Biosensors and Bioelectronics 44 (2013) 101–107

Contents lists available at SciVerse ScienceDirect

Biosensors and Bioelectronics

0956-56

http://d

n Corr

E-m

journal homepage: www.elsevier.com/locate/bios

Multiplex electrochemiluminescence immunoassay of two tumormarkers using multicolor quantum dots as labels and grapheneas conducting bridge

Zhiyong Guo n, Tingting Hao, Shuping Du, Beibei Chen, Zebo Wang, Xing Li, Sui Wang

Faculty of Materials Science and Chemical Engineering, Ningbo University, Ningbo, Zhejiang 315211, PR China

a r t i c l e i n f o

Article history:

Received 11 November 2012

Received in revised form

20 December 2012

Accepted 11 January 2013Available online 23 January 2013

Keywords:

Multiplex electrochemiluminescence

immunoassay

Tumor markers

Quantum dots

Graphene

Chitosan

63/$ - see front matter & 2013 Elsevier B.V. A

x.doi.org/10.1016/j.bios.2013.01.025

esponding author. Tel: þ86 574 8760 0798.

ail addresses: nbuguo@163.com, guozhiyong@

a b s t r a c t

A multiplex electrochemiluminescence (ECL) immunoassay for simultaneous determination of two

different tumor markers, alpha-fetoprotein (AFP) and carcinoembryonic antigen (CEA), using multicolor

quantum dots as labels and graphene as conducting bridge was developed. Herein, a typical sandwich

immune complex was constructed on the glass carbon electrode, with QDs525 and QDs625 labeled on

secondary anti-AFP and anti-CEA antibodies, respectively, thus to obtain distinguishable ECL signals.

Because most of those QDs labeled on secondary antibodies were beyond the space domain of the ECL

reaction, graphene was used as a conducting bridge to promote the electron transfer between QDs and

the electrode, leading to about 30-fold enhancement of the ECL intensity. Experimental results revealed

that the multiplex electrochemiluminescence immunoassay enabled the simultaneous monitoring of

AFP and CEA in a single run with a working range of 0.001–0.1 pg/mL. The detection limit (LOD) for

both analytes at 0.4 fg/mL was very low. No obvious cross-reactivity was found. Precision, recovery and

stability were satisfactory. This novel multiplex ECL immunoassay provided a simple, sensitive, specific

and reliable alternative for the simultaneous detection of tumor markers in clinical laboratory.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

Tumor markers are some kind of molecules in blood or tissuesand their determination enables early diagnosis of cancer(Freedland, 2011; Wickstrom et al., 2011). Carcinogenesis is acomplicated process and their diagnosis is hindered by the fact thatmost cancers have more than one marker associated with theirincidence and some tumor markers are associated with many typesof cancer, thus the measurement of a single tumor marker is usuallynot sufficient to diagnose a particular cancer (Kelloff et al., 1996; Wuet al., 2012). Compared to parallel single immune-detection, multi-plex immunoassay enables the simultaneous detection of two ormore target proteins in a single assay (Su et al., 2005; Wilson andNie, 2006) with remarkable advantages, including high samplethroughput, improved assay efficiency, shorter assay time, lowsample requirement and reduced cost per assay (Fu et al., 2007;Song et al., 2004). Therefore, developing multiplex immunoassaysfor several tumor markers have attracted much attention with afocus on improving the diagnostic efficiency and accuracy.

Recently, various immunoassays and immunosensors have beenreported to realize multiplex immunoassays of tumor markers.

ll rights reserved.

nbu.edu.cn (Z. Guo).

Among them, analytical techniques such as electrochemistry (Konget al., 2013; Lai et al., 2009, 2011a, 2011b; Qian et al., 2011; Tanget al., 2011), fluorescence (Cao et al., 2011; Hu et al., 2011; Tianet al., 2012), chemiluminescence (Fu et al., 2008; Pei et al., 2010;Yang et al., 2010), electrochemiluminescence (Wu et al., 2012) andsome others (Chon et al., 2011; Hwang et al., 2010; Lee et al., 2010)were the most widely employed. In addition, nano materials such asfunctionalized carbon nanotubes (CNTs) (Lai et al., 2009, 2011b),silica nanosphere (Lai, et al., 2011a; Pei et al., 2010) nanogold hollowmicrospheres (GHS) (Tang et al., 2011) and quantum dots (Cao et al.,2011; Kong et al., 2013; Qian et al., 2011; Tian et al., 2012) were themost used tags to distinguish different tumor markers. However,most of these methods are heterogeneous immunoassay, in whichthe quality of the separation of the free form from the immuno-complex would greatly affect the accuracy of the assay and samplecomposition. In addition, some methods require expensive arraydetectors and often suffer from potential cross-talk (Shi et al., 2006;Wilson and Nie 2006). Finally, the sensitivity of most methods werenot high enough to accomplish the early diagnosis of cancer sincethe detection limits being in the magnitude order of ng/mL–pg/mL.Therefore, it is still a great challenge to develop more simple,sensitive and accurate multiplex immunoassays for tumor markers.

