- Email: [email protected]
Redox cycling-based immunoassay for detection of carcinogenic embryonic antigen
Analytica Chimica Acta 971 (2017) 33e39
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Redox cycling-based immunoassay for detection of carcinogenic embryonic antigen Ga-Yeon Lee a, Jun-Hee Park a, Young Wook Chang a, Sungbo Cho b, Min-Jung Kang c, Jae-Chul Pyun a, * a b c
Department of Materials Science and Engineering, Yonsei University, 134 Shinchon-dong, Seodaemun-gu, Seoul 120-749, South Korea Department of Biomedical Engineering, Gachon University, South Korea Korea Institute of Science and Technology (KIST), Seoul, South Korea
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Redox cycling was developed for a highly sensitive immunoassay. Interdigitated electrodes (IDEs) were used as the generator and the collector electrodes. Redox cycling was optimized by controlling potential and distance of IDEs. Redox cycling with IDEs was applied to a commercial immunoassay for CEA.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 November 2016 Received in revised form 1 April 2017 Accepted 4 April 2017 Available online 10 April 2017
Redox cycling based on an interdigitated electrode (IDE) was used as a highly sensitive immunoassay for carcinogenic embryonic antigen (CEA) through the quantification of 3,30 ,5,50 -tetramethylbenzidine (TMB). For the redox cycling process, one pair of interdigitated finger electrodes was used as the first working electrode (generator) for cyclic voltammetry of TMB, and another pair of interdigitated finger electrodes was used as the second working electrode (collector) for sequential application of potentials for reduction and oxidation of TMB. The reduction (and oxidation) products of TMB at the collector were supplied to the generator, and following sequential oxidization (and reduction) at the generator, again supplied to the collector. Such redox recycling processes between the generator and collector allowed signal amplification. In this work, the influences of the following factors on the redox cycling of TMB were analyzed: (1) the redox potential at the collector, (2) the gap between the interdigitated finger electrodes, and (3) the scan rate of the generator. The redox potential and electrode gap influences were simulated with COMSOL software and compared with empirical results. At the optimum redox potentials and electrode gap, redox cycling was estimated to be five-fold more sensitive for the quantification of TMB than conventional cyclic voltammetry using one pair of interdigitated finger electrodes as the working electrode. Finally, redox cycling was applied to a commercial immunoassay for CEA, and the sensitivity of redox cycling was three-fold higher than that of conventional cyclic voltammetry using a single set of interdigitated finger electrodes as the working electrode. © 2017 Elsevier B.V. All rights reserved.
Keywords: Redox cycling Interdigitated electrode Immunoassay 3,30 ,5,50 -tetramethylbenzidine Carcinogenic embryonic antigen
Abbreviations: HRP, horseradish peroxidase; TMB, 3,30 ,5,50 -tetramethylbenzidine; IDE, interdigitated electrode; WE2, collector; WE1, generator; CV, cyclic voltammogram; CEA, carcinogenic embryonic antigen; DA, reactant diffusion coefficient; DB, product diffusion coefficient. * Corresponding author. E-mail address: [email protected] (J.-C. Pyun). http://dx.doi.org/10.1016/j.aca.2017.04.010 0003-2670/© 2017 Elsevier B.V. All rights reserved.
34
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39
1. Introduction Immunoassays are widely used for the detection of target analytes based on the highly specific interactions of antibodies with a specific antigen (analyte). Usually, antibodies are immobilized on a solid support and the antigen in a sample binds specifically to the antibodies [1,2]. The amount of antigen remaining is detected using secondary antibodies with the enzyme horseradish peroxidase (HRP), which undergoes a chromogenic reaction with the substrate 3,30 ,5,50 -tetramethylbenzidine (TMB). During this reaction, TMB molecules are oxidized in two steps, as shown in Fig. 1(a): TMB (transparent) / ox1-TMB (blue, lmax ¼ 650 nm) / ox2-TMB (yellow, lmax ¼ 490 nm) [3e5]. As this chromogenic reaction involves the oxidation of TMB molecules, quantification of the antigen (analyte) is also possible by using amperometric reduction of oxidized TMB [6e9]. In this work, redox cycling performed with an interdigitated electrode (IDE) was used for highly sensitive detection of the chromogenic reaction of TMB. IDEs have been employed for redox cycling because redox species generated at a finger electrode (generator, WE1) can undergo the reverse reaction at the adjacent finger electrode (collector, WE2) if the potentials of the two electrodes are set at an appropriate level for the reverse reaction [10]. As shown in Fig. 1(b), the redox species undergo reaction at the collector (WE2) and then diffuse back to the generator (WE1), and this redox cycling process can amplify currents at both the generator and collector electrodes [11]. Thus, one molecule only reacts
many times if the potentials of both the interdigitated finger electrodes are sufficient for consecutive oxidation and reduction reactions [12]. Additionally, the shape of cyclic voltammogram (CV) was consistent with that of a microelectrode owing to enhanced diffusion during redox cycling [13]. The influence of electrode geometry, such as electrode area and the electrode gap between interdigitated band electrodes, has been well studied during redox cycling [14e18]. In this work, the factors that affect redox cycling of TMB were analyzed by controlling (1) the potential of the collector, (2) the electrode gap between the generator and the collector, and (3) the scanning rate at the generator. Further, the behaviors of the electrodes during redox cycling of TMB were simulated using the commercial simulation software COMSOL®. Finally, redox cycling of TMB was used along with a commercial ELISA assay kit for the detection of the carcinogenic embryonic antigen (CEA), and the sensitivity of the measurements was compared with that of conventional cyclic voltammetry using a single set of interdigitated finger electrodes as the working electrode.
2. Materials and methods 2.1. Materials TMB, HRP, and other analytical grade chemicals were purchased from Sigma-Aldrich Korea (Seoul, Korea). CEA ELISA kits were purchased from Perfumed Group, Inc. (San Francisco, CA, USA). The photoresist (AZ-GXR601) was purchased from Merck Co.
Fig. 1. Redox cycling for the quantification of oxidized-TMB. (a) Redox species of TMB. (b) Redox cycling of TMB with an IDE.
