Microfluidic Enzyme Immunoassay Using Silicon ... - ACS Publications

Anal. Chem. 2002, 74, 2994-3004

Microfluidic Enzyme Immunoassay Using Silicon Microchip with Immobilized Antibodies and Chemiluminescence Detection Julia Yakovleva,† Richard Davidsson,‡ Anna Lobanova,† Martin Bengtsson,§ Sergei Eremin,† Thomas Laurell,§ and Jenny Emneús*,‡

Division of Chemical Enzymology, Department of Chemistry, M.V. Lomonosov, Moscow State University, Moscow 119899 Russia, Department of Analytical Chemistry, Lund University, P.O. Box 124, S-221 00, Lund, Sweden, and Department of Electrical Measurements, Lund Institute of Technology, Lund University, P.O. Box 118, S-221 00, Lund, Sweden

Silicon microchips with immobilized antibodies were used to develop microfluidic enzyme immunoassays using chemiluminescence detection and horseradish peroxidase (HRP) as the enzyme label. Polyclonal anti-atrazine antibodies were coupled to the silicon microchip surface with an overall dimension of 13.1 × 3.2 mm, comprising 42 porous flow channels of 235-µm depth and 25-µm width. Different immobilization protocols based on covalent or noncovalent modification of the silica surface with 3-aminopropyltriethoxysilane (APTES) or 3-glycidoxypropyltrimethoxysilane (GOPS), linear polyethylenimine (LPEI, MW 750 000), or branched polyethylenimine (BPEI, MW 25 000), followed by adsorption or covalent attachment of the antibody, were evaluated to reach the best reusability, stability, and sensitivity of the microfluidic enzyme immunoassay (µFEIA). Adsorption of antibodies on a LPEI-modified silica surface and covalent attachment to physically adsorbed BPEI lead to unstable antibody coatings. Covalent coupling of antibodies via glutaraldehyde (GA) to three different functionalized silica surfaces (APTES-GA, LPEI-GA, and GOPS-BPEI-GA) resulted in antibody coatings that could be completely regenerated using 0.4 M glycine/HCl, pH 2.2. The buffer composition was shown to have a dramatic effect on the assay stability, where the commonly used phosphate buffer saline was proved to be the least suitable choice. The best long-term stability was obtained for the LPEI-GA surface with no loss of antibody activity during one month. The detection limits in the µFEIA for the three different immuno surfaces were 45, 3.8, and 0.80 ng/L (209, 17.7, and 3.7 pM) for APTES-GA, LPEI-GA, and GOPS-BPEIGA, respectively. Following the rising demand for monitoring of a large number of bioaffinity interactions or biomarkers at cell level, the development of miniaturized chemical systems has grown dramatically in recent years, allowing the realization of the micro total analysis * Corresponding author: Jenny Emneús, Tel: +46-46-2224820. Fax: +4646-2224544. E-mail: [email protected]. † Moscow State University. ‡ Department of Analytical Chemistry, Lund University. § Lund Institute of Technology, Lund University.

2994 Analytical Chemistry, Vol. 74, No. 13, July 1, 2002

system (µTAS).1-3 The µTAS was introduced by Manz et al.4 and employs an integrated tool comprising all the components needed to perform analysis (sample handling, pretreatment, and detection) on a single platform. Today, many of the microfluidic components for realizing a wide range of µTASs are available, for example, micropumps,5 valves,6 and dispensers,7 as well as miniaturized separations systems.8 In the rapid development of µTAS in the life sciences, the need for new detection principles that enables the monitoring of higher-order biological analytes is evident. A generic way of introducing high selectivity and sensitivity in miniaturized biorecognition is the use of high-surface-area microstructures, which can be tethered with highly specialized probes that either provide a high selectivity or affinity as, for example, enzymes, antibodies or receptors. An inherent feature of miniaturization is the ability to generate array systems, enabling possibilities of miniaturized low-cost screening assays. Today, chemical and biochemical applications of microreactors are rapidly expanding areas of research.9 A most attractive way of realizing the high-surface-area µTAS components needed is by means of silicon microfabrication. The benefits of using silicon in this respect is multifold: (i) Silicon possesses excellent surface chemical properties for immobilizing ligands, that is, basically all coupling chemistry developed for glass (silica) matrixes can be employed; (ii) A wide spectrum of microfabrication technology is readily available for the design and development of microfluidic silicon components; and (iii) Silicon can be processed to a porous state, that is, porous silicon, which is a high-surface-area matrix comparable to any other conventional solid-phase matrixes. (1) McCreedy, T. Trends Anal. Chem. 2000, 19, 396-401. (2) Shoji, S. Electron. Commun. JPN 1999, 82. (3) van den Berg, A.; Lammerink, T. S. J. In Microsystem Technology in Chemistry and Life Science, 1998; Vol. 194, pp 21-49. (4) Manz, A.; Graber, N.; Widmer, H. M. Sensors Actuat. B 1990, 1, 244-248. (5) van der Schoot, B. H.; Jeanneret, S.; van den Berg, A.; de Rooij, N. F. Sensors Actuat. B 1992, 6, 57-60. (6) Oosterbroek, R. E.; Berenschot, J. W.; Schlautmann, S.; Krijnen, G. J. M.; Lammerink, T. S. J.; Elwenspoek, M. C.; van den Berg, A. J. Micromech. Microeng. 1999, 9, 194-198. (7) Nilsson, J.; Laurell, T.; Wallman, L.; Drott, J. Proceedings of the 2nd International Symposium on Miniaturized Analytical Systems, µTAS 1996, 17-19. (8) Dolnı´k, V.; Liu, S. R.; Jovanovich, S. Electrophoresis 2000, 21, 41-54. 10.1021/ac015645b CCC: $22.00