Electrochemiluminescence (ECL), a new method developed inrecent years, has been widely used in immunoassay due to itsversatility, simplicity, rapidity, stability, high sensitivity and low

Z. Guo et al. / Biosensors and Bioelectronics 44 (2013) 101–107102

background (Chi et al., 2006). Quantum dots (QDs), a popularnanostructured material with numerous advantageous featuressuch as broad excitation spectra for multicolor imaging, robustand narrowband emissions, high quantum yield, and feasibilityfor surface modification, have been widely used as ECL andluminescence labels for bioassay and bioimaging (Carroll-Portillo et al., 2009; Green, 2004; Klostranec and Chan, 2006; Liet al., 2009; Wang et al., 2008). Furthermore, all colors of QDs canbe excited by a single excitation source, light or electricity thatmakes it an ideal candidate than the traditional dyes to realizemultiplex detection (Liu et al., 2010).

Graphene (G), a ‘‘rising star’’ material, is a single layer of carbonatoms in a closely packed honeycomb two-dimensional lattice. Ithas recently attracted enormous attention in constructing biosen-sors due to its novel properties such as excellent biocompatibility,good mechanical strength, zero band gap, high carrier mobility,large specific surface area and outstanding electrical conductivity(Guo and Dong 2011; Lee et al., 2008; Li et al., 2008, 2011).Chitosan (CS), a natural polysaccharide biopolymer with amino andhydroxy groups, was often used to disperse QDs and provide anexcellent biocompatible microenvironment to construct biosensorsdue to its good water permeability and excellent film-formingability (Kang et al., 2009; Liu et al., 2012; Zhang and Gorski, 2005).The G–CS composite can accelerate the electron transfer on theelectrode surface to amplify the ECL signal due to the outstandingelectrical conductivity (Yin et al., 2010).

In this article, we report a multicolor quantum dots-based ECLmethod for simple, sensitive and selective detection of alpha-fetoprotein (AFP) and carcinoembryonic antigen (CEA) in humanserum and saliva. QDs labeled on secondary antibodies producedECL reaction after the immunoreaction and the intensity of ECLwas amplified about 30-fold using G as a conducting bridge. Thequantity of AFP and CEA was indicated by the ECL responses ofQDs525 and QDs625, respectively. The detection limit for both theanalytes using the developed ECL immunosensor was 0.4 fg/mL,much lower than the previously reported values (Kong et al., 2013;Tang et al., 2011). Thus, this work provided a promising platformfor a sensitive multiplex determination in clinical immunoassays.

2. Experimental

2.1. Chemicals and solutions

Standard grade antigens of AFP and CEA (Ag), primary anti-AFP1

and anti-CEA1 (Ab1), primary secondary anti-AFP2-biotin antibodiesand anti-CEA2-biotin antibodies (Ab2) were purchased from Key-BioCo. Ltd. (Beijing, China). Two kinds of streptavidin-QDs (SA-QDs,CdSe/ZnS core-shell structure, �10 nm diameter) with an initialconcentration of about 1 mM, SA-QDs525 (maximum emission wave-length lem¼525 nm) and SA-QDs625 (maximum emission wave-length lem¼625 nm), were purchased from Wuhan JiayuanQuantum Dots Co. Ltd. (Wuhan, China). Graphite powder(8000 mesh, 99.95%) was obtained from Aladdin (Shanghai, China),while chitosan (CS), glutaric dialdehyde (GLD) and bovine serumalbumin (BSA, 98–99%) were purchased from Sigma-Aldrich (St.Louis, MO). Potassium permanganate (KMnO4, 98 wt%), sulfuric acid(H2SO4, 98 wt%), phosphorus pentoxide (P2O5, 99 wt%), hydrogenperoxide (H2O2, 30 wt%), hydrochloric acid (HCl, 37 wt%) andpotassium peroxydisulfate (K2S2O8, 99 wt%) were obtained fromShanghai Reagent Company (Shanghai, China). All other reagentswere of analytical grade and were used as received. Ultrapurewater obtained from Millipore water purification system(Z18 MO, Milli-Q, Millipore) was used throughout the experiment.