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39
(Darmstadt, Germany). Polystyrene microplates were purchased from SPL Co. (Seoul, Korea). Ag/AgCl reference electrodes were purchased from Warner Instruments LLC (Hamden, CT, USA). Pieces of Pt wire with a diameter of 2 mm were used as counter electrodes. 2.2. Fabrication of interdigitated electrodes The IDEs were fabricated on a glass substrate using an ultraviolet (UV) photolithographic process [19,20]. A 2 mm thick layer of a positive photoresist (AZ-GXR-601) was spin-coated on a 4 inch quartz wafer [21e23]. After soft-baking the photoresist layer at 100 C for 1 min, UV-photolithography was performed using a photomask of the IDEs. The IDEs were designed to have 25e100 pairs of finger electrodes, each with a width of 5e20 mm (de), with a 5e20 mm space between each (dg). The IDEs were then baked at 180 C for 1 min. After the IDE pattern was developed, a 10 nm thick Ti layer was deposited as an adhesive layer, and then a 100 nm thick Au layer was deposited using a thermal evaporator. The IDE pattern was obtained after a lift-off process. As listed in Table 1, three types of IDEs were fabricated, with each with the same area (1.5 mm2). 2.3. Instrumentation and signal measurement The amperometric measurements were carried out using a commercial potentiostat from IVIUM Technologies (Netherlands). Cyclic voltammetry was performed at potentials between 200 and þ 600 mV versus Ag/AgCl at a scan rate of 50 mV s1. Conventional cyclic voltammetry for the oxidation of TMB molecules was carried out and the CV shown in Fig. 1(b) was analyzed. The signal for the TMB solution (reduction current) was calculated as the difference in the reduction current for the TMB stock solution (optical density (OD) ¼ 0 without oxidized molecules) and that for the TMB sample with oxidized TMB molecules at a reduction potential of 200 mV versus Ag/AgCl [6e9]. To analyze TMB using redox cycling, two distinct potentials were sequentially applied to the collector (WE2): an oxidative potential (þ600 mV versus Ag/ AgCl) and a reductive potential (200 mV versus Ag/AgCl). For each TMB sample, two CVs were obtained at the generator electrode, one for the oxidative potential and one for the reductive potential of the collector (WE2), as shown in Fig. 1(b). The signal for the TMB solution during redox cycling was calculated as the difference between the redox cycling signal of the TMB stock solution (OD ¼ 0, without oxidized molecules) and that for the TMB sample at the same reduction potential (200 mV versus Ag/AgCl). That is, the signal of the TMB solution during redox cycling ¼ TMB0 current TMB current. 2.4. Immunoassay with redox cycling of TMB Commercial ELISA tests were carried out according to the manufacturer's instructions. The CEA ELISA test from Perfumed Group, Inc. (San Francisco, CA, USA) was performed using a 96-well microplate, which was coated with anti-CEA antibodies. The cutoff value was estimated using the positive and negative samples included in the commercial ELISA kit. The standard CEA samples were prepared by diluting the positive CEA sample to
35
1e120 ng mL1. Each CEA sample was incubated for 1 h at room temperature. After repeating the washing steps three times, HRPconjugated secondary antibodies were incubated for 1 h at room temperature. The amount of CEA bound to the 96 wells was quantified by reaction with TMB and a hydrogen peroxide solution for 20 min. The OD was measured at a wavelength of 650 nm using an ELISA reader (Vesamax) from Molecular Devices (Sunnyvale, CA, USA) [24e26]. Redox cycling was performed as described in section 2.3 by dipping an IDE into each well of the microplate before the sulfuric acid quenching process. 2.5. Simulation of redox cycling The properties of the electrodes were simulated using COMSOL Multiphysics® software [27e31]. For the simulations, the electrodes were replicated using a two-dimensional model placed in an electrolyte solution. In the model electrodes, two pairs of working electrodes, corresponding to the generator and the collector, were separated by an electrode gap of 100 nm (80 pairs), 500 nm (40 pairs), 1 mm (20 pairs), 5 mm (4 pairs), 10 mm (2 pairs), and 20 mm (1 pair). The potential of the reference electrode was set as þ200 mV, and the scan range for the generator electrode during cyclic voltammetry was 200 to þ600 mV versus Ag/AgCl. The potential of the collector was controlled to be in the range of 200 to þ600 mV versus Ag/AgCl. The amperometric reaction was simulated using the ButlereVolmer equation [32,33] and the steady-state current was measured [31,34,35] after 15 s. The concentration of the electrolyte solution was set as 1 mM. To model the redox behavior of TMB, oxidized and reduced TMB were set as two species. The reactant diffusion coefficient (DA) and product diffusion coefficient (DB) were both set as 109 m2 s1. Finally, the current signal was calculated from the current density by considering the electrode area to be 1.08 107 mm2. 3. Results and discussion 3.1. Factors affecting redox cycling of TMB When conventional cyclic voltammetry was performed using one pair of interdigitated finger electrodes to oxidize TMB molecules, two-step oxidation of TMB molecules resulted in two oxidation peaks, as shown in Fig. 1(b): TMB (transparent) / ox1TMB (blue, lmax ¼ 650 nm) / ox2-TMB (yellow, lmax ¼ 490 nm). In this work, redox cycling performed with an IDE because the redox species generated at the generator (WE1) could undergo the reverse reaction at the collector (WE2). When two different potentials (oxidative potential: þ600 mV versus Ag/AgCl and reductive potential: 200 mV versus Ag/AgCl) were sequentially applied to the collector (WE2), two different CVs were obtained for TMB at the generator, as shown in Fig. 1(b). The position of the CV at the generator was determined by the potential of the collector. The TMB stock solution (OD ¼ 0, without oxidized molecules) did not exhibit a zero reduction current at the reduction potential of TMB (see a, b, and c in Fig. 2) because this solution did not contain any oxidized TMB molecules. When cyclic voltammetry was performed at potentials of 200 to þ600 mV
Table 1 Specification of interdigitated electrodes (IDEs). Electrode Width (de, mm)
Electrode Gap (dg, mm)
Electrode Length (dl, mm)
No. of Electrode Pairs (EA)
Total Electrode Area (mm2)
5 10 20
5 10 20
1 1 1
100 50 25
1.5 1.5 1.5
36
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39
versus Ag/AgCl, zero reduction currents were observed for reduction potentials of 200 to þ200 mV versus Ag/AgCl. Further, an oxidative current was observed at oxidative potentials of þ200 to þ600 mV versus Ag/AgCl, owing to the oxidation of TMB molecules at the generator. The oxidation products of TMB (ox-TMB) at the generator were forcibly reduced to produce the reductive wave (see a, b, and c in Fig. 2), and then supplied to the diffusion layer of the neighboring collector. When the oxidative current for TMB was applied to the collector (see d, e, f, and g in Fig. 2), the CVs shifted to the reductive-current region (negative region of the y-axis). First, TMB molecules were oxidized at the collector according to the Nernst equation E E0 ¼ RT/nF$log ([TMBox]/[TMBred]), where E is the potential at the collector electrode versus Ag/AgCl, E0 is the standard potential, R is the gas constant, F is the Faraday constant, T is the temperature, and n is the number of redox electrons. Hence, at the oxidation potential, TMB molecules were oxidized at the collector in proportion with ([TMBox]/[TMBred]) term of the Nernst equation, and the oxidized molecules were supplied to the generator electrode. When cyclic voltammetry was performed at potentials of 200 to þ600 mV versus Ag/AgCl, a reduction current was observed at reduction potentials of 200 to þ200 mV versus Ag/AgCl. This reduction current resulted in a shift in the positions of the CVs of the TMB solution. Further, an oxidative current was observed at oxidative potentials of þ200 to þ600 mV versus Ag/AgCl owing to the oxidation of TMB molecules on the generator electrode. Therefore, the maximum redox cycling signal for TMB was obtained when the oxidative potential (þ600 mV versus Ag/AgCl, f and g in Fig. 2) and reductive potential (200 mV versus Ag/AgCl, a, b, and c in Fig. 2) of TMB were applied to the collector electrode. These results confirmed that the maximum redox cycling signal for TMB could be obtained when the oxidative and the reductive potentials of TMB were applied to the collector. Redox cycling of TMB was also affected by the electrode gap between two interdigitated finger electrodes, which should be small enough to allow product movement between the generator and the collector. Therefore, the distance between the electrodes affected the transfer of oxidized TMB to the diffusion layer between the electrodes. In this work, the IDEs were fabricated using a labbased UV-lithography process, and the minimum distance between the finger electrodes was on the order of several micrometers. As shown in Fig. 3(a), the redox cycling signal for the TMB stock solution (OD ¼ 0, without oxidized molecules) decreased as the electrode gap between the finger electrode was increased from 5 to 10 mm and then to 20 mm. As previously mentioned, when the same oxidative potential was applied to the collector, TMB samples with different OD values resulted in different amounts of oxidized TMB molecules, with the amounts being proportional to the
[TMBox]/[TMBred] term in the Nernst equation. The sensitivity of the measurements was determined to be 861, 392, and 70 nA OD1 for electrode gaps of 5, 10, and 20 mm, respectively. From these results, the electrode gap of 5 mm was selected as optimal for redox cycling of TMB by considering the ease of fabrication and the sensitivity. In the case of conventional cyclic voltammetry using a single set of finger electrodes as the working electrode, the sensitivity was estimated to be 178 nA OD1. These results showed that the TMB measurements performed using redox cycling were more sensitive when using an IDE with a smaller electrode gap. Simulations were performed using COMSOL® to estimate the effect of electrode gaps of less than 5 mm on the redox cycling signal, as shown in Fig. 3(b). When the electrode gap was decreased to 1 mm, 500 nm, and 100 nm, the current was observed to increase exponentially to 9.5, 19.5, and 103.7 mA, respectively. Such results were considered to result from enhanced collector efficiency and confirm that the sensitivity of redox cycling could be increased by further decreasing the electrode gap between finger electrodes. Redox cycling of TMB was also affected by the scan rate of the generator. For effective redox cycling, oxidized (or reduced) TMB produced on the collector electrode should be transferred to the diffusion layer of the generator electrode. If the scan rate of the generator is much higher than the mass-transport rate of the oxidative products at the collector, the redox cycling efficiency decreases and the CV shape will change from that of a microelectrode to that of a conventional electrode. In this work, scan rates of 25, 50, and 100 mV s1 were used for an electrode gap of 5 mm. As shown in Fig. 4, for the same TMB solution (OD ¼ 0), the redox cycling signal decreased with increasing electrode gap. When TMB samples with same OD values were analyzed using redox cycling, the signal currents were calculated at scan rates of 25, 50, 75, and 100 mV s1 for an electrode gap of 5, 10, 20 mm. As shown in Fig. 4, the redox cycling signals were similar for IDEs with the same electrode gap. Thus, the scan rate did not have a significant effect on the redox cycling signals. These results showed that mass transport phenomena influenced the redox cycling signals, but mass transport from the collector to the generator was fast enough to maintain the concentration of oxidized (or reduced) TMB. Simulations were also performed using COMSOL® to estimate the effect of the scan rate for an IDE with an electrode gap of 5 mm. As shown in Fig. 4, the current signal was maintained at a similar level, irrespective of the scan rate. These results indicated that the scan rate does not have a significant influence on the current signal during redox cycling. Thus, the optimal conditions for performing redox cycling to analyze TMB with the highest sensitivity are (1) an IDE with an electrode gap of 5 mm, (2) a reduction potential of 200 mV versus Ag/AgCl and an oxidation potential of þ600 mV versus Ag/ AgCl, and (3) a scan rate in the range of 25e100 mV s1.