© 2002 American Chemical Society Published on Web 05/25/2002

The development of high-surface-area immobilized enzyme microreactors in silicon has been pursued by Laurell et al. since 1994.10,11 Considerable improvements were obtained as the surface was vastly enhanced by anodization of the silicon microstructure in hydrofluoric acid, providing a high-surface-area layer on the silicon microreactor structure.12,13 The use of porous silicon microstructures in other microanalytical applications in which surface area is a fundamental need is a natural development, and it seems promising, therefore, to expand this concept to the field of immunoassays. Immunoassays are powerful analytical tools for biochemical studies, clinical diagnostics, and environmental monitoring; however, conventional heterogeneous microplate-based immunoassays require relatively long assay times, and automation can be rather costly. The integration of immunoassays on microchips is a rapidly developing research area that has been shown to result in reduced assay time, lower sample and reagent consumption, and enhanced sensitivity.14-23 Most of the immunoassays published so far have been based on electroosmotic pumping and electrophoretic separation of the bound and nonbound fractions and have been carried out in both homogeneous noncompetitive20-22 and competitive format;16-18 however, heterogeneous assays have also been reported.23 From a purely scientific point of view, the downsizing of analytical systems is interesting because of its possibilities to increase the performance of the analysis, mainly due to the small dimensions, for example, processes in macroscopic systems that are limited by diffusion can be much more efficient in the microscale as a result of the shorter distances for the species to travel. The development of microfluidic assay systems has, in addition, several potential advantages. For example, many different detection principles can be used, sample preparation steps can be coupled on-line, and real time analysis is possible. Other important considerations in miniaturization are potentially lower power consumption, point of care/on site analysis, and more stringent law enforcement on the use of chemicals. The detection of analytes at microscale levels requires sensitive analytical techniques that are suitable for on-chip detection. Many techniques have been applied for detection of biomolecules in microfluidic devices, including refractive index, absorption, elec(9) Haswell, S.; Skelton, V. Trends Anal. Chem. 2000, 19, 389-395. (10) Laurell, T.; Rosengren, L. Sensors Actuat. B 1994, 19, 614-617. (11) Laurell, T.; Drott, J.; Rosengren, L. Biosens. Bioelectron. 1995, 10, 289299. (12) Laurell, T.; Drott, J.; Rosengren, L.; Lindstrom, K. Sensors Actuat. B 1996, 31, 161-166. (13) Drott, J.; Lindstrom, K.; Rosengren, L.; Laurell, T. J. Micromech. Microeng. 1997, 7, 14-23. (14) Sato, K.; Tokeshi, M.; Odake, T.; Kimura, H.; Ooi, T.; Nakao, M.; Kitamori, T. Anal. Chem. 2000, 72, 1144-1147. (15) Sato, K.; Tokeshi, M.; Kimura, H.; Kitamori, T. Anal. Chem. 2001, 73, 12131218. (16) Chiem, N.; Harrison, D. J. Anal. Chem. 1997, 69, 373-378. (17) Chiem, N. H.; Harrison, D. J. Clin. Chem. 1998, 44, 591-598. (18) Koutny, L. B.; Schmalzing, D.; Taylor, T. A.; Fuchs, M. Anal. Chem. 1996, 68, 18-22. (19) Bernard, A.; Michel, B.; Delamarche, E. Anal. Chem. 2001, 73, 8-12. (20) Mangru, S. D.; Harrison, D. J. Electrophoresis 1998, 19, 2301-2307. (21) Cheng, S. B.; Skinner, C. D.; Taylor, J.; Attiya, S.; Lee, W. E.; Picelli, G.; Harrison, D. J. Anal. Chem. 2001, 73, 1472-1479. (22) Wang, J.; Ibań ˜ez, A.; Chatrathi, M. P.; Escarpa, A. Anal. Chem. 2001, 73, 5323-5327. (23) Dodge, A.; Fluri, K.; Verpoorte, E.; de Rooij, N. F. Anal. Chem. 2001, 73, 3400-3409.

