Biosensors and Bioelectronics 44 (2013) 235–240

Contents lists available at SciVerse ScienceDirect

Biosensors and Bioelectronics

0956-56

http://d

n Corr

Ministry

E-m

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

Electrochemical immunosensor for ultrasensitive detection ofmicrocystin-LR based on graphene–gold nanocomposite/functionalconducting polymer/gold nanoparticle/ionic liquid composite filmwith electrodeposition

Li Ruiyi d, Xia Qianfang a, Li Zaijun a,b,n, Sun Xiulan c, Liu Junkang a,b

a The Key Laboratory of Food Colloids and Biotechnology, Ministry of Education, Chinab School of Chemical and Material Engineering, Chinac The School of Food Science and Technology, Jiangnan University, Wuxi 214122, Chinad National University of Ireland, Maynooth, Co. Kildare, Ireland

a r t i c l e i n f o

Article history:

Received 18 December 2012

Accepted 6 January 2013Available online 23 January 2013

Keywords:

Microcystin-LR

Sensor

Graphene–gold nanocomposite

Functional conducting polymer

Ionic liquid

Electrodeposition

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

x.doi.org/10.1016/j.bios.2013.01.007

esponding author at: The Key Laboratory of Foo

of Education, China. Fax: þ86 5105811863.

ail addresses: 1489539656@qq.com, zaijunli@

a b s t r a c t

The study developed an electrochemical immunosensor for ultrasensitive detection of microcystin-LR

in water. Graphene oxide and chloroauric acid were alternately electrodeposited on the surface of

glassy carbon electrode for 20 cycles to fabricate graphene–gold nanocomposite. The composite was

characterized and its apparent heterogeneous electron transfer rate constant (37.2870.16 cm s�1) was

estimated by Laviron’s model. To immobilize microcystin-LR antibody and improve the electrical

conductivity, 2,5-di-(2-thienyl)-1-pyrrole-1-(p-benzoic acid) and chloroauric acid were electrodepos-

ited on the modified electrode in sequence. The ionic liquid was then dropped on the electrode surface

and finally microcystin-LR antibody was covalently connected to the conducting polymer film.

Experiment showed the electrochemical technique offers control over reaction parameters and

excellent repeatability. The graphene–gold nanocomposite and gold nanoparticles enhance electron

transfer of Fe(CN)63�/4� to the electrode. The ionic liquid, 1-isobutyl-3-methylimidazolium bis(trifluor-

omethane-sulfonyl)imide, improves stability of the antibody. The sensor displays good repeatability

(RSD¼1.2%), sensitive electrochemical response to microcystin-LR in the range of 1.0�10-16–

8.0�10�15 M and detection limit of 3.7�10�17 M (S/N¼3). The peak current change of the sensor

after and before incubation with 2.0�10�15 M of microcystin-LR can retain 95% over a 20-weeks

storage period. Proposed method presents remarkable improvement of sensitivity, repeatability and

stability when compared to present microcystin-LR sensors. It has been successfully applied to the

microcystin-LR determination in water samples with a spiked recovery in the range of 96.3–105.8%.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

Microcystins are a family of stable heptapeptides producedmainly by the common cyanobacteria and found in water sourceat various times all over the world (Lawton et al., 2010).Microcystins-contaminated water bodies are increasing due tothe eutrophication of lake and reservoir (Hernandez et al., 2009).The toxicological mechanism of microcystins is via their stronginhibition of protein phosphatases 1 and 2 A that may lead tosevere liver damage (Dawson, 1998). Long-term exposure to lowconcentration of microcystin has implicated in tumor promotion(Khreich et al., 2009; Magalhaes et al., 2003). Among microcystinvariants, microcystin-LR is the most common and toxic congener.

ll rights reserved.

d Colloids and Biotechnology,

263.net (L. Zaijun).

To protect water quality and human health, World HealthOrganization proposed a provisional upper limit of 1.0 mg l�1 formicrocystin-LR in drinking water. Thus, there is a great need todevelop sensitive and reliable analytical method for the determi-nation of microcystin-LR. To date, many analytical technologieshave been developed for the determination of microcystin-LR,such as liquid chromatography–mass spectrometry (LC–MS) (Daiet al., 2008; Li et al., 2011; Ruangyuttikarn et al., 2004; Zhanget al., 2004), immunoassay method (Sassolas et al., 2011; Xiaet al., 2010; Xiao et al., 2009; Zhang et al., 2007) and sensor (Dingand Mutharasan, 2011; Ma et al., 2011; Sun et al., 2010; Xia et al.,2011; Xu et al., 2010). LC–MS was often used in routinemicrocystin-LR analysis, but it requires skilled operator, exten-sive sample pretreatment and expensive equipment. To furtherincrease the selectivity, various immunoadsorbent columns werealso used to clean up sample prior to the measurement. Recently,electrochemical sensor was given special attention due to its fast,simple, and low-cost. However, present electrochemical sensors

L. Ruiyi et al. / Biosensors and Bioelectronics 44 (2013) 235–240236

are difficult to be widely applied to routine analysis due to thelack of enough sensitivity and stability.

