A rapid competitive binding nonseparation electrochemical enzyme immunoassay (NEEIA) test strip for microcystin-LR (MCLR) determination

Biosensors and Bioelectronics 22 (2007) 1419–1425

A rapid competitive binding nonseparation electrochemical enzyme immunoassay (NEEIA) test strip for microcystin-LR (MCLR) determination Fenghua Zhang a , Soon Hye Yang a , Tae Young Kang a , Geun Sig Cha a , Hakhyun Nam a,∗ , Mark E. Meyerhoff b a

Chemical Sensor Research Group, Department of Chemistry, Kwangwoon University, Seoul 139-701, Republic of Korea b Department of Chemistry, The University of Michigan, Ann Arbor, MI 48109, USA Received 12 February 2006; received in revised form 7 June 2006; accepted 15 June 2006 Available online 2 August 2006

Abstract A competitive binding nonseparation electrochemical enzyme immunoassay (NEEIA) is described for the determination of microcystin-LR (MCLR) using a double-sided microporous gold electrode in cartridge-type cells. A gold film sputtered on one side of porous nylon membrane constitutes a working electrode, while another gold film formed on the opposite side serves as a pseudo reference electrode. After immobilizing MCLR antibody on working electrode by physical adsorption, the double-sided electrode was placed simply in a diffusion U-type or within a dry strip-type cell with a conjugate pad pre-loaded with a glucose oxidase labeled MCLR (GOx-MCLR) on working electrode side. Assays were performed in two steps: an MCLR-containing sample mixed with a known amount of GOx-MCLR conjugate either in buffer solution or in preloaded dry pad was incubated for an appropriate period (about 10 min) to induce competitive reaction with an immobilized anti-MCLR antibody on working electrode, and a fixed concentration of glucose solution (substrate) was then added to the backside of the working electrode. Due to the competitive nature of the assay, enzymatically generated product, hydrogen peroxide (H2 O2 ), was detected at the working gold electrode (at +800 mV versus Au) by oxidation, and the magnitude of amperometric current was inversely proportional to the concentration of MCLR in the sample. The response time after substrate addition was about 30 s. Mean recovery of MCLR added to tap water was 93.5%, with a coefficient of variation (CV) of 6.6%. The proposed competitive NEEIA system is in general comparable to existing heterogeneous enzyme immunoassays with a similar detection limit (100 pg/mL MCLR), and suitable for developing a disposable type biosensor for on-site monitoring of environment. © 2006 Elsevier B.V. All rights reserved. Keywords: Competitive nonseparation electrochemical enzyme immunoassay; Microcystin-LR; Glucose oxidase; Double-sided microporous gold electrode

1. Introduction Microcystins (MCs) are a family of hepatotoxic cyclic heptapeptides produced by freshwater cyanobacteria (blue green algae), and the one containing leucine (L) and arginine (R) (microcystin-LR or MCLR) was the first identified MC species and now known to be most toxic (McElhiney and Lawton, 2005; Fawell et al., 1993). Since the management of surface and drinking water is essential to protect both human and animal, it has become very important to develop rapid, reliable, and accurate analytical methods to detect MCLR (Kim et al., 2003). Of the several known methods (e.g., high performance liquid chromatography (HPLC), liquid chromatography/mass

∗

Corresponding author. Fax: +82 2 911 8584. E-mail address: [email protected] (H. Nam).

0956-5663/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2006.06.021

spectrometry (LC/MS), thin-layer chromatography (TLC) and capillary electrophoresis (CE)), enzyme-linked immunosorbant assay (ELISA) has been widely employed to monitor MCLR at levels below 1 ␮g L−1 , the provisional guideline of World Health Organization (WHO) (Yu et al., 2002; Pyo et al., 2005; Nagata et al., 1997). Although ELISA may be used for on-site environmental screening (Keay and McNeil, 1998), it is still a laboratory-based analytical tool (Meulenberg et al., 1995) that requires some instrumentation, trained technicians, and long assay time. An electrochemical enzyme immunoassay sensor that combines the highly specific ELISA with the sensitivity and relative simplicity of an electroanalytical measurement has been suggested as a possible on-site monitoring device (Rishpon and Ivnitski, 1997). However, the necessity to separate free from the bound analytes and labels via repeated washing steps has been the major obstacle in the development of a rapid elec-

