Tbb

Ha

Cb

h

�

�

�

�

a

ARRAA

KESMC

0h

Analytica Chimica Acta 746 (2012) 107– 113

Contents lists available at SciVerse ScienceDirect

Analytica Chimica Acta

j ourna l ho me page: www.elsev ier .com/ locate /aca

he sandwich-type electrochemiluminescence immunosensor for �-fetoproteinased on enrichment by Fe3O4-Au magnetic nano probes and signal amplificationy CdS-Au composite nanoparticles labeled anti-AFP

ankun Zhoua, Ning Gana,∗, Tianhua Lia, Yuting Caoa, Saolin Zenga, Lei Zhengb,∗∗, Zhiyong Guoa

The State Key Laboratory Base of Novel Functional Materials and Preparation Science, Faculty of Material Science and Chemical Engineering of Ningbo University, Ningbo 315211,hinaDepartment of Laboratory Medicine, Nanfang Hospital, Southern Medical University, Guangzhou 510515, China

i g h l i g h t s

Sandwich immunoreaction, testing alarge number of samples simultane-ously.The magnetic separation and enrich-ment by Fe3O4-Au magnetic nanoprobes.The amplification of detection signalby CdS-Au composite nanoparticleslabeled anti-AFP.Almost no background signal, whichgreatly improve the sensitivity ofdetection.

g r a p h i c a l a b s t r a c t

r t i c l e i n f o

rticle history:eceived 18 May 2012eceived in revised form 9 August 2012ccepted 17 August 2012vailable online 28 August 2012

eywords:lectrochemiluminescenceandwich immunoreactionagnetic capture probes

dS-Au signal tag

a b s t r a c t

A novel and sensitive sandwich-type electrochemiluminescence (ECL) immunosensor was fabricated on aglassy carbon electrode (GCE) for ultra trace levels of �-fetoprotein (AFP) based on sandwich immunore-action strategy by enrichment using magnetic capture probes and quantum dots coated with Au shell(CdS-Au) as the signal tag. The capture probe was prepared by immobilizing the primary antibody of AFP(Ab1) on the core/shell Fe3O4-Au nanoparticles, which was first employed to capture AFP antigens toform Fe3O4-Au/Ab1/AFP complex from the serum after incubation. The product can be separated fromthe background solution through the magnetic separation. Then the CdS-Au labeled secondary antibody(Ab2) as signal tag (CdS-Au/Ab2) was conjugated successfully with Fe3O4-Au/Ab1/AFP complex to forma sandwich-type immunocomplex (Fe3O4-Au/Ab1/AFP/Ab2/CdS-Au), which can be further separated byan external magnetic field and produce ECL signals at a fixed voltage. The signal was proportional to acertain concentration range of AFP for quantification. Thus, an easy-to-use immunosensor with magnetic

probes and a quantum dots signal tag was obtained. The immunosensor performed at a level of high sen-sitivity and a broad concentration range for AFP between 0.0005 and 5.0 ng mL−1 with a detection limit of0.2 pg mL−1. The use of magnetic probes was combined with pre-concentration and separation for tracelevels of tumor markers in the serum. Due to the amplification of the signal tag, the immunosensor is

n offe

highly sensitive, which ca effective biomonitoring for clin

∗ Corresponding author. Tel.: +86 574 87609987; fax: +86 574 87609987.∗∗ Corresponding author. Tel.: +86 20 61642147; fax: +86 20 62787681.

E-mail addresses: ganning@nbu.edu.cn, ssrs20060621@163.com (N. Gan), nfyyzl@163.

003-2670/$ – see front matter © 2012 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.aca.2012.08.036

r great promise for rapid, simple, selective and cost-effective detection of

ical application.

© 2012 Elsevier B.V. All rights reserved.

com (L. Zheng).