trochemical, and fluorescence, the latter of which has been generally shown to be very sensitive,24 especially implemented using laser-induced fluorescence and confocal microscopy. Another useful detection technique is chemiluminescence (CL), which was proved to be not only one of the most sensitive but also particularly suitable at microscale levels, since no external light source is needed.20,25,26 In this study, microfluidic enzyme immunoassays (µFEIA) based on antibody-coated silicon microchips were developed using enhanced CL detection of a horseradish peroxidase (HRP) label catalyzing the luminol/H2O2/p-iodophenol (PIP) reaction. The assay principle is based on a direct competitive heterogeneous format in which the analyte and an enzyme-labeled analyte (tracer) competes for immobilized antibody binding sites, following separation and detection of the bound tracer, generating the CL signal directly on the chip. The proper choice of immobilization methods for antibodies so that they will retain activity, stability, and specificity on the support surface, is one of the most important steps in the development of solid-phase immunoassays; thus, special attention was given to optimization of these conditions to reach the best stability and sensitivity of the µFEIA. 2. EXPERIMENTAL SECTION 2.1. Materials. The silicon microchips were fabricated by anisotropic chemical wet etching of silicon, described elsewhere.12,13 The microchips had overall dimensions of 13.1 × 3.2 mm, consisting of an inlet and an outlet basin at each end with 42 parallel, 10-mm-long, 235-µm-deep, and 25-µm-wide flow channels between (Figure 1c), through which the flow is directed. To increase the internal surface area, a porous layer was achieved by anodizing in a 1:1 mixture of 40% hydrofluoric acid and 96% ethanol with a current density of 50 mA/cm2 for 5 min.12,13 Affinity-purified polyclonal anti-atrazine immunoglobulin G (IgG) from sheep serum was generously provided by Dr. Ramadan Abuknesha (King’s College University of London, U.K.). Horseradish peroxidase (HRP) type VI-A (essentially salt-free), glutaraldehyde (GA) 25% v/v aqueous solution grade I, sodium cyanoborohydride, N,N′-dicyclohexylcarbodiimide (DCC), N-hydroxysuccinimide (NHS), and dimethylformamide (DMF) were purchased from Sigma (St. Louis, MO). Ammonia, hydrochloric acid, H2O2, acetonitrile, and toluene were from Merck (Darmstadt, Germany). Linear polyethylenimine (LPEI, MW 750 000; 50% w/v aqueous), branched polyethylenimine (BPEI, MW 25 000; 98%), 3-glycidoxypropyltrimethoxysilane (GOPS) 98%, luminol 97%, and PIP 99% were obtained from Aldrich (Milwaukee, WI), 3-aminopropyltriethoxysilane (APTES) 98%, was from ICN Biochemicals (Aurora, OH). Lysine monohydrate was obtained from Carl Roth Co. (Karlsruhe, Germany), and dimethyl sulfoxide (DMSO) was from Fisher Co. (Springfield, NJ). The atrazine hapten derivative 6-{{4-chloro-6-[(1-methylethyl) amino]-1,3,5-triazin-2-yl}amino}hexanoic acid (i-Pr/Cl/(CH2)5COOH) was synthesized according (24) Chabinyc, M. L.; Chiu, D. T.; McDonald, J. C.; Stroock, A. D.; Christian, J. F.; Karger, A. M.; Whitesides, G. M. Anal. Chem. 2001, 73, 4491-4498. (25) Wu, X. Z.; Suzuki, M.; Sawada, T.; Kitamori, T. Anal. Sci. 2000, 16, 321323. (26) Hashimoto, M.; Tsukagoshi, K.; Nakajima, R.; Kondo, K.; Arai, A. J. Chromatogr. A 2000, 867, 271-279.

Analytical Chemistry, Vol. 74, No. 13, July 1, 2002

2995

Figure 1. (a) Scheme of the µFEIA manifold; a syringe pump and a peristaltic pump were used for carrier buffer and regeneration buffer, respectively, at a flow rate of 50 µL/min. Sample injection was made through a six-port injection valve. The CL signal was measured with a PMT placed above the flow cell containing the silicon microchip with immobilized antibodies. (b) Detailed view of the plexiglass microchip flow cell. (c) A microchip beside a ruler and a magnified picture of the parallel channel network.

to Goodrow et al.27 Atrazine was obtained from Riedel-de-Hae¨n (Seelze, Germany). 2,4-Dichlorophenoxyacetic acid (2,4-D) coupled to HRP was generously provided by Prof. Milan Franek, Veterinary Research Institute, Brno, Czech Republic. All chemicals were of analytical grade. Milli-Q water was obtained using Milli-Q purification water system (Millipore, Bedford, MA). 2.2. Buffers and Solutions. The carrier buffer was 0.05 M Tris/HCl, pH 7.4. The substrate buffer was 0.05 M Tris/HCl, pH 9.0. When not in use, the microchips were kept in 0.1 M Tris/ HCl buffer, pH 7.0. Borate buffer (BB), 10 mM, was prepared from sodium tetraborate, and the pH was adjusted to 7.0 with 0.5 M boric acid. Phosphate buffer saline (PBS), 10 mM, was prepared by dissolving 8 g of NaCl, 0.2 g of KCl, 1.43 g of Na2HPO4‚2H2O, and 0.343 g of K2HPO4 in 1 L of water and adjusting the pH to 7.4 with HCl. Succinate buffer was prepared from 0.05 M succinic acid, and the pH was adjusted to 6.0 with NaOH. The solutions employed for regeneration of the immuno surfaces were as follows: (A) 50 mM H3PO4, pH 1.3; (B) 0.4 M glycine/HCl, pH 2.2; (C) 30% v/v acetonitrile; (D) 50% v/v DMSO; and (E) 0.5 M NaCl. The substrate solution was a mixture of 50 µL of luminol (50 mM in DMSO), 340 µL of PIP (150 mM in DMSO), 8.5 µL of 30% w/w H2O2, and 50 mL of substrate buffer. The working solution of the enzyme tracer (atrazine-HRP) was prepared from a 1 mg/L stock solution by diluting 20 000 times with Tris/HCl buffer, pH 7.4. Atrazine standard solutions of 0.001, 0.01, 0.1, 1, 10, and 1000 µg/L were prepared from a 5 mg/L water stock solution by diluting with Tris/HCl buffer, pH 7.4. 2.3. Immobilization of Proteins on Silicon Microchips. Antibodies were immobilized on silicon microchips employing five (27) Goodrow, M. H.; Harrison, R. O.; Hammock, B. D. J. Agric. Food Chem. 1990, 38, 990-996.

