Anal. Chem. 2004, 76, 1799-1803
Immunosensor for Detection of Yersinia enterocolitica Based on Imaging Ellipsometry Young Min Bae, Byung-Keun Oh, Woochang Lee, Won Hong Lee, and Jeong-Woo Choi*
Department of Chemical and Biomolecular Engineering, Sogang University, 1 Sinsu-Dong, Mapo-Gu, Seoul 121-742, Korea
An immunosensor for the detection of pathogens was developed using imaging ellipsometry (IE) as a detection method. Yersinia enterocolitica was selected as the target pathogen in this study. A gold surface deposited with a self-assembled layer of 11-mercaptoundecanoic acid (11-MUA) was used as a substrate. For the fabrication of the immunosensor, protein G spots were made on the substrate using an inkjet-type microarrayer, and monoclonal antibody (Mab) was adsorbed onto the protein G spots. Deposition of each layer onto the substrate was confirmed by the measurement of surface plasmon resonance. The ellipsometric image of the protein G spot and the Mab-adsorbed protein G spot were acquired using an off-null ellipsometry type of imaging ellipsometry system. By measuring the ellipsometric angles of the protein layers, the surface concentration of each protein layer was calculated. The change in the mean optical intensity of the protein spot to the various concentrations of Y. enterocolitica was estimated. The immunosensor using imaging ellipsometry could successfully detect Y. enterocolitica in concentrations varying from 103 to 107 cfu/ mL. The proposed immunosensor system has the advantage of allowing label-free detection, high sensitivity, and operational simplicity. Immunosensors based on antigen-antibody binding have recently been developed for the purpose of detecting analytes, including antigens, small molecules, and cells.1-3 This method can be quite rapid and offers very good detection limits. Techniques for the immobilization of antibodies on a solid surface and the detection of protein-protein binding were developed for the purpose of implementing such immunosensors. For the detection of protein-protein binding, labeling techniques have been widely employed. Typically, either an immunosensor is directly probed with a fluorescent molecule or the procedure is performed in two steps by first using a tagged probe, which can then be detected in a second step using a labeled affinity reagent. Because labeling techniques involving the use of * To whom correspondence should be addressed. Phone: (+82)-2-705-8480. Fax: (+82)-2-711-0439. E-mail:
[email protected]. (1) Darren, M. D.; David, C. C.; Hong-Xing, Y.; Christopher, R. L. Biosens. Bioelectron. 1998, 13, 1213-1225. (2) Sakai, G.; Ogata, K.; Uda, T.; Miura, N.; Yamazoe, N. Sens. Actuators, B 1998, 49, 5-12. (3) Toyama, S.; Shoji, A.; Yoshida, Y.; Yamauchi, S.; Ikariyama, Y. Sens. Actuators, B 1998, 52, 65-71. 10.1021/ac034748m CCC: $27.50 Published on Web 02/17/2004
© 2004 American Chemical Society
fluorescence,4-6 ELISA,7 and isotropic labeling8 employ protocols that require one or two intermediate steps for the binding of some labeled probes, the development of a label-free method would offer considerable advantages, in that it would provide a direct approach to the detection of protein-protein binding. Atomic force microscopy (AFM),9 electrochemical-impedance technique,10 and surface plasmon resonance (SPR)11 have all been cited as being labelfree detection methods. The SPR technique, an optical method based on the attenuation of surface plasmon generated between a metal surface and a dielectric layer, has matured to become a versatile detection tool for the study of the kinetics of receptorligand interactions, and the development of the array-based SPR chip has been reported.12 In addition, the SPR imaging technique, whose configuration is similar to that of SPR, has been applied to simultaneously detect the adsorption of biopolymers,13 peptideantibody binding,14 and protein-protein binding.15 However, because a metal surface, generally gold, must be used as the substrate, the immobilization strategy is confined to gold thiol self-assembly