Anal. Chem. 1998, 70, 1682-1685
Thick-Film Electrochemical Immunosensor Based on Stripping Potentiometric Detection of a Metal Ion Label Joseph Wang* and Baomin Tian
Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, New Mexico 88003 Kim R. Rogers*
National Exposure Research Laboratory, U.S. Environmental Protection Agency, P.O. Box 93478, Las Vegas, Nevada 89139
A disposable electrochemical immunosensor based on potentiometric stripping analysis (PSA) of a metal tracer and using an entirely on-chip assay format is demonstrated. Challenges associated with the adaptation of earlier stripping voltammetric immunoassays to an onchip operation, and with meeting the demands of decentralized testing, have been addressed. These include the surface immobilization of the antibody, the replacement of mercury drop electrodes, elimination of the separation and oxygen-removal steps, and the use of quiescent 30µL sample droplets. Human serum albumin (HSA) and anti-HSA antibody were used as a model system, while bismuth ion served as the metal label. The anti-HSA was immobilized onto the surface of a thick-film electrode, followed by a competition between the Bi-labeled analytetracer and the analyte (HSA) for the antibody binding sites. Upon removal of the unbound tracer, Bi3+ was released and detected by PSA. The dynamic concentration range for HSA (0.3-30 µg/mL) and the detection limit (0.2 µg/mL, i.e., 90 fmol in the 30-µL sample) indicate that the greatly simplified protocol does not compromise the performance characteristics of stripping immunoassays. Consequently, this on-chip operation offers great promise for decentralized (clinical and environmental) applications. Electrochemical immunosensors, based on the intimate coupling of specific immunochemical reactions with electrochemical transduction, have gained considerable attention in recent years.1,2 Such devices rely on the immobilization of one component of the immunological pair onto the electrode transducer and monitoring the electrochemical signals resulting from the antigen-antibody association. Most of the devices reported to date rely on the use of enzyme labels (particularly alkaline phosphatase) that generate electrochemically detectable species and offer a biocatalytic signal amplification. While various nonenzyme redox labels, including organic and inorganic tags, have been used in electrochemical immunoassays,3,4 their utility for immunosensor work has not been reported. (1) Morgan, C. L.; Newman, D.; Price, C. P. Clin. Chem. 1996, 42, 193. (2) Skladal, P. Electroanalysis 1997, 9, 737.
1682 Analytical Chemistry, Vol. 70, No. 9, May 1, 1998
The present article describes the development and attractive performance of a thick-film stripping potentiometric immunosensor, based on the use of a metal ion label. Because of its “builtin” preconcentration step, stripping analysis has been widely used for measuring trace metals down to the ppt concentration level.5 Pioneering studies by Henieman and co-workers6,7 illustrated the use of metal ion labels for heterogeneous immunoassays with anodic stripping voltammetric (ASV) detection. Such immunoassays involved covalently linking chelating agent to a protein to serve as a Chelon for the metal label. Following competitive equilibrium between the labeled and unlabeled protein for the antibody (immobilized on a polyester tube), the metal label was released and transferred to an electrochemical cell for an ASV detection with a hanging mercury drop electrode (HMDE) and a deaerated solution. The goal of the present work is to address various challenges associated with the conversion of the electrochemical stripping immunoassay protocol to a solid-state, user-friendly sensor format that does not compromise the attractive analytical performance of the former. For this purpose, we employed an antibody-coated, screen-printed sensor, performed the entire assay protocol directly on the surface of the disposable strip, and employed the highly sensitive potentiometric stripping mode for detecting the released metal ion label in microliter solutions (Figure 1). As desired for decentralized sensing applications, such an on-chip protocol offers several advantages compared to conventional ASV-based immunoassays, including simplified operation (e.g., the elimination of the separation and deaeration steps), greatly reduced costs and reagent volumes, elimination of toxic mercury drops, and a more sensitive stripping detection mode. These and other features of the new stripping immunosensing format are reported in the following sections. EXPERIMENTAL SECTION Apparatus. Potentiometric stripping analysis (PSA) was performed with a TraceLab system (PSU20, Radiometer Inc.) (3) Heineman, W. R.; Halsall, H. B. Anal. Chem. 1985, 57, 1321A. (4) Rapicault, S.; Limoges, B.; Degrand, C. Anal. Chem. 1996, 68, 930. (5) Wang, J. Stripping Analysis: Principles, Instrumentation and Applications; VCH Publishers: Deerfield Beach, FL, 1985. (6) Doyle, M. J.; Halsall, H. B.; Heineman, W. R. Anal. Chem. 1982, 54, 2318. (7) Hayes, F. J.; Halsall, H. B.; Heineman, W. R. Anal. Chem. 1994, 66, 1860. S0003-2700(97)01298-5 CCC: $15.00
© 1998 American Chemical Society Published on Web 03/19/1998
Figure 1. Schematic diagram of the screen-printed electrode immunoassay: (A) immobilization of antibody; (B) immunological incubation; (C) removal of unbound tracer; (D) release and PSA detection of Bi3+.
