Electrochemical Immunosensor for Cholera Toxin Using Liposomes

Jan 10, 2006 - A sensitive method for the detection of cholera toxin (CT) using an electrochemical immunosensor with liposomic magnification followed ...
0 downloads 0 Views 276KB Size
Download PDF
Anal. Chem. 2006, 78, 1115-1121

Electrochemical Immunosensor for Cholera Toxin Using Liposomes and Poly(3,4-ethylenedioxythiophene)-Coated Carbon Nanotubes Subramanian Viswanathan,† Li-chen Wu,‡ Ming-Ray Huang,‡ and Ja-an Annie Ho*,†,‡

BioAnalytical Chemistry Laboratory, Department of Chemistry, National Tsing Hua University, Hsinchu, 300 Taiwan, and BioChemistry Laboratory, Department of Applied Chemistry, National Chi-Nan University, Puli, Nantou, 545 Taiwan

A sensitive method for the detection of cholera toxin (CT) using an electrochemical immunosensor with liposomic magnification followed by adsorptive square-wave stripping voltammetry is described. Potassium ferrocyanideencapsulated and ganglioside (GM1)-functionalized liposomes act as highly specific recognition labels for the amplified detection of cholera toxin. The sensing interface consists of monoclonal antibody against the B subunit of CT that is linked to poly(3,4-ethylenedioxythiophene) coated on Nafion-supported multiwalled carbon nanotube caste film on a glassy carbon electrode. The CT is detected by a “sandwich-type” assay on the electronic transducers, where the toxin is first bound to the anti-CT antibody and then to the GM1-functionalized liposome. The potassium ferrocyanide molecules are released from the bounded liposomes on the electrode by lyses with methanolic solution of Triton X-100. The released electroactive marker is measured by adsorptive square-wave stripping voltammetry. The sandwich assay provides the amplification route for the detection of the CT present in ultratrace levels. The calibration curve for CT had a linear range of 10-14-10-7g mL-1. The detection limit of this immunosensor was 10-16 g of cholera toxin (equivalent to 100 µL of 10-15 g mL-1). There is an increasing interest in new analytical techniques capable of fast, sensitive, and reliable detection of pathogenic agents such as bacterial toxins and microbes. These methods would offer effective tools for use in clinical diagnostics, food safety monitoring, epidemic control, and, most recently, counter terrorism campaigns. Detection of bacterial toxins is of particular importance, as many toxins are major determinants of bacterial virulence and thus responsible for toxic activity and infection even without infestation by the microbes. For instance, cholera toxin (CT), secreted by the bacterium Vibrio cholerae, is a known causative agent for diarrhea in humans.1 In recent years, detection * Corresponding author. Fax: +886-3-571-1082. E-mail: [email protected]. † National Tsing Hua University. ‡ National Chi-Nan University. (1) Monyrvuvvo, V.; Pspini, R.; Dvhisbo, H. In Sourcebook of Bacterial Protein Toxins; Alouf, J. E., Freer, J. H., Eds.; Academic Press: New York, 1991; pp 45-56. 10.1021/ac051435d CCC: $33.50 Published on Web 01/10/2006

© 2006 American Chemical Society

of CT has drawn considerable research interest largely because (a) there exists a clinical need for developing such a biosensor and (b) the interaction between CT and its cell surface ligand is well characterized, making it ideal for testing various detection schemes that target bacterial toxins.2-6 CT binds to cell membrane ligand ganglioside (GM1), a glycosphingolipid containing a pentasaccharide headgroup and a ceramide tail. Upon binding, the catalytic domain of CT is inserted into the cell to trigger activation of adenylate cyclase, leading to elevated cAMP levels and, eventually, cell lysis.7 The binding constant of CT to GM1, as recently determined by surface plasmon resonance (SPR) measurement, is in the picomolar range.8 Several analytical methods have been developed for CT where the recognition is based on either GM1 or CT-specific antibody. Ligler and coworkers developed microarray sensors for both direct and “sandwich” immunoassays for CT and expanded them to simultaneous detection of both toxins and bacteria.5 Song and Swanson reported optical CT sensors utilizing fluoroscein and energytransfer fluorophore labeled GM1.9 Liposomes, spherical vesicles composed of a phospholipid bilayer surrounding an aqueous cavity, were originally developed to study cell membranes.10 However, because of their ability to carry different agents in the aqueous cavity, liposomes have been utilized in diagnostics, in drug delivery, and even by the cosmetics and food industries. Liposomes used in detection assay systems occur mostly as immunoliposomes with antibodies bound to the surface and also as DNA- or analyte-tagged liposomes.11,12 GM1-functionalized (2) Taitt, C. R.; Anderson, G. P.; Lingerfelt, B. M.; Feldstein, M. J.; Ligler, F. S. Anal. Chem. 2002, 74, 6114-6120. (3) Zayats, M.; Raitman, O. A.; Chegel, V. I.; Kharitonov A. B.; Willner, I. Anal. Chem. 2002, 74, 4763-4773. (4) Puu, G. Anal. Chem. 2001, 73, 72-79. (5) Rowe-Taitt, C. A.; Cras, J. J.; Patterson, C. H.; Golden, J. P.; Ligler, F. S. Anal. Biochem. 2000, 281, 123-133. (6) Cooper, M. A.; Hansson, A.; Lofas S.; Williams, D. H. Anal. Biochem. 2000, 277, 196-205. (7) Guidebook to Protein Toxins and Their Use in Cell Biology; Rappuoli, R., Montecucco, C., Eds.; Oxford University Press: Oxford, 1997; pp 30-39. (8) Kuziemko, G. M.; Stroh, M.; Stevens, R. C. Biochemistry 1996, 35, 63756384. (9) Song, X.; Swanson, B. I. Anal. Chem. 1999, 71, 2097-2107. (10) Papahadjopoulos, D.; Portis, A.; Pangborn, W. Ann. N. Y. Acad. Sci. U,S,A, 1978, 308, 50-66. (11) Yoon, S. A.; DeCory, T. R.; Baeumner, A. J.; Durst, R. A. Anal. Chem. 2003, 75, 2256-2261.