1420

F. Zhang et al. / Biosensors and Bioelectronics 22 (2007) 1419–1425

trochemical enzyme immunoassay sensing device (Lin and Ju, 2005). Although several nonseparation electrochemical enzyme immunoassay methods have been proposed (Ho et al., 1995; Ivnitski and Rishpon, 1996; Killard et al., 2000; Duan and Meyerhoff, 1994, 1995; Meyerhoff et al., 1995; Ducey et al., 1997), they have not yet been applied in practice for on-site monitoring. For example, the nonseparation enzyme immunoassay system developed by Meyerhoff et al. used an analyte-specific antibody immobilized on a microporous working gold electrode formed on a nylon membrane that was mounted in an U-type diffusion cell. Analytes and the corresponding enzyme conjugate were applied to the solution bathing the working electrode (front) side while the substrate was introduced from the backside. Diffusion of substrate from the back to the working electrode on the front, provided a means to determine enzyme activity immunobound to the working electrode surface, without need to wash away any unbound conjugate in the bulk sample solution (Duan and Meyerhoff, 1994, 1995; Meyerhoff et al., 1995; Ducey et al., 1997). However, the proposed system may be used mainly in laboratory setting as it requires a special cell and a separate reference electrode. We recently reported that the microporous gold electrodes formed on both sides of nylon membrane constitute a thin-layertype two-electrode electrochemical cell that does not require a separate reference electrode, and they could be used readily in the preparation of a self-sampling-and-flow biosensor system for long term monitoring and also a new generation of mediatorless biosensors (Zhang et al., 2005, 2006). In this work, we employ the same type of double-sided microporous gold electrode and describe a competitive binding nonseparation electrochemical enzyme immunoassay (NEEIA) system for the quantitative determination of MCLR in an U-type diffusion cell arrangement, and further applied this detection principle to develop a rapid NEEIA strip-test for detecting MCLR.

0.1% (w/v) Tween-20 and 0.2% (w/v) BSA for enzyme-linked immunosorbent assay (ELISA), and a Tris buffered gelatin (TBG) (0.01 M Tris, pH 7.4 and 0.1%, w/v gelatin) and tap water for NEEIA. The p-nitrophenyl phosphate solution was prepared in a diethanolamine buffer (1.0 M, pH 9.8) containing 50 mM MgCl2 and the glucose solution was prepared in an acetate buffer (1 M, pH 5.1). All aqueous solutions were prepared with deionized water (18 M cm). 2.2. Preparation of double-sided microporous gold electrodes Double-sided microporous gold electrodes were prepared according to a procedure reported previously (Zhang et al., 2005, 2006). Briefly, microporous nylon membranes (15 mm × 30 mm) were placed under a mask having an appropriate aperture pattern; a circular shaped working elec˚ is connected to a lead trode (o.d. = 4 mm; thickness ≈ 600 A) line (2 mm × 14 mm) with rectangular electrical contact end (4 mm × 8 mm). Both surfaces of nylon membranes were coated with gold using plasma sputter (LVC-76 Sputter Coater; Torr International, Inc., New Windsor, NY, USA) under the following experimental conditions: sputtering time 500 s, pressure 75 m Torr, plasma current 25–30 mA, potential 350–500 V. Deposition of gold on both sides of nylon membrane under the given condition did not result in an electrical shunt between the two electrodes, while preserving the microporosity of the electrodes (Zhang et al., 2005, 2006). The entire area of the double-sided electrodes was covered with a thin PVC film [33% PVC and 67% bis(2-ethylhexyl)sebacate (all w/w, %) dissolved in tetrahydrofuran (1:6 w/v)], exposing only the working and counter electrode sites, and the electrical contacts. This allows the sample components and or substrate to pass only through the boundaries defined by the areas of the circular working and pseudo reference electrodes.

