T3 antibody- Ru(bpy)33+ at anode. The excited compound then reacts with co-reactant TPrA, accompanied with light production. The signal is captured by a luminometer. The use of GO in this design offers several advantages. First, GO has a large surface area and therefore can accommodate a large quantity of sensing molecules, thus increases the sensitivity of the assay. Second, the carboxylated form of GO possesses high negative surface charges which allows it move to the anode rapidly under a DC electric field. Third, its good electric conductivity could enhance the efficiency of the ECL assay, in which transfer of electrons from an anode to the transition metal complex is required. This study also decorated Ag nanoparticles (AgNP) on GO to further improve the conductivity of the nanocarrier. 2.2 Characterization of Ag@fGO and Ag@fGO-T3 nanocarrier This study adopted a simple procedure to synthesize GO functionalized with carboxylic groups for the use of conjugating amino group-containing molecules via carbodiimide reaction. To examine the successful functionalization of GO, FTIR analysis was performed. The FTIR spectrum of the modified GO sample, but not that

of untreated GO, displays a peak at 1384 cm−1 indicating the presence of carboxylic groups (-COO-, acid salt form) (Fig. 2a). The C-OH peaks, including those of ester, hydroxyl, and epoxide groups in the range from 1080 to 1120 cm-1 were diminished in modified GO, reflecting that these functional groups have been converted to carboxylic groups. The functionalization process reduced the size of GO, from original 200 nm - 5 m to approximately 300 nm (Fig. 2b). The homogeneous size of functionalized GO can improve test reproducibility in the subsequent analysis. This work employed a facile process to decorate Ag nanoparticles (AgNP) onto the fGO surface (Y. Li et al., 2013). Fig. 2c depicts the SEM images of the topology of fGO after AgNP decoration. Uniform deposition of AgNPs on fGO could be clearly observed. Further analysis of AgNP decorated fGO sample with EDX has revealed that approximately 2.3 wt% of AgNPs was loaded onto the fGO (Fig. 2c inset). The high abundance of AgNP loading is beneficial for signal amplification and sensing in the ECL immunoassay proposed in this study. The height and phase of the AgNP on fGO were measured using atomic force microscopy (Fig. 2 d-f) and the result indicates that AgNP are distributed uniformly on fGO and their height and width are mostly smaller than 20 nm and 10 nm, respectively. Together, these data demonstrate that the carboxylic functionalization process is successful and results in a large number of active sites for AgNP nucleation and growth.

This study employed a two-step process to prepare T3-conjugated fGO for the competitive ECL assay. First, the EDC/NHS coupling protocol was used to immobilize SA on fGO. Then biotinylated T3 was added to form Ag@fGO-T3 via binding to SA. This study utilized biotin (5-fluorescein) as a probe to evaluate the degree of SA conjugation on Ag@fGO. Results shown in Fig.S1 indicate that biotin (5-fluorescein) binds strongly to Ag@fGO-SA, but not to either Ag@fGO or Ag@fGO-O-acylisourea. The significant increase in fluorescence intensity of Ag@fGO-SA suggests that a large number of binding sites for biotinylated T3 immobilization are available on Ag@fGO-SA. Our experimental design demands that the Ag@fGO-T3 is electrophoresed to the anode where ECL could be induced. In this regard, the surface charge of Ag@fGO-T3 is a crucial factor determining the success of the approach. As shown in Fig. S2, the surface charge of Ag@fGO-T3 is about -45.1 V, suggesting that it would move to anode at a low voltage DC electric field. 2.3 ECL behavior of the immunosensors In the absence of free T3 molecule, ECL signal could be clearly detected employing fGO-T3 as a carrier, although it decreased rapidly (Fig. S3, red curve). In, contrast, the signal intensity under the same assay condition using Ag@fGO-T3 as the carrier (Fig. S3, black curve) was twice as high as that of fGO-T3. The ECL signal

enhancement effect of Ag@fGO-T3 is presumably attributed to AgNP’s surface plasmon resonance, which creates strong local electric fields to enhance the ECL response of the Ru(bpy)32+ (Y. T. Li et al., 2013; Wan et al., 2011; Wu et al., 2013; Zhang et al., 2014). The surface plasmon resonance effect of AgNP could also result in several luminescence peaks, although most of them were overlapped with the ECL spectrum.

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 2. Characterization of nanomaterials used in this study. (a) The FTIR spectra of GO and fGO. (b) The zeta potential measurement for the estimation of the lateral size of fGO (black line) and GO (red line). (c) FE-SEM images of Ag@fGO. (d) AFM image of Ag@fGO for height and topology measurement. (e) AFM image of Ag@fGO for phase analysis. (f) Analysis of diameter of AgNP on fGO using AFM.