equipped with an IBM PS/55SX computer. Voltammetric stripping measurements were carried out with an electrochemical analyzer (model 620, CH Instruments) in connection with a Packard Bell pentium computer. Thick-film sensor strips, containing the screen-printed carbon working and silver reference electrodes (along with the immobilized antibody), were used in most experiments. Few experiments employed a large-volume (10 mL) cell, utilizing the printed carbon strip, a BAS RE-1 Ag/ AgCl reference electrode, and a platinum wire auxiliary electrode. Screen-Printing Fabrication. The thick-film sensors were fabricated with a semiautomatic screen printer (model TF-100, MPM, Franklin, MA). All inks were received from Electro-Science Laboratories (ESL Inc., King of Prussia, PA). The carbon ink (DRS12113, modified by adding 30% w/w graphite powder (Aesar)) was printed through a patterned stencil onto 10 × 10 cm2 alumina ceramic plates containing 30 strips with 3.33 cm × 1.00 cm dimensions (defined by a pre/semi laser cut). The resulting printed working electrodes were cured for 30 min at 180 °C. The reference electrodes were then printed (parallel to the working electrodes) using the ESL type 1110-s silver ink. Subsequently, an insulator layer was printed (using the ESL 240-SB ink), defining the working electrode (1.5 mm × 5 mm) and reference electrode (1.2 mm × 5 mm) areas and corresponding electrical contacts on the other end. The insulator layer was cured at 180 °C for 4 h. Chemicals. Human serum albumin (HSA, 96-99%) and antihuman albumin (anti-HSA IgG developed in rabbit) were purchased from Sigma. Bismuth(III) nitrate pentahydrate (99%) and diethylenetriaminepentaacetic acid (DTPA, 98%) were obtained from Aldrich. All other chemicals were of reagent grade and supplied by Aldrich. The HSA standard stock solution was prepared with 10 mM phosphate-buffered saline (PBS, pH 7.2), containing 0.15 g/L sodium azide (NaN3, added as a preservative), and stored at 4 °C. The anti-HSA was stored (as received) in the frozen state, and its standard solutions were prepared daily with PBS solution as in use. All solutions were prepared with distilled, deionized water (>18 MΩ). Labeling and Characterization of the HSA Tracer. DTPA was coupled to HSA as previously described.7 In brief, the bicyclic anhydride method was used to prepare the anhydride of DTPA.8 The anhydride was then reacted (1 h, room temperature) with
100 mg of HSA at a molar ratio of 50:1 (DTPA:HSA) in 0.1 M HEPES, pH 7.0. The reaction mixture was then dialyzed against 3 L of citrate buffer (0.25 M, pH 5.5) overnight at 4 °C. A 50 M excess of Bi(NO3)3‚5H2O was added to the dialysate and incubated for 1 h at room temperature prior to dialysis overnight against 3 L of PBS (with two buffer changes). As was previously reported,7 the HSA-DTPA-Bi(III) complex could be stored at 4 °C for a month without loss of the label. The concentration of protein in HSA-DTPA-Bi was determined by absorption spectroscopy at 280 nm using the standard addition method (for HSA). The amount of Bi(III) released from a given amount of HSA-DTPABi (measured with PSA) was used to calculate a DTPA:HAS molar labeling ratio of 4. Immunoassay Procedure. The on-chip immunosensor protocol is shown in Figure 1. Anti-HSA was noncovalently adsorbed onto the electrode working area (exposed electrodes and surrounding ceramic) by casting a 30-µL droplet of the 10 µg/mL anti-HAS solution over the surface (Figure 1A). The electrode was incubated for 1 h in a chamber saturated with water vapor to prevent sample evaporation. Following a 15-min incubation in the blocking buffer (PBS containing 1% (w/v) dried skim milk and 0.02% Tween 20), 30 µL of the sample or standard solution, containing 10 µg/mL HSA-DTPA-Bi and a given concentration of HSA, was pipetted onto the resulting electrode tip (Figure 1B). After a 30-min incubation, the electrode was rinsed with water to remove unbound tracer (Figure 1C) and was stored in water until the PSA measurement. Subsequently, for the PSA measurement, 30 µL of release reagent, containing 1.0 M HCl and 20 µg/mL of Hg2+, was pipetted onto the working area (Figure 3D). The bismuth deposition (along with the in situ mercury plating) proceeded from the quiescent droplet for 10 min at -0.70 V. The potentiogram was recorded in connection with chemical oxidation. The normalized signal peak areas obtained from the immunoassay data using different batches of electrodes were calculated using the following relationship: (8) Krejcarek, G. E.; Tucker, K. L. Biochem. Biophys. Res. Commun. 1977, 77, 581.