Analytical Chemistry, Vol. 78, No. 4, February 15, 2006 1115

liposomes have advantages over immunoliposomes because of their specific strong affinity for biological toxins comparable to those of antibodies and amphiphilicity of the GM1. GM1 contains the hydrophobic ceramide, which can be spontaneously incorporated into a lipid bilayer structure, while antibodies require several chemical steps for covalent conjugation to the liposome structure. Colorimetric sensors utilizing conjugated polymers (e.g., polydiacetylenes) have been successfully developed by incorporating GM1 into the optically sensitive lipid membranes prepared by forming bilayer liposomes or through Langmuir-Blodgett technique.13,14 Singh and co-workers reported a method of fluoroimmunoassay for detecting CT with GM1-bearing liposomes.15 Willner and co-workers reported an electrochemical sensor using a quartz crystal microbalance and GM1-functionalized liposomes in which high detection sensitivity was achieved.16 The same group recently reported an ion-selective field-effect transistor sensor coupled with SPR for CT analysis.17 Durst and co-workers developed a bioassay for cholera toxin detection using gangliosideincorporated liposomes.11 Cheng and co-workers reported an electrochemical biosensor for CT using redox lipid microstructures on a gold electrode.18 Another promising method in electrochemical sensing is to explore voltammetric detection of toxins by using surface-modified electrodes. The difficulty in voltammetric signaling, however, is that most toxins are nonelectroactive. Therefore, a specific interface that links binding events to direct or lateral electron transfer must be developed. One important development in CT detection is fabricating electrochemical sensors with high sensitivity and low cost. Recent electrochemical studies have shown the ability of carbon nanotubes (CNTs) to promote certain types of electron-transfer reactions,19 to minimize electrode surface fouling, and to enhance electrocatalytic activity.20 The immobilization of biomolecules such as enzyme or antibody on the surface of CNT nanoelectrodes as biosensors has been reported.21,22 Wang et al.23,24 reported that glassy carbon electrodes modified with single-wall and multiwalled carbon nanotubes (MWCNT) showed a remarkable decrease in the overvoltage for nicotinamide adenine dinucleotide (NADH) oxidation and also circumvented NADH surface-fouling effects during amperometric measurements. Composite materials based on nanotubes (NTs) embedded in polymers have attracted enormous interest because of the expected enhanced strength and improved electronic properties of such materials compared (12) Edwards, A. J.; Durst, R. A. Electroanalysis 1995, 7, 38-845. (13) Charych, D.; Cheng, Q.; Reichert, A.; Kuziemko, G.; Stroh, M.; Nagy, J. O.; Spevak, W.; Stevens, R. C. Chem. Biol. 1996, 3, 113-120. (14) Pan, J. J.; Charych, D. Langmuir 1997, 13, 1365-1367. (15) Singh, A. K.; Harrison S. H.; Schoeniger, J. S. Anal. Chem. 2000, 72, 60196024. (16) Alfonta, L.; Willner, I.; Throckmorton, D. J.; Singh, A. K. Anal. Chem. 2001, 73, 5287-5295. (17) Zayats, M.; Raitman, O. A.; Chegel, V. I.; Kharitonov A. B.; Willner, I. Anal. Chem. 2002, 74, 4763-4773. (18) Cheng, Q.; Zhu, S.; Song J.; Zhang, N. Analyst 2004, 129, 309-314. (19) Baughman, R. H.; Cui, C.; Zakhidov, A. A.; Iqbal, Z.; Barisci, J. N.; Spinks, G. M.; Wallace, G. G.; Mazzoldi, A.; Rossi, D. D.; Rinzler, A. G.; Jaschinski, O.; Roth, S.; Kertesz, M. Science 1999, 284, 1340-1344. (20) Banks, C. E.; Moore, R. R.; Davies, T. J.; Compton, R. G. Chem. Commun. 2004, 16, 1804-1805. (21) Wang, J. Analyst 2005, 130, 421-426. (22) Tsai, Y. C.; Li, S. C.; Chen, J. M. Langmuir 2005, 21, 3653-3658. (23) Musameh, M.; Wang, J.; Merkoci, A.; Lin, Y. Electrochem. Commun. 2002, 4, 743-746. (24) Wang, J.; Musameh, M. Anal. Chem. 2003, 75, 2075-2079.