Graphene has attracted increasing attention since its discovery in2004 due to unique electronic, mechanical, physical and chemicalproperties (Miller et al., 2009). Graphene has been employed as theblock for new materials for many applications, including solar cell(Yin et al., 2010), lithium-ion battery (Lian et al., 2010), cell culture(Agarwal et al., 2010) and sensing materials (Ding et al., 2010; Guet al., 2011a, 2011b; Singh et al., 2012). The use of graphene and itscomposite as sensing materials has received much attention inrecent two years (Luo et al., 2011). Present graphene-based sensorswere mostly fabricated by drop-casting graphene dispersion, whichwas obtained from chemical reduction of graphene oxide (Luo et al.,2011). The method has intrinsic limitations such as the lack controlof film thickness and use of toxic chemicals. Recently, electrochemi-cal reduction of graphene oxide to graphene has drawn greatattention due to its fast and green natures (Guo et al., 2009; Shaoet al., 2010). Typically, graphene oxide was coated on the electrodesurface. The electrode was then immersed in a metallic precursorsolution to perform one-step coelectrochemical reduction (Chenet al., 2011). This method lacks control over the film thickness, morenotably, the resulting graphene sheets were not separated by metalnanoparticles since they are mostly located on surface of thegraphene film. Moreover, the enzyme/antibody was often immobi-lized on the graphene-based electrode by physical adsorption. Themethod would result in poor analytical characteristics in precisionand repeatability due to the loss of enzyme/antibody during theelectrochemical measurement. To overcome the problem, Liu andRoy researched covalent immobilization of the enzyme (Liu et al.,2010; Roy et al., 2011). Here, the enzyme was covalently connectedwith carboxyl group in graphene oxide sheets. The biosensor offersbroad linearity, excellent reproducibility and storage stability, but itssensitivity is lower than that of the graphene-based biosensor dueto poor electrical conductivity of the graphene oxide (Wang et al.,2009).

In this study, we focus on developing a green and controllablestrategy to fabricate well-dispersed graphene–gold nanocompo-site with electrodeposition and its use as sensing materialsfor microcystin-LR sensor. The graphene–gold nanocomposite(G–AuNP) and gold nanoparticles (AuNP), functional conductingpolymer (polyDPB) and 1-isobutyl-3-methylimidazolium bis(tri-fluoromethane-sulfonyl)imide ionic liquid (IL) were employedfor improving electron transfer of Fe(CN)6

3�/4� to the electrode,antibody (Ab) immobilization and stability of the Ab on theelectrode surface, respectively. The main parameters governingthe current responses were optimized, including operating pHand incubation time. Proposed immunoassay presents remarkableimprovement of sensitivity, repeatability and stability whencompared to present microcystin-LR sensors.

2. Experimental

2.1. Materials and reagents

Graphite, chlorauric acid (HAuCl4), chitosan, bovine serum albu-min, tetrabutylammonium perchlorate, 1-ethyl-3-(3-dimethyl-ami-nopropyl)carbodiimide (EDC), N-hydroxyosuccinimide (NHS) andmicrocystin-LR, were purchased from Sigma-Aldrich Chemical Com-pany (Mainland, China) and used without further purification.Polyclonal antibody (Ab) was generated in rabbits by immunizingthe animals with microcystin-LR-bovine serum albumin (Zhao et al.,2006). Phosphate-buffered saline (PBS, pH 7.0, Na2HPO4–KH2PO4–NaCl–KCl, 0.01 M) was prepared. The 2,5-Di(2-hienyl)-1-pyrrole-1-(p-benzoic acid) (DPB) was synthesized by using a similar proceduredescribed in Kim et al. (2010). Graphite oxide was prepared from

natural graphite by Hummers’ method (Hummers and Offeman(1958)). The 1-isobutyl-3-methylimidazolium bis(trifluoromethane-sulfonyl)imide ([i-C4mim][TFSI]) was prepared using a similarmethod described in Cai et al. (2010), Shan et al. (2008). Thechitosan solution was prepared by dissolving 5 g of chitosan in100 ml of 1.0% (v/v) acetic acid. All other reagents employed were ofanalytical reagent grade or with highest quality and were purchasedfrom Shanghai Chemical Company (Shanghai, China). Ultra purewater (18.2 MO cm) purified from a Milli-Q purification system wasused throughout the experiment.