Phosphate buffer solutions (PBS) were prepared with Na2H-PO4 �12H2O and NaH2PO4 �2H2O, and the pH of PBS was 7.4 unless

otherwise stated. A 0.1 M PBS, containing 0.1 M K2S2O8 and 0.1 MKCl served as the electrolyte in ECL analysis.

2.2. Apparatus

A laboratory-built ECL detection system was used. All electro-chemical experiments were performed with a CHI 660D electro-chemical workstation (Shanghai, China) using a conventionalthree-electrode system. A glassy carbon electrode (GCE,F¼3 mm) was used as the working electrode, with Ag/AgCl andplatinum wire acting as the reference electrode and the counterelectrode, respectively. The spectral width of the photomultipliertube (PMT) was 200–800 nm and the voltage was 800 V duringthe detection. Two band-pass filters (525750 nm and625750 nm, central wavelength7semi-bandwidth), obtainedfrom Giai Photnics Co., Ltd. (Shenzhen, China), were placedbetween the ECL cell and the photomultiplier tube as requiredto separate the ECL of different wavelengths emitted by the twokinds of QDs. Hitachi SU-70 scanning electron microscope (SEM,Tokyo, Japan) was used to characterize the sensor.

2.3. Preparation of G–CS composite

The G was synthesized as described by Li et al. (2011), whilethe G–CS composite solutions were prepared as reported by Yinet al. (2010). In short, 0.5 wt% of CS solution was prepared bydissolving CS powder in 1.0% (v/v) acetic acid solution withstirring for 1 h at room temperature and adjusting the pH to5.0 with NaOH solution. Then, 1 mg of G was added into a certainvolume of 0.5 wt% CS solution, ultrasonicated for 2 h, and filteredusing a 0.45 mm Millex-HA syringe filter unit (Millipore). Finally,homogeneous and stable G–CS composite solutions at differentconcentrations were prepared.

2.4. Fabrication of colloidal anti-AFP2/QDs525 and

anti-CEA2/QDs625 conjugates

As shown in Scheme 1A, 100 mL of 0.5 mM SA-QDs525 in 0.01 MPBS was mixed with 100 mL of 250 mg/mL anti-AFP2-biotin in0.01 M PBS, followed by incubation at 4 1C for 2 h and then five-fold diluted with 0.01 M PBS. Due to the specific binding betweenstreptavidin and biotin, anti-AFP2-biotin and SA-QDs525 wereconjugated. To block non-specific binding sites, 100 mL of 2 wt%BSA was added into the solution obtained and incubated at 4 1Cfor another 4 h. The resulting conjugates were separated bycentrifugation and resuspended in 200 mL of 0.01 M PBS, anddenoted as anti-AFP2/QDs525. Similarly, anti-CEA2/QDs625 wasobtained. Finally, a mixture solution was obtained by mixinganti-AFP2/QDs525 solution and anti-CEA2/QDs625 solution at avolume ratio of 1:1 for the ECL detection.

2.5. Preparation of the ECL immunosensor

Prior to the fabrication of the ECL immunosensor, a 3 mmdiameter GCE was polished successively with 1.0, 0.3 and 0.05 mma-Al2O3 powder; cleaned ultrasonically with ethanol and water;rinsed thoroughly with water and allowed to dry at roomtemperature. Then, 10 mL of 0.25 mg/mL G–CS composite solutionwas coated onto the working electrode and dried in the air. Afteractivating with 2.5% GLD solution for 2 h and washing with 0.01 MPBS, the electrode was rinsed with water and dipped in themixture solution containing 0.5 mg/mL anti-AFP1 and 0.5 mg/mLanti-CEA1 at 37 1C for 12 h. Finally, the immunosensor wasobtained after incubating in 20 mL of 2 wt% BSA at 37 1C for 1.5 hto block non-specific binding sites and rinsing with 0.01 M PBS.

Scheme 1. Schematic representation of (A) the preparation of colloidal anti-AFP2/QDs525 and anti-CEA2/QDs625 conjugates, and (B) fabricating steps of the ECL

immunosensor.

Z. Guo et al. / Biosensors and Bioelectronics 44 (2013) 101–107 103

Scheme 1B outlines the detailed fabrication process of the ECLimmunosensor, which included the formation of the G–CS com-posite film on GCE, the linkage of GLD with the film, theimmobilization of the first antibodies on the electrode by GLDand the specific immunoreaction.