1116

Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

to bare polymer matrixes.25 Polymer wrapping, peptide coating, and encapsulation in amphiphilic copolymer micelles and cyclodextrins were used to enhance the dispersion of NTs in polar and nonpolar solvents and in polymer matrixes.26 Either chemical or electrochemical techniques were used to obtain conductingpolymer growth onto NTs or NT-doped conducting polymers.27 In particular, NT/polymer composites are popular materials used for sensing applications. Bioelectrochemical sensors have been approached by producing the immobilization of metalloproteins and enzymes within or on NTs by either physical adsorption or covalent binding, often involving the carboxylic functionalities introduced onto the NT surface by oxidizing procedures.28 Nafion, a perfluorinated sulfonated cation exchanger, has been widely used as an electrode modifier due to the chemical inertness, thermal stability, mechanical strength, and antifouling properties. Solubilization of integral NTs by Nafion was also used for the preparation of amperometric biosensors.26,29 Poly(3,4-ethylenedioxythiophene) (PEDOT) is one of the most promising conducting polymers of recent times and has been reported to exhibit better stability and conductivity compared to other conducting polymers.30,31 PEDOT and its composite films have been used as the amperometric sensor for detection of glucose.32,33 Unlike polyaniline, PEDOT is electroactive and conducting in neutral buffer solutions and has been used as a conductometric sensor in an immunoassay, exhibiting superior stability compared to polypyrrole.34 Following the first report of the preparation of a CNT/ polymer composite by Ajayan et al.,35,36 many efforts have been directed toward combining CNT and polymers to produce functional composite materials with superior properties.37-38 In this paper, a new approach for electrochemical immunoassay based on the utilization of encapsulated electrochemical signalgenerating liposomes is described. GM1 liposomes encapsulated with an electroactive redox marker, potassium ferrocyanide, were used for the detection of CT. Conjugated PEDOT-coated MWCNTs were used as the immobilization matrix-cum-transducer. Antibodies against CT were immobilized on a Nafion-MWCNT-PEDOTcoated glassy carbon electrode. CT first binds to anti-CT antibody on the electrode surface and is then sandwiched by the GM1 liposomes. The GM1 elctrochemical immunoassay can be an (25) S. Lefranta, S.; Baibaraca, M.; Baltogb, I.; Mevelleca, J. Y.; Godona, C.; Chauveta, O. Diamond. Relat. Mater. 2005,14, 867-872. (26) Wang, J.; Musameh, M.; Lin, Y. J. Am. Chem. Soc. 2003, 125, 2408-2409. (27) Zhang, X.; King, R. C. Y.; Jose, A.; Manohara, S K. Synth. Met. 2004, 145, 23-29 (28) Tsai, Y. C.; Chen, J. M.; Li, S. C.; Marken, F. Electrochem. Commun. 2004, 6, 917-922. (29) Fan, Z.; Harrison, D. J. Anal. Chem. 1992, 64, 1304-1311. (30) Groenendal, L.; Jonas, F.; Freitag, D.; Pielartzik, H.; Reynolds, J. R. Adv. Mater. 2000, 12, 481-494. (31) Manisankar, P.; Viswanathan, S., Pusphalatha, A. M.; Rani, C. Anal. Chim. Acta 2005, 528, 57-163. (32) Kros, A.; van Ho ¨vell, S. W. F. M.; Sommerdijk, N. A. J. M.; Nolte, R. J. M. Adv. Mater. 2001, 13, 1555-1557. (33) Piro, B.; Dang, L. A.; Dham, M. C.; Fabiano M. C. S.; Canh, T. M. J. Electroanal. Chem. 2001, 512, 101-109. (34) Kanungo, M.; Srivastava, D. N.; Kumar, A.; Contractor, A. Q. Chem. Commun. 2002, 680-681. (35) Ajayan, P. M.; Stephan, O.; Colliex, C.; Trauth, D. Science 1994, 265, 12121214. (36) Rege, K.; Raravikar, N. R.; Kim, D. Y.; Schadler, L. S.; Ajayan, P M.; Dordick, J. S. Nano Lett. 2003, 3, 829-832. (37) Baughman, R. H.; Zakhidov, A. A.; Heer, W. A. Science 2002, 197, 787792. (38) Dai, L.; Mau, A. W. H. Adv. Mater. 2001, 13, 899-913.