Fig. 2. Influence of applied potential at the collector on redox cycling of TMB. CVs were obtained at the generator at different potentials at the collector.
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39
37
Fig. 3. Influencing of the electrode gap between two interdigitated band electrodes on redox cycling of TMB. (a) CVs were obtained using IDEs with different electrode gaps (dg) between two band electrodes, and (b) COMSOL® simulation with different electrode gaps (dg).
Fig. 4. Influencing of scan rate at the generator on redox cycling of TMB. (a) CVs were obtained at different scan rates of the generator. (b) COMSOL® simulation for different scan rates.
3.2. Immunoassay for CEA based on redox cycling The current signals for redox cycling of TMB samples with different ODs were measured using an IDE with an electrode gap of 5 mm. As mentioned in section 3.1, the amounts of oxidized TMB molecules produced by samples with different OD values are proportional to the [TMBox]/[TMBred] term in the Nernst equation, when the same oxidative potential is applied to the collector electrode. As shown in Fig. 5(a), the signal currents for redox cycling of oxidized TMB samples were smaller than that for the TMB stock solution (OD ¼ 0, without oxidized molecules), and the signal current decreased as the OD value increased. A standard curve for quantification of TMB was obtained by performing redox cycling under the optimal conditions, that is, by using collector potentials of 200 and þ 600 mV versus Ag/AgCl as the reduction and oxidation potentials, respectively. The potential range of the generator was set as 200 to þ600 mV versus Ag/AgCl at a scan rate of 50 mV s1 using an IDE with an electrode gap of 5 mm. For comparison, conventional cyclic voltammetry using one set of interdigitated finger electrodes as the working electrode was carried out using a potential range of 200 to þ600 mV versus Ag/AgCl
at a scan rate of 50 mV s1. As shown in Fig. 5(a), standard curves were obtained using standard samples with different OD values. The sensitivities of the redox cycling and conventional cyclic voltammetry measurements were calculated as 861 and 178 nA OD1, respectively. These results showed that the sensitivity of redox cycling is five-fold higher than that of conventional cyclic voltammetry using a single interdigitated band electrode without the collector. Quantification of TMB by redox cycling was applied to a commercial immunoassay kit for the detection of CEA. The immunoassay was performed using a 96-well plate, which was coated with anti-CEA antibodies, as shown in Fig. 5(c). The standard curve for CEA was calibrated using standard CEA samples of known concentrations. The amount of bound CEA was quantified by treating the secondary antibodies with HRP and then with TMB. As shown in Fig. 5(d), the limit of detection for CEA detection was determined to be 1.08 ng mL1 for redox cycling and conventional cyclic voltammetry, respectively (n ¼ 3). The cutoff value for positive determination of CEA was established to be an OD value of 0.1226 at a wavelength of 650 nm, which corresponds to a CEA concentration of 5 ng mL1 according to the manufacturer's instructions. As shown in Fig. 5(d), the current signal at the cut-off concentration of CEA (5 ng mL1) was determined to be 0.7 mA and 0.2 mA for redox cycling and conventional cyclic voltammetry, respectively (n ¼ 3). These results showed that the sensitivity of redox cycling is more than three-fold higher than that of single-mode amperometry using a single interdigitated band electrode without the collector. Usually, test results within the standard deviation at the cutoff value are regarded to be in the grey zone for the medical diagnosis and require retesting. In this work, the standard deviations at the cut-off concentration of CEA (5 ng mL1) were determined to be 51 mA (71.3% of mean value) and 58 mA (8.9% of mean value) for redox cycling and conventional cyclic voltammetry, respectively (n ¼ 3). These results showed that the grey zone for positive determination of CEA could be improved considerably by using redox cycling.
4. Conclusions Redox cycling based on IDEs was applied to the quantification of TMB during immunoassays. During the redox cycling process, the
38
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39
Fig. 5. Application of redox cycling to a commercial immunoassay kit for the detection of CEA. (a) CVs obtained for redox cycling of TMB samples with different OD values. (b) Standard curves for quantification of oxidized TMB using redox cycling and a single interdigitated band electrode without the collector. (c) Schematic view of ELISA test for the detection of CEA. (d) Standard curves for detection of CEA using redox cycling and a single interdigitated band electrode without the collector.
first working electrode (generator) was used for cyclic voltammetry of TMB by changing the potential between 200 and þ 600 mV versus Ag/AgCl, while two fixed potentials were sequentially applied at the second working electrode (collector) for reduction (and oxidation) of oxidized (and reduced) TMB. The optimum conditions for maximizing sensitivity during redox cycling of TMB were determined to be (1) an interdigitated finger electrode gap of 5 mm, (2) reduction and oxidation potentials of 200 and þ 600 mV versus Ag/AgCl, respectively, and (3) a scan rate in the range of 25e100 mV s1. The sensitivity of TMB measurements performed using redox cycling was five-fold higher than that of single-mode amperometry using a single interdigitated band electrode without the collector. Finally, the redox cycling of TMB was applied to a commercial immunoassay kit for detecting CEA. Based on the assay results, the sensitivity of redox cycling was found to be threefold higher than that of conventional cyclic voltammetry using one set of interdigitated finger electrodes as a working electrode.
Acknowledgements This work was supported by Nano-Convergence Foundation (R201602210) funded by the Ministry of Science, ICT and Future Planning (MSIP, Korea) and the Ministry of Trade, Industry and Energy (MOTIE, Korea), and by the Industry Technology Development Program (10063335) funded by the Ministry of Trade, Industry and Energy (MOTIE, Korea).