2.6. ECL detection

As shown in Scheme 1B, the ECL immunosensor was incubatedin 40 mL of a sample solution containing AFP and CEA antigens for1 h at 37 1C, thoroughly washed with 0.01 M PBS to removeunbound AFP and CEA, and then incubated in 40 mL of the mixturesolution containing anti-AFP2/QDs525 and anti-CEA2/QDs625 asdescribed in Section 2.4 at 37 1C for 1 h to form the final sandwichimmunocomplex anti-AFP1/AFP/anti-AFP2/QDs525 and anti-CEA1/CEA/anti-CEA2/QDs625. Finally, 10 mL of 0.5 mg/mL G–CS compo-site solution was coated, using G as conducting bridge whichensured the ECL of the QDs labeled on the secondary antibodies.The electrodes, in 0.1 M PBS at pH 7.4 containing 0.1 M K2S2O8

and 0.1 M KCl were scanned from 0 to �1.6 V with a scan rate of100 mV/s. With the help of the two band-pass filters above-mentioned, the ECL signals related to the AFP (525 nm) and CEA(625 nm) concentrations were measured.

3. Results and discussion

3.1. Characteristics

3.1.1. ECL behavior

The ECL behaviors of the immunosensor were recorded step bystep in 0.1 M PBS containing 0.1 M K2S2O8 and 0.1 M KCl. Asshown in Fig. 1A, the ECL signal of the GCE electrode coatedby G–CS composite solution was not found (curve a). When anti-

AFP1 and anti-CEA1 were immobilized onto the electrode (curveb) and BSA was used to block the nonspecific binding sites, theimmunoreaction occurred (curve c), however, the ECL intensitywas still very weak due to the absence of luminophore and thepresence of the antigen–antibody complex and the BSA protein onthe electrode, which usually acted as an inert electron and mass-transfer blocking layer. After incubating in colloidal anti-AFP2/QDs525 and anti-CEA2/QDs625 conjugates solution, a smallincrease in the ECL intensity was observed (curve d). Theseobservations indicate that most of those QDs labeled on thesecondary antibodies were so far away from the electrode surfacethat they were beyond the space domain of the ECL reaction (Guoet al., 2012), and G coated on the electrode surface promoted theelectron transfer between the QDs and the electrode in a certaindegree. Therefore, more G would result in higher ECL intensity. Itwas verified by using higher concentration of G–CS compositesolution and an enhancement of the ECL intensity was obtained(curve e). However, concentration of G–CS composite solutioncoated on the electrode surface could not increase unlimitedly;else the stability of the ECL immunosensor would reduceobviously with the decrease of the adhesive ability of G–CScomposite solution. To further improve the electron transferbetween the QDs and the electrode, and the consequent enhance-ment of the ECL signal, a layer of G–CS composite solution couldbe added after the final sandwich immunocomplex was formed.In this case, G would act as a conducting bridge between QDs andthe electrode, which could eliminate the inert electron and mass-transfer blocking layer of Ab1–Ag–Ab2-biotin–streptavidin–QDseffectively and thus amplify the ECL signal by 20–30 times (curvef). Finally, when the ECL immunosensor was prepared withsecondary antibodies unlabeled by QDs and the sandwich immu-nocomplex was formed exactly as before, no ECL signal could befound (curve g). It proved that the ECL signals were originatedfrom the QDs again.

Fig. 1. (A) ECL–potential curves of (a) 0.25 mg/mL G–CS, (b) (a)þanti-AFP1þanti-

CEA1, (c) (b)þ0.05 pg/mL AFPþ0.05 pg/mL CEAþBSA, (d) (c)þanti-AFP2-biotin/

SA-QDs525þanti-CEA2-biotin/SA-QDs625, (e) same as (d) except G–CS coated on

the electrode surface was 0.5 mg/mL and (f) (d) with G as conducting bridge and

(g) (d) without QDs labeled. Concentration of all solutions used was as described

in Sections 2.5. and 2.6. unless stated otherwise. Experimental conditions: GC

electrode in 0.1 M PBS at pH 7.4 containing 0.1 M KCl and 0.1 M K2S2O8. The

voltage of the photomultiplier tube was set at 800 V. Scan rate: 100 mV/s.