2.4 ECL behavior of the immunosensors In the absence of free T3 molecule, ECL signal could be clearly detected employing fGO-T3 as a carrier, although it decreased rapidly (Fig. S3, red curve). In, contrast, the signal intensity under the same assay condition using Ag@fGO-T3 as the carrier (Fig. S3, black curve) was twice as high as that of fGO-T3. The ECL signal enhancement effect of Ag@fGO-T3 is presumably attributed to AgNP’s surface plasmon resonance, which creates strong local electric fields to enhance the ECL response of the Ru(bpy)32+ (Y. T. Li et al., 2013; Wan et al., 2011; Wu et al., 2013; Zhang et al., 2014). The surface plasmon resonance effect of AgNP could also result in several luminescence peaks, although most of them were overlapped with the ECL spectrum. In addition, the Ag@fGO-T3 assay system remained functioning for more than 20 consecutive cycles under cyclic potential scans. The reason why AgNP decoration on GO could produce a signal stabilizing effect is not clear. It is believed that the deposition of nanocarrier on the anode would increase with reaction time. While the electron transfer activity of GO accumulated at the anode would gradually decrease with time, Ag@fGO is more efficient than GO in electron transfer and hence more stable in reduction of Ru(bpy)32+. On the basis of the competitive analysis, the ECL signal of Ag@fGO-T3/anti-T3

antibody-Ru(bpy)32+ would decrease with the increase of the sample T3 concentrations. As expected, an inverse linear relationship between ECL signals and the testing T3 concentrations was obtained in the range from 0.1 pg/mL to 25 pg/mL (Fig. 3 b). The linear regression equation was y = 1301.8 x + 21.957 with a correlation coefficient of R2 = 0.9682, where y is the ECL intensity and x is the T3 concentration. The detection limit of the Ag@fGO-based ECL assay is approximately 0.05 pg/mL (Fig. S4). Judging the physiological concentration of T3 in human serum, ranging from 0.8 to 2 ng/ml, the currently developed assay is well suited for clinical diagnosis of thyroid function. In addition, the ultra-high sensitivity of the Ag@fGO-based ECL assay can be adopted and used favorably for the detection of trace biomarkers, in particular for early detection of diseases. A comparison of performance, signal amplification strategy, and detection targets between this GO-based ECL assay and other commonly used sandwich immunoassays is shown in Table 1 (Qian et al., 2014; Z. Li et al., 2013; Yang et al., 2010). Additionally, the performance of the currently reported ECL assay and other reported studies on T3 detection is shown in Table 2 (Han et al., 2013; Q. Zhang et al., 2014; Stefan et al., 2004).

(a)

(b)

Fig. 3. (a) Profiles of Ag@fGO-based ECL production with time to detect different concentrations of T3. Curves (a–j) Concentrations of T3 used in the study were: 0, 0.1, 1, 10, 25, 50, 100, 200, 400, and 800 pg/mL in PBS (pH 7.4) in the presence of 0.005 M TPrA. The scan rate was 100 mV/s and the potential range was 0.2–1.6 V. (b) The calibration curves between ECL intensity and T3 concentration. Each point is the average of 8 measurements. Even though the detection range of the ECL assay described in this study is not as wide as that of the electrochemical biosensors (Table 2), several ways can be used to improve this property. The assay curve starts to saturate when the testing T3 concentration is above 50 pg/mL (Fig. 2), indicating that little anti-T3 antibody-Ru(bpy)32+ is available for binding the Ag@fGO-T3. Therefore, addition of more anti-T3 antibody-Ru(bpy)32+ should improve the linear dynamic range of the ECL assay (Martin et al., 2008). Increasing amount of Ag@fGO-T3 did not improve the sensitivity of the assay, presumably because GO can absorb light emitted from the assay. Optimization of SA and AgNP ratio on functionalized GO, the synthesis of

both relies on the presence of carboxylic groups, is a future direction to improve the ECL assay. Table 1. Performance compared with those of other ECL sandwiched immunosensors luminophore system

test solution

linear range for target (ng/mL)

detection limit (pg/mL)

Reference

Ru(bpy)32+/ C2O42-

PBS/ C2O42-

0.05-20 AFP

3.5 pg/mL

Qian et al. (2010)

Ru(bpy)32+/ TPrA

MeCN/ TPrA

10-10000 CRP

3.3 ng/mL

Miao & Bard (2004)

Ru(bpy)32+/ TPrA

PBS/ TPrA

0.05-200 mouse IgG

17 pg/mL

Yang et al. (2010)

Ru(bpy)32+/ NADH

PBS/ NADH

5.0×10−9 5.0×10−4 M LDH

0.4 nM

Qian et al. (2014)

Ru(bpy)32+/ TPrA

PBS/ TPrA

0.05-10 SAL

17 pg/mL

Li et al. (2013)

Ru(bpy)32+/ TPrA

PBS/ PrA

0.0001-0.8 T3

0.05

Our work

2.5 Long-term performance and specificity of the immunosensor The long-term performance of the Ag@fGO-T3 immunosensor was evaluated under continuous cyclic potential scans for 400 s. As shown in Fig. 4a, the ECL intensity yielded under the standard assay condition did not change significantly and even increase slightly, in the time period from 175 s to 400 s. Some irregular signals in the beginning of test could be observed, presumably due to the different sizes and surface charges of the Ag@fGO-T3/T3/ anti-T3 antibody-Ru(bpy)32+ complex that resulted in different electrophoretic speeds (Dong et al., 2011). One of the future aims of this work is to speed up the electrophoresis process and reduce the noises.