2.2. Apparatus

Electrochemical experiment was performed with an IM6eelectrochemical system (ZAHNER Elektrik, German) and CHI660Belectrochemical workstation (Shanghai, China) and CHI400BElectrochemical quartz crystal microbalance (Shanghai, China).A conventional three-electrode system was used: a glassy carbonelectrode (GCE, 1 mm in diameter) as the working electrode, anAg/AgCl (saturated KCl) electrode as the reference electrode, and aplatinum wire electrode as the counter electrode. The circularAT-cut 7.995 MHz quartz crystals coated with polished goldelectrodes on both sides was used for electrochemical quartzcrystal microbalance (QCM) analysis. The quartz crystals were14 mm in the diameter and 0.2 mm thick and the gold electrodeswere 0.196 cm2 in the area and 100 nm thickness. For cleaningpurpose, the quartz crystals were immersed in 1 M of sodiumhydroxide for 20 min and then washed with ultra pure water.After that, they were immersed in 1 M of hydrochloride for10 min and then rinsed with ethanol. The upper and lower piecesof the cell were held together by two screws and sealed betweentwo O-rings. The quartz crystal microbalance crystal had one sideof the electrode exposed to the solution. Scanning electronmicroscope (SEM) image, Raman spectrum and X-ray diffraction(XRD) patterns were obtained using HITACHI S4800 field emissionscanning electron microscope, HR800 Raman microprobe with a514 nm laser excitation and X-ray D/max-2200vpc (Rigaku Cor-poration, Japan) instrument operated at 40 kV and 20 mA andusing Cu Ka radiation (k¼0.15406 nm) respectively. Circulardichroism analysis was performed with a MOS-450 circulardichroism spectrometer using a 0.01 cm quartz cell at roomtemperature.

2.3. Sensor preparation

Procedure of the sensor preparation includes five assembleprocesses (shown in Fig. 1), i.e. the immobilizations of G–AuNP,polyDPB, AuNP, IL and Ab. (1) The GCE was immersed in0.01 mg l�1 graphene oxide dispersion to electrodeposit graphenewith constant potential at �1.2 V for 50 s at 10 1C. The electrodewas rinsed with ultra pure water and allowed to dry at roomtemperature. The modified electrode was then immersed in a0.01 mM chlorauric acid solution containing 0.5 M of sulfuric acidto electrodeposit gold with constant potential at �0.25 V for 50 sat 10 1C. The electrode was rinsed with ultra pure water andallowed to dry at room temperature. Above process was repeatedfor 20 cycles to obtain a G–AuNP/GCE. (2) The G–AuNP/GCEwas subjected to cyclic scanning for 2 times in a 1.0 mM DPBin dichloromethane containing 0.1 M of tetrabutylammoniumperchlorate with potential range from 0.0 V to þ1.4 V. (3) ThepolyDPB–G–AuNP/GCE was immersed in a 0.005 mM chlorauricacid solution containing 0.5 M sulfuric acid to electrodepositgold with constant potential at �0.25 V for 200 s at 10 1C. Theelectrode was rinsed with ultra pure water and allowed to dry atroom temperature. (4) A 50 ml of the IL was dispersed in 0.2 mlchitosan solution. After the mixture was sonicated for 30 min, its

Fig. 1. Procedure for preparation of the immunosensor.

L. Ruiyi et al. / Biosensors and Bioelectronics 44 (2013) 235–240 237

1.0 ml was dropped on surface of the AuNP–polyDPB–G–AuNP/GCE using a microsyringe and the electrode was then dried in thestream of hot air. (5) The IL–AuNP–polyDPB–G–AuNP/GCE wassoaked in 1 ml of 5 mM EDC and 5 mM NHS solution to activatecarboxylic groups on the polyDPB film for 1 h. Then, a 20 ml of0.01 g ml�1 Ab was added onto the activated electrode surface for2 h. The electrode was cleaned with distilled water and theremaining activated carboxylic groups were blocked by 10 ml of1% bovine serum albumin solution. For immunity of the sensor,the modified electrode was incubated in 20 ml of the pH 7.0 PBScontaining different concentrations of microcystin-LR for 15 minat 37 1C.

2.4. Electrochemical measurements

Prior to each experiment, the buffer solutions were purgedwith highly purified nitrogen for at least 30 min and nitrogenatmosphere environment was kept in all electrochemical mea-surements. Cyclic voltammetry (CV), differential pulse voltam-metry (DPV) and impedance measurements were performed in10 ml of pH 7.0 PBS containing 1.0 mM K3[Fe(CN)6]/K4[Fe(CN)6](1:1 mixture as redox probe) ([Fe(CN)6]3�/4�).