(B) Representative SEM images of the sandwich ECL immunosensor with G as

conducting bridge. (C) Electrochemical impedance spectroscopy (EIS) of

(a) 0.5 wt% CS, (b) 0.25 mg/mL G–CS, (c) (b)þanti-AFP1þanti-CEA1, (d)

(c)þ0.05 pg/mL AFPþ0.05 pg/mL CEAþBSA, (e) (d)þanti-AFP2-biotin/SA-

QDs525þanti-CEA2-biotin/SA-QDs625 and (f) (e) with G as conducting bridge.

Concentration of all solutions used was as described in Sections 2.5. and 2.6.

unless stated otherwise. Experimental conditions: GC electrode in 0.01 M PBS

(5 mM Fe(CN)64�/3�

þ0.1 M KCl, pH 7.4). The frequency range was between 0.01

and 100,000 Hz with a signal amplitude of 5 mV.

Z. Guo et al. / Biosensors and Bioelectronics 44 (2013) 101–107104

3.1.2. SEM image

The surface topography of the composite-formed film wasinvestigated using scanning electronic microscope (SEM) and theimage is presented in Fig. 1B. After G–CS composite wasassembled onto the surface of sandwich immunocomplex anti-AFP1/AFP/anti-AFP2/QDs525 and anti-CEA1/CEA/anti-CEA2/QDs625,

many QDs were coated by thin films of G with the typicalcrumpled and wrinkled structure. By this way, G would act as aconducting bridge between QDs and the electrode to increase theelectron transfer significantly, thus enhancing the ECL intensityeffectively.

3.1.3. EIS behavior

Electrochemical impedance spectroscopy (EIS) is an effectivemethod to monitor the interface properties of the electrodes inthe assembly process and was widely used to characterize thefabrication process of the ECL immunosensor. EIS of the immu-nosensor was measured in 0.01 M PBS containing 5.0 mM[Fe(CN)6]3�/4� and 0.1 M KCl. Fig. 1C shows the impedancespectroscopy of stepwise assembly procedures. Results revealedthat G–CS modified electrode (curve b) showed a lower chargetransfer resistance, Ret, than the CS modified one (curve a),implying that G–CS composites possessed better electrical con-ductivity and were able to accelerate the electron transfer. Afteranti-AFP1 and anti-CEA1 were subsequently conjugated (curve c),Ret increased due to the increase in the thickness of the insulatingfilm on the substrate. Similarly, AFP and CEA antigens (curve d),and anti-AFP2/QDs525 and anti-CEA2/QDs625 (curve e) significantlyresisted the electron transfer kinetics of the redox probe at theelectrode interface, resulting in gradual increase in impedance ofthe electrode. However, a great decrease of Ret was observed afterG–CS composite film was formed (curve f) further confirming thatthe G–CS coated onto the surface of the sandwich immunocom-plex could act as a conducting bridge.

3.2. Optimization of experimental conditions

In order to establish optimal conditions for ECL detection of AFPand CEA, density of G as a conducting bridge, pH of the supportingelectrolyte, incubation temperature and incubation time wereinvestigated systematically in the presence of 0.05 pg/mL AFP and0.05 pg/mL CEA.

The density of G as a conducting bridge highly influences theperformance of the ECL. The conducting effect and the consequentthe ECL intensity would be low if the density of G is low; however,it could not be too high, because the ECL emitted might besheltered by excess G. Fig. 2A illustrates the ECL intensity versusdifferent concentration of G–CS solution (0.1 mg/mL, 0.3 mg/mL,0.5 mg/mL, 0.7 mg/mL, 1.0 mg/mL). When the density of Gincreased, the ECL intensity also increased at the up to aconcentration of 0.5 mg/mL; after that the ECL intensitydecreased. Hence, the density of 0.5 mg/mL of G was selectedfor the following optimization.

The pH of the solution could greatly affect ECL behavior of theimmunosensor because the activity of the immobilized proteinmay be influenced by the acidity of the solution. Fig. 2B shows theeffect of pH of the detection solution on the ECL responses of theimmunosensor. It could be seen that ECL intensity increased withthe increase of pH from 6.0 to 7.4 and decreased thereafter (up topH 8.0). The results indicate that the optimal pH was 7.4. Highlyacidic or alkaline environment would damage the immobilizedprotein; therefore, the ECL detection was performed in pH 7.4 PBScontaining 0.1 M K2S2O8 and 0.1 M KCl.