alternative method to enzyme-linked immunosorbent assay or other conventional assays for CT, having the advantages of sensitivity, speed, and simplicity. To the best of our knowledge, this is the first time that PEDOT-coated MWCNTs have been used as a voltammetric sensor in an immunoassay for CT. EXPERIMENTAL SECTION Reagents and Materials. Dipalmitoylphosphatidylcholine (DPPC) and dipalmitoylphosphatidylglycerol (DPPG) were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL). The highpurity MWCNTs (diameter 20-40 nm and length 5-15 µm) were purchased from Conyuan Biochemistry Technology Co., Ltd. (Taipei, Taiwan). Mouse monoclonal antibodies to the cholera toxin B subunit were purchased from Biodesign International (Saco, ME). Monosialoganglioside (GM1), cholesterol, N-acetylneuraminic acid, 5% Nafion in a mixture of lower aliphatic alcohols and water, 3,4-ethylenedioxythiophene (EDOT), cholera toxin B subunit, and all other chemicals were purchased from Sigma Aldrich Chemical Co. (St. Louis, MO). The reagents used were of analytical grade or the highest purity commercially available and were used as received. All solutions were prepared with deionized water of a resistivity not less than 18 Ω cm-1 (Milli-Q, Bedford, MA). Different concentrations of cholera toxin were prepared in 0.1 M phosphate-buffered saline (PBS, pH 7.4) and stored at 4 °C. The Nafion solution used in this work was prepared by diluting with PBS. Apparatus. Electrochemical experiments were carried out with a model 660B electrochemical workstation (CH Instruments, Austin, TX). A bare glassy carbon (GCE) working electrode (3mm diameter), a platinum wire counter electrode, an Ag|AgCl reference electrode with 3 M KCl, and a conventional threeelectrode electrochemical cell were all purchased from CH Instruments. The electrode surface morphologies were studied by using a scanning electron microscopy (model JSM 5200, JEOL Ltd.). Conductivity measurements were carried out by the fourpoint probe method. A Beckman Coulter N4 Plus (Beckman Coulter Inc., Fullerton, CA) particle-size analyzer was used for characterization of the lipomoses. Preparation of Electrode. MWCNTs were added into 1 mL of diluted Nafion solution, and this solution was ultrasonicated for 45 min to form a uniform MWCNT-Nafion black suspension. Prior to the surface modification, the bare GCE was activated in two ways: by mechanical polishing over a velvet microcloth with 0.3- and 0.05-µm alumina slurries, and electrochemical treatment by applying potentials of +1.5 and -0.2 V in 0.1 M H2SO4 for 5 and 3 s, respectively. The surface modification was preceded by casting a 5-µL aliquot of MWCNT-Nafion solution. The solvent was allowed to evaporate at room temperature in the air. The immunosensing layers were fabricated on the modified electrode surface by immobilizing different concentrations of cholera toxin antibody during polymerization. Antibody mixed with 50 mM EDOT solution in pH 7.4 PBS (contains 6:1 water/acetonitrile (v/ v)) was electrochemically polymerized by a cyclic voltammetric potential range of -0.5 to 1.2 V versus Ag|AgCl on NafionMWCNT-coated GCE. The thickness of the PEDOT coatings was controlled by integration of the electric charge passed during the polymerization, and (8.0 ( 0.2) × 10-5 C (∼13-15 cycles) was fixed for this study. This resulted in formation of antibodyimmobilized polymeric film on the GC|Nafion-MWCNT elec-