1 imica

1

scicehm[[ctbodaib(totaUadtaiarbElok(dtotrtRwtkcewfiatpmd

cbfTrc

08 H. Zhou et al. / Analytica Ch

. Introduction

It is well known that the protein �-fetoprotein (AFP) in humanerum is clinically measured as a biomarker for hepatocellular can-er. The average concentration of AFP is approximately 25 ng mL−1

n healthy human serum but rises greatly in patients with liverancer [1]. Thus, sensitive detection of AFP levels of cancer biomark-rs plays an important role in the early detection of disease andighly reliable predictions. Until now, conventional immunoassayethods include the enzyme-linked immunosorbent assay (ELISA)

2,3], fluoroimmunoassay [4], flow injection chemiluminescence5], chemiluminescence enzyme immunoassay [6]. However, thehallenge remains for obtaining a rapid sample, sensitive detec-ion and low-cast for early and ultrasensitive screening of canceriomarkers [7]. The conventional sandwich-type immunoassay isne of the major analytical techniques for sensitive and selectiveetection of protein and has wide application in clinical diagnosisnd biomedical research [8]. It is employed as a sandwich formatn which analytes are captured and detected by an excess immo-ilized primary antibody (Ab1) and labeled secondary antibodyAb2). The general sandwich-type immunoassay protocol requireshe primary antibody’s (Ab1) immobilization of the bare electroder modified electrode, then its conjugation with the antigen byhe immunoreaction. Then, a fluorescence-labeled antibody wasssembled to construct an immunosensor for protein detection.nfortunately, long incubation time and washing steps for sep-ration of bound free antibodies and antigens are the two mainrawbacks [9]. It is easy to change the fluorescence intensity duringhe elution condition. Now, electro-immunosensors are importantnalytical tools designed to tumorous processes and tumor markersn human serum can be used in screening for a disease [10,11]. As

valuable detection method, electrochemiluminescence (ECL) haseceived considerable attention due to its versatility, low-cost, low-ackground, easy operation procedure and high sensitivity [12–14].lectrochemiluminescence immunoassay has been applied in bio-ogical detection and quantifying it combines the high sensitivityf ECL detection and the specificity of immunosensors [15]. Someinds of ECL reagents such as, Tris (2,2′-bipyridyl), ruthenium (II)Ru(bpy)3

2+) [16–18], luminol and its derivatives [19], quantumots or semiconductor nanocrystals (NCs) [20–22], have been usedo construct ECL immunosensors. However, the bioanalysis basedn these conventional luminescent reagents possesses some limi-ation. A high degree of ruthenium labeling at multiple sites mayesult in the loss of biological activity of the biomolecules [23] andhe luminol ECL system is weak in the neutral solution [14,24].ecently, quantum dots (QDs), as a new kind of ECL luminophore,ith size-tunable, optical, narrow emission spectra and broad exci-

ation spectra [25–28], have been widely used in fabricating allinds of photoluminescence probes in biological analysis. Zhu ando-workers used semiconductor nanoparticles to develop a series oflectrochemiluminescence biosensors for various biosystem assayshere the immunosensor provides a convenient specific method

or protein detection [15,20,29]. Liu’s group presents a versatilemmunosensor using a quantum dot coated silica nanosphere as

label for IgG detection [30]. Among these strategies, the quan-um dot based amplification has received special attention for itsossible outstanding optical, electronic, and biocompatible perfor-ance of fabricated electrochemiluminescence sensor for clinical

iagnosis [31].In many applications, gold nanoparticles serve as an attractive

andidate for biosensors and modified electrodes due to their goodiocompatibility and stability. They have also facilitated the trans-

er of electrons between the electrode and the biomolecules [32].here is growing interested in developing a new, advanced mate-ial for using a novel biosensor construct. Nanocomposite materialsonstitute a rapidly evolving field of science and technology. They

Acta 746 (2012) 107– 113

have attracted the interest of researchers, due to their differentnanoscale functionalities and capability of endowing the substratewith enhanced properties [33]. Actually, the core/shell compositenanoparticles Fe3O4-Au, with Au coating the magnetic nanoparti-cles, exhibits suitable intrinsic properties of the magnetic core andAu shell. It is anticipated that incorporation of Au coating on a mag-netic core could attain both the advantages of chemical stability,biocompatibility of Au and magnetic separation of Fe3O4 [34–37].

Herein, in our work, the protocol for a novel sandwichimmunoassay, with quantum dots labeled and excellent mag-netic/luminescent properties, is fabricated. For the primaryantibody and secondary antibody there was simultaneous incuba-tion with the Fe3O4-Au and CdS-Au, respectively. A sandwich-typeimmunosensor based on CdS-Au, labeled secondary antibody, wasconstructed with the following process. The core/shell magneticmaterials Fe3O4-Au was labeled as the first antibody (Fe3O4-Au/Ab1) by the Au-S band of the Au-antibody and the CdS-Aucomposite nanoparticles was labeled as the secondary antibody(CdS-Au/Ab2), which carried huge amounts of ECL signal probe.With the Fe3O4-Au/Ab1 capturing antigen onto the antibody siteby the specific bond of the antigen–antibody, the CdS-Au, labeledthe secondary antibody, could be successively conjugated with theantigen via the immunoreaction. Based on the ECL signal relatedto the concentration of detected antigen, the immunoassay couldbe accomplished successfully. Therefore, this ECL immunosensorcan effectively amplify the detected signal, avoid the inactivationof proteins and exhibit attractive sensitivity.