2. Materials and methods 2.1. Reagents

2.3. Electrochemical properties of double-sided microporous gold electrode

Microcystin-LR (MCLR), glucose oxidase (GOx; EC 1.1.3.4, type VII-S, 245 900 units/g, from Aspergillus Niger), alkaline phosphatase (ALP; EC 3.1.3.1, 10–30 DEA units/mg, from bovine intestinal mucosa), ␤-d(+)-glucose, p-nitrophenyl phosphate, tris(hydroxymethyl)aminomethane (Tris), 1-ethyl3,3-dimethylaminopropyl carbodiimide (EDAC), N-hydroxysulfosuccinimide (NHS), bovine serum albumin (BSA) and Tween-20 were purchased from Sigma (St. Louis, MO, USA); anti-MCLR antibodies were from BoDiTech Inc. (Chunchon Kangwondo, Korea); sodium hydrogen carbonate and sodium carbonate were from Yakuri Pure Chemicals (Kyoto, Japan); acetic acid was from Kanto Chemical (Tokyo, Japan); Nytran neutral microporous nylon membranes (pore size: 0.4 ␮m) were from Schleicher & Schuell (Keene, New Hampshire); CytoSep fibers (thickness: 640 ␮m) were from Pall Corporation (Ann Arbor, MI, USA). All other chemicals employed were of analytical grade. Sample solutions were prepared in a carbonate buffer (0.1 M, pH 9.6) containing 0.01% (w/v) NaN3 ,

The scanning electron micrograph (SEM) of a typical nylon ˚ (SEM image membrane coated with gold (thickness ≈ 600 A) not shown) showed that the gold-coated nylon membrane remains completely microporous (average pore size ≈ 0.2 ␮m), and that the two coated sides form a thin-layer cell with a faceto-face electrode structure (Zhang et al., 2005, 2006). The two microscopically rough electrodes patterned on the porous membrane also provide greatly increased surface area for electron transfer reactions and for the immobilization of various bioreagents. In accordance with a previously reported procedure (Zhang et al., 2005, 2006), the electrochemical behavior of the doublesided microporous gold electrode system was examined by cyclic voltammograms (CVs) using 1 mM [Fe(CN)6 ]3− ions (EG&G PAR Model 273A potentiostat/galvanostat, Princeton, NJ, USA). A pair of well-defined reversible waves for [Fe(CN)6 ]4−/3− couple was obtained and the separation of the anodic and cathodic peak potential was about 60 mV. The CVs

F. Zhang et al. / Biosensors and Bioelectronics 22 (2007) 1419–1425

1421

Scheme 1. Scheme for the activation of MCLR using EDAC and NHS for the conjugation with GOx.

obtained with two (gold/porous nylon/gold) and three electrode (gold/porous nylon/gold + Ag/AgCl external reference) systems by varying the concentrations of representative redox couple (Fe2+ /Fe3+ ) and scanning speed at fixed redox concentration were similar in terms of their electrochemical behavior, supporting the notion that the proposed two-electrode system may be used for constructing practical electroanalytical devices.

and 2.3 mg NHS overnight at 4◦ C. The ALP-MCLR and GOxMCLR conjugate were then dialyzed against Tris–HCl buffer (0.05 M, pH 7.4) at 4◦ C for 2 d while changing the buffer solution twice a day. The dialyzed ALP-MCLR and GOx-MCLR conjugate were stored at 4◦ C prior to use.

2.4. Immobilization of anti-MCLR antibody

In order to develop a rapid NEEIA strip-test device, the sample and substrate chambers of an U-type cell (see Fig. 1) were replaced with CytoSep fiber pads. The conjugate loaded pads were prepared by treating CytoSep glass fiber pad with a 0.01 M Tris–HCl buffer, pH 7.4 containing 2% BSA and 0.1% Tween-20, followed by 40 ␮L of 1.25–100 ␮g/mL GOx-MCLR conjugate; they were stored overnight at 4◦ C prior to use in the strip-type cell shown in Fig. 2b. The untreated CytoSep glass fiber was used as the substrate pads. Both substrate and conjugate pads were cut out in circular shape (o.d. = 10 mm) with a borer. The strip-type NEEIA was assembled by stacking up, in order, from the bottom block, a pierce of substrate pad-fixing adhesion tape, a circular shaped glass fiber pad (for substrate applica-