References [1] E. Marulli, A. Aloisi, P.D. Giuseppe, R. Rinaldi, Micro and nanotechnology for early diagnosis and detection of rheumatic diseases-molecular markers, Biochip J. 10 (2016) 189e197. [2] J.H. Choi, J.W. Lee, B.K. Oh, Nanomaterial-based in vitro analytical system for diagnosis and therapy in microfluidic device, Biochip J. 10 (2016) 331e345. [3] S.G. Rhee, H.A. Woo, I.S. Kil, S.H. Bae, Peroxiredoxin functions as a peroxidase and a regulator and sensor of local peroxides, J. Biol. Chem. 287 (2012) 4403e4410. [4] P.D. Josephy, T. Eling, R.P.J. Mason, The horseradish peroxidase-catalyzed oxidation of 3,5,3',5'-tetramethylbenzidine. Free radical and charge-transfer complex intermediates, Biol. Chem. 257 (1982) 3669e3675. [5] C.A. Groom, J.H.T. Luong, R. Thatipalmala, Dual functionalities of 4aminodiphenlyamine in enzymatic assay and mediated biosensor construction, Anal. Biochem. 231 (1995) 393e399. [6] J.C. Pyun, H.H. Lee, C.S. Lee, Application of an amperometric detector to the conventional enzyme immunoassay, Sensors Actuator B 78 (2001) 232e236. [7] J.C. Pyun, S.D. Kim, J.W. Chung, Gold film array electrode for an ELISA analyser based on an amperometric detector, Sensors Actuators B 111e112 (2005) 416e422. [8] J.I. Kim, A. Bordeanu, J.C. Pyun, Diamond-like carbon (DLC) microelectrode for electrochemical ELISA, Biosens. Bioelectron. 24 (2009) 1394e1398. [9] J.K. Lee, G. Yoo, M. Park, J. Jose, M.J. Kang, J.C. Pyun, Electrochemical ELISA based on E. coli with autodisplayed Z-domains, Sensors Actuators B 175 (2012) 46e52. [10] O. Niwa, M. Morita, H. Tabei, Electrochemical behavior of reversible redox species at interdigitated array electrodes with different geometries: consideration of redox cycling and collection efficiency, Anal. Chem. 62 (1990) 447e452. [11] K. Ueno, M. Hayashida, J.Y. Ye, H. Misawa, Fabrication and electrochemical characterization of interdigitated nanoelectrode arrays, Electrochem. Comm. 7 (2005) 161e165. [12] M. Paeschke, U. Wollenberger, T. Lisec, U. Schnakenberg, R. Hintsche, Highly sensitive electrochemical microsensors using submicrometer electrode arrays,
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39 Sensors Actuators B 27 (1995) 394e397. [13] A.J. Bard, J.A. Crayston, G.P. Kittlesen, T. Varco Shea, M.S. Wrighton, Digital simulation of the measured electrochemical response of reversible redox couples at microelectrode arrays: consequences arising from closely spaced ultramicroelectrodes, Anal. Chem. 58 (1986) 2321e2331. [14] V.A.T. Dam, W. Olthuis, Redox cycling with facing interdigitated array electrodes as a method for selective detection of redox species, van den Berg A. Anal. 132 (2007) 365e370. [15] D.G. Sanderson, L.B. Anderson, Filar electrodes: steady-state currents and spectroelectrochemistry at twin interdigitated electrodes, Anal. Chem. 57 (1985) 2388e2393. [16] T.B. Tran, S.J. Son, J.H. Min, Nanomaterials in label-free impedimetric biosensor: current process and future perspectives, Biochip J. 10 (2016) 318e330. [17] T. Lee, J.W. Choi, Fabrication of fusion protein-based heterolayers composed of redox protein/myoglobin for bioelectronic device, Biochip J. 10 (2016) 103e110. [18] K.S. Lee, S.Y. Shin, Y.S. Yoon, Fe3O4 nanoparticles on MWCNTs backbone for lithium ion batteries, J. Korean Ceram. Soc. 53 (2016) 376e380. [19] W. Smirnov, N. Yang, R. Hoffmann, J. Hees, H. Obloh, W. MVuller-Sebert, C.E. Nebel, Integrated all-diamond ultramicroelectrode arrays: optimization of faradaic and capacitive currents, Anal. Chem. 83 (2011) 7438e7443. [20] J. Duay, J.M. Goran, K.J. Stevenson, Facile fabrication of carbon ultramicro- to nanoelectrode arrays with tunable voltammetric response, Anal. Chem. 86 (2014) 11528e11532. [21] H.W. Jung, Y.W. Chang, G.Y. Lee, S.B. Cho, M.J. Kang, J.C. Pyun, Capacitive biosensor based on an interdigitated electrode with nanoislands, Anal. Chim. Acta 844 (2014) 27e34. [22] B.G. An, Y.W. Chang, H.R. Kim, M.J. Kang, J.K. Park, J.C. Pyun, In-situ synthesized CdS nanowires for the application to photosensor, Sensors Actuators B 221 (2015) 884e890. [23] Y.Y. Noh, K.B. Kim, M.K. Choi, O.S. song properties of blocking layer with Ag nano powder in a dye sensitized solar cell, J. Korean Ceram. Soc. 53 (2016) 105e109. [24] J.C. Pyun, M.Y. Cheong, S.H. Park, H.Y. Kim, J.S. Park, Modification of short peptides using 3-aminocaproic acid for improved coating efficiency in indirect enzyme-linked immunosorbent assays (ELISA), J. Immunol. Methods 208
39
(1997) 141e149. [25] J.W. Chung, R. Bernhardt, J.C. Pyun, Sequential analysis of multiple analytes using a surface plasmon resonance (SPR) biosensor, J. Immunol. Methods 311 (2006) 178e188. [26] K. Petermann, S. Vordenb€ aumen, J.C. Pyun, A. Braukmann, E. Bleck, M. Schneider, J. Jose, Autodisplay of 60-kDa Ro/SS-A antigen and development of a surface display enzyme-linked immunosorbent assay for systemic lupus erythematosus patient sera screening, Anal. Biochem. 407 (2010) 72e78. [27] J. Guo, E. Lindner, Cyclic voltammetry at shallow recessed microdisc electrode:Theoretical and experimental study, J. Electroanal. Chem. 629 (2009) 180e184. [28] A. Lavacchi, U. Bardi, C. Borri, S. Caporali, A. Fossati, I. Perissi, Cyclic voltammetry simulation at microelectrode arrays with COMSOL multiphysics, J. Appl. Electrochem 39 (2009) 2159e2163. [29] C. Amatore, C. Pebay, L. Thouin, A. Wang, Cyclic voltammetry at microelectrodes. Influence of natural convection on diffusion layers as characterized by in situ mapping of concentration profiles, Electrochem. Comm. 11 (2009) 1269e1272. [30] M. Odijk, W. Olthuis, V.A.T. Dam, A. Berg, Simulation of redox-cycling phenomena at interdigitated array (IDA) electrodes:amplification and selectivity, Electroanalysis 20 (5) (2008) 463e468. [31] E. Goluch, B. Wolfrum, P.S. Singh, M. Zevenbergen, S.G. Lemay, Redox cycling in nanofluidic channels using interdigitated electrodes, Anal. Bioanal. Chem. 394 (2) (2009) 447e456. [32] S.J. Hood, D.K. Kampouris, R.O. Kadara, N. Jenkinson, F.J. Campo, F.X. Munoz, C.E. Banks, Why ‘the bigger the better’ is not always the case when utilizing microelectrode arrays: high density vs. low density arrays for the electroanalytical sensing of chromium(VI), Analyst 134 (2009) 2301e2305. [33] C.H. Hamann, A. Hamnett, W. Vielstich, Electrochemistry, second ed., WileyVCH Verlag GmbH, Weinheim, 1998, pp. 172e179. [34] F. Liu, G. Kolesov, B.A. Parkinson, Time of flight electrochemistry: diffusion coefficient measurements using interdigitated array (IDA) electrodes, J. Electrochem. Soc. 161 (13) (2014) H3015eH3019. [35] G. Lewis, S. Dale, B. Kasprzyk-Hordern, A. Lubben, E. Barnes, R. Compton, F. Marken, Cavity transport effects in generatorecollector electrochemical analysis of nitrobenzene, Phys. Chem. Chem. Phys. 16 (2014) 18966e18973.