In the sandwich-type immunoassays, incubation temperatureand incubation time for the antigen–antibody interaction greatlyinfluence the analytical performance of immunoassays. Fig. 2Cshows the effect of incubation temperature, from 20 to 45 1C, onthe performance of the immunosensor; and the maximumimmunoreaction was observed at 37 1C. As shown in Fig. 2D, withincreasing incubation time, the ECL intensity for AFP and CEAincreased and reached a plateau after 50 min, which indicated the

Fig. 2. Effect of (A) concentration of G–CS composite as conducting bridge, (B) pH, (C) incubation temperature and (D) incubation time on the ECL intensity of the ECL

immunosensor toward 0.05 pg/mL AFP and 0.05 pg/mL CEA.

Z. Guo et al. / Biosensors and Bioelectronics 44 (2013) 101–107 105

saturation of the sandwich immunocomplex formation. Therefore,37 1C and 50 min were selected as the optimum incubationtemperature and time, respectively, for the subsequent study.

3.3. Analytical performance

To demonstrate the simultaneous quantitative immunoassaysof AFP and CEA, measurements were carried out after incubationwith various analyte compositions under the optimum condi-tions. As indicated in inset (1) of Fig. 3, the ECL signals ofimmunosensor, under consecutive potential scans from 0 to�1.6 V for 16 cycles, were high and stable (RSD was 1.1% forAFP and 1.4% for CEA), suggesting that the sensor is suitable forECL detection.

As presented in Fig. 3, the response ECL intensity increasedwith an increase in the concentration of AFP and CEA (0.001–0.1 pg/mL, from bottom to top). Both calibration plots shows goodlinear relationship between the ECL intensity (y) and the analyteconcentrations (cAFP and cCEA) in the range of 0.001–0.1 pg/mL(n¼6) for both AFP (Fig. 3A) and CEA (Fig. 3B). The regressionequations were y¼148.71þ52118.14� cAFP (pg/mL) with a cor-relation coefficient r of 0.991 for AFP and y¼106.02þ62169.04� cCEA (pg/mL) with a correlation coefficient r of 0.994for CEA. The detection limit for both AFP and CEA was 0.4 fg/mL,which was lower than the limits reported in simultaneousdetermination method for tumor markers (Kong et al., 2013;Tang et al., 2011). The excellent analytical performance, includingvery low detection limits and wide linear ranges over two orders

of magnitude for multiplex analysis of two analytes, was verysignificant for practical applications.

3.4. Specificity, reproducibility and stability of the immunosensor

Simultaneous multianalyte detection must exclude cross-reactivity between analytes. For the evaluation of cross-reactivity,the immunosensor array was incubated with solution containingeither AFP or CEA only. As expected, only the immunosensorprepared with corresponding antibody showed an obvious ECLresponse and cross reactivities were not observed. Further, thesimultaneous addition of AFP and CEA resulted in the simulta-neous increase of ECL intensity at specific wavelengths, and theECL intensities were as similar as that when AFP or CEA waspresent alone. These indicated that the cross-reactivity betweenthe two analytes was negligible. Thus, simultaneous multianalyteimmunoassay could be achieved in a single run using theproposed method.

To investigate the reproducibility of analysis, we assayedsamples containing 0.05 pg/mL AFP and 0.05 pg/mL CEA with fivesimilarly prepared immunosensors. The relative standard devia-tions (RSD) of the measurements obtained were 8.3% for AFP and4.7% for CEA, indicating the excellent precision and reproduci-bility of the immunosensor.

The stability of the immunosensor stored in 0.01 M PBS (pH7.4) containing 0.1% NaN3 at 4 1C was investigated by periodicalchecking of its relative activity. The ECL intensity for the detectionof 0.05 pg/mL was 89.7% and 93.6% of the initial value for AFP and

Fig. 3. (A) ECL profiles of the immunosensor in the presence (a–g) of different

concentrations of AFP in pH 7.4 PBS containing 0.1 M KCl and 0.1 M K2S2O8. AFP

concentration (pg/mL): (a) 0.001, (b) 0.002, (c) 0.005, (d) 0.01, (e) 0.02, (f) 0.05 and

(g) 0.1. The detection wavelength was set at 525750 nm. The insets: (1) ECL

emission from the immunosensor for the determination of 0.05 pg/mL AFP under

continuous cyclic voltammetry for 16 cycles, and (2) calibration curve for AFP

determination. (B) ECL profiles of the immunosensor in the presence (a–g) of

different concentrations of CEA in pH 7.4 PBS containing 0.1 M KCl and 0.1 M

K2S2O8. CEA concentration (pg/mL): (a) 0.001, (b) 0.002, (c) 0.005, (d) 0.01,

(e) 0.02, (f) 0.05 and (g) 0.1. The detection wavelength was set at 625750 nm.