2. Experimental

2.1. Materials

The AFP antibody (anti-AFP, 12 mg mL−1) and antigen (AFP)were bought from Biocell Company (Zhengzhou, China). Gold chlo-ride (HAuCl4) and bovine serum albumin (BSA) was obtainedfrom Sinopharm Chemical Reagent Co. Ltd. The phosphate-bufferedsaline (PBS), 0.1 M with various pH value was prepared by mixingthe stock solution of NaH2PO4 and Na2HPO4 and then adjustingthe pH with 0.1 M NaOH and H3PO4. PBS (0.1 M, pH 7.4) containing0.1 M K2S2O8 and 0.1 M KCl was used as the electrolyte. All otherreagents were of analytical reagent grade and used without furtherpurification.

2.2. Apparatus

Electrochemiluminescence measurements were performedusing a Model MPI-B electrochemiluminescence analyzer (Xi’anRemax, China) with the voltage of the photomultiplier tube (PMT)set at 700 V in the detection process. A three-electrode configura-tion was employed consisting of a modified glassy carbon electrode(4 mm diameter) as a working electrode, a platinum wire as thecounter electrode and an Ag/AgCl (sat. KCl) reference electrode. Thetransmission electron microscope (TEM) image was performed ona TEM (H-7650, Japan) operating at an acceleration voltage of 60 kV.The XRD patterns of prepared powder samples were collected usinga Bruker AXS (D8) X-ray diffractometer with a Cu target (40 kV,40 mA). The magnetization curve was measured with a vibratingsample magnetometer (Lake Shore 7410) at room temperature.All electrochemiluminescence experiments were performed in a5.0 mL quartz cell at room temperature.

2.3. Preparation of the core/shell Fe3O4-Au nanoparticles and

CdS-Au composite nanospheres

The Fe3O4-Au nanoparticles were prepared by chemical copre-cipitation methods according to the reference [33]. The Fe3O4-Au

H. Zhou et al. / Analytica Chimica Acta 746 (2012) 107– 113 109

tion p

nfi

lF2astttfg

2n

(FP4tapaa5ttds

2

abbatpap

Scheme 1. Scheme of the prepara

anoparticles were obtained and dispersed in distilled water to anal volume 5.0 mL.

Monodisperse CdS nanospheres were prepared according to theiterature [28] and the final product was dried to yield a powder.or the preparation of gold NPs (GNPs) assembler CdS nanospheres,0 mg of CdS nanospheres were first dispersed in 2.0 mL of ethanolnd treated with 0.2 g cysteine. After stirring for 6 h, the suspen-ion was centrifuged and washed with ethanol repeatedly for threeimes, and thiol-functionalized nanoparticles were obtained. Thenhe thiol-functionalized CdS nanospheres were dispersed in a mix-ure of 1.0 mL of gold NPs, the mixed suspension was stirred at 4 ◦Cor 12 h, unbound gold NPs were removed by successive centrifu-ation and washed with water several times.

.4. Immobilization of anti-AFP on to the Fe3O4-Au and CdS-Auanoparticles (Fe3O4-Au/Ab1, CdS-Au/Ab2)

The process for the preparation of magnetic capture probesFe3O4-Au/Ab1) and signal tag of CdS-Au/Ab2 is shown Scheme 1.irst, 20 mg of Fe3O4-Au nanoparticles were reached with 1.0 mLBS (pH 7.4), 1.0 mL of 50 �g mL−1 Ab1. The solution was stirred at◦C for 24 h. Unbound Ab1 was removed by successively washing

he Fe3O4-Au nanoparticles with pH 7.4 PBS. After magnetic sep-ration and washing with PBS three times, the magnetic capturerobes (Fe3O4-Au/Ab1) were redispersed in 1.0 mL of pH 7.4 PBSnd stored at 4 ◦C for a later experiment. Simultaneously, (same asbove), 20 mg CdS-Au was added to 1.0 mL PBS (pH 7.4), 1.0 mL of0 �g mL−1 Ab2 solution and stirred at 4 ◦C for 24 h. Subsequently,he mixture was centrifuged and washed with water several timeso obtain the signal tag (CdS-Au/Ab2). Finally, the CdS-Au/Ab2 wereispersed with 0.1 M of pH 7.4 PBS to a final volume of 1 mL andtored at 4 ◦C for later usage.