2996


different modified silica surfaces: (1) glutaraldehyde-activated APTES (APTES-GA surface), (2) glutaraldehyde-activated physically adsorbed BPEI (BPEI-GA surface), (3) glutaraldehydeactivated BPEI covalently attached to GOPS silanized surface (GOPS-BPEI-GA surface), (4) glutaraldehyde-activated physically adsorbed LPEI (LPEI-GA surface), and (5) adsorption of antibodies on a physically adsorbed LPEI layer (LPEI surface) (see Figure 2). In addition, bovine serum albumin (BSA) was immobilized to the APTES-GA surface and was used as a reference for nonspecific binding. To estimate the immobilization efficiency, the antibody concentration in the protein solution was measured spectrophotometrically before and after attachment to the microchip, assuming that an antibody solution of 1 mg/L corresponds to 1.35 absorbance units at 280 nm.28 Prior to immobilization, the silicon microchips were cleaned in a mixture of 25% NH3, 30% H2O2, and H2O (1:1:5 by volume), followed by cleaning in a mixture of 37% HCl, 30% H2O2 and H2O (1:1:5, by volume) at 100 °C for 5 min. The microchips were then carefully rinsed with water, ethanol, and acetone and dried under a stream of air. Finally, the microchips were dried in an oven at 110 °C for 1 h. 2.3 1. Immobilization of Antibodies onto APTES-GA Surface. The cleaned microchips were washed with sodium-dried toluene and then immersed in a solution of 10% APTES in dried toluene.29 To remove air bubbles from the microchip pores, vacuum was applied for 1-3 min. Then the reaction mixture was refluxed overnight at room temperature (RT) in a sealed vessel, protected from moisture with a drying tube filled with silica gel. After removal of the solution, the microchips were rinsed several times with toluene and acetone and dried in an oven at 110 °C for 1 h. The dried chips were immersed in 100 mL ethanol and sonicated for 10 min to introduce the solvent into the pores. After washing with BB, the amine groups of the APTES silanized chips were reacted with 2.5% v/v GA in BB for 1 h at RT, followed by thorough rinsing with Milli-Q water in order to remove traces of glutaraldehyde to avoid cross-linking after addition of antibodies. Antibody (1 mg/mL) in BB was added to the GA-activated chips and reacted overnight at +4 °C under gentle shaking. After 12 h, the residual aldehyde groups remaining after antibody attachment were blocked with 10 mg/mL of L-lysine. The Schiff bases were reduced with 20 mg/mL NaBH3CN solution in BB, and the reaction mixture was allowed to proceed for 1-2 h under stirring at RT. The microchips were then carefully washed and stored in 0.1 M Tris/HCl buffer at +4 °C until use. 2.3.2. Immobilization of Antibodies on BPEI-GA Surface. The cleaned microchips were immersed in 0.5% v/v solution of BPEI in BB and kept under stirring at RT overnight and then thoroughly washed with BB. To incorporate active aldehyde groups, the microchips were reacted with 2.5% v/v GA in BB for 2 h at RT under stirring. After careful washing with Milli-Q water and BB, the aldehyde-functionalized microchips were reacted with antibody solution at different concentrations, ranging from 0.5 to 1 mg/ mL. The final coupling step was performed overnight at +4 °C. Then residual aldehyde groups on the microchip were blocked and the Schiff bases reduced as described under section 2.3.1. (28) Harlow, E.; Lane, D. Antibodiessa Laboratory Manual; Cold Spring Harbor Laboratory: Cold Spring Harbor, 1988. (29) Weetall, H. H. Methods Enzymol. 1976, 44, 134-148.

Figure 2. (a) Structures of the five tested immuno surfaces. (b) Chemical structures of LPEI and BPEI.

2.3.3. Immobilization of Antibodies on GOPS-BPEI-GA Surface. The cleaned microchips were reacted with GOPS in dry toluene, containing 2% v/v GOPS and 0.2% triethylamine at RT.30 After 1 h, the GOPS-coated chips were first rinsed with toluene, then with acetone, and then dried in an oven at 110 °C for 1 h. To introduce a solvent in the pores, the chips were sonicated in 100 mL of ethanol for 10 min and then were washed with Milli-Q water. A solution of 0.5% v/v BPEI in succinate buffer was added, and the reaction mixture was gently shaken for 5 h at RT. After careful washing with Milli-Q water, the microchips were treated with 2.5% v/v GA in BB. After 2 h, the microchips were rinsed and then immersed in antibody solution in BB of different IgG concentration, ranging from 0.5 to 1 mg/mL. The reaction was allowed to proceed overnight at +4 °C, after which the microchips were blocked and reduced as described under section 2.3.1. 2.3.4. Immobilization of Antibodies on LPEI-GA Surface. The cleaned microchips were immersed in 0.5% v/v solution of LPEI in BB and kept under stirring at RT overnight and then thoroughly washed with BB. The glutaraldehyde activation and IgG attachment steps were carried out as described in section 2.3.1.