The insets: (1) ECL emission from the immunosensor for the determination of

0.05 pg/mL CEA under continuous cyclic voltammetry for 16 cycles, and (2) cali-

bration curve for CEA determination.

Table 1Recovery tests for AFP and CEA in spiked human serum and saliva samples. (x7s,

n¼3).

Samples Added (pg/mL) Found (pg/mL) Recovery (%)

AFP CEA AFP CEA AFP CEA

Serum 1 0.002 0.002 0.002470.0005 0.002370.0005 120 115

Serum 2 0.02 0.02 0.01870.002 0.01970.002 90 95

Serum 3 0.08 0.08 0.08470.007 0.08370.008 105 104

Saliva 1 0.002 0.002 0.002270.0004 0.001970.0004 110 95

Saliva 2 0.02 0.02 0.01970.002 0.02170.002 95 105

Saliva 3 0.08 0.08 0.07570.006 0.08370.007 94 104

Z. Guo et al. / Biosensors and Bioelectronics 44 (2013) 101–107106

CEA, after a storage period of two weeks. Thus, the immunosensorhas acceptable storage stability.

3.5. Application in analysis of serum and saliva samples

The elevated concentration of tumor marker in serum andsaliva may be an early indication of certain cancer. Therefore, it isnecessary and important to develop a method with high sensi-tivity and selectivity for the determination of low levels of tumormarkers in serum and saliva. To evaluate the reliability andapplication potential of the proposed multianalyte immunosen-sor, a series of samples were prepared by spiking AFP and CEA in

different concentrations to the human serum and saliva samplesand the results are presented in Table 1. The recoveries were in anacceptable range from 90 to 120% for AFP and from 95 to 115% forCEA, indicating that the proposed method has good accuracy inthe sample matrix.

4. Conclusions

In this work, a novel highly sensitive multiplex ECL immu-noassay for the simultaneous detection of AFP and CEA wasdeveloped using QDs as trace tag and G as a conducting bridge.Through a sandwich immunoreaction and G conducting bridge,QDs525 and QDs625 labels were coupled to electrode surface. ECLanalysis of the coupled QDs525 and QDs625 was used to quantifythe concentration of AFP and CEA. This immunosensor showedexcellent performance for simultaneous detection AFP of and CEAwith wide linear ranges, low detection limits, good specificity andacceptable accuracy. Therefore, the proposed method providespotential opportunities of multianalyte determination for clinicaldiagnostics and other biosensor applications.

Acknowledgments

Financial supports from the National Natural Science Founda-tion of China (Nos. 81273130, 81072336), Science and TechnologyDepartment of Zhejiang Province of China (2012R405061,2011R405051), Ningbo Science and Technology Bureau(2011C50037) and Scientific Research Foundation of GraduateSchool of Ningbo University are gratefully acknowledged. Thiswork was also sponsored by K.C. Wong Magna Fund in NingboUniversity.

References

Cao, Z.J., Li, H., Lau, C., Zhang, Y.H., 2011. Analytica Chimica Acta 698, 44–50.Carroll-Portillo, A., Bachand, M., Greene, A.C., Bachand, G.D., 2009. Small 5,

1835–1840.Chi, Y.W., Duan, J.P., Lin, S.D., Chen, G.N., 2006. Analytical Chemistry 78,

1568–1573.Chon, H., Lee, S., Yoon, S.Y., Chang, S.I., Lim, D.W., Choo, J., 2011. Chemical

Communications 47, 12515–12517.Freedland, S.J., 2011. Cancer 117, 1123–1135.Fu, Z.F., Yan, F., Liu, H., Yang, Z.J., Ju, H.X., 2008. Biosensors and Bioelectronics 23,

1063–1069.Fu, Z.F., Yang, Z.J., Tang, J.H., Liu, H., Yan, F., Ju, H.X., 2007. Analytical Chemistry 79,

7376–7382.Green, M., 2004. Angewandte Chemie International Edition 43, 4129–4131.Guo, S.J., Dong, S.J., 2011. Journal of Materials Chemistry 21, 18503–18516.Guo, Z.Y., Hao, T.T., Wang, S., Gan, N., Li, X., Wei, D.Y., 2012. Electrochemistry

Communications 14, 13–16.Hu, W.H., Liu, Y.S., Yang, H.B., Zhou, X.Q., Li, C.M., 2011. Biosensors and