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Redox cycling-based immunoassay for detection of carcinogenic embryonic antigen Ga-Yeon Lee a, Jun-Hee Park a, Young Wook Chang a, Sungbo Cho b, Min-Jung Kang c, Jae-Chul Pyun a, * a b c
Department of Materials Science and Engineering, Yonsei University, 134 Shinchon-dong, Seodaemun-gu, Seoul 120-749, South Korea Department of Biomedical Engineering, Gachon University, South Korea Korea Institute of Science and Technology (KIST), Seoul, South Korea
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Redox cycling was developed for a highly sensitive immunoassay. Interdigitated electrodes (IDEs) were used as the generator and the collector electrodes. Redox cycling was optimized by controlling potential and distance of IDEs. Redox cycling with IDEs was applied to a commercial immunoassay for CEA.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 November 2016 Received in revised form 1 April 2017 Accepted 4 April 2017 Available online 10 April 2017
Redox cycling based on an interdigitated electrode (IDE) was used as a highly sensitive immunoassay for carcinogenic embryonic antigen (CEA) through the quantification of 3,30 ,5,50 -tetramethylbenzidine (TMB). For the redox cycling process, one pair of interdigitated finger electrodes was used as the first working electrode (generator) for cyclic voltammetry of TMB, and another pair of interdigitated finger electrodes was used as the second working electrode (collector) for sequential application of potentials for reduction and oxidation of TMB. The reduction (and oxidation) products of TMB at the collector were supplied to the generator, and following sequential oxidization (and reduction) at the generator, again supplied to the collector. Such redox recycling processes between the generator and collector allowed signal amplification. In this work, the influences of the following factors on the redox cycling of TMB were analyzed: (1) the redox potential at the collector, (2) the gap between the interdigitated finger electrodes, and (3) the scan rate of the generator. The redox potential and electrode gap influences were simulated with COMSOL software and compared with empirical results. At the optimum redox potentials and electrode gap, redox cycling was estimated to be five-fold more sensitive for the quantification of TMB than conventional cyclic voltammetry using one pair of interdigitated finger electrodes as the working electrode. Finally, redox cycling was applied to a commercial immunoassay for CEA, and the sensitivity of redox cycling was three-fold higher than that of conventional cyclic voltammetry using a single set of interdigitated finger electrodes as the working electrode. © 2017 Elsevier B.V. All rights reserved.
Keywords: Redox cycling Interdigitated electrode Immunoassay 3,30 ,5,50 -tetramethylbenzidine Carcinogenic embryonic antigen
Abbreviations: HRP, horseradish peroxidase; TMB, 3,30 ,5,50 -tetramethylbenzidine; IDE, interdigitated electrode; WE2, collector; WE1, generator; CV, cyclic voltammogram; CEA, carcinogenic embryonic antigen; DA, reactant diffusion coefficient; DB, product diffusion coefficient. * Corresponding author. E-mail address: [email protected] (J.-C. Pyun). http://dx.doi.org/10.1016/j.aca.2017.04.010 0003-2670/© 2017 Elsevier B.V. All rights reserved.
34
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39
1. Introduction Immunoassays are widely used for the detection of target analytes based on the highly specific interactions of antibodies with a specific antigen (analyte). Usually, antibodies are immobilized on a solid support and the antigen in a sample binds specifically to the antibodies [1,2]. The amount of antigen remaining is detected using secondary antibodies with the enzyme horseradish peroxidase (HRP), which undergoes a chromogenic reaction with the substrate 3,30 ,5,50 -tetramethylbenzidine (TMB). During this reaction, TMB molecules are oxidized in two steps, as shown in Fig. 1(a): TMB (transparent) / ox1-TMB (blue, lmax ¼ 650 nm) / ox2-TMB (yellow, lmax ¼ 490 nm) [3e5]. As this chromogenic reaction involves the oxidation of TMB molecules, quantification of the antigen (analyte) is also possible by using amperometric reduction of oxidized TMB [6e9]. In this work, redox cycling performed with an interdigitated electrode (IDE) was used for highly sensitive detection of the chromogenic reaction of TMB. IDEs have been employed for redox cycling because redox species generated at a finger electrode (generator, WE1) can undergo the reverse reaction at the adjacent finger electrode (collector, WE2) if the potentials of the two electrodes are set at an appropriate level for the reverse reaction [10]. As shown in Fig. 1(b), the redox species undergo reaction at the collector (WE2) and then diffuse back to the generator (WE1), and this redox cycling process can amplify currents at both the generator and collector electrodes [11]. Thus, one molecule only reacts
many times if the potentials of both the interdigitated finger electrodes are sufficient for consecutive oxidation and reduction reactions [12]. Additionally, the shape of cyclic voltammogram (CV) was consistent with that of a microelectrode owing to enhanced diffusion during redox cycling [13]. The influence of electrode geometry, such as electrode area and the electrode gap between interdigitated band electrodes, has been well studied during redox cycling [14e18]. In this work, the factors that affect redox cycling of TMB were analyzed by controlling (1) the potential of the collector, (2) the electrode gap between the generator and the collector, and (3) the scanning rate at the generator. Further, the behaviors of the electrodes during redox cycling of TMB were simulated using the commercial simulation software COMSOL®. Finally, redox cycling of TMB was used along with a commercial ELISA assay kit for the detection of the carcinogenic embryonic antigen (CEA), and the sensitivity of the measurements was compared with that of conventional cyclic voltammetry using a single set of interdigitated finger electrodes as the working electrode.
2. Materials and methods 2.1. Materials TMB, HRP, and other analytical grade chemicals were purchased from Sigma-Aldrich Korea (Seoul, Korea). CEA ELISA kits were purchased from Perfumed Group, Inc. (San Francisco, CA, USA). The photoresist (AZ-GXR601) was purchased from Merck Co.
Fig. 1. Redox cycling for the quantification of oxidized-TMB. (a) Redox species of TMB. (b) Redox cycling of TMB with an IDE.