3. Results and discussion

3.1. Synthesis and structure characterization of the graphene–gold

nanocomposite

Graphene oxide sheet contains large number of carboxyl andhydroxyl groups, the dissociation of these groups would producevarious anionic graphene oxide sheets. With the increase of pHvalue, the concentration of anionic graphene oxide would rapidlyincrease. Because of the repulsion of cathode with anion, anionicgraphene oxide is more difficult to be moved, reduced anddeposited on the electrode surface (as cathode) when comparedto neutral and cationic graphene oxides. The use of neutral oracidic aqueous medium is beneficial to electrodeposit graphenedue to relatively high graphene oxide concentration on theworking electrode, but graphene oxide sheets are easy to pre-cipitate from strong acidic aqueous phase due to the protonationof carboxyl and hydroxyl groups. In this study, a graphene oxidedispersion of pH 7.0 was employed as electrolyte for the grapheneelectrodeposition without adding any base, acid or supportingelectrolyte.

Effect of the precursor concentration in the electrolyte onthe electrodeposition was investigated in the study. Fig. S1shows that graphene nanosheets are very rare on the electrodesurface when the graphene oxide concentration is smaller than0.01 mg l�1. With the increase of graphene oxide concentration,the number of graphene nanosheets would rapidly increase.When the graphene oxide concentration increases to 0.01 mg l�1,whole surface of the electrode would be well covered with single-layer graphene nanosheets. However, obvious agglomerationof the graphene sheets was observed with further increase ofgraphene oxide concentration, which leads to significantlydecreased surface area of the modified graphene layer. Thus, a0.01 mg l�1 of graphene oxide was selected as electrolyte for thegraphene electrodeposition. After graphene was deposited on theelectrode, the modified electrode was immersed in a chlorauricacid solution to electrodeposit gold. Fig. S2 indicates that goldnanoparticles are very rare with a small particle size on surface ofthe modified electrode when the chlorauric acid concentration islower than 0.01 mM. With the increase of chlorauric acid con-centration, particle size and number of gold particles rapidlyincreases. When the concentration of chlorauric acid increases to0.01 mM, whole surface of the modified electrode would be wellcovered with single-layer gold particles with an average particlesize of about 10 nm. However, obvious agglomeration of goldparticles was observed with further increase concentration of thechlorauric acid, which leads to significantly decreased surfacearea of the modified gold nanoparticles layer. Thus, a 0.01 mM ofchlorauric acid was selected as electrolyte for the gold electro-deposition. To test repeatability of the above electrodeposition, acircular AT-cut 7.995 MHz quartz crystal was also employed asthe working electrode for the electrodeposition. After each elec-trodeposition completed, the modified-layer was investigatedby electrochemical quartz crystal microbalance. Fig. S3 showsthe each deposition would bring similar decrease of the QCMfrequency, indicating that the electrodeposition offers excellentrepeatability.

Graphene oxide and chloroauric acid were alternately electro-deposited on the GCE surface for 20 cycles under optimal condi-tions to fabricate G–AuNP. The as-prepared G–AuNP wascharacterized by SEM, XRD, Raman and infrared (IR) spectrum.Fig. S4 indicates gold nanoparticles are very regular and welldispersed on the wrinkled graphene nanosheets with an averageparticle size of about 10 nm. On the XRD pattern there are foursharp diffraction peaks at scattering angles of 38.281, 44.481, 64.61and 77.641, corresponding to crystal planes Au (111), Au (200),

Fig. 3. Cyclic voltammograms of the IL–Ab–AuNPs–polyDPB–G–AuNPs/GCE in pH

7.0 PBS containing 1.0 mM [Fe(CN)6]3� /4� at various scan rates (from a to t) of 0.1,

0.2, 0.3, 0.4, 0.5, 0.6,0.7, 0.8, 0.9, 1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9 and 2.0 V/s.

Inset: Plot of Ip vs. square root of scan rate.

L. Ruiyi et al. / Biosensors and Bioelectronics 44 (2013) 235–240238

Au (220) and Au (311) of gold face-centered cubic crystallo-graphic structure (JCPDS card no. 65-2870). On the Ramanspectrum of GO there are two prominent peaks, including D band(1355 cm�1) and G band (1600 cm�1), which represent theamount of sp3 carbons in the surroundings and E2g phonon ofsp2 C atoms. During the reduction process, the most of oxygen-containing groups were removed, and the conjugated G network(sp2 carbon) would be re-established. However, the size of the re-established G network is smaller than the original one, whichwould lead to an increase in the ID/IG ratio (Zhou et al., 2010). Arelatively broad peak at 3414 cm�1 and relatively sharp peak at1620 cm�1 on the IR spectrum demonstrated that the samplecontains adsorbed water. The peaks at 1397 cm�1 and 1081 cm�1

can be assigned to the deformation vibration of O–H and stretch-ing vibration of C–O, respectively. The characteristic bands ofC¼O carbonyl stretching and C–O–C vibration located at1720 cm�1 and 1250 cm�1 are very weak, indicating smallamount of these two functional groups.