trode. The electrode was then treated with 0.5% poly(vinyl alcohol) (PVA) aqueous solution followed by rinsing with PBS and stored at 4 °C until use. Preparation of Liposomes. GM1-tagged liposomes were prepared by the extrusion method from a lipid mixture.11,39 In short, the lipid mixture consisted of a 10:10:1:0.3 molar ratio of DPPC, cholesterol, DPPG, and GM1. The total lipid mixture was dissolved in 4 mL of a solvent mixture consisting of a 6:6:1 volume ratio of chloroform, isopropyl ether, and methanol, followed by the addition of an aqueous solution of 1 mL of 150 mM potassium ferrocyanide. After the sonication of the mixture for 5 min, the organic solvent was removed by rotary evaporation under reduced pressure, leaving a yellow jelly of liposomes. To this, another aliquot of ferrocyanide was added, followed by an additional 5-min sonication and vortexing at 45 °C. Liposome sizes were regulated with extrusion by passing them through 2- and 0.4-µm polycarbonate filters 20 times and then gel filtration and dialysis to remove free, unencapsulated ferrocyanide. This suspension of liposomes was stored in 0.1 M Tris-buffered saline (TBS, pH 7.4) at 4 °C. Analytical Procedure. The antibody-functionalized electrode interacted with varying concentrations of the CT in 100 µL of 0.1 M PBS solution for 30 min, pH 7.4, at ambient temperature. After attachment of the toxin to the sensing interface, the electrode was incubated with 10 µL of solution of the GM1-functionalized liposomes encapsulated with ferrocyanide for 30 min at 4 °C. The electrode was covered by an electrode cap to avoid drying during the period of incubation. After incubation of the respective electrodes in the probe solution, the electrodes were rinsed with PBS. A 10-µL solution of methanolic solution of Triton X-100 was added onto the electrode surface and allowed to dry for 5 min. Finally, the electrode was introduced into the electrochemical cell containing 5 mL of PBS buffer for the adsorptive square-wave stripping voltammetric analyses. It should be noted that, except for the final step, the electrode in all other steps was rinsed carefully with PBS solution and all the electrochemical experiments were carried out under nitrogen atmosphere. The control experiments were conducted to eliminate the current due to the nonspecific binding of liposomes on the electrode. The squarewave stripping voltammetry was carried out without CT, and it was subtracted from the corresponding stripping signal. Determination of CT in Water Samples. Tap water and domestic kitchen wastewater were boiled, centrifuged, and spiked with different concentrations of CT and stored at 4 °C. The CT analyses were carried out with five replicates for each sample. The peak current was taken from the average five measurements. Safety Note. Cholera toxin is harmful to humans and should be handled with care. Organic solvents used in this study are suspected carcinogens. Handle them with extreme care. RESULTS AND DISCUSSION Liposome Characteristics. The homogeneous ferrocyanideentrapped GM1 liposomes of similar size range are very important in diagnostic applications to obtain better reproducibility. Extrusion of the liposome preparations through polycarbonate filters was found to reduce the size heterogeneity. Liposomes used in the present study were extruded 20 times through 2- and 0.4-µm polycarbonate filters. A mean diameter of 238 nm was determined (39) Ho, J-a. A.; Huang, M. R. Anal. Chem. 2005, 77, 3431-3436.

Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

1117

Table 1. Characteristics of Liposomes mean diameter (nm)

238

entrapped volume of liposome (L) population of liposomes (number/µL) K4[Fe(CN)6] concentration (M) K4[Fe(CN)6] molecules per liposome GM1 on the surface of liposomes

6.3 × 10-18 1.5 × 1012 0.15 5.6 × 104 4200

Scheme 1. Schematic outlines of Immunosensor for Cholera Toxin

by analysis with a Coulter N4 particle-size analyzer. The characteristics of liposomes are listed in Table 1. The total volume of the liposomes (7.1 × 10-18 L) is calculated from the diameter of 238 nm. The entrapped volume (6.3 × 10-18 L) is calculated from an inner diameter of 230 nm (assuming bilayer thickness of 4 nm). Assuming that the ferrocyanide concentration inside the liposomes was equal to the original solution used, and comparing the square-wave voltammetry of lysed liposomes to that of standard potassium ferrocyanide solutions, it is possible to calculate that there were ∼1.5 × 1012 liposomes/µL and that each liposome contained ∼5.6 × 104 molecules of potassium ferrocyanide. A 5-µL aliquot of liposomes was ruptured and the amount of ferrocyanide molecules released was measured. The relative standard deviation of these experiments was 2.0% for five replicates. RSD value indicates that the variance of ferrocyanide molecules in each liposome is very small and also negligible. These results confirm that the size population of the liposome is uniformly distributed in bulk. If the average surface area of the DPPC molecules was 71 Å, and that of cholesterol was 19 Å, it is estimated that 4200 molecules of GM1 was on the outer surface of each liposome, given that 0.7 mol % GM1 was successfully incorporated in the formation of liposomes. The potassium ferrocyanide-containing liposomes during storage were found to be stable at 4 °C in TBS for one month. The goal of the present work is to transfer the stripping immunoassay protocol to a solid-state sensor format by performing the entire heterogeneous immunoassay on the electrode surface. Scheme 1 depicts the outlines of an analytical technique that we developed for the amplified detection of the CT. The surfacemodified electrode fabrication technology is commonly used to increase the electrode performance over the analytes. This technology has been used recently for the development of electrochemical immunosensors. Recent reports have illustrated that the use of conventional electrodes does not compromise the attractive performance of the surface-modified electrode. The oxidative polymerization of EDOT on GC or MWCNTs was carried out by a cyclic voltammetric (CV) range of -0.5 to 1.2 V (vs Ag|AgCl) with a scan rate of 50 mV s-1 at 25 °C. With the increase 1118 Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

