Anal. Chem. 2001, 73, 4388-4392
Development of a Flow Injection Capillary Chemiluminescent ELISA Using an Imprinted Polymer Instead of the Antibody Ioana Surugiu, Juraj Svitel, Lei Ye, Karsten Haupt,* and Bengt Danielsson*
Department of Pure and Applied Biochemistry, Chemical Center, Lund University, P.O. Box 124, S-22100 Lund, Sweden
A flow injection competitive assay analogous to enzyme immunoassays has been developed using a molecularly imprinted polymer instead of the antibody. A glass capillary was modified by covalently attaching an imprinted polymer to the inner capillary wall. The herbicide 2,4dichlorophenoxyacetic acid was used as a model analyte. The analyte was labeled with tobacco peroxidase, and chemiluminescence was used for detection in combination with a photomultiplier tube or a CCD camera. In a competitive mode, the analyte-peroxidase conjugate was passed together with the free analyte through the polymercoated capillary mounted in a flow system. After a washing step, the chemiluminescent substrate was injected and the bound fraction of the conjugate was quantified by measuring the intensity of the emitted light. Calibration curves corresponding to analyte concentrations ranging from 0.5 ng mL-1 to 50 µg mL-1 (2.25 nM-225 µM) were obtained. A lowered detection limit by 2 orders of magnitude was obtained when detection was done in discontinuous mode and the chemiluminescence light was conducted inside the photomultiplier tube by an optical fiber bundle, thus yielding a dynamic range of 5 pg mL-1100 ng mL-1 (22.5 pM-450 nM). Enzyme-linked immunosorbent assays (ELISAs) are among the most extensively used types of immunoassays, and their application in environmental, food, and medical analysis is an area with an enormous potential for growth. In comparison with classical analytical methods, ELISA methods offer the possibility of highly sensitive, relatively rapid, and cost-effective measurement.1,2 Due to the increasing number of samples that have to be measured in particular in combination with high-throughput screening and unattended monitoring, there is an ever-increasing demand for automated, high-performance assay formats. Standardized immunochemical methods profit from (monoclonal) antibodies because of defined selectivities and affinities toward the analyte.2 It has been suggested that molecularly imprinted polymers (MIPs) could provide an alternative to antibodies for use as recognition elements in such assays, owing to their high chemical and physical stability, * Corresponding authors:
[email protected]; bengt.danielsson@ tbiokem.lth.se. (1) Nunes, G. S.; Toscano, I. A.; Barcelo´, D. Trends Anal. Chem. 1998, 17, 79-87. (2) Hock, B.; Danwardt, A.; Kramer, K.; Marx, A. Anal. Chim. Acta 1995, 311, 393-405.
4388 Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
ease of preparation, and low price.3 This is especially true in cases where an antibody or another natural receptor is difficult to obtain. We have recently reported the development of a scalable highthroughput MIP assay based on chemiluminescence imaging with a charge-coupled device.4 Another aspect in assay and sensor development is the possible use of the system for unattended monitoring. For such applications, flow systems are well suited,5 due to easier sample handling. Their combination with chemically and physically stable, regenerable MIP materials should be a particular advantage. In the present paper, we describe the design of a flow injection ELISA-type MIP assay for the herbicide 2,4-dichlorophenoxyacetic acid (2,4-D). MIPs specific for this compound have been reported earlier by us6 and by others.7,8 The competitive assay we describe here uses peroxidase-catalyzed chemiluminescence for detection, similar to the earlier reported imaging format. A glass capillary was coated with the imprinted polymer and mounted in a flow system. A photomultiplier tube (PMT) was used for detection. EXPERIMENTAL SECTION Materials. 