Bioelectronics 26, 3683–3687.Hwang, H., Chon, H., Choo, J., Park, J.K., 2010. Analytical Chemistry 82, 7603–7610.Kang, X.H., Wang, J., Wu, H., Aksay, I.A., Liu, J., Lin, Y.H., 2009. Biosensors and

Bioelectronics 25, 901–905.Kelloff, G.J., Boone, C.W., Crowell, J.A., Nayfield, S.G., Hawk, E., Malone, W.F., Steele,

V.E., Lubet, R.A., Sigman, C.C., 1996. Journal of Cellular Biochemistry 25, 1–14.Klostranec, J.M., Chan, W.C.W., 2006. Advanced Materials 18, 1953–1964.Kong, F.Y., Xu, B.Y., Xu, J.J., Chen, H.Y., 2013. Biosensors and Bioelectronics 39,

177–182.Lai, G.S., Wu, J., Ju, H.X., Yan, F., 2011a. Advanced Functional Materials 21,

2938–2943.Lai, G.S., Wu, J., Leng, C., Ju, H.X., Yan, F., 2011b. Biosensors and Bioelectronics 26,

3782–3787.Lai, G.S., Yan, F., Ju, H.X., 2009. Analytical Chemistry 81, 9730–9736.Lee, C.G., Wei, X.D., Kysar, J.W., Hone, J., 2008. Science 321, 385–388.Lee, H.J., Lee, S.H., Yasukawa, T., Ramon-Azcon, J., Mizutani, F., Ino, K., Shiku, H.,

Matsue, T., 2010. Talanta 81, 657–663.Li, D., Muller, M.B., Gilje, S., Kaner, R.B., Wallace, G.G., 2008. Nature Nanotechnol-

ogy 3, 101–105.Li, L.L., Liu, K.P., Yang, G.H., Wang, C.M., Zhang, J.R., Zhu, J.J., 2011. Advanced

Functional Materials 21, 869–878.

Z. Guo et al. / Biosensors and Bioelectronics 44 (2013) 101–107 107

Li, J., Zhao, X.W., Zhao, Y.J., Gu, Z.Z., 2009. Chemical Communications 17,2329–2331.

Liu, B., Lian, H.T., Yin, J.F., Sun, X.Y., 2012. Electrochimica Acta 75, 108–114.Liu, J., Lau, S.K., Varma, V.A., Kairdolf, B.A., Nie, S.M., 2010. Analytical Chemistry 82,

6237–6243.Pei, X.P., Chen, B.A., Li, L., Gao, F., Jiang, Z., 2010. Analyst 135, 177–181.Qian, J., Dai, H.C., Pan, X.H., Liu, S.Q., 2011. Biosensors and Bioelectronics 28,

314–319.Shi, M.H., Peng, Y.Y., Zhou, J., Liu, B.H., Huang, Y.P., Kong, J.L., 2006. Biosensors and

Bioelectronics 21, 2210–2216.Song, S.P., Li, B., Hu, J., Li, M.Q., 2004. Analytica Chimica Acta 510, 147–152.Su, X., Zhang, J.W., Sun, L., Koo, T.W., Chan, S., Sundararajan, N., Yamakawa, M.,

Berlin, A.A., 2005. Nano Letters 5, 49–54.Tang, J., Tang, D.P., Niessner, R., Chen, G.N., Knopp, D., 2011. Analytical Chemistry

83, 5407–5414.

Tian, J.N., Zhou, L.J., Zhao, Y.C., Wang, Y., Peng, Y., Zhao, S.L., 2012. Talanta 92,72–77.

Wang, J., Liu, G.D., Wu, H., Lin, Y.H., 2008. Small 4, 82–86.Wickstrom, M., Larsson, R., Nygren, P., Gullbo, J., 2011. Cancer Science 102,

501–508.Wilson, M.S., Nie, W.Y., 2006. Analytical Chemistry 78, 6476–6483.Wu, M.S., Shi, H.W., He, L.J., Xu, J.J., Chen, H.Y., 2012. Analytiacl Chemistry 84,

4207–4213.Yang, Z.J., Zong, C., Yan, F., Ju, H.X., 2010. Talanta 82, 1462–1467.Yin, H.S., Ma, Q., Zhou, Y.L., Ai, S.Y., Zhu, L.S., 2010. Electrochimica Acta 55,

7102–7108.Zhang, M., Gorski, W., 2005. Journal of the American Chemical Society 127,

2058–2059.