Table 2. Comparison of Square-wave Voltammetric Behavior of Electrodes in 1.5 × 10-5 M Potassium Ferrocyanide in 0.1M PBS Buffer at Frequency 15 Hz, Amplitude 0.025V

electrode

peak current ×106 A

peak potential (V)

GC GC|Nafion GC|Nafion MWCNT GC|PEDOT GC|Nafion-MWCNT-PEDOT GC|Nafion-MWCNT-PEDOT-anti CT

-1.014 -0.309 -2.400 -2.428 -3.991 -3.474

0.230 0.257 0.200 0.229 0.221 0.227

in the scanning cycle, the magnitude of the CV redox waves increased, indicating that the PEDOT film is constructed on the surface of the electrode. The CV pattern is similar to that obtained from electrochemical polymerization of EDOT itself under the same reaction conditions. Table 2 compares the square-wave voltammetric responses of bare GC and different modified GCs. All the electrodes give one well-defined oxidation peak around +0.2 V for [Fe(CN)6]4- in neutral PBS buffer. There was a drastic decrease in the peak current if the glassy carbon electrode was coated only with Nafion. This result can easily be understood because Nafion film becomes a barrier in diffusion on the surface of the electrode. The MWCNTs and PEDOT films enhanced the peak currents due to their particular electrical properties. As expected, for a Nafion-MWCNT-PEDOT, electrode behavior exhibits a more-favorable response when using voltammetric analysis. The conductivity of MWCNT/PEDOT composite films (1.5 × 10-1 S cm-1) is comparatively 50 times higher than PEDOT (2.7 × 10-3 S cm-1). This may also be attributed to the doping effect associated with Nafion and MWCNTs, which were believed to help induce the formation of a more efficient matrix for the charge transport, thus enhancing the conductivity of the films. Nanocomposite film behaves as arrays of very small electrodes; hence, the Nafion-MWCNT-PEDOT-modified GC electrode has an efficient surface area to take part in this analysis. The concept of the conducting polymer electrode approach was that, when a positive charge occurred on the electrode surface during electropolymerization, the negative charge protein was entrapped in the cationic polymer network. CT antibody molecules are immobilized on the Nafion-MWCNT-coated glassy carbon electrode during the electropolymerization of EDOT. It is apparent that the conventional polymer protein entrapment usually settled in bulk thus creates steric hindrance to the assessment of either antigen or antibody to the entrapped counterimmunological moiety. But the electron-induced 3,4-ethylenedioxythiophene polymer coated on the electrode surface can only be accumulated no more than a few layers. Hence, the electrogenerated polymers cause only a slight loss of biological activity for antibody. However, this slight decrease was compensated for by the polymer thin film grown on the carbon nanotube. The higher surface area of the carbon nanotube-modified electrode provides more accommodation for antibodies on the electrode. This antibody immobilization strategy shows the good reproducibility of immunosensors. The MWCNTNafion-PEDOT, anti-CT antibody composite film is able to bind CT, and its morphology is directly observable by scanning electron microscope (SEM). The SEM images of PEDOT and Nafion-

Figure 1. SEM images of the surfaces of the (A) GC|PEDOT and (B) GC|Nafion-MWCNT-PEDOT electrodes.

Figure 2. Square-wave voltammetric responses before (solid circle) and after (solid line) immobilization of anti-CT antibody on GC|NafionMWCNT-PEDOT electrodes in 1.5 × 10-5 M potassium ferrocyanide in 0.1 M PBS buffer at frequency 15 Hz, amplitude 0.025 V.

MWCNT-PEDOT coated on a glassy carbon surface are shown in Figure 1. The SEM image features an entangled aggregate of structures (Figure 1B) that is totally different from circular aggregates observed for the PEDOT (Figure 1A). There are no significant changes in the square-wave voltammetric response of selected modified electrodes after immobilization of anti-CT antibody (Figure 2). Figure 3 shows the response of the immunosensor as a function of percent Nafion, MWCNT, anti-CT antibody concentration for the modified electrode surface composition. In these cases, different concentrations of anti-CT antibody were immobilized into the polymer matrix during