2,4-Dichlorophenoxyacetic acid, 4-chlorophenoxyacetic acid, phenoxyacetic acid, and luminol were obtained from Sigma (St. Louis, MO). Tobacco peroxidase (TOP) from transgenic tobacco plants9 was a gift from Irina G. Gazaryan, Chemical Faculty, Moscow State University, Russia. All other chemicals were of analytical grade, and solvents were of HPLC quality. Glass capillaries of 76-mm length, outer/inner diameter 1.2 mm/0.9 mm, were obtained from World Precision Instruments Inc. (Sarasota, FL). Brand-Haematocrit sealing compound was obtained from Kebo Lab (Lund, Sweden). The 2,4-D-TOP conjugate was prepared using the periodate oxidation method10 with minor modifications.11 In brief, TOP was activated with NaIO4, and subsequently coupled with diamino(3) Takeuchi, T.; Haginaka, J. J. Chromatogr., B 1999, 728, 1-20. (4) Surugiu, I.; Danielsson, B.; Ye, L.; Mosbach, K.; Haupt, K. Anal. Chem. 2001, 73, 487-491. (5) Franek, M.; Deng, A. P.; Kolar, V. Anal. Chim. Acta 2000, 412, 1-2. (6) Haupt, K.; Dzgoev, A.; Mosbach, K. Anal. Chem. 1998, 70, 628-631. (7) Nomura, Y.; Mugumura, H.; Yano, K.; Kugimiya, A.; McNiven, S.; Ikebukuro, K.; Karube, I. Anal. Lett. 1998, 31, 973-980. (8) Scho ¨llhorn, B.; Maurice, C.; Flohic, G.; Limoges, B. Analyst 2000, 125, 665-667. (9) Gazaryan, I. G.; Rubtsova, M. Y.; Kapeliuch, Y. L.; Rodriguez-Lopez, J. N.; Lagrimini, L. M.; Thorneley, R. N. F. Photochem. Photobiol. 1998, 67, 106110. (10) Nakane, P. K. J. Histochem. Cytochem. 1974, 22, 1084. (11) Dzantiev, B. Int. J. Environ. Anal. Chem. 1996, 65, 95. 10.1021/ac0101757 CCC: $20.00
© 2001 American Chemical Society Published on Web 08/02/2001
propane. 2,4-D was converted into the N-hydroxysuccinimide ester with N-hydroxysuccinimide and dicyclohexylcarbodiimide. The activated 2,4-D was allowed to react with the TOP-diaminopropane conjugate, resulting in 2,4-D-TOP. After each step, the modified TOP was purified by gel filtration. Activation of Glass Capillaries. Glass capillaries were coated with an imprinted polymer layer using a protocol similar to the one described by Bru¨ggemann et al.12 The capillaries were rinsed with 1 M NaOH for 5 min and incubated for another 10 min in the same solution, followed by consecutive rinses with water (2 min), 1 M HCl (2 min), and water (5 min), and dried in an air stream (5 min). Afterward the capillaries were rinsed twice with dry toluene, filled with a 10% solution of (3-methacryloxypropyl)trimethoxysilane in toluene, and incubated for 2 h at room temperature. Excess silane was then removed by rinsing three times with toluene, whereafter the capillaries were dried in an air stream. Preparation of Imprinted Polymer-Coated Capillaries. A polymerization solution was prepared containing 2 mmol of trimethylolpropane trimethacrylate (cross-linker), 2 mmol of 4-vinylpyridine (functional monomer), 0.08 mmol of 2,2′-azobisisobutyronitrile (polymerization initiator), and 5 mmol of 2,4-D (imprint molecule) in 40 mL of methanol/H2O (4/1, v/v). The solution was sonicated for 5 min, placed on ice, and bubbled with nitrogen for 3 min. The activated capillaries were filled with the polymerization solution, and polymerization was carried out at 60 °C in a water bath for 16 h. The capillaries were