in SPR analysis. On the other hand, because imaging ellipsometry (IE) is based on ellipsometry, the other optical technique that involves measuring the change of the polarization state of an elliptically polarized beam reflected from thin films,16 it is sensitive enough to detect the adsorption of a molecular monolayer on a solid surface, such as a silicon wafer or gold surface. Using IE, ellipsometric images of the patterned Langmuir-Blodgett film on silicon17 and of carbon monoxide oxidation on platinum18 have previously been acquired. Additionally, the application of IE to biosensors has the advantage (4) Chudinova, G. K.; Chudinov, A. V.; Savransky, V. V.; Prokhorov, A. M. Thin Solid Films 1997, 307, 294-297. (5) Lepesheva, G. I.; Azeva, T. N.; Knyukshto, V.N.; Chashchin, V. L.; Usanov, S. A. Sens. Actuators, B 2000, 68, 27-33. (6) Seong, S.-Y. Clin. Diagn. Lab. Immunol. 2002, July, 927-930. (7) Natalya V.; Avseenko, T. Y.; Morozova, F. I.; Ataullakhanov, V. N. Anal. Chem. 2002, 74, 927-933. (8) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760-1763. (9) Raiteri, R.; Grattarola, M.; Butt, H.-J.; Skladal, P. Sens. Actuators, B 2001, 79, 115-126. (10) Lillie, G.; Payne, P.; Vadgama, P. Sens. Actuators, B 2001, 78, 249-256. (11) Mullet, W. M.; Lai, E. P. C.; Yeung, J. M. Methods 2000, 22, 77-91. (12) Myszka, D. G.; Rich, R. L. Pham. Sci. Technol. Today 2000, 3, 310-317. (13) Jordan, C. E.; Corn, R. M. Anal. Chem. 1997, 69, 1449-1456. (14) Wegner, G. J.; Lee, H. J.; Corn, R. M. Anal. Chem. 2002, 74, 5161-5168. (15) O’Brien, M. J.; Perez-Luna, V. H.; Brueck, S. R. J.; Lopez, G. P. Biosens. Bioelectron. 2001, 16, 97-98. (16) Azzam, R. M. A.; Bashara, N. M. Ellipsometry and Polarized Light, 1st ed.; North-Holland Publishing Company: Amsterdam, 1977; Chapter 3. (17) Neumaier, K. R.; Elender, G.; Sackmann, E.; Merkel, R. Europhys. Lett. 2000, 49, 14-19. (18) Rotermund, H. H.; Haas, G.; Franz, R. U.; Tromp, R. M.; Ertl, M. G. Science 1995, 270, 608-610.
Analytical Chemistry, Vol. 76, No. 6, March 15, 2004 1799
of offering label-free detection, high sensitivity, and operational simplicity. Jin et al. reported that the IE technique based on offnull ellipsometry could be applied to be detection of proteinprotein binding on a solid surface.19 Moreover, the development of a biochip has been attempted using the IE based on phasemodulated ellipsometry, which is an alternative technique used for measuring the polarization of light.20 However, the biosensor system using IE has not yet been applied to the detection of microorganisms, such as pathogens. Yersinia enterocolitica is one of the human pathogenic species of Yersinia genus, and has emerged as a significant foodborne pathogen.21 Human infections due to Y. enterocolitica, which causes a form of pediatric enterocolitis known as yersiniosis that is characterized by fever, diarrhea, and abdominal pain, have increased dramatically over the past two decades. A number of the techniques based on the polymerase chain reaction (PCR) assay have been developed for the detection of Y. enterocolitica.22,23 Although the PCR-based method has the potential to detect the pathogen with high sensitivity, this method requires the use of expensive equipment and the associated specialist skills needed to perform the analyses. Therefore, it would be extremely beneficial if an alternative method of detecting Y. enterocolitica which offered high sensitivity with a short detection time and greater simplicity could be developed. In this study, an immunosensor system for the detection of Y. enterocolitica using IE as detection method is reported. A protein G layer adsorbed onto a self-assembled 11-mercaptoundecanoic acid layer was adopted for immobilizing the antibody. The deposition of each layer was investigated using the SPR technique. The protein spots were fabricated using a microarrayer, and ellipsometric images of protein spots were acquired. Various concentrations of Y. enterocolitica were measured using IE. MATERIALS AND METHODS Materials. Protein G (MW 22 600 Da), which is