polymerization. It is clear from Figure 3 that the sensor response increased significantly when the concentration of antibody loaded on the electrode surface was changed from 2 to 40 µg mL-1. With further increase in the antibody concentration to 40-70 µg mL-1, the increase of the sensor response was not observed. It is likely that the anti-CT antibody strongly binds this modified electrode through a covalent bond by sulfonic acid groups of Nafion films, and the electron-deficient center of redox polymer PEDOT also promotes the orientation and affinity toward the anti-CT antibody. Direct immobilization minimizes the organic overload on the electrode surface.38 The use of a blocking agent, 0.5% PVA, reduced the nonspecific binding of GM1 liposomes on the electrode surface during immunological incubation. The proposed immunoassay using GM1 liposomes showed strong specific affinity toward CT. The reported binding affinity of cholera toxin for the ganglioside is ∼4.6 × 10-12 M.8 This high specificity promotes the immunological determination of CT over other methods. With the presence of CT in the analyzed sample, the toxin binds to the anti-CT antibody-immobilized electrode. The effect of the incubation time was also studied. The anti-CT immobilized electrode was allowed to incubate with 100 µL of solution of CT for different periods of time. The results show that the SWSV signal increased rapidly with the incubation time up to 30 min, and then the electrode did not show any significant increases in signal as incubation time increased (Table 3). Thus, 30 min was selected as the optimum. Association of the CT with the sensing interface is then amplified by the binding of GM1 liposomes encapsulated with ferrocyanide for 30 min. The amount of bound liposomes on the electrode surface is proportional to the CT concentration. A range of concentrations of Triton X-100 and percentage methanol were examined for the ability to lyse

Figure 3. Influence of the electrode compositions on the CT electrochemical immunosensor.

Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

1119

Table 3. Optimum Parameters for Electrode Composition and Adsorptive Square-Wave Stripping Voltammetry for Cholera Toxin variables

unit

range studied optimum

Nafion MWCNT antibody concentration incubation time for electrode with CT lyse solution % of Triton X-100 Lyse solution % of methanol square-wave variables accumulation time accumulation potential quiet time quiet potential initial scan potential scan increment amplitude frequency

% (v/v) mg mL-1 µg mL-1 min

0.10-1.0 2.0-16 2.0-70 0-120

0.50 10. 40. 30.

% (v/v) % (v/v)

0.050-0.50 10.-50

0.10 25

s V s V V V V Hz

0-120 -0.10 to 0.10 1.0-10. 0.10 to -0.10 -0.30 to 0.00 0.0010-0.010 0.010-0.050 5.0-50

60. 0.020 5.0 0.00 -0.20 0.0040 0.025 15

the liposomes and release the K4[Fe(CN)6] molecules into the electrode matrix. Optimal lysis occurred at 0.1% (v/v) Triton X-100 in 25% (v/v) methanol-water solution concentration levels within 5 min. The released ferrocyanide markers are easily adsorbed on the electrode surface. Optimization of Adsorptive Square-Wave Stripping Voltammetry. AdSWSV follows nonelectrolytic (adsorptive) preconcentration and stripping. The optimization of both steps is important for analytical procedures. Many preconcentration stripping experiments were performed for accumulation potentials (Eacc) varying from -0.10 to 0.10V and at an accumulation time (tacc) of 90 s to evaluate the accumulation behavior of potassium ferrocyanide at the active sites of the electrode surface. Maximum peak current is observed for an accumulation potential of 0.02 V. This may be due to the electrostatic interaction between the nature of the electrode at this potential and the charge of the substrate. Hence, an Eacc of 0.02 V is chosen as optimum for further studies. The dependence of the stripping peak current on the accumulation time showed linearity for an oxidation peak at 0.20 V up to 60 s. There is no significant increase in peak current after a 60-s accumulation period. Hence, an accumulation time of 60 s is selected as optimum for stripping studies. During the accumulation step, the usual electrostatic interaction occurs between the negatively charged [Fe(CN)6]4- and the positively charged conducting polymer would result in a biosensor with a more efficient electron transfer. The applied Eacc potential prevents the leaching of already-accumulated ferrocyanide from the electrode surface. The initial scan potential (Eis) is also an important parameter similar to accumulation potential. It also controls both peak potential and peak height in the stripping voltammogram. The Eis was varied between -0.30 and 0.00 V. The maximum stripping current was observed at -0.20 V. For the optimization of square-wave conditions, the square-wave frequency, the scan increment, and square-wave amplitude were examined, varying one of them and maintaining the others constant. The variable ranges studied were 0.01-0.05 V for the square-wave amplitude, 0.001-0.010 V for the scan increment, and 5-50 Hz for the frequency. The peak current increases with increase in these parameters. The peak width increases at higher scan-increment values. The background current and the potential increase at 1120 Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

Figure 4. Calibration plot for cholera toxin. The inset shows a linear part of the main curve.