.5. The sandwich immunoreaction

As shown in Scheme 1, 1.0 mL Fe3O4-Au immobilized Ab1 (A)nd CdS-Au immobilized Ab2 (B) were first prepared and incu-ated in 1.0 mL of 1% BSA at room temperature (20 ◦C) for 1 h tolock nonspecific binding sites. A concentration of AFP was firstdded in the Fe3O4-Au immobilized Ab1 and incubated at room

emperature for 1 h to obtain the Fe3O4-Au/Ab1/AFP (C). The pre-ared Fe3O4-Au/Ab1/AFP and CdS-Au/Ab2 was mixed in a pipend incubated at room temperature for 1 h to form the sandwichroducts, which were magnetically separated from the background

rocedures of the immunosensor.

solution. The product was washed with PBS (pH 7.4) for threetimes, and then re-dispersed in 50 �L of PBS (pH 7.4). As a result,the magnetic separation and quantum dots labeled sandwich-type(Fe3O4-Au/Ab1/AFP/CdS-Au/Ab2) construction was obtained (D).

2.6. Preparation of ECL immunosensor

A glassy carbon electrode (GCE) with 4 mm diameter was firstpolished carefully to a mirrorlike surface with 0.3–0.05 �m alu-mina slurry, then rinsed and ultrasonically in ethanol and distilledwater. Before modification, the bare electrode was cyclic-potentialscanned in the potential range of −0.2 to 0.6 V in 5.0 × 10−3 MK3[Fe(CN)6] solution containing 0.1 M KCl supporting electrolyteuntil a pair of well-defined redox peaks was obtained. After theelectrode was dried under nitrogen at room temperature, 10 �Lof solution (which was prepared) dropped onto the surfaced ofthe clean glassy carbon electrode, and allowed to air-dry at roomtemperature, to obtain the ECL sensor.

2.7. ECL detection

The ECL measurements of the modified electrodes above wereperformed in 5 mL of 0.1 M PBS (pH 8.0) containing 0.1 M K2S2O4and 0.1 M KCl and the potential scanned from −1.4 to 0.2 V withscan rate of 100 mV s−1. The ECL emission intensity was recordedby a MPI-B multifunctional chemiluminescence analyzer, with thePMT set at 700 V, and the AFP concentrations were measuredrelated to the ECL signals.

3. Results and discussion

3.1. Characterization of core/shell magnetic NPs

Fig. 1 shows the property characterization of the Fe3O4-AuMNPs. Fig. 1A shows the typical image of Fe3O4-Au MNPs preparedusing about 40 nm. Fig. 1B shows the vibration sample magnetome-ter (VSM) magnetization curves of Fe3O4 and Fe3O4-Au MNPs atroom temperature. It is evident that the sample presented a satu-ration magnetization (Ms) of 23.89 emu g−1 and 15.84 emu g−1 forFe3O4 and Fe3O4-Au MNPs. The saturation magnetization of the

nanoparticles reduced after coating with a layer of gold. This maybeascribed to the nonmagnetic gold layer.

The coated nano Au shell of Fe3O4-Au magnetic nano-probeswas confirmed by the X-ray diffraction (XRD) pattern. Fig. 1C shows

110 H. Zhou et al. / Analytica Chimica Acta 746 (2012) 107– 113

Fig. 1. TEM image of the Fe3O4-Au (A) MNPs; (B) magnetization measurements ofapplied field for pure Fe3O4 (a) and core/shell Fe3O4-Au MNPs (b); (C) XRD patternsof Fe3O4 (a) and Fe3O4-Au MNPs (b).