Anti-MCLR antibody was immobilized on the surface of the working gold electrode by pipetting 10 ␮L of the antibody (1 ␮g/mL) in 0.1 M carbonate buffer, pH 9.6, directly onto the electrode and allowing physical adsorption. Electrodes were stored at 4◦ C for 24 h. Electrodes were then rinsed with Tris buffered gelatin (TBG) (0.01 M Tris, pH 7.4 and 0.1%, w/v gelatin) to remove unbound antibodies, and then allowed to dry prior to mounting either in a diffusion U- or strip-cell for competitive NEEIA measurements. Anti-MCLR antibody was also immobilized on the surface of the 96-well polystyrene-based flat bottom microtiter plate (Franklin Lakes, NJ, USA) by pipetting 100 ␮L of antibody (33 ␮g/mL) in a carbonate buffer (0.1 M, pH 9.6) and incubated at 4◦ C for 16 h. After washing with a carbonate buffer (0.1 M, pH 9.6) using Multi Wash PLUS (Grass Valley, CA, USA), the unbound surfaces of the microtiter plate wells were blocked with 3% (w/v) BSA and 0.01% NaN3 in carbonate buffer at room temperature for 30 min. The competitive ELISA was performed using ALP-MCLR conjugate.

2.6. Preparation of strip-type NEEIA cell

2.5. Preparation of ALP-MCLR and GOx-MCLR conjugates As shown in Scheme 1, the conjugation of MCLR to ALP and GOx were achieved by dissolving 1 mg of MCLR in 160 mL of dimethylformamide (DMF) and dimethyl sulfoxide (DMSO) (1:2.3), and cross-linking with ALP (20 mg in 100 mL of carbonate buffer (0.1 M, pH 9.6)) and GOx (403 mg in 100 mL of phosphate buffer (0.05 M, pH 7.4)) by adding 3.8 g EDAC

Fig. 1. Competitive nonseparation electrochemical enzyme immunoassay (NEEIA) in an U-type diffusion cell constructed with a double-sided microporous gold electrode.

1422

F. Zhang et al. / Biosensors and Bioelectronics 22 (2007) 1419–1425

Fig. 2. The NEEIA-based strip-test cell: (a) expanded structure and (b) assembled cell.

tion), anti-MCLR immobilized double-sided microporous gold electrode, GOx-MCLR-containing pad, a pierce of conjugate pad-fixing adhesion tape and the upper block. A schematic of the disassembled structure is shown in Fig. 2a. There are four NEEIA strips in the assembled block for multiple tests at the same time. 2.7. Competitive ELISA In order to perform ELISA, 50 ␮L of sample and 50 ␮L of 0.25–20 ␮g/mL ALP-MCLR conjugate in carbonate buffer (0.1 M, pH 9.6) containing 0.01% (w/v) NaN3 , 0.1% (w/v) Tween-20 and 0.2% (w/v) BSA were pipetted into the antiMCLR antibody-immobilized microtiter plate wells and incubated at room temperature for 2 h. After washing with carbonate buffer (0.1 M, pH 9.6) using Multi Wash PLUS, 100 ␮L

of 10 mM p-nitrophenyl phosphate in diethanolamine buffer (1.0 M, pH 9.8) containing 50 mM MgCl2 was then pipetted into the microtiter plate wells at room temperature for 30 min. After washing with distilled water containing 0.1% Tween-20, the competitive ELISA measurements were carried out by measuring the absorbance changes of p-nitrophenol, an enzymatic product of ALP, at the wavelength of 405 nm using EmaxTM precision microplate reader (Sunnyvale, CA, USA). 2.8. Competitive NEEIA Competitive NEEIA measurements were also carried out in a conventional diffusion cell (similar to that described by Ducey et al.) by pipetting 2 mL of TBG, pH 7.4, containing free and GOx labeled MCLR into the sample chamber of the cell. Following a 10 min incubation of the sample/GOx-MCLR conjugate