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39
(Darmstadt, Germany). Polystyrene microplates were purchased from SPL Co. (Seoul, Korea). Ag/AgCl reference electrodes were purchased from Warner Instruments LLC (Hamden, CT, USA). Pieces of Pt wire with a diameter of 2 mm were used as counter electrodes. 2.2. Fabrication of interdigitated electrodes The IDEs were fabricated on a glass substrate using an ultraviolet (UV) photolithographic process [19,20]. A 2 mm thick layer of a positive photoresist (AZ-GXR-601) was spin-coated on a 4 inch quartz wafer [21e23]. After soft-baking the photoresist layer at 100 C for 1 min, UV-photolithography was performed using a photomask of the IDEs. The IDEs were designed to have 25e100 pairs of finger electrodes, each with a width of 5e20 mm (de), with a 5e20 mm space between each (dg). The IDEs were then baked at 180 C for 1 min. After the IDE pattern was developed, a 10 nm thick Ti layer was deposited as an adhesive layer, and then a 100 nm thick Au layer was deposited using a thermal evaporator. The IDE pattern was obtained after a lift-off process. As listed in Table 1, three types of IDEs were fabricated, with each with the same area (1.5 mm2). 2.3. Instrumentation and signal measurement The amperometric measurements were carried out using a commercial potentiostat from IVIUM Technologies (Netherlands). Cyclic voltammetry was performed at potentials between 200 and þ 600 mV versus Ag/AgCl at a scan rate of 50 mV s1. Conventional cyclic voltammetry for the oxidation of TMB molecules was carried out and the CV shown in Fig. 1(b) was analyzed. The signal for the TMB solution (reduction current) was calculated as the difference in the reduction current for the TMB stock solution (optical density (OD) ¼ 0 without oxidized molecules) and that for the TMB sample with oxidized TMB molecules at a reduction potential of 200 mV versus Ag/AgCl [6e9]. To analyze TMB using redox cycling, two distinct potentials were sequentially applied to the collector (WE2): an oxidative potential (þ600 mV versus Ag/ AgCl) and a reductive potential (200 mV versus Ag/AgCl). For each TMB sample, two CVs were obtained at the generator electrode, one for the oxidative potential and one for the reductive potential of the collector (WE2), as shown in Fig. 1(b). The signal for the TMB solution during redox cycling was calculated as the difference between the redox cycling signal of the TMB stock solution (OD ¼ 0, without oxidized molecules) and that for the TMB sample at the same reduction potential (200 mV versus Ag/AgCl). That is, the signal of the TMB solution during redox cycling ¼ TMB0 current TMB current. 2.4. Immunoassay with redox cycling of TMB Commercial ELISA tests were carried out according to the manufacturer's instructions. The CEA ELISA test from Perfumed Group, Inc. (San Francisco, CA, USA) was performed using a 96-well microplate, which was coated with anti-CEA antibodies. The cutoff value was estimated using the positive and negative samples included in the commercial ELISA kit. The standard CEA samples were prepared by diluting the positive CEA sample to
35
1e120 ng mL1. Each CEA sample was incubated for 1 h at room temperature. After repeating the washing steps three times, HRPconjugated secondary antibodies were incubated for 1 h at room temperature. The amount of CEA bound to the 96 wells was quantified by reaction with TMB and a hydrogen peroxide solution for 20 min. The OD was measured at a wavelength of 650 nm using an ELISA reader (Vesamax) from Molecular Devices (Sunnyvale, CA, USA) [24e26]. Redox cycling was performed as described in section 2.3 by dipping an IDE into each well of the microplate before the sulfuric acid quenching process. 2.5. Simulation of redox cycling The properties of the electrodes were simulated using COMSOL Multiphysics® software [27e31]. For the simulations, the electrodes were replicated using a two-dimensional model placed in an electrolyte solution. In the model electrodes, two pairs of working electrodes, corresponding to the generator and the collector, were separated by an electrode gap of 100 nm (80 pairs), 500 nm (40 pairs), 1 mm (20 pairs), 5 mm (4 pairs), 10 mm (2 pairs), and 20 mm (1 pair). The potential of the reference electrode was set as þ200 mV, and the scan range for the generator electrode during cyclic voltammetry was 200 to þ600 mV versus Ag/AgCl. The potential of the collector was controlled to be in the range of 200 to þ600 mV versus Ag/AgCl. The amperometric reaction was simulated using the ButlereVolmer equation [32,33] and the steady-state current was measured [31,34,35] after 15 s. The concentration of the electrolyte solution was set as 1 mM. To model the redox behavior of TMB, oxidized and reduced TMB were set as two species. The reactant diffusion coefficient (DA) and product diffusion coefficient (DB) were both set as 109 m2 s1. Finally, the current signal was calculated from the current density by considering the electrode area to be 1.08 107 mm2. 3. Results and discussion 3.1. Factors affecting redox cycling of TMB When conventional cyclic voltammetry was performed using one pair of interdigitated finger electrodes to oxidize TMB molecules, two-step oxidation of TMB molecules resulted in two oxidation peaks, as shown in Fig. 1(b): TMB (transparent) / ox1TMB (blue, lmax ¼ 650 nm) / ox2-TMB (yellow, lmax ¼ 490 nm). In this work, redox cycling performed with an IDE because the redox species generated at the generator (WE1) could undergo the reverse reaction at the collector (WE2). When two different potentials (oxidative potential: þ600 mV versus Ag/AgCl and reductive potential: 200 mV versus Ag/AgCl) were sequentially applied to the collector (WE2), two different CVs were obtained for TMB at the generator, as shown in Fig. 1(b). The position of the CV at the generator was determined by the potential of the collector. The TMB stock solution (OD ¼ 0, without oxidized molecules) did not exhibit a zero reduction current at the reduction potential of TMB (see a, b, and c in Fig. 2) because this solution did not contain any oxidized TMB molecules. When cyclic voltammetry was performed at potentials of 200 to þ600 mV
Table 1 Specification of interdigitated electrodes (IDEs). Electrode Width (de, mm)
Electrode Gap (dg, mm)
Electrode Length (dl, mm)
No. of Electrode Pairs (EA)
Total Electrode Area (mm2)
5 10 20
5 10 20
1 1 1
100 50 25
1.5 1.5 1.5
36
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39
versus Ag/AgCl, zero reduction currents were observed for reduction potentials of 200 to þ200 mV versus Ag/AgCl. Further, an oxidative current was observed at oxidative potentials of þ200 to þ600 mV versus Ag/AgCl, owing to the oxidation of TMB molecules at the generator. The oxidation products of TMB (ox-TMB) at the generator were forcibly reduced to produce the reductive wave (see a, b, and c in Fig. 2), and then supplied to the diffusion layer of the neighboring collector. When the oxidative current for TMB was applied to the collector (see d, e, f, and g in Fig. 2), the CVs shifted to the reductive-current region (negative region of the y-axis). First, TMB molecules were oxidized at the collector according to the Nernst equation E E0 ¼ RT/nF$log ([TMBox]/[TMBred]), where E is the potential at the collector electrode versus Ag/AgCl, E0 is the standard potential, R is the gas constant, F is the Faraday constant, T is the temperature, and n is the number of redox electrons. Hence, at the oxidation potential, TMB molecules were oxidized at the collector in proportion with ([TMBox]/[TMBred]) term of the Nernst equation, and the oxidized molecules were supplied to the generator electrode. When cyclic voltammetry was performed at potentials of 200 to þ600 mV versus Ag/AgCl, a reduction current was observed at reduction potentials of 200 to þ200 mV versus Ag/AgCl. This reduction current resulted in a shift in the positions of the CVs of the TMB solution. Further, an oxidative current was observed at oxidative potentials of þ200 to þ600 mV versus Ag/AgCl owing to the oxidation of TMB molecules on the generator electrode. Therefore, the maximum redox cycling signal for TMB was obtained when the oxidative potential (þ600 mV versus Ag/AgCl, f and g in Fig. 2) and reductive potential (200 mV versus Ag/AgCl, a, b, and c in Fig. 2) of TMB were applied to the collector electrode. These results confirmed that the maximum redox cycling signal for TMB could be obtained when the oxidative and the reductive potentials of TMB were applied to the collector. Redox cycling of TMB was also affected by the electrode gap between two interdigitated finger electrodes, which should be small enough to allow product movement between the generator and the collector. Therefore, the distance between the electrodes affected the transfer of oxidized TMB to the diffusion layer between the electrodes. In this work, the IDEs were fabricated using a labbased UV-lithography process, and the minimum distance between the finger electrodes was on the order of several micrometers. As shown in Fig. 3(a), the redox cycling signal for the TMB stock solution (OD ¼ 0, without oxidized molecules) decreased as the electrode gap between the finger electrode was increased from 5 to 10 mm and then to 20 mm. As previously mentioned, when the same oxidative potential was applied to the collector, TMB samples with different OD values resulted in different amounts of oxidized TMB molecules, with the amounts being proportional to the
[TMBox]/[TMBred] term in the Nernst equation. The sensitivity of the measurements was determined to be 861, 392, and 70 nA OD1 for electrode gaps of 5, 10, and 20 mm, respectively. From these results, the electrode gap of 5 mm was selected as optimal for redox cycling of TMB by considering the ease of fabrication and the sensitivity. In the case of conventional cyclic voltammetry using a single set of finger electrodes as the working electrode, the sensitivity was estimated to be 178 nA OD1. These results showed that the TMB measurements performed using redox cycling were more sensitive when using an IDE with a smaller electrode gap. Simulations were performed using COMSOL® to estimate the effect of electrode gaps of less than 5 mm on the redox cycling signal, as shown in Fig. 3(b). When the electrode gap was decreased to 1 mm, 500 nm, and 100 nm, the current was observed to increase exponentially to 9.5, 19.5, and 103.7 mA, respectively. Such results were considered to result from enhanced collector efficiency and confirm that the sensitivity of redox cycling could be increased by further decreasing the electrode gap between finger electrodes. Redox cycling of TMB was also affected by the scan rate of the generator. For effective redox cycling, oxidized (or reduced) TMB produced on the collector electrode should be transferred to the diffusion layer of the generator electrode. If the scan rate of the generator is much higher than the mass-transport rate of the oxidative products at the collector, the redox cycling efficiency decreases and the CV shape will change from that of a microelectrode to that of a conventional electrode. In this work, scan rates of 25, 50, and 100 mV s1 were used for an electrode gap of 5 mm. As shown in Fig. 4, for the same TMB solution (OD ¼ 0), the redox cycling signal decreased with increasing electrode gap. When TMB samples with same OD values were analyzed using redox cycling, the signal currents were calculated at scan rates of 25, 50, 75, and 100 mV s1 for an electrode gap of 5, 10, 20 mm. As shown in Fig. 4, the redox cycling signals were similar for IDEs with the same electrode gap. Thus, the scan rate did not have a significant effect on the redox cycling signals. These results showed that mass transport phenomena influenced the redox cycling signals, but mass transport from the collector to the generator was fast enough to maintain the concentration of oxidized (or reduced) TMB. Simulations were also performed using COMSOL® to estimate the effect of the scan rate for an IDE with an electrode gap of 5 mm. As shown in Fig. 4, the current signal was maintained at a similar level, irrespective of the scan rate. These results indicated that the scan rate does not have a significant influence on the current signal during redox cycling. Thus, the optimal conditions for performing redox cycling to analyze TMB with the highest sensitivity are (1) an IDE with an electrode gap of 5 mm, (2) a reduction potential of 200 mV versus Ag/AgCl and an oxidation potential of þ600 mV versus Ag/ AgCl, and (3) a scan rate in the range of 25e100 mV s1.