rinsed with methanol and placed in a sonication bath in order to remove loosely adhering polymer. Afterward the capillaries were washed by 2-h incubation with methanol/acetic acid (7/3, v/v), followed by acetone, to remove the imprint molecule. After a final wash with water, the capillaries were stored in water at 4 °C until use. Control capillaries were prepared in the same way but without the addition of 2,4-D. Measurements with CCD Camera. All solutions were drawn into the capillary by capillary force. To the coated capillaries was added a solution containing 5.5 ng of 2,4-D-TOP and the analyte, if appropriate, in 50 µL of 0.01 M phosphate buffer, pH 7, containing 0.1% Tween-20. The capillaries were incubated for 1 h. After three washing steps with buffer, the chemiluminescent substrate was added. SuperSignal ELISA Pico Chemiluminescent Substrate from Pierce (Rockford, IL) was prepared immediately prior to measurements by mixing equal volumes of stable peroxide solution with luminol/enhancer solution as indicated by the manufacturer. After substrate addition the capillaries were sealed immediately with Brand-Haematocrit sealing paste and fixed into a capillary holder under the objective of the CCD camera (Figure 1a). The capillary was imaged using a Photometrix 200 cooled CCD camera (Photometrix, Tucson, AZ). The camera was thermoelectrically cooled to -45 °C and was fitted with a 50-mm AF Nikkor objective (Nikon, Japan). The background value was obtained by imaging an equally sized region outside the region of interest and was subtracted from each measurement. During these studies, we used a binning factor such that the final image contained 256 × 256 superpixels. The exposure time was 90 s. The intensity of the spots was determined using the ROI function
of the software, which combines the pixel intensities. The intensities (analog-to-digital units, ADU) were plotted as a function of analyte concentration. Measurements with the Photomultiplier Tube. The coated capillaries were filled with a solution containing 1.1 ng 2,4-DTOP and the analyte, if appropriate, in 50 µL of 0.01 M phosphate buffer pH 7 containing 0.1% Tween-20 and incubated for 1 h. The capillaries were then rinsed three times with 0.01 M phosphate buffer, pH 7, containing 0.1% Tween-20. The Pierce chemiluminescent substrate (50 µL) was then added, and the light emitted was quantified using a photomultiplier tube. In some experiments, a different chemiluminescent substrate was used, which was prepared as follows:4,13 To 18 mL of 0.1 M phosphate buffer, pH 7, were added 4 mg of luminol dissolved in 2 mL 0.1 M NaOH and 8 µL of 30% H2O2. The PMT (Sensor Module HC 135-01) with embedded microcontrol system for chemiluminescence intensity detection was from Hamamatsu (Bridgewater, NJ). It was mounted
(12) Bru ¨ ggemann, O.; Freitag, R.; Whitcombe, M.; Vulfson, E. J. Chromatogr., A 1997, 781, 43-53.
(13) Surugiu, I.; Ye, L.; Yilmaz, E.; Dzgoev, A.; Danielsson, B.; Mosbach, K.; Haupt, K. Analyst 2000, 125, 13-16.
Figure 1. Schematic representation of the instrumental setups: (a) measurements with CCD camera, (b) discontinuous measurements with a photomultiplier tube, and (c) FIA system with a photomultiplier tube.
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
4389
Figure 2. Scanning electron microscope image of the inner capillary wall at 3000× magnification: (a) uncoated glass capillary, (b) MIPcoated capillary before sonication, and (c) MIP-coated capillary after sonication. (d) Atomic force microscopy image of the polymer-coated capillary wall after sonication of the capillary.