3.2. Electrochemical properties of the modified electrodes

Cyclic voltammograms of the bare and modified electrodes wererecorded in a pH 7.0 PBS containing 1.0 mM of [Fe(CN)6]3�/4� at ascan rate of 100 mV s�1 (Fig. 2). It can be clearly seen from Fig. 2the cyclic voltammogram of bare GCE (curve a) shows a well-defined reversible redox behavior attributed to highly electrontransfer of [Fe(CN)6]3�/4� to the electrode. After the attachmentof G–AuNP, AuNP or IL on the surface of the electrode, peakpotential of the cyclic voltammogram reduces with enhancementin the peak current (curve b, d and e) due to excellent electro-catalytic property, this would improve the sensitivity of theimmunosensor. However, the introduction of the polyDPB or Abinto the film would result in the decrease of the peak current (curvec and f). This is because the polyDPB or Ab on surface of themodified electrode partially blocks electron transfer of [Fe(CN)6]3�/4�

to the electrode.Electrochemical stability of the sensor was checked by repe-

titive potential sweep. It was found that the peak currents almostmaintain the same intensity after 100 scans, indicating that thesensor has fairly good electrochemical stability, and the leakageof G–AuNP, polyDPB, AuNP and IL from the electrode surface canbe neglected. Moreover, the effect of varying scan rate on the

Fig. 2. Cyclic voltammograms of bare GCE (a) and G–AuNPs/GCE (b), polyDPB–G–

AuNPs/GCE (c), AuNPs-polyDPB–G–AuNPs/GCE (d), Ab–AuNPs–polyDPB–G–AuNPs/

GCE (e), and IL–Ab–AuNPs–polyDPB–G–AuNPs/GCE (f) in pH 7.0 PBS containing

1.0 mM [Fe(CN)6]3�/4� . The scan rate was 0.1 V s�1.

electrochemical performance was also studied in the work. Fig. 3shows both cathodic and anodic peak currents increase with theincrease of scan rate. The peak currents are linear to the squareroot of scan rate, this reveal the electrochemical kinetics was thesolution phase quasi-reversible process.

3.3. Electrochemical response of the sensor to microcystin-LR

The composite film exhibits stable and reversible electrochemis-try, it can be used as electron transfer mediator to shuttle electronsbetween the redox in aqueous solution and electrode. Fig. 4 presentsDPV curves of the sensor before and after incubated microcystin-LR. Itcan be clearly seen from Fig. 4 the peak current would decrease afterincubated microcystin-LR. The higher microcystin-LR concentration,the smaller peak current. This is because the specific interaction ofmicrocystin-LR and Ab produces an insulating bioconjugate, whichpartially blocks electron transfer of [Fe(CN)6]3�/4� to the electrode.

Fig. 4. DPVs of the sensor in pH 7.0 PBS containing 0.0, 8.0�10�16 and

2.0�10�15 M of microcystin-LR. The DPV parameters were set to a scan rate of

4 mV/s, 50 mV pulse amplitude, 20 ms pulse width and �0.2 V initial potential.

L. Ruiyi et al. / Biosensors and Bioelectronics 44 (2013) 235–240 239

3.4. Optimization of the assay conditions

Influence of the buffer pH is very essential to the sensitivity ofthe sensor, because the pH affects not only the electrochemicalbehavior of the sensor but also the bioactivity of the Ab. Theoptimal pH reported for the Ab is usually in the range of 6.5–7.5(Xu et al., 2010), which varies with immobilization method andmicroenvironment around the Ab. Thus, effect of the pH wasexamined in the range of 5.0–8.0. Fig. S5 presents the relationshipof DPV peak current change (DIp) of the sensor with the buffer pH,in which the sensor was incubated 4.0�10�15 M of microcystin-LR for 15 min. It can be seen from Fig. S5 the DIp would increasewith the increase of pH, attain maximum at about pH 7.0, andthen decrease with further increase of the pH. This shows thespecific interaction between microcystin-LR and Ab takes placemore effectively at pH 7.0.