Fig. 2. ECL – potential curves of (a) the Fe3O4/GCE, (b) the Fe3O4-Au/GCE, the

CdS/GCE, and the CdS-Au/GCE in 0.1 M PBS (pH 8.0) containing 0.1 M KCl and0.1 M K2S2O8. Scan rate: 100 mV s−1, the voltage of the PMT was 700 V.

the XRD spectra of synthesized Fe3O4 (a) and core/shell Fe3O4-Au (b) MNPs. It reveals the face-centered-cubic magnetite (Fe3O4)structure (JCPDS Card No. 19-06290), and Fe3O4-Au MNPs exhib-ited diffraction peaks (at 2� = 38.25, 44.46, 64.69 and 77.72), whichwere indexed to (1 1 1), (2 0 0), (2 2 0) and (3 1 1) planes of gold cubicphase, respectively. Compared with the XRD spectra of two sam-ples, the Fe3O4-Au MNPs displayed all the diffraction peaks of pureFe3O4 MNPs; it can be infered that a thin gold was successfullycoated on the Fe3O4 core [37].

3.2. Electrochemical and ECL behaviors of CdS-Au nanospheres

Fig. S1, Supporting information shows the TEM image of theCdS (A) and CdS-Au (B) nanoparticles. According to the TEM obser-vation, the average size of the CdS nanoparticles is 50 nm andtheir size distribution is relatively uniform, and the gold NPs withthe diameter of 10 nm was successfully assembled on the surfaceof CdS. As shown in the insert of Fig. S1(B), displayed the XRDspectra of prepared CdS-Au assembler nanoparticles. It reveals thetypical diffraction peaks of CdS (at 2� = 28.20, 36.65, 43.73, 50.93and 58.34) and which also include the peaks of gold nanoparti-cles.

Fig. 2 shows the ECL intensity of the Fe3O4/GCE (curve a),Fe3O4-Au/GCE (curve b), CdS/GCE (curve c), and CdS-Au/GCE (curved) electrode in 0.1 M PBS (pH 7.4) containing 0.1 M KCl and0.05 M K2S2O8, respectively. It can be seen that the Fe3O4 MNPs(a) modified the electrode in the solution almost had no ECL sig-nal, and the ECL signal was slightly enhanced with the Fe3O4-Aumodified electrode, but still weak. Curve c and d shows the ECL-potential curves of the pure CdS NCs and CdS-Au composite film,respectively and the strong ECL signal obtained. The results indicatethe ECL signal from the reaction between CdS and S2O8

2−. Com-pared with curve c and d, it should be noted that the ECL intensityfrom the CdS-Au composite film is about 2.5-fold higher than thatobserved from the pure CdS NCs film. The reason may be that thegold particles exhibit great catalytic activity and enhanced electri-cal conductivity, which results in the occurrence of enhanced ECLsignal.

Furthermore, as shown in the Fig. 2 (curve a and b), the ECL signal

was very low, in the absence of the CdS NCs. The results indicatethat the Fe3O4 and Au could not generate the ECL signal, and thuswas generated from the CdS NCs.

H. Zhou et al. / Analytica Chimica Acta 746 (2012) 107– 113 111

Fig. 3. ECL – potential curves of (a) the bare GCE, (b) Fe3O4-Au/Ab1/AFP, (c) Fe3O4-Acr

3t

woTninesfwActtdrfiPbtwEpFmoiab

3

csotAe

presence of AFP, respectively. The standard calibration curve wasfound to be log(I) = 2.7638 + 0.2906 log C, with a correlation coeffi-cient of 0.9939, covering the AFP concentration range from 0.0005

u/Ab1/AFP/Ab2/CdS, and (d) Fe3O4-Au/Ab1/AFP/Ab2/CdS-Au modified glassyarbon electrodes in 0.1 M PBS (pH 8.0) containing 0.1 M KCl and 0.1 M K2S2O8. Scanate: 100 mV s−1, the voltage of the PMT was 700 V.

.3. ECL behavior of sandwich immunoassay using CdS-Au/Ab2 ashe label

The process of sandwich immunoassay used to determine AFPas shown in Scheme 1. Specifically, the Ab1 was first anchored

n the surface of the Fe3O4-Au nanospheres by the Au–S bound.he products with an Ab1 immobilized on the surface of theanosphere to capture AFP (antigen) from a solution, contain-