in a light-tight case that was connected to the capillary holder by an optical fiber bundle (Figure 1b). The MIP capillaries were also used in a FIA system in combination with the PMT. In this configuration, the capillary was placed close to the light inlet and perpendicular to the axis of the PMT (Figure 1c). The solutions were transported through the FIA system by a peristaltic pump at a flow rate of 0.44 mL min-1. Samples were introduced into the carrier stream through a sixport injection valve using a 34-µL sample loop. The system was connected to a computer and data were analyzed using the DASYLab software (Data Acquisition System Laboratory, Dasytec, Amherst, MA), which translates the signal from the PMT into counts per second that are plotted versus time. RESULTS AND DISCUSSION In our model system, we have used the same MIP recipe and chemiluminescence detection method as previously described for a MIP-based chemiluminescent ELISA.13 This assay used imprinted polymer microspheres obtained by precipitation polymerization. Since the aim of the present work was to develop a flow injection assay, it seemed to be most appropriate to use as the reaction chamber, a glass capillary with the inner wall coated with imprinted polymer. For covalent coating, the capillary was silanized with (3-methacryloxypropyl)trimethoxysilane. It was then filled with the polymerization solution, and during polymerization, a polymer layer formed at the capillary wall consisting of 4390
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
precipitated MIP microparticles (Figure 2). The microparticles originally formed a thick layer of aggregates, as can be seen from Figure 2b. Loosely adhering polymer was removed by sonication, leaving a surface covered by a thin, uniform layer (Figure 2c and d). Optimization of Chemiluminescence Measurements. The experimental conditions of the chemiluminescence measurements were adapted from a previous publication.4 However, for better sensitivity and to be able to use longer exposure times (this was particularly important for measurements with the CCD camera), we used a commercial chemiluminescent substrate (Pierce) instead of luminol as described previously.4,13 The Pierce substrate yields a constant signal over 30 min, which is more than enough time since one measurement takes only 90 s (integration time). We have used CCD camera measurements to optimize the conditioning of the capillary that is crucial for good capillary-tocapillary reproducibility. First, a titration experiment of the polymer-coated capillary with the enzyme-analyte conjugate (2,4D-TOP) was performed, which showed a clear difference in binding of 2,4-D-TOP to the imprinted and control polymers (Figure 3a). A 2,4-D-TOP concentration of 110 ng mL-1 was found to be optimal for these experiments. When the capillaries were used without sonication as described above, we observed a low reproducibility of the signal obtained with different capillaries that were analyzed under identical conditions. Thus, the sonication
Figure 4. Competitive assay with the PMT in discontinuous mode: filled squares, 2,4-D; filled triangles, 4-chlorophenoxyacetic acid; open circles, phenoxyacetic acid. Measurements were done in triplicate and points represent mean values. A new capillary was used for each measurement.
Figure 3. (a) Saturation curve for capillaries with 2,4-D-TOP. Filled squares represent capillaries coated with MIP and open squares capillaries coated with a nonimprinted control polymer. (b) Calibration curve obtained for 2,4-D with the imprinted capillary. Measurements were done with the CCD camera. A new capillary was used for each measurement.
step is needed after polymerization in order to create a uniform surface (Figure 2c and d). In that way, the preparation of the capillaries was reproducible with a variation in the signal of less than 10% between capillaries. When unlabeled 2,4-D as a competitor was injected together with the 2,4-D-TOP, a calibration curve for 2,4-D could be recorded (Figure 3b). Measurements with the PMT. PMT-based assays were performed using essentially the same experimental conditions as with the CCD camera, except that a lower concentration of conjugate (27.5 ng mL-1 or 1.1 ng/assay) could be used due to the superior sensitivity of the PMT over the CCD camera. Since the emitted light was guided directly into the PMT by an optical fiber bundle, light harvesting was very efficient. This resulted in a detection limit that was several orders of magnitude lower than that achieved with the CCD camera, with a dynamic range for 2,4-D from 5 pg mL-1 to 100 ng mL-1 (Figure 4). When compounds structurally related to 2,4-D were tested in the competitive assay, their cross-reactivities (IC50 of the related compounds relative to that of 2,4-D) were in the same range as the ones previously reported for isotope- and enzyme-linked assays based on the same polymer.4,13 For example, cross-reactivities of 30 and 10% for 4-chlorophenoxyacetic acid and phenoxyacetic acid, respectively, were obtained (Figure 4). Flow Injection Assay. Optimization of the FIA system was done in view of reaching a compromise between reagent economy, reproducibility, sensitivity, and sampling rate. The system was maintained under constant buffer flow. The chemiluminescent
Figure 5. Competitive assay with the PMT in FIA mode: (a) typical readout of the PMT (2,4-D concentrations 0.05, 0.5, 5, 50, 500, 5000, 50000, 500000 ng mL-1); (b) plot of peak height vs 2,4-D concentration.