F. Zhang et al. / Biosensors and Bioelectronics 22 (2007) 1419–1425

mixture in the working electrode side (with mixing) 2 mL of substrate solution (400 ␮g/mL of glucose in 0.01 M acetate buffer) was added to the opposite side. The enzymatic reaction between the diffused substrate and bound GOx-MCLR conjugate at the surface of the working electrode was monitored by applying +800 mV versus Au with an eight-channel potentiostat (Elbio, Seoul, Korea). The applied potential was determined by measuring the current changes versus varying potentials from 0 to +1 V at 50 mV/s of scan rate using an EG&G PAR Model 273A potentiostat/galvanostat (Princeton, NJ, USA). A similar procedure was followed for the strip-test cell except that a different concentration of substrate (800 ␮g/mL of glucose in 0.01 M acetate buffer) was added to the backside pad of the strip-cell. A steady-state current was measured after about 30 s using this arrangement. A calibration curve for MCLR was obtained by plotting the steady-state currents versus MCLR concentrations from 10−13 to 10−6 M. Calibration curves were also obtained for tap water samples spiked with various concentrations of MCLR. 3. Results and discussion The NEEIA sensors for MCLR determination were prepared either using a conventional diffusion (U-type) or strip-test cell configuration. As described in Section 2, the dry strip-test

1423

NEEIA sensors evolved from the U-type cell with the doublesided microporous gold electrodes. The amounts of necessary reagents, i.e., anti-MCLR antibody, GOx-MCLR conjugate and glucose substrate, were optimized first with the conventional diffusion cell and further developed into the strip-test sensor system. The optimized sensors were then used for the determination of MCLR in drinking water. Each process is described in detail as follows. 3.1. Optimization of reagent concentrations The reagent concentrations for NEEIA strips were optimized in three steps. The concentration of substrate was determined first by measuring the change in amperometric signal (I = Imax − Ibackground ) in the presence of excess GOx-MCLR conjugate (50 ␮g/mL) with the electrodes immobilized with sufficient amount of anti-MCLR (5 ␮g/mL). When the glucose concentrations were varied from 100 to 1000 ␮g/mL, I gradually increased and nearly leveled over 400 ␮g/mL in the conventional diffusion cell. On the other hand, the strip-test cell required a higher concentration, 800 ␮g/mL. Then, the amount of antibody immobilized on the working electrode was optimized using a fixed concentration of substrate (as optimized) and with sufficient amount of conjugate (50 ␮g/mL). When the concentration of anti-MCLR was varied from 0.125 to 5 ␮g/mL (in 0.1 M

Fig. 3. Principle of the competitive NEEIA for the detection of MCLR.