2.3.5. Adsorption of Antibodies on LPEI Surface. The cleaned microchips were immersed in 0.5% v/v solution of LPEI in BB and kept under stirring at RT overnight. The microchips were then carefully washed with Milli-Q water, immersed in 1 mg/mL solution of IgG in BB, and allowed to proceed overnight at +4 °C. 2.4. Synthesis of Enzyme Tracer. The enzyme tracer was synthesized by attachment of the i-Pr/Cl/(CH2)5COOH atrazine hapten derivative to HRP by the NHS ester method.31 The hapten (1 mg), NHS (1.7 mg), and DCC (6.2 mg) were mixed with 130 µL of DMF. The mixture was agitated overnight at RT. Then the activated ester solution was added dropwise to a solution containing 1 mg of HPR in 0.5 mL of 130 mM NaHCO3. The coupling reaction was performed for 3 h, after which the tracer was purified by dialysis against 3 L of PBS (five changes of buffer) using a Slide-A-Lyzer cassette (Pierce, Rockford, IL) with a molecular cutoff of 10 kDa. 2.5. System Setup and Assay Procedure. The µFEIA system is depicted in Figure 1a. The carrier buffer, 0.05 M Tris/ HCl, pH 7.4, was delivered by a CMA Microdialysis 100 syringe

(30) Lin, J. N.; Chang, I. N.; Andrade, J. D.; Herron, J. N.; Christensen, D. A. J. Chromatogr. 1991, 542, 41-54.

(31) Giersch, T. J. Agric. Food Chem. 1993, 41, 1006-1011.


2997

one assay cycle, including three repetitive injections of substrate, regeneration and washing steps, was ∼10 min. The data were extracted from the CL signal curves as peak heights and were always corrected for the background signal from the luminol/PIP/H2O2 mixture by subtraction.

Figure 3. Scheme of the direct competitive assay format performed in the µFEIA system shown in Figure 1.

pump (CMA/Microdialysis, Solna, Sweden) at 50 µL/min. The samples were injected via a Rheodyne six-port valve (Rohnert Park, CA) equipped with a 7-µL injection loop. For regeneration of the immuno surface, the flow was switched to 0.4 M glycine/ HCl buffer, pH 2.2, via a Gilson Minipuls 2 peristaltic pump (Villiers-le-Bell, France) at 50 µL/min. The microchip was incorporated in the flow system via a specially designed flow cell unit made of plexiglass with the inlet and outlet tubing glued into holes drilled in the top cover (Figure 1b). A thin transparent silicon rubber membrane was placed between the microchip and the top cover to prevent leakage. The CL signal was monitored via a photomultiplier tube (PMT, model no. HC135-01 UV to visible, Hamamatsu Photonics K. K., Japan), aligned right above the microchip flow cell, exposing the active immuno surface to the PMT. The PMT and microchip flow cell were placed in a holder to ensure correct positioning of the two units, and the whole detection was performed in a “black box” to prevent light from the surroundings from interfering, as seen in Figure 1a. The total volume of the µFEIA system was 22 µL with connecting PEEK tubing (i.d. 0.25 mm, Alltech, Deerfield, IL). Data acquisition was performed using homemade software. The assay procedure is shown in Figure 3 and is based on the principle of competitive direct immunoassay according to the following steps: Prior to introduction of antigen and tracer, the substrate mixture of luminol/PIP/H2O2 was injected into the flow system to monitor the inherent background luminescence. Next, a solution of enzyme tracer and the standard sample were mixed off-line and injected into the flow system at a flow rate of 50 µL/ min 20 s after injection; that is, when the sample plug had reached the chip surface, the flow was stopped for 2 min incubation. The flow was then started again, and the chip was washed with carrier buffer for 2 min. The amount of bound enzyme tracer was detected by injection of the substrate mixture (luminol, PIP, and H2O2). The HRP catalyzed CL reaction took place on the chip surface, and the emitted light was registered by the PMT. To complete the assay cycle, the immuno surface was regenerated using solution A, B, C, D, or E, pumped with the Gilson peristaltic pump at a flow rate of 50 µL/min for 2 min. The total operation time for 2998