ng a concentration of AFP. Finally, the protein-labeled CdS-Auanoparticles were introduced to the immunoreaction with thexposed part of AFP by an incubation period. This produced aandwich-type complex, with the magnetic separation for per-orming the following electrochemiluminescence analyses. In thisork, Fe3O4-Au/Ab1/AFP, Fe3O4-Au/Ab1/AFP/Ab2/CdS and Fe3O4-u/Ab1/AFP/Ab2/CdS-Au sandwich-type complex were used toonstruct an ECL immunosensor. Before using CdS and CdS-Au ashe label for preparation of immunosensor, the performance ofhe bare electrode and Fe3O4-Au/Ab1/AFP modified electrode inetecting of ECL signal was investigated and compared with the ECLesponse of bare electrode (Fig. 3a) and Fe3O4-Au/Ab1/AFP modi-ed electrode (Fig. 3b) toward the same condition as detected in theBS solution. As shown in Fig. 3, they were the ECL profiles of theare electrode and Fe3O4-Au/Ab1/AFP modified electrode, respec-ively. It could be concluded, that both electrodes of the ECL signalere very low, without CdS or CdS-Au as the label. However, the

CL intensity was increased obviously after the CdS or CdS-Au pre-ared immunocomplex Fe3O4-Au/Ab1/AFP/Ab2/CdS (Fig. 3c) ande3O4-Au/Ab1/AFP/Ab2/CdS-Au (Fig. 3d) sandwich-type complexodified on the electrode. The results show that the ECL response

f CdS-Au as the label is much higher then the pure CdS labeledmmunocomplex. It can also be concluded that the CdS-Au labeledntibody was successfully conjugated with the Fe3O4-Au/Ab1/AFPy the specific bond of the antigen–antibody.

.4. Optimization of major parameters for immunoreaction

One of the most important issues in the development of an effi-iency ECL sensor is the dependence of ECL signals on the CdS-Auignal tag. The detection sensitivity of the immunosensor depended

n the immunocomplex of the CdS-Au/Ab2 suspension. Therefore,he effect of incubation time and pH were studied in this study.s seen in Fig. 4, at room temperature, electrochemiluminescencemission intensity increased along with an increase in incubation

Fig. 4. Effects of incubation time on the response of ECL emission.

time and reached the maximum when incubation time was 60 min.Afterwards, the emission intensity was slightly decreased and vari-ation was mild. The results indicated a saturated binding of theimmobilized CdS-Au/Ab2. The optimal incubation time was 60 minin this experiment. The pH was another important issue for theformation of immunocomplex, as shown in Fig. 5, the ECL intensityincreased with an increase in pH values from 6.0 to 8.0, and thendecreased with pH higher than 8.0, indicating that pH played animportant role in building the immunocomplex process. Thus, theECL measurements were performed in pH 8.0 PBS solution.

3.5. Performance of the sandwich-type ECL immunosensor forAFP detection

Under optimal conditions, Fig. 6 displays the ECL response of theimmunosensor before (a) and after (b–n) reacting with differentconcentrations of AFP. It is found that the ECL intensity enhancedlinearly with the concentration of AFP. As shown in the insert, alinear relationship between the logarithm of the �I (�I = Is − I)and the logarithm of concentration of AFP wherein Is and I rep-resent the ECL intensity of the immunosensor in the absence and

Fig. 5. Effect of pH value of substrate solution on the response signals.

112 H. Zhou et al. / Analytica Chimica Acta 746 (2012) 107– 113

Table 1The recovery of the proposed immunosensor in human serum.

Serum samples 1 2 3 4 5

Proposed method (pg mL−1) 0.0552a 5.39b 52.9c 7.80b 8.22b

Reference method (ng mL−1) 6.02 6.02 6.02 7.12 7.24Relative error (%) 8.6 10.5

a–c The serum samples were diluted at 1.0 × 104, 1000 and 100 times, respectively.

Fig. 6. ECL profiles of the immunosensor in the absence (a) and increasing of AFPconcentration (b–n) in 0.1 M PBS (pH 8.0) containing 0.1 M KCl and 0.1 M K2S2O8.T02

tbaAcd

ttP

Finh11

he inset is the calibration curve, AFP concentration (ng mL−1): (a) 0, (b) 0.0005, (c).001, (d) 0.002, (e) 0.005, (f) 0.01, (g) 0.02, (h) 0.05, (i) 0.1, (j) 0.2, (k) 0.5, (l) 1.0, (m).0, (n) 5.0. Scan rate: 100 mV s−1, the voltage of the PMT was 700 V.

o 5.0 ng mL−1. The detection limit is calculated to be 0.2 pg mL−1

ased on a signal-to-noise of 3. The results indicate an accept-ble quantitative behavior of the proposed method used for theFP detected. Moreover, the ultrasensitive detection at this level ofoncentration was much lower than the previously reported ECLetection of AFP [13,18,32].