Fig. 2. Influence of applied potential at the collector on redox cycling of TMB. CVs were obtained at the generator at different potentials at the collector.
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39
37
Fig. 3. Influencing of the electrode gap between two interdigitated band electrodes on redox cycling of TMB. (a) CVs were obtained using IDEs with different electrode gaps (dg) between two band electrodes, and (b) COMSOL® simulation with different electrode gaps (dg).
Fig. 4. Influencing of scan rate at the generator on redox cycling of TMB. (a) CVs were obtained at different scan rates of the generator. (b) COMSOL® simulation for different scan rates.
3.2. Immunoassay for CEA based on redox cycling The current signals for redox cycling of TMB samples with different ODs were measured using an IDE with an electrode gap of 5 mm. As mentioned in section 3.1, the amounts of oxidized TMB molecules produced by samples with different OD values are proportional to the [TMBox]/[TMBred] term in the Nernst equation, when the same oxidative potential is applied to the collector electrode. As shown in Fig. 5(a), the signal currents for redox cycling of oxidized TMB samples were smaller than that for the TMB stock solution (OD ¼ 0, without oxidized molecules), and the signal current decreased as the OD value increased. A standard curve for quantification of TMB was obtained by performing redox cycling under the optimal conditions, that is, by using collector potentials of 200 and þ 600 mV versus Ag/AgCl as the reduction and oxidation potentials, respectively. The potential range of the generator was set as 200 to þ600 mV versus Ag/AgCl at a scan rate of 50 mV s1 using an IDE with an electrode gap of 5 mm. For comparison, conventional cyclic voltammetry using one set of interdigitated finger electrodes as the working electrode was carried out using a potential range of 200 to þ600 mV versus Ag/AgCl
at a scan rate of 50 mV s1. As shown in Fig. 5(a), standard curves were obtained using standard samples with different OD values. The sensitivities of the redox cycling and conventional cyclic voltammetry measurements were calculated as 861 and 178 nA OD1, respectively. These results showed that the sensitivity of redox cycling is five-fold higher than that of conventional cyclic voltammetry using a single interdigitated band electrode without the collector. Quantification of TMB by redox cycling was applied to a commercial immunoassay kit for the detection of CEA. The immunoassay was performed using a 96-well plate, which was coated with anti-CEA antibodies, as shown in Fig. 5(c). The standard curve for CEA was calibrated using standard CEA samples of known concentrations. The amount of bound CEA was quantified by treating the secondary antibodies with HRP and then with TMB. As shown in Fig. 5(d), the limit of detection for CEA detection was determined to be 1.08 ng mL1 for redox cycling and conventional cyclic voltammetry, respectively (n ¼ 3). The cutoff value for positive determination of CEA was established to be an OD value of 0.1226 at a wavelength of 650 nm, which corresponds to a CEA concentration of 5 ng mL1 according to the manufacturer's instructions. As shown in Fig. 5(d), the current signal at the cut-off concentration of CEA (5 ng mL1) was determined to be 0.7 mA and 0.2 mA for redox cycling and conventional cyclic voltammetry, respectively (n ¼ 3). These results showed that the sensitivity of redox cycling is more than three-fold higher than that of single-mode amperometry using a single interdigitated band electrode without the collector. Usually, test results within the standard deviation at the cutoff value are regarded to be in the grey zone for the medical diagnosis and require retesting. In this work, the standard deviations at the cut-off concentration of CEA (5 ng mL1) were determined to be 51 mA (71.3% of mean value) and 58 mA (8.9% of mean value) for redox cycling and conventional cyclic voltammetry, respectively (n ¼ 3). These results showed that the grey zone for positive determination of CEA could be improved considerably by using redox cycling.
4. Conclusions Redox cycling based on IDEs was applied to the quantification of TMB during immunoassays. During the redox cycling process, the
38
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39
Fig. 5. Application of redox cycling to a commercial immunoassay kit for the detection of CEA. (a) CVs obtained for redox cycling of TMB samples with different OD values. (b) Standard curves for quantification of oxidized TMB using redox cycling and a single interdigitated band electrode without the collector. (c) Schematic view of ELISA test for the detection of CEA. (d) Standard curves for detection of CEA using redox cycling and a single interdigitated band electrode without the collector.
first working electrode (generator) was used for cyclic voltammetry of TMB by changing the potential between 200 and þ 600 mV versus Ag/AgCl, while two fixed potentials were sequentially applied at the second working electrode (collector) for reduction (and oxidation) of oxidized (and reduced) TMB. The optimum conditions for maximizing sensitivity during redox cycling of TMB were determined to be (1) an interdigitated finger electrode gap of 5 mm, (2) reduction and oxidation potentials of 200 and þ 600 mV versus Ag/AgCl, respectively, and (3) a scan rate in the range of 25e100 mV s1. The sensitivity of TMB measurements performed using redox cycling was five-fold higher than that of single-mode amperometry using a single interdigitated band electrode without the collector. Finally, the redox cycling of TMB was applied to a commercial immunoassay kit for detecting CEA. Based on the assay results, the sensitivity of redox cycling was found to be threefold higher than that of conventional cyclic voltammetry using one set of interdigitated finger electrodes as a working electrode.
Acknowledgements This work was supported by Nano-Convergence Foundation (R201602210) funded by the Ministry of Science, ICT and Future Planning (MSIP, Korea) and the Ministry of Trade, Industry and Energy (MOTIE, Korea), and by the Industry Technology Development Program (10063335) funded by the Ministry of Trade, Industry and Energy (MOTIE, Korea).