Analytical Chemistry, Vol. 70, No. 9, May 1, 1998
1683
normalized peak area ) (Ax - A0)/(A100 - A0) × 100
where Ax is the peak area at various HSA concentrations, A100 is the peak area in the absence of HSA, and A0 is the peak area at the highest (i.e., saturating) concentrations of HSA. RESULTS AND DISCUSSION From Immunoassays to Immunosensors. Previous immunoassays based on stripping measurements of metal ion tracers have relied on the use of polystyrene tubes, containing the immobilized antibody, and a conventional electrochemical cell with the hanging mercury drop electrode and large-volume deaerated solutions.6,7 The goal of the present work is to transfer the stripping immunoassay protocol to a solid-state sensor format by performing the entire heterogeneous competitive immunoassay on the surface of a disposable, thick-film, screen-printed electrode (Figure 1). The HSA-specific antibody was first immobilized on the surface (Figure 1A), followed by a blocking step using skim milk and Tween 20 in the PBS assay buffer. The use of blocking buffer reduced the nonspecific binding of the bismuth-labeled antigen tracer to the surface during the immunological incubation (Figure 1B). Following removal of the unbound tracer (Figure 1C), Bi3+ was released from the HSA-DTPA chelate using 1.0 M HCl and was detected by PSA (Figure 1D). The screen-printing microfabrication technology is commonly used for mass production of highly reproducible yet inexpensive strip electrodes. Such technology has been used recently for the development of amperometric enzyme immunosensors.9 Recent studies in our laboratory have illustrated that the use of disposable screen-printed electrodes does not compromise the attractive stripping performance of the common mercury drop or mercurycoated disk electrodes.10,11 Figure 2 compares the potentiometric and voltammetric stripping responses of the mercury-coated, screen-printed (B,D) and hanging mercury drop (A,C) electrodes for an 8 µg/L Bi3+ solution. As expected for a mercury film behavior, the coated, printed electrode exhibits a more favorable response when using both stripping modes. Note, in particular, the substantial improvement at the strip electrode upon using quiescent solutions (dotted line). Such improvement, which is attributed to the microelectrode array character (accrued from the deposition of mercury droplets), holds great promise for simplifying the sensor operation. The data of Figure 2 indicate also that the potentiometric stripping mode would be advantageous over the differential pulse stripping operation that was employed in previous immunoassay procedures.6,7 In particular, the PSA operation is not affected by the oxygen background contribution, hence resulting in a well-defined bismuth signal. Similar improvements of the PSA detection scheme were observed in comparison to square-wave stripping voltammetry and following the immunoassay protocol and release of the bismuth tracer (not shown). Such use of nondeaerated solutions further simplifies the operation of the stripping immunosensor. Optimization of Assay Conditions. The optimal concentration of analyte tracer was determined by incubating the antibodycoated electrode with increasing concentrations of HSA-DTPA(9) Skladal, P.; Kalab, T. Anal. Chim. Acta 1995, 316, 73. (10) Wang, J.; Tian, B. Anal. Chem. 1992, 64, 1706. (11) Wang, J.; Tian, B. Anal. Chem. 1993, 65, 1529.
1684 Analytical Chemistry, Vol. 70, No. 9, May 1, 1998
Figure 2. Comparison of the response of the hanging mercury drop (A,C) and screen-printed (B,D) electrodes to an 8 µg/L Bi3+ solution. Potentiometric (A,B) and differential-pulse voltammetric (C,D) stripping analysis with a 5-min deposition at -0.7 V: (B,D) 8 µg/L Bi3+, 20 mg/L Hg2+ in 1.0 M HCl; (A,C) 8 µg/L Bi3+ in 1.0 M HCl solution. Solid lines, stirred solution (300 rpm; 10 mL) during the deposition; dotted line, quiescent solution. (C,D) Pulse width, 0.2s; pulse height, 50 mV; scan rate, 20 mV/s.
Figure 3. Titration of immobilized antibody with Bi-labeled HSA. Electrodes were incubated in a 10 µg/mL anti-HSA solution for 1 h and for 30 min in solutions containing different concentrations of Bilabeled HSA. Sample volume, 300 µL; deposition time, 10 min. Other conditions are the same as in Figure 2.