1424

F. Zhang et al. / Biosensors and Bioelectronics 22 (2007) 1419–1425

carbonate buffer, pH 9.6), the maximum current response was observed at 0.25 ␮g/mL for both the diffusion and strip-test cells. This result indicates that an excessive amount of antibody immobilized on the surface of the electrode can impede the electron transfer process for the oxidation of hydrogen peroxide, the optimal concentration of the glucose oxidase reaction. Likewise, the optimal concentration of GOx-MCLR conjugate was also determined with the electrode that contain optimized amount of antibody using optimized glucose concentrations. The conjugate concentrations that yielded the optimal zero-dose analyte signals were 20 ␮g/mL in diffusion cell arrangement, and 12 ␮g/mL in strip-test configuration. 3.2. NEEIA for MCLR determination in U-type diffusion cell The original principle of the NEEIA measurement, as described by Duan and Meyerhoff, is depicted in Fig. 3 for the quantitative determination of MCLR using anti-MCLR antibody immobilized on one electrode of the double-sided microporous electrode configuration. Since glucose is scarcely present in the environment, while GOx is one of the most robust enzymes, the system in Fig. 3 may constitute a highly sensitive and interference-free detection method for MCLR. After optimization of reagents, this approach was first tested with the conventional diffusion cell shown in Fig. 1. Initially, the sample chamber was filled with 1.8 mL of 0.01 M TBG buffer (pH 7.4) while the substrate chamber with 1.9 mL of 0.01 M acetate buffer. To initiate the competitive NEEIA measurement, 100 ␮L of 20 ␮g/mL GOx-MCLR and 100 ␮L of standard MCLR were simultaneously pipetted into the sample chamber while stirring the mixture. As the competitive binding reaction was proceeding, a constant potential of +800 mV (versus Au) was applied to the working electrode. A steady-state current was observed after 8 min of incubation, indicating the competitive binding reaction between the conjugated and free MCLR over immobilized antibody sites was near completion. To initiate the diffusion controlled enzymatic reaction between the glucose and the immuno-bound MCLR-GOx at the surface of the working electrode, stirring was stopped for 2 min before adding 100 ␮L of 400 ␮g/mL glucose into the substrate chamber in contact with the backside of the membrane. The amperometric measurements were made at 30 s after adding the glucose at various MCLR concentrations from 10−13 to 10−6 M. The typical calibration plot is shown in Fig. 4a; a standard curve for MCLR in TBG buffer using the conventional diffusion cell exhibited a linear range from 3.16 × 10−9 to 1 × 10−11 M, proving that the NEEIA method is comparable to other existing heterogeneous MCLR assays, in terms of assay sensitivity (Kim et al., 2003; Lei et al., 2004). The results of MCLR determination with U-type NEEIA cell were directly compared with those from conventional ELISA. The ELISA determination yields a sigmoid calibration curve similar to Fig. 4a with a little wider detection range from 1.38 × 10−7 to 2.14 × 10−11 . Considering the differences in conjugated enzyme activities (GOx and ALP), assay times (11 min versus 2.5 h) and the methods of determination (electro-

Fig. 4. Typical calibration curves of the competitive NEEIA for MCLR determination obtained with: (a) U-type and (b) with strip-type cells (n = 4 for each point).

chemical versus optical), we may conclude that the NEEIA and ELISA are essentially identical in their analytical performances. 3.3. NEEIA for MCLR determination with dry strip-cell Since the U-type NEEIA cell is not readily applicable for field testing, we replaced the two liquid-holding chambers with fiber pads, making the U-type NEEIA cell as a dry strip device. A predetermined amount of GOx-MCLR conjugate (20 ␮g/mL) was loaded onto the conjugate pad and dried. It was then mounted on the antibody-immobilized side of the porous electrode as shown in Fig. 2. The pad placed on the counter electrode side (backside) does not contain any reagent. We initially prepared the dry strip with the substrate-loaded pad, assuming that the sample

F. Zhang et al. / Biosensors and Bioelectronics 22 (2007) 1419–1425 Table 1 Recovery of MCLR added to tap water MCLR added (pg/mL)

3160 1000 316 100

Recovery

CV (%)

pg/mL

%

3006 937 294 52

95.1 93.7 93.0 92.0 93.5

4.9 6.3 7.0 8.0 6.6

solution added on the conjugate-containing pad would diffuse into the substrate-containing pad and induce the dissolution and back diffusion of the substrate to the working electrode side. However, the assay results were poorly reproducible when the glucose substrate-loaded pad was used. A similar sequence described in Section 3.2 was used for the dry strip NEEIA cell, but with greatly reduced sample volumes. Thirty microliters of standard MCLR was pipetted onto the pad containing GOx-MCLR conjugate and allowed to make a competitive reaction for 10 min. About 10 ␮L of 800 ␮g/mL glucose solution was then applied to the backside pad of the dry strip. The chronoamperometric signals were read 30 s after adding the glucose solution. The resulting signals yielded a calibration plot (Fig. 4b) which is almost identical to that obtained with the Utype diffusion cell (Fig. 4a). It shows that the NEEIA can be performed conveniently with dry strip-cell, providing quantitative analytical results in short time. To test the practical applicability of the dry strip NEEIA cell, it was used to determine MCLR in drinking water samples. Simulated samples were prepared by adding known amount of MCLR in tap water. The accuracy of the dry NEEIA strips was examined by estimating the recovery of the measurement with respect to the calibration obtained with the buffered samples in the range from 100 to 3160 pg/mL. As summarized in Table 1, the overall recovery was 93.5% with an average coefficient of variation (CV) of 6.6% (n = 4). The result indicates that the sample matrix effect on dry NEEIA strip was not so significant since the conjugate-containing pad is already loaded with buffer reagents. The dry strip NEEIA cells may be used conveniently as a quantitative on-spot rapid assay kit with no cumbersome sample pretreatment process. 4. Conclusion The nonseparation electrochemical enzyme immunoassay (NEEIA) method has been applied for the determination of microcystin-LR (MCLR), a family of hepatotoxic cyclic heptapeptides produced by freshwater cyanobacteria, using a newly