The incubation time also played an important role in thedetection of microcystin-LR. Fig. S6 shows that with the increaseof incubation time the DIp rapidly increases until the incubationtime increases to 15 min. When the incubation time exceeds15 min, the DIp reaches maximum and remains almost unchanged.To obtain good sensitivity and rapidity, the operating pH of 7.0and incubation time of 15 min were adopted in the followingexperiments.

3.5. Analytical characteristics of the sensor

Fig. 5 presents the relationship of DPV response withmicrocystin-LR concentration. With the increase of microcystin-LR concentration the DIp increases linearly in the range from1.0�10�16 M to 8.0�10�15 M (inset in Fig. 5). The linearequation was DIp¼0.0317Cþ0.019, with a statistically significantcorrelation coefficient of 0.9991, which DIp is in mA and concen-tration (C) in 10�16 M. Detection limit of the proposed methodwas 3.7�10�17 M that was obtained from the signal-to-noisecharacteristics of these data (S/N¼3). The microcystin-LR sensorwas repeatedly measured 25 times in 2.0�10�15 M of microcystin-LR standard solution under the same conditions. A relative standarddeviation of 1.2% for the measurements was obtained, this indicatesthe measurement has high precision for the sensor. The fabrication

Fig. 5. DPVs of different concentration microcystin-LR. The DPV parameters were

set to a scan rate of 4 mV/s, 50 mV pulse amplitude, 20 ms pulse width and �0.2 V

initial potential. Insets: Insets: calibration plots of concentration of microcystin-LR

vs. peak current change (DIp).

reproducibility was also estimated by the measurements of 2.0�10�15 M of microcystin-LR in duplicate, with ten different electro-des made at the same GCE independently. The relative standarddeviation was 2.2%, this shows good reproducibility. The proposedsensor was stored in air at ambient conditions and its sensitivitywas checked every week. DPV response of the sensor was 95% of itsinitial value after 20 weeks, this shows an excellent long-termstability. These analytical parameters are better than that of thereported microcystin-LR sensors (Xu et al., 2010).

3.6. Regeneration of the electrode

Regeneration of the sensor was investigated by repeatedlymonitoring the binding of 2.0�10�15 M microcystin-LR under sameexperimental conditions. Following each binding event, the elec-trode was regenerated for 30 min in 0.1 M glycine buffer solutionto desorb the combined Ab. The relative standard deviation ofDIp DPV response changes for twenty binding and regenerationcycles was found to be 2.0%. This shows that the method usingthe sensor for determination of microcystin-LR was of excellentreproducibility.

3.7. Functions of graphene–gold nanocomposite and the ionic liquid

To understand functions of G–AuNP and the ionic liquid,electrocatalytic activity of the G–AuNP and secondary structurechange of the Ab in the ionic liquid were investigated, respec-tively. Fig. S7A shows the effect of scan rate on the CV response ofG–AuNP/GCE in a 1.0 mM K4Fe(CN)6 in PBS. It can be found thatall scan rates results in a well-defined reduction and oxidationpeaks with little shift in both cathodic and anodic peak potentialswith respect to scan rate. Further, Fig. S7B shows that bothcathodic and anodic peak currents are linearly proportional withsquare root of the scan rate, which demonstrates that theelectrode reaction corresponds to the solution phase quasi-reversible process. The small peak-to-peak separation (DEp)illustrates a fast electron transfer rate. Because the nDEp is lessthan 200 mV, the apparent heterogeneous electron transfer rateconstant (ks) can be estimated by the following equation:ks¼mnFv/RT (Laviron, 1979), where m is a parameter related topeak-to-peak separation, T is the temperature, n is the number ofelectrons, and v is the scan rate. The ks was 37.2870.16 cm s�1,which is obviously higher than that of the graphene-modified GCE(28.5570.22 cm s�1) and bare GCE (5.4770.16 cm s�1). Higherks value revealed that the introduction of G–AuNP enhanceselectron transfer of Fe(CN)6

3�/4� solution to the electrode, whichwould lead to improve the sensitivity. Moreover, the circulardichroism spectrum was employed for the investigation ofsecondary structure change of the Ab in blank and added[i-C4mim][TFSI] ionic liquid PBS. Fig. S8 shows that main second-ary structure of the Ab such as a-Helix, b-Beta and Random in theionic liquid has better stability than in blank PBS. For the abovereasons, we suggest ultra sensitive response of the sensor wouldbe mainly attributed that the G–AuNP enhances electron transferof the redox in aqueous solution to the electrode, and the ionicliquid provides a benign microenvironment for the Ab whichkeeps a high activity of the Ab.