3. RESULTS AND DISCUSSION The performance of any kind of immunoassay is primarily a function of antibody affinity and selectivity;32,33 however, the careful design of labeled competitor, optimization of assay format, type of support, immobilization procedure, and detection technique have also shown to contribute to the performance of solid-phase immunoassays. Herewith we are reporting on the development of a µFEIA, that is, a heterogeneous competitive flow immunoassay employing competition between the analyte and a HRP-labeled analyte derivative for antibodies immobilized on a silicon microchip surface. The microchips were fabricated from pure silicon; however, since this is not stable to oxygen in the atmosphere, the surface will oxidize to silicon dioxide. Thus, when discussing the properties of the microchip surface to which the antibodies are attached, it is more correct to denote it as “the silicon dioxide surface”, or in a more convenient way, “the silica surface”. Several challenges are associated with immobilizing antibodies on silica surfaces. The ability of silica to exhibit strong nonspecific protein adsorption via ionic interactions, van der Waals forces, polar-polar interactions, are a well-known phenomenon.34 If the interaction forces and contact area are large enough, the protein may become irreversibly adsorbed in a completely or partially denatured conformation.30 Physical adsorption of antibodies is the most simple and probably one of the most employed immobilization methods; however, as noted in a number of publications,30,35 a low surface activity was detected for antibodies physically adsorbed on silanized silica slides, of which 2 times higher than that observed for the LPEI-GA immuno surface. The RSD of the zerodose signals for both surfaces was calculated using the same number of points (N ) 27 binding/desorption cycles). For reasons such as the hydrodynamic effects mentioned above, the stability of the zero analyte dose signal only cannot represent a complete and adequate overview of assay stability. Therefore, monitoring the stability of assay parameters, for example, IC50 obtained from the midpoint of the calibration curve (the point where assay precision is the best) and estimated as the analyte concentration that inhibits 50% of the tracer binding, can give some additional (61) Gonza´lez-Martıńez, M. A.; Morais, S.; Puchades, R.; Maquieira, A.; Abad, A.; Montoya, A. Anal. Chem. 1997, 69, 2812-2818. (62) Skla´dal, P.; Deng, A. P.; Kolar, V. Anal. Chim. Acta 1999, 399, 29-36. (63) Marquette, C. A.; Blum, L. J. Talanta 2000, 51, 395-401. (64) Yulaev, M. F.; Sitdikov, R. A.; Dmitrieva, N. M.; Yazynina, E. V.; Zherdev, A. V.; Dzantiev, B. B. Sensors Actuat. B 2001, 75, 129-135. (65) Ko ¨nig, B.; Gra¨tzel, M. Anal. Chim. Acta 1993, 280, 37-41. (66) Gao, Z.; Chao, F.; Chao, Z.; Li, G. Sensors Actuat. B 2000, 66, 193-196. (67) Turiel, E.; Fernandez, P.; Perez-Conde, C.; Gutierrez, A. M.; Camara, C. Fresenius’ J. Anal. Chem. 1999, 365, 658-662. (68) Vianello, F.; Signor, L.; Pizzariello, A.; Di Paolo, M. L.; Scarpa, M.; Hock, B.; Giersch, T.; Rigo, A. Biosens. Bioelectron. 1998, 13, 45-53.


3001

Figure 7. The long-term stability of the LPEI-GA and GOPSBPEI-GA immuno surfaces is shown. The CL signals at zero analyte doze were measured on different days and plotted vs time over a period of 44 and 16 days, for the LPEI-GA and GOPS-BPEI-GA immuno surfaces, respectively. The conditions were the same as in Figure 5.

information. Figure 8 presents calibration curves run on different days; the normalized signals were plotted vs atrazine concentration and the experimental points were fit to a four-parameter logistic equation,69 also including the change in IC50 value between days. For the LPEI-GA surface (Figure 8a), the curves coincide fairly well with a between-assay precision, expressed as the RSD between curves, of 7.88% (see Table 2); however, as seen, there is a trend of increasing IC50 with time. For the GOPS-BPEI-GA surface (Figure 8b), the IC50 values vary randomly between days, and the between-assay RSD is 30%, which cannot be explained in simple terms, but in general, we found that it was more difficult to obtain reproducible result with the GOPS-BPEI-GA immuno surface. As already mentioned, physical adsorption of LPEI is a practically irreversible process that in this work gave rise to a relatively stable immuno surface, whereas the physically adsorbed branched PEI counterpart could not form a correspondingly stable immuno surface, that is, the BPEI had to be covalently attached. One reason for the better long-term stability of the LPEI-GA immuno surface compared to that of GOPS-BPEI-GA could simply be that the larger and also linear PEI polymer (Figure 2b LPEI) forms a very uniform and dense layer on top of the silica surface, that protects the antibody, whereas the covalently attached smaller, branched PEI (Figure 2b BPEI) gives rise to a more bulky nonuniform and maybe less-protective layer. 3.4. Antibody Immobilization Efficiency, Calibration and Assay Characteristics. The amount of immobilized antibody is an important parameter for the development of a heterogeneous competitive immunoassay, since a low antibody concentration should be used for the development of a sensitive assay.33,70 This approach, however, cannot be directly extended to the immuno surfaces reported here, since the support material, coupling chemistry, surface modification, and length of spacer used for immobilizing the antibody will also influence the antibody activity and, thus, the assay performance. The immuno surfaces obtained by different antibody immobilization procedures were quantitatively characterized in terms (69) Howes, L. In Immunoassays: Essential Data; Edwards, R., Ed.; Wiley: Chichester, 1996. (70) Garcinun õ, R. M.; Fernańdez, P.; Pe´rez-Conde, C.; Gutie´rrez, A. M.; Ca´mara, C. Talanta 2000, 52, 825-832.