a recombinant protein G capable of binding the Fc region of the antibody, was purchased from Prozyme Inc, (San Leandro, CA). Monoclonal antibody against Y. enterocolitica (Mab) was obtained from Fitzgerald Industries International, Inc. (Concord, MA). The other chemicals used in this study were reagent grade and were obtained commercially. Immobilization of Antibody. A substrate was prepared by DC magnetron sputtering of gold on a P-type Si wafer. Before the gold sputtering operation was performed, chromium was sputtered onto the wafer to promote the adhesion of the gold. The gold and chromium films had thicknesses of 150 nm and 5 nm, respectively. Before the fabrication of the immunosensor, the gold surface was cleaned using piranha solution (30 vol % H2O2 and 70 vol % H2SO4) at 60 °C for 5 min, and then rinsed with ethanol and deionized water. A monolayer of 11-mercaptoundecanoic acid (11-MUA) was deposited on the gold surface by submerging the substrate in a (19) Jin, G.; Tengvall, P.; Lundstrom, I.; Arwin, H. Anal. Biochem. 1995, 232, 69-72. (20) Noort, D.; Rumberg, J.; Jager, E. W. H.; Mandenius, C.-F. Meas. Sci. Technol. 2000, 11, 801-808. (21) Durisin, M. D.; Ibrahim, A.; Griffiths, M. W. Int. J. Food Microbiol. 1997, 37, 103-112. (22) Bhaduri, S.; Cottrell, B. Mol. Cell. Probes 1998, 12, 79-83. (23) Kong, R. Y. C.; Lee, S. K. Y.; Law, T. W. F.; Law, S. H. W.; Wu, R. S. S. Water Res. 2002, 36, 2802-2812.
1800 Analytical Chemistry, Vol. 76, No. 6, March 15, 2004
Figure 1. Imaging ellipsometry system based on null-type ellipsometry. 1, He-Ne laser beam; 2, polarizer; 3, compensator; 4, objective lens; 5, analyzer; 6, CCD camera.
glycerol/ethanol (1:1, v/v) solution containing 150 mM 11-MUA for at least 12 h.24 To allow the chemical binding between the 11-MUA layer adsorbed on the gold and the free amine of protein G, the carboxyl group of 11-MUA was activated by submerging the substrate modified with 11-MUA into a solution of 10% 1-ethyl3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDAC) in water/ethanol (10/1, v/v) for 2 h at room temperature. The protein G solution was spotted onto the activated surface using an inkjet-type microarrayer (Nano-Plotter model 1.2, GeSiM mbH, Groβerkmannsdorf, Germany). The spotted amount per spot was 0.4 nL of a solution of 0.1 mg/mL protein G in a mixed solution of 10 mM phosphate buffer saline (PBS, pH 7.4) buffer and 10 vol % glycerol. The spotted substrate was incubated in a humid chamber at 4 °C for at least 24 h, taking into consideration the diffusivity delay of the protein molecules, which are hindered from entering into the surface due to the viscosity of glycerol. After the incubation period, the chip was washed with PBS buffer for 20-30 min. Before the immobilization of the Mab, the residue carboxyl groups of 11-MUA on the chip were inactivated by blocking them with 3 wt % bovine serum albumin (BSA). A solution containing the Mab (10 µg/mL Mab against Y. enterocolitica) in PBS buffer was applied to the blocked chip. After being incubated at 4 °C for 3 h, the substrate was washed with PBS buffer containing 0.1% Tween 20. Imaging Ellipsometry. The imaging ellipsometry configuration, based on off-null ellipsometry, which has a component sequence of polarizer-compensator-sample-analyzer (PCSA) as shown in Figure 1 (MultiskopTM, Optrel Gbr, Kleinmachnow, Germany), was used. For the acquisition of ellipsometric images, an objective lens (×10) was inserted between the sample and the analyzer. After the reflected beam passed through the objective and the analyzer, the intensity profile of the cross section of it was recorded in the form of an image with a resolution of 640 × 480 pixels by means of a CCD camera. The light source was a He-Ne laser beam (632.8 nm). The incident angle of the laser beam was set to 40°. The mean optical intensity (MOI) values of the ellipsometric images of the protein spots were calculated using (24) Patel, N.; Davies, M. C.; Hartshorne M.; Heaton, R. J.; Roberts, C. J.; Tendler, J. B.; Williams P. Langmuir 1997, 13, 6485-6490.