Table 2. Performances compared with other immunosensors used to detect thyroxine. Immunosensor System

Amplification Strategy

linear range (ng/mL) and target

detection limit (pg/mL)

Reference

Electrochemiluminescence

Poly-L-lysine modified electrode as signal amplified enhancer

0.001-10 T3

0.03

Liao et al.(2013)

0.3-5.9 T3

195

Roche company

0.00005-5 T4

0.015

Han et al.(2013)

Ru(bpy)32+/ Poly-L-lysine Electrochemiluminescence Ru(bpy)32+ /TPrA (Cobas e T3/Roche )

Electrochemical immunoassay

cascade catalysis of cytochrome c and glucose oxidase used for signal amplification

Electrochemical immunoassay

Au@ macroporous CS / Fe3O4 @MWCNTs

0.000000710.00115 T4

0.0002

Zhang et al. (2014)

Sequential injection analysis immunosensor

anti-L-T(4) immobolized carbon paste use as electrode

36-1080 T4

24600

Stefan et al. (2008)

Thyroxine (T4) is a prohormone of triiodothromine. Both T3 and T4 are thyroid hormones produced by the thyroid gland, being primarily responsible for regulation of metabolism. T4 converts to the active T3 within cells by deiodinases before exhibiting its function. This study also tested whether this novel ECL assay could distinguish T3 from T4. In this regard, Ag@fGO-T3 was separately incubated with T4 at 10 ng/ mL and 25 ng/ mL in the presence of BSA (10 pg/mL). As shown in Fig. 4b, no obvious variation in ECL intensity was observed when T4 was added, showing high specificity

of the Ag@fGO-T3 ECL assay. To further demonstrate the feasibility of the Ag@fGO-T3 ECL assay in clinical use, FBS (1.0%, w/v in PBS) was used to simulate human serum in the assay. The results depicted in Fig. 5 show that addition of FBS did not produce obvious variations in the produced ECL signals, suggesting that the Ag@fGO-T3 assay system has high potential to be used for the detection of T3 in clinical samples. 2.6 Reproducibility We have tested five different electrode sets, each was used to detect 0.1, 10, 50, 200, and 800 pg/mL T3. The relative standard deviation of the measurements for each of these concentrations were 9%, 6%, 7%, 3%, and 4%, respectively, showing high reproducibility of the assay.

(a)

(b)

Fig. 4. (a) The ECL intensity at different T3 concentrations vs. time profiles during 400 s in the standard assays. Curves (a)-(j) depicts T3 concentration of 0, 0.1, 1, 10, 25, 50, 100, 200, 400, 800 pg/mL, respectively, under a scan rate of 100 mV/s in the potential range of 0.2 – 1.6 V. (b) The specificity of the developed ECL immunosensors toward different targets in 10 pg/mL BSA: Bar (a) Blank (no T3), bar (b) T3 (10 pg/mL), bar (c) a mixture containing T3 (10 pg/mL) and T4 (10 pg/mL), bar (d) T3 (25 pg/mL), bar (e) a mixture containing T3 (25 pg/mL) and T4 (25 pg/mL). *p < 0.05, ***p < 0.001.

Fig. 5. Comparison of the ECL immunosensor performance in the presence of 1% FBS for detecting 0, 10, and 25 pg/mL of T3. *p < 0.05, *** p < 0.001. 3. Conclusions This study demonstrates that Ag@fGO is an ideal nanocarrier for ECL-based immunosensing assay. According the proposed method, quantitative measurement of T3 could be achieved in the range from 0.1 pg/mL to 0.8 ng/mLwith a detection limit of 0.05 pg/mL. The assay system is highly sensitive and specific and is not interfered by serum. The assay system is inexpensive to set up and in principle can be adopted to all sandwich immunoassays. We believe that the ECL assay has great potential to be used for early detection of diseases after simple optimization. Acknowledgments This work was supported in part by Ministry of Science and Technology, Republic of China (102-2320-B-007 -007 -MY3), and National Tsing-Hua University

(101N7049E1). Appendix A. Supplementary data Supplementary data associated with this article can be found in this section. References