References [1] E. Marulli, A. Aloisi, P.D. Giuseppe, R. Rinaldi, Micro and nanotechnology for early diagnosis and detection of rheumatic diseases-molecular markers, Biochip J. 10 (2016) 189e197. [2] J.H. Choi, J.W. Lee, B.K. Oh, Nanomaterial-based in vitro analytical system for diagnosis and therapy in microfluidic device, Biochip J. 10 (2016) 331e345. [3] S.G. Rhee, H.A. Woo, I.S. Kil, S.H. Bae, Peroxiredoxin functions as a peroxidase and a regulator and sensor of local peroxides, J. Biol. Chem. 287 (2012) 4403e4410. [4] P.D. Josephy, T. Eling, R.P.J. Mason, The horseradish peroxidase-catalyzed oxidation of 3,5,3',5'-tetramethylbenzidine. Free radical and charge-transfer complex intermediates, Biol. Chem. 257 (1982) 3669e3675. [5] C.A. Groom, J.H.T. Luong, R. Thatipalmala, Dual functionalities of 4aminodiphenlyamine in enzymatic assay and mediated biosensor construction, Anal. Biochem. 231 (1995) 393e399. [6] J.C. Pyun, H.H. Lee, C.S. Lee, Application of an amperometric detector to the conventional enzyme immunoassay, Sensors Actuator B 78 (2001) 232e236. [7] J.C. Pyun, S.D. Kim, J.W. Chung, Gold film array electrode for an ELISA analyser based on an amperometric detector, Sensors Actuators B 111e112 (2005) 416e422. [8] J.I. Kim, A. Bordeanu, J.C. Pyun, Diamond-like carbon (DLC) microelectrode for electrochemical ELISA, Biosens. Bioelectron. 24 (2009) 1394e1398. [9] J.K. Lee, G. Yoo, M. Park, J. Jose, M.J. Kang, J.C. Pyun, Electrochemical ELISA based on E. coli with autodisplayed Z-domains, Sensors Actuators B 175 (2012) 46e52. [10] O. Niwa, M. Morita, H. Tabei, Electrochemical behavior of reversible redox species at interdigitated array electrodes with different geometries: consideration of redox cycling and collection efficiency, Anal. Chem. 62 (1990) 447e452. [11] K. Ueno, M. Hayashida, J.Y. Ye, H. Misawa, Fabrication and electrochemical characterization of interdigitated nanoelectrode arrays, Electrochem. Comm. 7 (2005) 161e165. [12] M. Paeschke, U. Wollenberger, T. Lisec, U. Schnakenberg, R. Hintsche, Highly sensitive electrochemical microsensors using submicrometer electrode arrays,
G.-Y. Lee et al. / Analytica Chimica Acta 971 (2017) 33e39 Sensors Actuators B 27 (1995) 394e397. [13] A.J. Bard, J.A. Crayston, G.P. Kittlesen, T. Varco Shea, M.S. Wrighton, Digital simulation of the measured electrochemical response of reversible redox couples at microelectrode arrays: consequences arising from closely spaced ultramicroelectrodes, Anal. Chem. 58 (1986) 2321e2331. [14] V.A.T. Dam, W. Olthuis, Redox cycling with facing interdigitated array electrodes as a method for selective detection of redox species, van den Berg A. Anal. 132 (2007) 365e370. [15] D.G. Sanderson, L.B. Anderson, Filar electrodes: steady-state currents and spectroelectrochemistry at twin interdigitated electrodes, Anal. Chem. 57 (1985) 2388e2393. [16] T.B. Tran, S.J. Son, J.H. Min, Nanomaterials in label-free impedimetric biosensor: current process and future perspectives, Biochip J. 10 (2016) 318e330. [17] T. Lee, J.W. Choi, Fabrication of fusion protein-based heterolayers composed of redox protein/myoglobin for bioelectronic device, Biochip J. 10 (2016) 103e110. [18] K.S. Lee, S.Y. Shin, Y.S. Yoon, Fe3O4 nanoparticles on MWCNTs backbone for lithium ion batteries, J. Korean Ceram. Soc. 53 (2016) 376e380. [19] W. Smirnov, N. Yang, R. Hoffmann, J. Hees, H. Obloh, W. MVuller-Sebert, C.E. Nebel, Integrated all-diamond ultramicroelectrode arrays: optimization of faradaic and capacitive currents, Anal. Chem. 83 (2011) 7438e7443. [20] J. Duay, J.M. Goran, K.J. Stevenson, Facile fabrication of carbon ultramicro- to nanoelectrode arrays with tunable voltammetric response, Anal. Chem. 86 (2014) 11528e11532. [21] H.W. Jung, Y.W. Chang, G.Y. Lee, S.B. Cho, M.J. Kang, J.C. Pyun, Capacitive biosensor based on an interdigitated electrode with nanoislands, Anal. Chim. Acta 844 (2014) 27e34. [22] B.G. An, Y.W. Chang, H.R. Kim, M.J. Kang, J.K. Park, J.C. Pyun, In-situ synthesized CdS nanowires for the application to photosensor, Sensors Actuators B 221 (2015) 884e890. [23] Y.Y. Noh, K.B. Kim, M.K. Choi, O.S. song properties of blocking layer with Ag nano powder in a dye sensitized solar cell, J. Korean Ceram. Soc. 53 (2016) 105e109. [24] J.C. Pyun, M.Y. Cheong, S.H. Park, H.Y. Kim, J.S. Park, Modification of short peptides using 3-aminocaproic acid for improved coating efficiency in indirect enzyme-linked immunosorbent assays (ELISA), J. Immunol. Methods 208
39
(1997) 141e149. [25] J.W. Chung, R. Bernhardt, J.C. Pyun, Sequential analysis of multiple analytes using a surface plasmon resonance (SPR) biosensor, J. Immunol. Methods 311 (2006) 178e188. [26] K. Petermann, S. Vordenb€ aumen, J.C. Pyun, A. Braukmann, E. Bleck, M. Schneider, J. Jose, Autodisplay of 60-kDa Ro/SS-A antigen and development of a surface display enzyme-linked immunosorbent assay for systemic lupus erythematosus patient sera screening, Anal. Biochem. 407 (2010) 72e78. [27] J. Guo, E. Lindner, Cyclic voltammetry at shallow recessed microdisc electrode:Theoretical and experimental study, J. Electroanal. Chem. 629 (2009) 180e184. [28] A. Lavacchi, U. Bardi, C. Borri, S. Caporali, A. Fossati, I. Perissi, Cyclic voltammetry simulation at microelectrode arrays with COMSOL multiphysics, J. Appl. Electrochem 39 (2009) 2159e2163. [29] C. Amatore, C. Pebay, L. Thouin, A. Wang, Cyclic voltammetry at microelectrodes. Influence of natural convection on diffusion layers as characterized by in situ mapping of concentration profiles, Electrochem. Comm. 11 (2009) 1269e1272. [30] M. Odijk, W. Olthuis, V.A.T. Dam, A. Berg, Simulation of redox-cycling phenomena at interdigitated array (IDA) electrodes:amplification and selectivity, Electroanalysis 20 (5) (2008) 463e468. [31] E. Goluch, B. Wolfrum, P.S. Singh, M. Zevenbergen, S.G. Lemay, Redox cycling in nanofluidic channels using interdigitated electrodes, Anal. Bioanal. Chem. 394 (2) (2009) 447e456. [32] S.J. Hood, D.K. Kampouris, R.O. Kadara, N. Jenkinson, F.J. Campo, F.X. Munoz, C.E. Banks, Why ‘the bigger the better’ is not always the case when utilizing microelectrode arrays: high density vs. low density arrays for the electroanalytical sensing of chromium(VI), Analyst 134 (2009) 2301e2305. [33] C.H. Hamann, A. Hamnett, W. Vielstich, Electrochemistry, second ed., WileyVCH Verlag GmbH, Weinheim, 1998, pp. 172e179. [34] F. Liu, G. Kolesov, B.A. Parkinson, Time of flight electrochemistry: diffusion coefficient measurements using interdigitated array (IDA) electrodes, J. Electrochem. Soc. 161 (13) (2014) H3015eH3019. [35] G. Lewis, S. Dale, B. Kasprzyk-Hordern, A. Lubben, E. Barnes, R. Compton, F. Marken, Cavity transport effects in generatorecollector electrochemical analysis of nitrobenzene, Phys. Chem. Chem. Phys. 16 (2014) 18966e18973.