the image processing software (Image-Pro, verison 4.7, Media Cybernetics, Silver Spring, MD). The measurement of the ellipsometric angles (∆, Ψ), based on traditional null-ellipsometry, was performed using the PCSA null-ellipsometry system with a photodiode sensor as the optical detector. In addition, the measurement of the angle of the surface plasmon resonance (SPR) was performed using the SPR system based on Kretschmann configuration.25 RESULTS AND DISCUSSION Immobilization of Mab. A gold substrate modified with a selfassembled monolayer of 11-MUA, followed by the chemical binding of protein G to it, was used for immobilizing Mab. The deposition of each layer on the gold substrate was investigated using SPR. The BK 7 glass (18 × 18 mm, Superior, LaudaKoenigshofen, Germany), on which chromium (5 nm thick) and gold films (43 nm thick) were sequentially deposited via DC magnetron sputtering, was used as a substrate for the SPR measurements. The SPR angle for a bare gold surface is 43.02 ( 0.07°. Following the deposition of the 11-MUA layer onto the gold surface via self-assembly, the SPR angle was shifted to 43.24 ( 0.06°. It is known that the strong and the closely packed 2D molecular layer of 11-MUA is formed on the substrate as a result of the van der Waals attractive force among the long alkyl chains.26 The deposition of a thiol monolayer on the gold surface via selfassembly can modify the surface properties. Here, the 11-MUA layer, with a carboxyl group as the terminal group, was deposited in order to create a biocompatible surface for the chemical binding of protein G. The gold substrate modified with the 11-MUA layer was treated with EDAC to allow the chemical binding between the lysine residues of protein G and the carboxyl group via amide bonding. After the protein G solution was applied to the modified surface, the SPR angle of the substrate was shifted to 43.46 ( 0.06°. It is known that since protein G has two domains that can bind the Fc region of antibody, an antibody layer is immobilized on the protein G layer with a well-orientated configuration.27 When such a well-oriented configuration of the antibody layer is obtained, the sensitivity of the immunosensor can usually be improved, because the Fab regions, to which the specific antigen can be bound, face away from the surface.28 Last, Mab was applied to the surface containing the protein G layer. After being incubated for 2 h at room temperature, the substrate was washed with PBS buffer and water. Due to the immobilization of the Mab layer, the SPR angle was shifted to 43.71 ( 0.08°. Such a shift in the SPR angle implied that the Mab layer was immobilized onto the substrate. Ellipsometric Images of the Protein Spots. On the basis of the strategy for the immobilization of antibody, the immunosensor was fabricated using the inkjet-type microarrayer. Protein G solution was spotted onto the substrate modified with the 11MUA layer. To confirm the configuration of the protein G spots, FITC-labeled protein G solution was spotted. The fluorescence image of the FITC-labeled protein G spots and the ellipsometric image of the protein G spots blocked with BSA are shown in (25) Kretschmann, E. Z. Phys. 1971, 241, 313-324. (26) Ulman, A. Chem. Rev. 1996, 96, 1533-1554. (27) Oh, B.-K.; Kim, Y. K.; Bae, Y. M.; Lee, W. H.; Choi, J.-W. J. Microbiol. Biotechnol. 2002, 12, 780-786. (28) Schramm, W.; Paek, S.-H.; Voss, G. ImmunoMethods 1993, 3, 93-103.