An, S.J., Zhu, Y., Lee, S.H., Stoller, M.D., Emilsson, T., Park, S., Velamakanni, A., An, J., Ruoff, R.S., 2010. J. Phys. Chem. Lett. 1, 1259–1263. Bottini, M., Magrini, A., Di Venere, A., Bellucci, S., Dawson, M.I., Rosato, N., Bergamaschi, A., Mustelin, T., 2006. J. Nanosci. Nanotechnol. 6, 1381–1386. Chavez-Valdez, a., Shaffer, M.S.P., Boccaccini, a. R., 2013. J. Phys. Chem. B 117, 1502–1515. Dong, R., Yan, X., Liu, J., 2011. Anal. Chem 9100–9106. Han, J., Zhuo, Y., Chai, Y., Yu, Y., Liao, N., Yuan, R., 2013. Anal. Chim. Acta 790, 24–30. Jin, G., Lu, L., Gao, X., Li, M.J., Qiu, B., Lin, Z., Yang, H., Chen, G., 2013. Electrochim. Acta 89, 13–17. Lai, Y., Bai, J., Shi, X., Zeng, Y., Xian, Y., Hou, J., Jin, L., 2013. Talanta 107, 176– 182. Li, Q., Fan, F., Wang, Y., Feng, W., Ji, P., 2013. Ind. Eng. Chem. Res. 52, 6343– 6348. Li, Y., Chen, Y., Tai, N., 2013. Langmuir 29, 8433-8439. Li, Y.T., Qu, L.L., Li, D.W., Song, Q.X., Fathi, F., Long, Y.T., 2013. Biosens. Bioelectron. 43, 94–100. Li, Z., Wang, Y., Kong, W., Li, C., Wang, Z., Fu, Z., 2013. Biosens. Bioelectron. 39, 311–314.

Liao, N., Zhuo, Y., Chai, Y.Q., Xiang, Y., Han, J., Yuan, R., 2013. Biosens. Bioelectron. 45, 189–194. Martin, H., Murray, C., Christeller, J., Mcghie, T., 2008. Anal. Biochem. 381, 107– 112. Miao, W., Bard, A.J., 2004. Anal. Chim. Acta 76, 7109–7113. Qian, J., Wang, K., Jin, Y., Yang, X., Jiang, L., Yan, Y., Dong, X., Li, H., Qiu, B., 2014. Biosens. Bioelectron. 57, 149–156. Qian, J., Zhou, Z., Cao, X., Liu Songqin, S., 2010. Anal. Chim. Acta 665, 32–38. Razmi, H., Mohammad-Rezaei, R., 2013. Biosens. Bioelectron. 41, 498–504. Shiddiky, M.J. a, Rauf, S., Kithva, P.H., Trau, M., 2012. Biosens. Bioelectron. 35, 251–257. Song, W., Li, H., Liu, H., Wu, Z., Qiang, W., Xu, D., 2013. Electrochem. commun. 31, 16–19. Srivastava, S., Kumar, V., Ali, M.A., Solanki, P.R., Srivastava, A., Sumana, G., Saxena, P.S., Joshi, A.G., Malhotra, B.D., 2013. Nanoscale 5, 3043–51. Stefan, R., Staden, J.F. Van, Aboul-enein, H.Y., 2004. Talanta 64, 151–155. Wan, Y., Wang, Y., Wu, J., Zhang, D., 2011. Anal. Chem. 83, 648–653. Wu, L., Wang, J., Ren, J., Li, W., Qu, X., 2013. Chem. Commun. (Camb). 49, 5675– 7. Xu, S., Liu, Y., Wang, T., Li, J., 2011. Anal. Chem. 83, 3817–3823. Yang, X., Yuan, R., Chai, Y., Zhuo, Y., Mao, L., Yuan, S., 2010. Biosens. Bioelectron. 25, 1851–1855. Yu, Y., Cao, Q., Zhou, M., Cui, H., 2013.. Biosens. Bioelectron. 43, 137–142. Zhang, H., Wu, M., Xu, J., Chen, H., 2014. Anal. Chem. 86, 3834–40. Zhang, J., Zhang, J., Zhang, F., Yang, H., Huang, X., Liu, H., Guo, S., 2010. Langmuir 26, 6083–6085.

Zhang, Q., Chen, X., Tu, F., Yao, C., 2014. Biosens. Bioelectron. 59, 377–383.

Highlights 1. Ag@fGO-T3 nanoprobe for T3 detection in a sandwich-type immunosensor is designed. 2. High-loading silver decoration and antigen immobilization on nanoprobe is achieved. 3. Ag@fGO-T3 nanoprobe is used in the Ru(bpy)32+/TPrA electrochemiluminescent system. 4. Ag@fGO-T3 captures the target and attaches to the electrode through electrophoresis. 5. Ag@fGO-T3 own excellent sensitivity and low detection limit in the ECL immunosensor.