The selectivity of the proposed immunosensor for AFP detection,he as-prepared capture probe (Fe3O4-Au/Ab1), was incubated inhe AFP solution containing different interfering agents, such asSA, CEA, BSA and HIgG, and the results are shown in Fig. 7. As can

ig. 7. Selectivity of the ECL immunosensor to AFP (0.5 ng mL−1) by comparingt to the interfering protein: prostate protein antigen (PSA, 10 ng mL−1), carci-oembryonic antigen (CEA, 10 ng mL−1), bovine serum albumin (BSA, 1.0 �g mL−1),uman IgG (HIgG, 10 ng mL−1) and the mixed sample containing 0.5 ng mL−1 AFP,0 ng mL−1 PSA, 10 ng mL−1 CEA, 1.0 �g mL−1 BSA and 10 ng mL−1 HIgG. Scan rate:00 mV s−1, the voltage of the PMT was 700 V.

12.2 −9.5 −13.5

be seen from Fig. 7, the ECL response of the proposed immunosen-sor showed no remarkable change in the AFP mixed with interferingagents compared to AFP only. On the contrary, many weak ECLresponses were exhibited by replacing AFP with PSA, CEA, BSA andHIgG in the solution. The results showed a good selection of theproposed immunosensor for AFP detection.

3.6. Application

To determine the feasibility of the immunoassay system forclinical applications, the proposed method was evaluated by com-paring the assay results of real serum samples using the present ECLimmunosensor with reference values obtained by the commercialELISA method. The serum samples were appropriately diluted with0.05 M pH 7.4 PBS solutions, a level at which the level of the serumtumor marker was over the calibration range. Table 1 describes thecorrelation between the partial results obtained by the proposedmethod and the ELISA method, indicating an acceptable agreement,with relative errors less than 13.5%. The proposed method could beused for satisfactory clinical determination of AFP levels in humanserum.

4. Conclusions

In this work, we demonstrated a sensitive ECL-sensing proto-col for detecting AFP using a sandwich immunoreaction strategyfor magnetic separation and quantum dots labeled sandwich-typeimmunocomplex. Since the Fe3O4-Au and CdS-Au NPs displayedgood biocompatibility and strong ECL intensity, a simple andsensitive sandwiched ECL immunosensor was fabricated success-fully. Without almost any background intensity, the developedimmunosensor highly enhanced sensitivity compared to previousreports. In particular, this novel strategy could open a new andappealing approach for bioassays.

Acknowledgements

This work was supported by the Natural Science Foundation ofZhejiang (Y3110479), the Natural Science Foundation of Ningbo(2011A610018 and 2011A610006), the Social Development Projectof Ningbo (2011C50037, 2011C50038 and 2011B82014), and theKC Wong Magna Fund in Ningbo University, Science and Tech-nology Planning Project of Guangdong Province (2010A030300006and 2008A050200006).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.aca.2012.08.036.

References

[1] J.H. Lin, C.Y. He, L.J. Zhang, S.S. Zhang, Anal. Biochem. 384 (2009) 130–135.[2] Y.X. Ci, Y. Qin, W.B. Chang, Y.Z. Li, Anal. Chim. Acta 300 (1995) 273–276.

[3] Y. Liu, H.X. Wang, J.Y. Huang, J. Yang, B.H. Liu, P.Y. Yang, Anal. Chim. Acta 650

(2009) 77–82.[4] L.M. Ao, F. Gao, B.F. Pan, R. He, D.X. Cui, Anal. Chem. 78 (2006) 1104–1106.[5] Y.F. Wu, Y.F. Zhuang, S.Q. Liu, L. He, Anal. Chim. Acta 630 (2008) 186–193.[6] Q.Y. Zhang, X. Wang, Z.J. Li, J.M. Lin, Anal. Chim. Acta 631 (2009) 212–217.

imica

[[[[[

[

[

[[

[[[[

[

[[

[[

[

[

[[[

[

[[

1447–1453.

H. Zhou et al. / Analytica Ch

[7] S.J. Xu, Y. Liu, T.H. Wang, J.H. Li, Anal. Chem. 83 (2011) 3817–3823.[8] H. Li, Q. Wei, J. He, T. Li, Y.F. Zhao, Y.Y. Cai, B. Du, Z.Y. Qian, M.H. Yang, Biosens.