3.8. Application to microcystin-LR analysis in real water samples

The feasibility of the newly developed method for possibleapplications was investigated by analyzing real water samples.Lake water and river water were collected from Wuxi station inTaihu Lake. Tap water and ground water were collected from thecampus at Jiangnan University. The sample and the spiked samplewith different amounts of microcystin-LR standard solution were

Table 1The results for the determination of microcystin-LR in water samplesa

Sample Microcystin-LR

found by

proposed method

(pM)

Microcystin-LR

found by

LC–MS (Li et al.,

2011) (pM)

Recovery

(%)

Tap water 0.1970.03 0.1870.05 96.3

F¼2.78, t¼0.66

Lake water 10.0270.19 10.1270.17 103.2

F¼1.25, t¼1.16

River water 1.2970.13 1.2170.11 104.9

F¼1.40, t¼1.44

Ground water 3.5670.16 3.3970.21 105.8

F¼1.72, t¼2.11

a Results expressed as: X7st/n1/2 (n¼5), where X is the mean of n observa-

tions of x, s is the standard deviation, t is distribution value chosen for the desired

confidence level, the t- and F-values refer to comparison of the proposed method

with the HPLC method. Theoretical values at 95% confidence limits: F¼6.39,

t¼2.78.

L. Ruiyi et al. / Biosensors and Bioelectronics 44 (2013) 235–240240

testified. The sample was used for the analysis to replace the PBSsolution of microcystin-LR as mentioned above. The concentrationof microcystin-LR in water sample was determined from thecalibration curve and the value was used to calculate theconcentration in the original sample. The mean7SD of eachsample and recovery of each spiked sample were calculated,and the values are reported in Table 1. The analytical data forvarious samples given in Table 1 indicate a high degree ofcorrelation between the results of LC–MS described in thereference (Li et al., 2011) and proposed sensor. The recovery ofmicrocystin-LR for spiked samples analysis is in the range from96.3% to 105.8%. These indicated proposed method has a goodaccuracy and precision.

4. Conclusions

The study developed an electrochemical immunosensor for thedetection of microcystin-LR based on G–AuNP/polyDPB/AuNP/ILcomposite film on the GCE with electrodeposition. Experimentdemonstrated the electrochemical technology for the sensor fabrica-tion provides control over reaction parameters and excellent repeat-ability. The G–AuNP and AuNP, polyDPB and IL remarkably improveelectron transfer of the redox in aqueous solution to the electrode,Ab immobilization and stability of the Ab on the electrode surface,respectively. The proposed sensor offers the best sensitivity, repeat-ability and stability for the determination of microcystin-LR upto now.

Acknowledgments

The authors acknowledge the financial support from the Funda-mental Research Funds for the Central Universities (JUSRP51314B),the National Natural Science Foundation of China (Nos. 21176101,20771045 and 20676052), the Natural Science Foundation ofZhejiang Province (Y4080404), the country 12th Five-Year Plan toSupport Science And Technology Project (No. 2011BAK10B03) andsponsored by Qing Lan Project.

Appendix A. Supporting information

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.bios.2013.01.007.

References

Agarwal, S., Zhou, X.Z., Ye, F., He, Q.Y., Chen, G.C.K., Soo, J., Boey, F., Zhang, H., Chen,P., 2010. Langmuir 26, 2244–2249.

Cai, Y., Li, Z.J., Zhang, H.L., Fang, Y.J., Fan, X., Liu, K., 2010. Electrochimica Acta 55,4728–4733.

Chen, L.Y., Tang, Y.H., Wang, K., Liu, C.B., Luo, S.L., 2011. Electrochemistry Commu-nications 13, 133–137.

Dai, M., Xie, P., Liang, G.D., Chen, J., Lei, H.H., 2008. Journal of Chromatography B862 (1/2), 43–50.

Dawson, R.M., 1998. Toxicon 36, 953–962.Ding, Y.J., Mutharasan, R., 2011. Environmental Science & Technology 45 (4),

1490–1496.Ding, Z., He, Y., Xue, Z., Zhao, X., Lu, X., Liu, X., 2010. Colloids and Surfaces B 79,

27–32.Gu, Z.G., Yang, S.P., Li, Z.J., Sun, X.L., Wang, G.L., Fang, Y.J., Liu, J.K., 2011a. Analytica

Chimica Acta 701, 75–80.Gu, Z.G., Yang, S.P., Li, Z.J., Sun, X.L., Wang, G.L., Fang, Y.J., Liu, J.K., 2011b. Electrochimica

Acta 56, 9162–9167.Guo, H.L., Wang, X.F., Qian, Q.Y., Wang, F.B., Xia, X.H., 2009. ACS Nano 3, 2653–2659.Hernandez, J.M., Lopez-Rodas, V., Costas, E., 2009. Medical Hypotheses 72, 539–540.Hummers, W.S., Offeman, R.E., 1958. Journal of the American Chemical Society 801339Khreich, N., Lamourette, P., Renard, P.Y., Clave, G., Fenaille, F., Creminon, C., Volland,