substrate was injected 1 min after sample injection. A total of 20 s was necessary for the sample to reach the waste, and after 1 min, all of the unbound or loosely adsorbed conjugate seemed to be washed away. In fact, when after sample injection chemiluminescent substrate was added into the flow every 20 s, the peak area decreased from injection to injection during 1 min, whereafter reproducible signals were obtained with a variation of (10% upon three further injections. Another important issue was the regeneration protocol that is needed to allow for a repeated use of the capillary. The completeness of regeneration was determined by Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
4391
doing repeated injections of conjugate with and without injections of regeneration solution and comparing the responses obtained. Three injections of a 0.2 M glycine-HCl buffer, pH 2.5, were found to be optimal for capillary regeneration, resulting in reproducible signals upon repeated measurements. The amount of conjugate was varied from 13.75 to 220 ng mL-1 and the best results were obtained with a 2,4-D-TOP concentration in the sample of ∼20 ng mL-1. Figure 5a shows the signal obtained from the PMT measurements with a constant concentration of 18.33 ng mL-1 2,4-D-TOP and a varying concentration of unlabeled 2,4-D as the competitor. A secondary plot of peak height versus 2,4-D concentration yielded a calibration curve for 2,4-D with a dynamic range from 0.5 ng mL-1 to 50 µg mL-1 (Figure 5b), which is slightly better than with the CCD camera (Figure 3b). To reduce costs, we also tested the more economical, homemade chemiluminescent substrate as described above and found that in combination with the PMT it yielded equally good results as the PIERCE substrate. The higher detection limit of the FIA system as compared to the discontinuous measurements with the PMT can be explained by a less efficient light harvest when the capillary is placed perpendicularly to the PMT (some light is lost by refraction and total reflection). Another possible explanation is the presence of the flow as an additional parameter that results in shorter contact times. In fact, in the flow injection assay, the system is not in equilibrium. Work is under way in our laboratory to place the capillary in the axis of the PMT even in the flow system. CONCLUSIONS We have developed an imprinted polymer-based capillary assay using chemiluminescence and a PMT for detection. The target (14) Gerdes, M.; Spener, F. Meusel, M. Quim. Anal. 2000, 19, 8-14. (15) Dzgoev, A.; Mecklenburg, M.; Xie, B.; Miyabashi, A.; Larsson, P. O.; Danielsson, B. Anal. Chim. Acta 1997, 347, 87-93. (16) Rogers, K. R.; Apostol, A. B.; Brumley, W. C. Anal. Lett. 2000, 33, 443453.
4392
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
analyte is labeled with the enzyme tobacco peroxidase, and in a competitive format, the bound fraction of the conjugate is quantified. The fact that the MIP capillary can be regenerated after each measurement allows for consecutive measurements of large numbers of samples. This could be one of the advantages of using imprinted polymers instead of biomolecules. The assay can be easily transformed into a FIA system and automated for the design of unattended monitoring systems where long-term stability and costs are important factors. Even though in FIA mode the assay shows a higher detection limit than in discontinuous mode, we believe that we have shown that high sensitivities in the picomolar range can be obtained when using a PMT for detection, which is on a par with antibody-based systems.5,14-16 In principle, chargecoupled devices can also be used for detection, although they require cooling to lower the background signal, which might be less convenient for unattended monitoring applications. On the other hand, a CCD-based imaging format would allow us to monitor several capillaries in parallel, and thus to increase sample throughput, or to construct a multisensor for several different analytes. Work is currently ongoing in our laboratory to improve the sensitivity of the FIA system and to extend the system to other target analytes. Tests for long-term stability of the MIP capillaries and compatibility with real samples are also under way. ACKNOWLEDGMENT This work was in part supported by the Swedish Research Council for Engineering Sciences (TFR Grant 230-99-685). I.S. acknowledges financial support by the Swedish International Development Cooperation Agency. J.S. acknowledges financial support by the European Commission (Marie Curie Fellowship QLK1-CT-1999-51347). The authors thank Dr. Igor Boudachov for performing the AFM measurements. Received for review February 8, 2001. Accepted June 11, 2001. AC0101757