3002 Analytical Chemistry, Vol. 74, No. 13, July 1, 2002

Figure 8. Calibration curves obtained on different days and the variability of the IC50 values with time for the LPEI-GA and GOPSBPEI-GA immuno surfaces are shown. A four parameter logistic equation was used to fit the standard curves according to the formula y ) (A - D)/[1 + (x/C)B] + D, where A and D are the maximal and minimal signals of the assay, respectively. B corresponds to the slope of the sigmoid and C is the concentration of atrazine resulting in 50% inhibition of binding of the enzyme tracer to antibody (IC50). (a) Calibration curves for the LPEI-GA immuno surface obtained on different days in which each curve is an average of three replicates, performed on each day. (×) represents the IC50 values (right y axis) plotted versus time (top x axis) extracted from the corresponding fourparameter curves. (b) Calibration curves for the GOPS-BPEI-GA immuno surface in the same way as in part a.

of the loaded amount of IgG, which was estimated spectrophotometrically by the depletion of antibody in the solution during the immobilization procedure. Table 2 shows the total amount of antibody covalently attached to APTES-GA, GOPS-BPEI-GA, and LPEI-GA on the microchip surface. It should be stated that the antibodies were immobilized in the batch mode, meaning that the whole microchip surface, not only the microchannels, was modified. However, the total microchip surface and the microchannel surface exposed during assay performance are approximately the same for all microchips. As can be seen from these data, the LPEI-GA surface results in the highest amount of immobilized IgG, whereas attachment to the APTES-GA surface gives the lowest value. In Table 2, the analytical characteristics of the curves seen in Figure 8 are presented as well as data obtained for the APTESGA surface: the LOD was calculated in two different ways, that is, using three times the standard deviation of the analyte zero

Table 2. Analytical Characteristics for Atrazine Assays Obtained Using the APTES-GA, LPEI-GA, and GOPS-BPEI-GA Immuno Surfacesa µFEIA

APTES-GA

LPEI-GA

GOPS-BPEI-GA

Ab loading mg LOD (3*SD) ng/L LOD (90%) ng/L IC50 µg/L Dynamic range µg/L RSD

0.09 (N ) 1)b 45c 24c 1.264c 0.130-30c

0.23 ( 0.09 (N ) 4) 3.83 ( 1.88 1.63 ( 0.83 0.049 ( 0.012 0.003-5.56 7.88

0.15 ( 0.02 (N ) 3) 0.80 ( 0.35 0.83 ( 0.15 0.061 ( 0.057 0.002-2.09 30.08

a The limit of detection (LOD), the midpoint of the calibration curves (IC ) and the dynamic ranges were determined as average values obtained 50 from all curves in Figure 8a or b, and the resulting mean values and standard deviations are given in the table. The assay precision is calculated as the relative standard deviation (RSD) in all points of the dynamic range of the curves (between 20 and 80% tracer inhibition) shown in Figure 8a or b, where each binding-desorption cycle is accompanied by three substrate injections. b N denotes the number of different microchips used. c Data based on one assay only.

dose signal and the analyte concentration corresponding to 90% tracer binding. The dynamic range was defined by the analyte concentration inhibiting the maximum signal by 20-80%. The hydrophilic polymer-based immuno surfaces (LPEI-GA and GOPS-BPEI-GA) resulted in average IC50 values at least 10 times lower than that of the ordinarily silanized immuno surface (APTES-GA), whereas the latter had a two times lower amount of antibodies immobilized. This might be explained from the steric point of view, considering that when antibody molecules are covalently attached to silica through a short spacer arm such as APTES, a large portion of its binding sites comes very close to the silica surface. The vicinity of a solid surface to the antigen binding sites may hinder the correct binding, leading to a decreased antibody activity. Spacer arms constituted by the long polymeric chains of LPEI and BPEI keep the antibody away from the support surface and protect it from the interaction with nonspecific adsorption sites on the silica surface, making the recognition sites more accessible to the antigen. If the LPEI-GA and GOPS-BPEI-GA immuno surfaces are compared, the latter has the lower antibody density and also the best assay sensitivity. Despite the difference in assay parameters, all three µFEIAs provide LOD values below the official European Union (E.U.) requirements for drinking water (0.1 µg/L), and for the PEI-based surfaces, the IC50 values are also below 0.1 µg/L. Such sensitivities are achieved not only as a result of the high antibody affinity, which usually is the main factor governing assay sensitivity.71,72 Other contributing factors are most likely the highly sensitive CL detection and the direct assay format used. The PIP-enhanced CL detection of peroxidase labels is a very sensitive technique73 previously used for development of pesticide immunoassays with low LOD.74 Immunoassays based on immobilized antibodies are generally known to be more sensitive than those based on immobilized antigens.50,61 It should be mentioned that the same antibody used in this work was previously used by other researchers in the flow mode, indirect format, but with different labels and detection techniques, reaching LODs that were more than 1 (71) Gosling, J. P. In Immunoassay; Diamandis, E. P., Christopoulos, T. K., Eds.; Academic Press: San Diego, 1996. (72) Ekins, R. P.; Chu, F. W. Clin. Chem. 1991, 37, 1955-1967. (73) Thorpe, G. H. G.; Kricka, L. J.; Moseley, S. B.; Whitehead, T. P. Clin. Chem. 1985, 31, 1335-1341. (74) Navas Dıáz, A.; Sanchez, F. G.; Lovillo, J.; Garcia, J. A. G. Anal. Chim. Acta 1996, 321, 219-224.