Bioelectron. 26 (2011) 3590–3595.[9] S. Katoh, T. Miura, J. Chromatogr. A 852 (1999) 97–104.10] D. Faraggi, A. Kramar, Urol. Oncol. 5 (2000) 211–213.11] J.H. Lin, H.X. Ju, Biosens. Bioelectron. 20 (2005) 1461–1470.12] M.M. Richter, Chem. Rev. 104 (2004) 3003–3036.13] J. Qian, Z.X. Zhou, X.D. Cao, S.Q. Liu, Anal. Chim. Acta 665 (2010) 32–38.14] H. Niu, R. Yuan, Y.Q. Chai, L. Mao, Y.L. Yuan, Y. Zhuo, S.R. Yuan, X. Yang, Biosens.

Bioelectron. 26 (2011) 3175–3180.15] G.F. Jie, J.J. Zhang, D.C. Wang, C. Cheng, H.Y. Chen, J.J. Zhu, Anal. Chem. 80 (2008)

4033–4039.16] L. Mao, R. Yuan, Y.Q. Chai, Y. Zhuo, X. Yang, Sens. Actuators B 149 (2010)

226–232.17] M. Su, W. Wei, S.Q. Liu, Anal. Chim. Acta 704 (2011) 16–32.18] H.Y. Wang, D.Y. Sun, Z.A. Tan, W. Gong, L. Wang, Colloids Surf. B 84 (2011)

515–519.19] H. Cui, G.Z. Zou, X.Q. Lin, Anal. Chem. 75 (2003) 324–331.20] G.F. Jie, B. Liu, H.C. Pan, J.J. Zhu, H.Y. Chen, Anal. Chem. 79 (2007) 5574–5581.

21] W.R. Algar, A.J. Tavares, U.J. Krull, Anal. Chim. Acta 673 (2010) 1–25.22] H.C. Nian, J. Wang, H. Wu, J.G. Lo, K.H. Chiu, J.G. Pounds, Y.H. Lin, Anal. Chim.

Acta 713 (2012) 50–55.23] M. Zhou, J. Roovers, G.P. Robertson, C.P. Grover, Anal. Chem. 75 (2003)

6708–6717.

[

[

Acta 746 (2012) 107– 113 113

24] K.A. Fähnrich, M. Pravda, G.G. Guilbault, Talanta 54 (2001) 531–559.25] G.F. Jie, H.P. Huang, X.L. Sun, J.J. Zhu, Biosens. Bioelectron. 23 (2008)

1896–1899.26] N. Myung, Z.F. Ding, A.J. Bard, Nano Lett. 2 (2002) 1315–1319.27] Y.H. Tennico, D. Hutanu, M.T. Koesdjojo, C.M. Bartel, V.T. Remcho, Anal. Chem.

82 (2010) 5591–5597.28] F. Wang, G.Y. Xu, Z.Q. Zhang, X. Xin, Eur. J. Inorg. Chem. (2006)

109–114.29] J.J. Zhang, T.T. Zheng, F.F. Cheng, J.R. Zhang, J.J. Zhu, Anal. Chem. 83 (2011)

7902–7909.30] J. Qian, C.Y. Zhang, X.D. Cao, S.Q. Liu, Anal. Chem. 82 (2010) 6422–6429.31] Y. Chai, D.Y. Tian, H. Cui, Anal. Chim. Acta 715 (2012) 86–92.32] S.R. Yuan, R. Yuan, Y.Q. Chai, L. Mao, X. Yang, Y.L. Yuan, H. Niu, Talanta 82 (2010)

1468–1471.33] X.L. Zhao, Y.Q. Cai, T. Wang, Y.L. Shi, G.B. Jiang, Anal. Chem. 80 (2008)

9091–9096.34] Z.C. Xu, Y.L. Hou, S.H. Sun, J. Am. Chem. Soc. 129 (2007) 8698–8699.35] J.D. Qiu, H.P. Peng, R.P. Liang, X.H. Xia, Biosens. Bioelectron. 25 (2010)

36] J.P. Li, H.L. Gao, Z.Q. Chen, X.P. Wei, C.F. Yang, Anal. Chim. Acta 665 (2010)98–104.

37] Y.R. Cui, C. Hong, Y.L. Zhou, Y. Li, X.M. Gao, X.X. Zhang, Talanta 85 (2011)1246–1252.