higher frequency values. Hence, square-wave frequency of 15 Hz is chosen as optimum. The optimum square-wave stripping parameters for the CT are summarized in Table 3. Calibration Curve. The analytical calibration for CT was conducted to determine the sensitivity of the immunosensor to CT, and the immunosensors were tested with different concentrations of the target analyte (10-17-10-5 g mL-1). The calibration curve for the voltammetric detection of CT was carried out by recording, for each concentration, the response of a modified electrode at optimum experimental conditions (Figure 4). Each point of the calibration graph corresponds to the mean value obtained from five independent measurements. The limit of detection (LOD) was defined as the lowest concentration of toxin producing a peak current 3 times higher than the standard deviation of the peak current in the absence of CT under identical conditions. A linear current variation was observed over the range 10-14-10-7 g mL-1 while the very sensitive LOD was obtained, namely, 10-15 g mL-1. The immunosensor response and hence its construction was also quite reproducible; five immunosensors were prepared by following identical electropolymerization steps, and their responses toward CT (10-11 g mL-1) led to a relative standard deviation of only 4.2%. At higher concentrations of CT, the peak current measured by the electrode reaches saturation, followed by poor performance of the electrode due to the overload of organics. Determination of CT in Water Samples. The proposed method was applied to the analysis of the water samples spiked with CT. For this purpose, the standard addition method was used to eliminate the matrix effects. The method of multiple standard additions was applied for synthetic samples prepared with tap water and domestic kitchen wastewater. To evaluate the analytical applicability of the proposed method, the immunosensor was used to determine three spiked samples with different CT concentrations. Figure 5 shows typical stripping voltammograms for CT present in tap water and wastewater. The precision of CT was calculated from seven repeated analyses at different intervals. The percentage of recoveries with standard error is presented in Table 4. The accuracy of this method is understood from low RSD values. The results are listed in Table 4. The recoveries were all in the range 90-98%, which demonstrates that the immunosensor meets the test requirements for environmental samples.

Figure 5. CT-spiked water samples: (A) tap water and (B) domestic kitchen wastewater with 10-11 and 10-14 g mL-1, respectively. Table 4. Results of Recovery Studies from Water Samples Spiked with CT

samples

CT g mL-1 added found CT

CT spiked in tap water 1 × 10-8 1 × 10-11 1 × 10-13 CT spiked in kitchen 1 × 10-8 wastewater 1 × 10-11 1 × 10-13

% RSD recovery (n ) 7)

0.98 × 10-8 0.96 × 10-11 0.94 × 10-13 0.90 × 10-8

98 96 94 90

3.5 4.3 4.7 4.9

0.92 × 10-11 0.90 × 10-13

92 90

5.1 5.5

CONCLUSION We have demonstrated herein the development of a sensitive detection immunoassay for cholera toxin. Since CT is toxic in very small quantities, sensitive detection methods are required. Various detection assays for CT have been reported, including immunological assays and polymerase chain reaction methods, and some of the recent assays using amplification techniques have shown high sensitivity. However, most of these assays are costly, laborious, and time-consuming. We have demonstrated that the combination of MWCNT-PEDOT-modified electrodes with an anti-CT-immobilized immunosensor and GM1-functionalized, ferrocyanide-encapsulated liposome operation yields an analytically attractive performance. It was expected that when MWCNTs were fabricated by electron-conductive polymers, the resultant composites would provide a convenient method to prepare an electrochemical immunosensor with new electrochemical functions. The fibrous morphology of MWCNTs-PEDOT composites is novel, reflecting that of MWCNTs. Furthermore, the ability of the anti-CT antibody, trapped in PEDOT, to bind CT was well retained. In this system, therefore, all functions of Nafion, MWCNTs, and electron-conductive PEDOT are cooperatively and (40) Ionescu, R. E.; Gondran, C.; Cosnier, S.; Gheber, L. A.; Marks, R. S. Talanta 2005, 66, 15-20.

efficiently utilized. We believe that this fabrication method, as well as the resultant composites, will be useful in the design of novel functional materials. In this study, ganglioside GM1 liposomes were prepared to replace the secondary antibody or marker enzyme used in previous reports.40 It has been well demonstrated that the ganglioside GM1 is a natural cell membrane receptor for CT, and the interaction between CT and GM1 is very strong and specific. Since gangliosides have both hydrophilic and hydrophobic portions, GM1 liposomes can be prepared without the complicated chemical steps that are required for the preparation of immunoliposomes. The technique developed in this study showed much higher sensitivity than others previously reported, with a limit of detection of 10-16 g of cholera toxin (equivalent to 100 µL of 10-15 g mL-1) and an exceptional dynamic range of at least 5 orders of magnitude. This bioassay could also be applied to the detection of CT in water samples with an acceptable loss of sensitivity. These results show this electrochemical immunosensor, using GM1 liposomes, can be a simple, sensitive assay system for CT detection. ACKNOWLEDGMENT This work was supported by the National Science Council in Taiwan, ROC, under Grants NSC 93-2113-M-007-051 and 94-2811M-007-041. Note Added after ASAP Publication. The article was posted on the Web on 1/10/06. The author affiliation for Ho was changed to include National Chi-Nan University. The paper was reposted on 1/12/06.

Received for review August 10, 2005. Accepted November 30, 2005. AC051435D

Analytical Chemistry, Vol. 78, No. 4, February 15, 2006

1121