Figure 2(A). To acquire the ellipsometric images, the azimuth of the polarizer and analyzer were set to 126.0 ( 0.3° and 136.0 ( 0.3°, respectively. At the azimuth points of the polarizer and analyzer, the BSA layer region of the ellipsometric images was in the null condition, in which the optical intensity was at its lowest. From the fluorescence image, it was observed that the protein patterns were roughly circular, their diameter was ∼130 µm, and the distance between the spots was 300 µm. In the ellipsometric image, the protein patterns appeared to have more of an elliptical shape, rather than circular. Such a distortion of the protein patterns was due to the oblique angle of incidence.29 On the basis of the principle of off-null ellipsometry, the optical intensity of the light beam reflected from a film onto a solid surface is dependent on the thickness of the film, according to the Fresnel equation.30 Therefore, the difference in optical intensity of the magnified cross section of the beam reflected from a thin film implies the local difference in the thin film thickness in the irradiated region. Since the diameter of the cross section of the laser beam used in this study was ∼600 µm, four protein spots could be simultaneously captured in the ellipsometric image using an objective lens with a magnification of 10×. However, a blurred region was observed in the low margin of the ellipsometric image of Figure 2A. The reason for this is that the region is out of focus due to the limited depth of field. This problem can be compensated for by tilting the CCD camera by an incident angle with respect to the reflected beam.29 Detection of Y. enterocolitica. Following the sequence of steps that involved applying the Mab solution to the chip with protein G spots blocked with BSA, incubation, and washing with PBS buffer, the immunosensor was completed. The ellipsometric image of the Mab spot immobilized on the protein G spot is shown in the part a of Figure 2B. It is easy to see that the optical intensity of the Mab spot is darker than that of protein G spot. The MOIs were 243.8 ( 0.2 (a.u.) and 177.0 ( 4.3 (a.u.) for the protein G spot and the Mab spot, respectively. Since the azimuths of the polarizer and analyzer were set for the null condition of the BSA layer, the difference of optical intensities in the ellipsometric images corresponds to the difference between the thicknesses of the protein spots and those of the BSA layer. To estimate the amount of protein adsorbed onto the surface, the ellipsometric angles (∆, Ψ) of each layer, a BSA layer, a protein G layer, and the Mab layer on the protein G layer were measured with the traditional null-ellipsometry system. Figure 3 shows the change in the ∆ and Ψ of each layer compared with that of the 11-MUA layer. In the thin film model, the adsorption of a protein onto a solid surface results in the change of ∆, and the change of Ψ is dependent on the nature of the substrate.31 The surface concentration of protein, Γ(ng/mm2) was calculated using the following equation32
Γ ) K × df where K ()1.36 g/mL) is the density of the proteins, and df (nm) (29) Harke, M.; Teppner, R.; Shulz, O.; Motschmann, H.; Orendi, H. Rev. Sci. Instrum. 1997, 68, 3130-3134. (30) Arwin, H.; Welin-Klintstrom, S.; Jansson, R. J. Colloid Interface Sci. 1993, 156, 377-382. (31) Tronin, A.; Dubrovsky, T.; Nicolini, C. Langmuir 1995, 11, 385-389. (32) Elwing, H. Biomaterials 1998, 19, 397-406.
Analytical Chemistry, Vol. 76, No. 6, March 15, 2004
1801
Figure 2. Ellipsometric images of protein spots. (A) Fluorescence image (a) and ellipsometric image (b) of protein G spots injected by means of an inkjet-type microarrayer. The distance between the spots is 300 µm, and the diameter of the spots is ∼130 µm. (B) Ellipsometric images of the binding of Y. enterocolitica with various concentrations [0 cfu/mL (a), 105 cfu/mL (b), and 107 cfu/mL (c)] to the Mab spot.