1425

developed double-sided microporous gold electrode in a new dry strip configuration. The dry strip-type NEEIA cell was constructed by stacking the substrate pad, MCLR antibodyimmobilized double-sided gold electrode, and conjugate pad loaded with GOx-MCLR between the assembly blocks (see Fig. 2). The amount of immobilized antibody, concentration of substrate and reaction times were optimized first with the U-type diffusion cell, and further with the dry strip-cell. Calibration plots obtained with the standard MCLR samples prepared in TBG (pH 7.4) were highly comparable to those obtained with the conventional ELISA. The practical utility of the dry strip-type NEEIA cell was also demonstrated by measuring the MCLR in tap water. Acknowledgments G.S. Cha gratefully acknowledges the grant from the Korea Health 21 R&D Project, Ministry of Health & Welfare (project no. 02-PJ3-PG6-EV05-0001). H. Nam also gratefully acknowledges the support from the grant no. (R01-2006-000-10240-0) from the Basic Research Program of the Korea Science & Engineering Foundation. References Duan, C., Meyerhoff, M.E., 1994. Anal. Chem. 66, 1369–1377. Duan, C., Meyerhoff, M.E., 1995. Mikrochim. Acta 117, 195–206. Ducey Jr., M.W., Smith, A.M., Guo, X., Meyerhoff, M.E., 1997. Anal. Chim. Acta 357, 5–12. Fawell, J.K., Hart, H.A., Parr, W., 1993. Water Supply 11, 109–121. Ho, W.S., Athey, D., McNeil, C.J., 1995. Biosens. Bioelectron. 10, 683–691. Ivnitski, D., Rishpon, J., 1996. Biosens. Bioelectron. 11, 409–417. Keay, R.W., McNeil, C.J., 1998. Biosens. Bioelectron. 13, 963–970. Killard, A.J., Smyth, M.R., Grennan, K., Micheli, L., Palleschi, G., 2000. Biochem. Soc. Trans. 28, 81–84. Kim, Y.M., Oh, S.W., Jeong, S.Y., Pyo, D.J., Choi, E.Y., 2003. Environ. Sci. Technol. 37, 1899–1904. Lei, L.-M., Wu, Y.-S., Gan, N.-Q., Song, L.-R., 2004. Clin. Chim. Acta 348, 77–180. Lin, J., Ju, H., 2005. Biosens. Bioelectron. 20, 1461–1470. McElhiney, J., Lawton, L.A., 2005. Toxicol. Appl. Pharmacol. 203, 219–230. Meulenberg, E.P., Mulder, W.H., Stokes, P.G., 1995. Environ. Sci. Technol. 29, 553–561. Meyerhoff, M.E., Duan, C., Meusel, M., 1995. Clin. Chem. 41, 1378–1384. Nagata, S., Tsutsumi, T., Hasegawa, A., Yoshida, F., Ueno, Y., Watanabe, M.F., 1997. J. AOAC Int. 80, 408–417. Pyo, D.J., Lee, J.A., Choi, E.Y., 2005. Microchem. J. 80, 165–169. Rishpon, J., Ivnitski, D., 1997. Biosens. Bioelectron. 12, 195–204. Yu, F.-Y., Liu, B.-H., Chou, H.-N., Chu, F.S., 2002. J. Agric. Food Chem. 50, 4176–4182. Zhang, F., Kim, J.S., Cui, G., Cha, G.S., Nam, H., 2005. Electroanalysis 17, 668–673. Zhang, F., Cho, S.S., Yang, S.H., Seo, S.S., Cha, G.S., Nam, H., 2006. Electroanalysis 18, 217–222.