Bi. The resulting plot of the Bi3+ peak area vs the log of the tracer concentration (Figure 3) displays a typical sigmoidal binding profile. To maximize the sensitivity of the assay to HSA while maintaining a sufficient working signal, a metal-labeled HSA concentration of 10 µg/mL (based on HSA protein) was routinely used for these assays. This tracer concentration yielded peak areas well above the detection limit for Bi3+, even at the highest concentrations of HSA (which correspond to the lowest signals). Another parameter which affected the observed peak (in the absence of HSA analyte) was the amount of time in which the tracer was allowed to bind to the immobilized antibody. Figure 4A indicates that saturation of the immunochemical binding begins after about 20 min. To maximize the signal and minimize the assay time, 30 min was selected for all subsequent assays.
Figure 5. Calibration data. Chronopotentiograms for solutions containing increasing levels of HSA: (a) 0, (b) 0.3, (c) 1, (d) 3, (e) 10, (f) 30, and (g) 100 µg/mL. Other conditions are as in Figure 3. Error bars represent SD, n ) 3.
Figure 4. Effect of immunoreaction time (A) and PSA deposition time (B) on the response. (A) One hour of incubation in 10 µg/mL anti-HSA and different incubation periods in the 10 µg/mL Bi-labeled HSA solution containing 10 µg/mL HSA; (B) 8 µg/L Bi3+ and 20 mg/L Hg2+ in 1.0 M HCl solution. Other conditions are the same as in Figure 3.
With respect to a competitive immunoassay format, the Bi3+ released from the HSA-DTPA-Bi bound to the immobilized antibody will be greatest in the absence of HSA and will decrease to nonspecific binding levels with increasing concentrations of the analyte. Consequently, it is critical for the assay that the Bi3+ released from the tracer be sufficiently higher than the electrochemical detection limit. Parameters involved with the electrochemical analysis were thus characterized and optimized. As might be expected, the deposition time (for the bismuth accumulation onto the electrode surface) had a profound effect on the peak area. Using an 8 µg/L bismuth (1 M HCl) solution and an unmodified (antibody-free) strip electrode, the peak area increased linearly with the deposition time over the 0-900 s range (Figure 4B). Such linearity indicates that exhaustive deposition of the bismuth from the droplet does not occur over this time scale. A deposition time of 600 s yielded sufficient peak areas generated from release of the Bi3+ from the antibody tracer and was selected for all assays. Such deposition time resulted in a detection limit of 0.2 µg/L Bi3+ (∼1 nM), as was indicated from the signal-to-noise characteristics (S/N ) 3) of the response for 0.5 µg/L Bi3+ (not shown). This value corresponds to 3 × 10-14 mol (6 pg) in the 30-µL samples. Analytical Performance. Figure 5 displays representative PSA signals resulting from the release of Bi3+ from the HSADTPA-Bi tracer bound to the immobilized antibody in the presence of increasing concentrations of HSA. The resulting calibration curve (Figure 5 inset), was fit to a standard fourparameter model,12 with an r2 value of 0.996. The dynamic measurement range extended from 0.3 to 30 µg/mL with a
detection limit of 0.2 µg/mL HSA (as was determined from 3 × SD of the zero analyte response subtracted from the zero analyte response and calculated from the calibration curve). Such value corresponds to 3 × 10-9 M (or to 90 fmol in the 30-µL sample) and is about an order of magnitude lower than that reported for a stripping immunoassay for HSA in connection with the same Bi tracer.7 A relative standard deviation of 7.4% was calculated for 14 repetitive measurements of a 1 µg/mL HAS solution. Such good precision reflects the reproducibility of the screen-printing fabrication as well as of the immobilization, immunobinding, and stripping procedures. Conclusions. We have demonstrated that the combination of screen-printed electrodes with an PSA immunosensor operation yields an analytically attractive performance. A simplified heterogeneous immunoassay for HSA can be conducted entirely on the surface of a disposable electrode and can be completed within less than 2 h. The coupling of the mercury-coated electrode with PSA facilitates stripping measurements of the metal ion label in unstirred, nondeaerated sample droplets. Such an on-chip sensing protocol does not compromise the attractive performance of stripping immunoassay measurements and offers a dynamic range of 0.3-30 µg/mL HSA, with a detection limit of 0.2 µg/mL (90 fmol). While the concept of single-use stripping immunosensors has been demonstrated using HSA as a model compound, it could be adapted for decentralized measurements of numerous compounds of environmental and clinical significance. Multiple analytes may be detected simultaneously in connection with different metal labels.7 Portable, low-power PSA instruments, developed recently for on-site detection of trace metals,13 could be adapted for use with the herein described immunosensor. ACKNOWLEDGMENT K.R. acknowledges EPA support for an IPA assignment at NMSU. Received for review December 1, 1997. Accepted January 30, 1998. AC971298N (12) Edwards, R., Ed. Immunoassays Essential Data; J. Wiley: New York, 1996. (13) Wang, J. Analyst 1994, 119, 763.
Analytical Chemistry, Vol. 70, No. 9, May 1, 1998
1685