Figure 3. Changes in ellipsometric angles ∆ and Ψ of each layer (BSA layer, protein G layer, and the Mab layer on the protein G layer) compared with those of the 11-MUA layer deposited onto the gold surface. The angles were measured with a traditional null-type ellipsometry system having PCSA configuration.
is the thickness of the film. With the three-phase optical model (ambient-film-gold), the df values of the protein layers were calculated by means of the software designed for ellipsometric analysis obtained from Optrel Gbr. The refractive index of the gold layer and that of the film in the optical model were assumed to be 0.183 + j3.090,33 and 1.50 + j0,32 respectively. The Γ values were 3.77, 1.89, and 2.39 ng/mm2 for the BSA layer, the protein (33) Pelik, E. D. Handbook of Optical Constants of Solids I; Academic Press: Orlando, 1985; p 286.
1802
Analytical Chemistry, Vol. 76, No. 6, March 15, 2004
G layer, and the Mab layer immobilized on the protein G layer, respectively. This indicates that the BSA layer was denser than the protein G spot and that the difference in the surface concentration between the BSA layer and the Mab spot was reduced as result of the immobilization of the Mab onto the protein G spot. Pathogen solutions with various concentrations were applied to the completed immunosensor chips. After incubation and being washed with PBS buffer, they were dried at room temperature.
Typical ellipsometric images of the protein spots, on which the binding of the pathogen to the Mab spot was completed, are shown in the parts b and c of Figure 2B. It was observed that as the pathogen concentration increased, the white regions in the protein spot increased. Y. enterocolitica is rod-shaped, and its general size is ∼0.5 to 5 µm.34 Thus, since the region where the pathogen was bound to the substrate became thicker than the BSA layer, it was considered that the optical intensity of this region increased. In part b of Figure 2B, which shows the ellipsometric image of the region where the 105 cfu/mL solution of the pathogen was applied, the diameter of the white regions within the protein spot was estimated to be ∼7 µm, which is larger than the size of bacteria. This might be due to the aggregation of the pathogen. The MOI of each protein spot was calculated. The changes in the MOIs against the concentration of the pathogen are shown in Figure 4. The change in the MOI appeared beginning with the protein spot with a binding of 103 cfu/mL pathogen, and the MOI increased up to a binding of 107 cfu/mL in a manner that was proportional to the logarithm of the concentration of the pathogen. The MOI of the protein spot with a binding of 107 cfu/mL was 232.7 ( 2.1 (a.u.), which meant that the image was almost saturated, because the output signal from an eight-bit CCD camera ranges from 0 to 255. However, a large standard deviation was observed. The reason for this was that residue salts from the PBS buffer affected the ellipsometric image. In particular, at low concentration, the signal noise due to the residue salts might result in errors in the immunosensor based on the IE. To solve this problem, it will be necessary to operate the immunosensor in an in situ flow system. Recently, the development of immunosensors capable of detecting pathogens based on several detection methods has been reported. In these studies, the detection limit (34) Joklik, W. K. Zinsser Microbiology; Prentice Hall: London, 1985; Chapter 38. (35) Ruan, C.; Yang, L.; Li, Y. Anal. Chem. 2002, 74, 4814-4820. (36) Gehring, A. G.; Patterson, D. L.; Tu, S. Anal. Biochem. 1998, 258, 293298.
Figure 4. Changes in the mean optical intensity of the protein spots as a function of the concentration of Y. enterocolitica.
was reported to be 6 × 102 cfu/mL for Escherichia coli 0157 with impedance spectroscopy,35 105 cfu/mL for Legionella pneumophila with surface plasmon resonance,27 and 2.5 × 104 cfu/mL for E. coli 0157 with a light-addressable potentiometric sensor.36 In view of these results, it was considered that the sensitivity of detection of the pathogen using the immunosensor based on IE was comparable to those of the other methods. ACKNOWLEDGMENT This work was supported in part by the Korea Science and Engineering Foundation (KOSEF) through the Advanced Environmental Monitoring Research Center at Kwangju Institute of Science and Technology. Received for review July 7, 2003. Accepted January 13, 2004. AC034748M
Analytical Chemistry, Vol. 76, No. 6, March 15, 2004
1803