H., 2009. Toxicon 53, 551–559.Kim, Y.H., Hwang, J.Y., Ik Son, J., Shim, Y.B., 2010. Synthetic Metals 160, 413–418.Laviron, E., 1979. Journal of Electroanalytical Chemistry 101, 19–22.Lawton, L.A., Chambers, H., Edwards, C., Nwaopara, A.A., Healy, M., 2010. Toxicon

55, 973–978.Li, W., Duan, J.M., Niu, C.Y., Qiang, N.C., Dennis, M., 2011. Journal of Chromatographic

Science 49 (9), 665–670.Lian, P.C., Zhu, X.F., Liang, S.Z., Li, Z., Yang, W.S., Wang, H.H., 2010. Electrochimica

Acta 55, 3909–3914.Liu, Y., Yu, D.S., Zeng, C., Miao, Z.C., Dai, L.M., 2010. Langmuir 26, 6158–6160.Luo, J., Jiang, S.S., Zhang, H.Y., Jiang, J.Q., Liu, X.Y., 2011. Analytica Chimica Acta

709, 47–53.Ma, W., Chen, W., Yan, W., Kuang, H., Wang, L., Xu, C., Zhu, Y., Xu, L., 2011. Biosensors

and Bioelectronics 26 (11), 4393–4398.Magalhaes, V.F., Marinho, M.M., Domingos, P., Oliveira, A.C., Costa, S.M., Azevedo,

L.O., Azevedo, S.M.F.O., 2003. Toxicon 42, 289–295.Miller, D.L., Kubista, K.D., Rutter, G.M., Ruan, M., Heer, W.A.D., First, P.N., Stroscio,

J.A., 2009. Science 324, 924–927.Roy, S., Soin, N., Bajpai, R., Misra, D.S., Mclaughlin, J.A., 2011. Journal of Materials

Chemistry 21, 14725–14731.Ruangyuttikarn, W., Miksik, I., Pekkoh, J., Peerapornpisal, Y., Deyl, Z., 2004. Journal

of Chromatography B 800 (1/2), 315–319.Sassolas, A., Catanante, G., Fournier, D., Marty, J.L., 2011. Talanta 85 (5), 2498–2503.Shan, H.X., Li, Z.J., Li, M., Ren, G.X., Fang, Y.J., 2008. Journal of Chemical Technology

& Biotechnology 83, 886–891.Shao, Y.Y., Wang, J., Engelhard, M., Wang, C.M., Lin, Y.M., 2010. Journal of Materials

Chemistry 20, 743–748.Singh, G., Choudhary, A., Haranath, D., Joshi, A.G., Singh, N., Singh, S., Pasricha, R.,

2012. Carbon 50, 385–394.Sun, X.L., Shi, H., Wang, H.X., Li, L.N., Xiao, L.W., 2010. Analytical Letters 43 (4/6),

533–544.Wang, Z.J., Zhou, X.Z., Zhang, J., Boey, F., Zhang, H., 2009. Journal of Physical

Chemistry C 11, 14071–14075.Xia, Y., Zhang, J., Jiang, L., 2011. Colloids and Surfaces B 86 (1), 81–86.Xia, Y.T., Deng, J.L., Jiang, L., 2010. Sensors and Actuators B: Chemical 145 (2),

713–719.Xiao, F.G., Zhao, X.L., Tang, J., Gu, X.H., Zhang, J.P., Niu, W.M., 2009. Bulletin of

Environmental Contamination and Toxicology 82 (2), 230–233.Xu, C.L., Ding, L., Ju, H.X., Lei, J.P., Zhang, J., 2010. Analytical Chemistry 82 (3),

1117–1122.Yin, Z.Y., Wu, S.X., Zhou, X.Z., Huang, X., Zhang, Q.H., Boey, F., Zhang, H., 2010.

Small 6, 307–312.Zhang, F.H., Yang, S.H., Kang, T.Y., Cha, G.S., Nam, H., Meyerhoff, M.E., 2007. Biosensors

and Bioelectronics 22 (7), 1419–1425.Zhang, L.F., Ping, X.F., Yang, Z.G., 2004. Talanta 62 (1), 193–200.Zhao, X.L., Sun, W.Y., Liu, Y., Liu, X.J., 2006. Journal of Hygiene Research 35 (1),

76–78.Zhou, G.M., Wang, D.W., Li, F., Zhang, L.L., Li, N., Wu, Z.S., Wen, L., Lu, G.Q., Cheng,

H.M., 2010. Chemistry of Materials 22, 5306–5312.