order of magnitude less sensitive than what is reported here.75,76 The obtained LODs for the µFEIAs are also better than those reached for microtiter-plate-based ELISA, even though flow immunoassays are claimed to be less sensitive than batch ones. For instance, a highly sensitive ELISA was reported for atrazine by Gascoń et al.,77 reaching an IC50 of 0.06 µg/L and a LOD of 9 ng/L, although the flow immunoassay based on the same antibody had an IC50 of 0.47 µg/L and a LOD of 75 ng/L. Winklmair et al.78 reported on monoclonal antibodies toward s-triazine herbicides, reaching an IC50 of 0.045 µg/L and a LOD of 6 ng/L for an atrazine plate-based ELISA. Bruun et al.79 developed a sensitive ELISA for a number of s-triazines, reaching an IC50 of 0.271 µg/L and a LOD of 24 ng/L for atrazine. CONCLUSIONS Silicon microchips were applied for the development of µFEIAs based on a heterogeneous competitive direct assay format, with HRP as the enzyme label in combination with CL detection. The developed µFEIAs had a total assay time of ∼10 min for atrazine, including incubation, detection, and regeneration. Though numerous immobilization chemistries have been reported for antibody immobilization, most efforts were devoted to obtain good assay sensitivity by optimizing the assay conditions, while the search for the best immobilization procedure and comparison of different techniques in terms of short- and longterm stability has gained less attention. For miniaturized systems such as this, nonspecific binding effects can easily become a dominant and limiting factor. In addition, the amount of reagents immobilized is relatively low, as compared to “macro” systems, and thus, the stability of the biological reagent becomes of paramount importance. Our research was focused on testing different antibody immobilization chemistries in search of the best short- and long-term stability and sensitivity of the µFEIA. (75) Brecht, A.; Klotz, A.; Barzen, C.; Gauglitz, G.; Harris, R. D.; Quigley, G. R.; Wilkinson, J. S.; Sztajnbok, P.; Abuknesha, R.; Gascon, J.; Oubina, A.; Barcelo, D. Anal. Chim. Acta 1998, 362, 69-79. (76) Wilson, R.; Barker, M. H.; Schiffrin, D. J.; Abuknesha, R. Biosens. Bioelectron. 1997, 12, 277-286. (77) Gascon, J.; Oubina, A.; Ballesteros, B.; Barcelo, D.; Camps, F.; Marco, M. P.; Gonzalez-Martinez, M. A.; Morais, S.; Puchades, R.; Maquieira, A. Anal. Chim. Acta 1997, 347, 149-162. (78) Winklmair, M.; Weller, M. G.; Mangler, J.; Schlosshauer, B.; Niessner, R. Fresenius’ J. Anal. Chem. 1997, 358, 614-622. (79) Bruun, L.; Koch, C.; Jakobsen, M. H.; Pedersen, B.; Christiansen, M.; Aamand, J. Anal. Chim. Acta 2001, 436, 87-101.


3003

The same immuno reaction showed different features of reproducibility and sensitivity, depending on how the specific antibody was immobilized on the sensing surface. Adsorption of antibodies on physically adsorbed LPEI (LPEI surface) was fast and easy to implement, but the interactions between antibodies and LPEI are weak and did not allow the regeneration of the immuno surface. Similar results were obtained for covalently attached antibodies on physically adsorbed BPEI (BPEI-GA surface) and resulted in a very quick loss of antibody activity. Covalent coupling of antibodies to a covalently attached BPEI layer (GOPS-BPEI-GA surface) gave rise to the most sensitive assay; however, the precision between assays was low (RSD 30%). Covalent attachment of antibodies on physically adsorbed LPEI (LPEI-GA surface) proved to give a very sensitive as well as the most stable immuno surface without any significant loss of antibody activity after operation and storage for 28 days and with a RSD between assays of 7.8%. The ordinary silanized APTESGA surface proved to give an assay with a sensitivity >1 order of magnitude lower than the PEI-based surfaces. The buffer used as carrier and for storage proved to be a very important factor, that is, phosphate buffer proved in general to be detrimental to the µFEIAs and needed to be replaced with the organic amine buffer Tris/HCl to obtain any kind of assay stability. The positioning of the microchips in the flow cell in a reproducible manner from day to day proved to be a problem and is something that will be looked into for the future.

3004


For the future, the authors envisage the development of a multiassay platform by the use of channel arrays with different immobilized reagents and by construction of parallel or sequential analysis lines in the flow system, for example, multiple information about a particular biomarker could potentially be obtained with a single system using a channel array with different immobilized biomarker targets (antibody, receptor, enzyme, and cells). ACKNOWLEDGMENT The authors acknowledge financial support from the European Commission (INCO Copernicus project ERBIC15-CT98-0910), the Swedish Foundation for Strategic Environmental Research (MISTRA), the Swedish Council for Forestry and Agricultural Research (SJFR), and the Swedish Research Council (Vetenskapsrådet). Authors also thank Dr. Ramadan Abuknesha (King’s College University of London, U.K.) for providing us with anti-atrazine antibodies. Sven Ha¨gg (Department of Analytical Chemistry, Lund University) is acknowledged for providing us with homemade software for µFEIA data acquisition.

Received for review October 5, 2001. Accepted April 4, 2002. AC015645B

Microfluidic Enzyme Immunoassay Using Silicon ... - ACS Publications

Recommend Documents