Anal. Chem. 2006, 78, 5525-5531
On-Chip Enzyme Immunoassay of a Cardiac Marker Using a Microfluidic Device Combined with a Portable Surface Plasmon Resonance System Ryoji Kurita, Yoshimi Yokota, Yukari Sato, Fumio Mizutani,† and Osamu Niwa*
National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba Central 6, 1-1-1 Higashi, Tsukuba, Ibaraki, Japan
This paper reports a miniaturized immunosensor designed to determine a trace level cardiac marker, B-type natriuretic peptide (BNP), using a microfluidic device combined with a portable surface plasmon resonance (SPR) sensor system. Sample BNP solution was introduced into the microchannel after an immunoreaction with acetylcholine esterase-labeled antibody (conjugate), and only unbound conjugate was trapped on the BNPimmobilized surface in the flow channel. Then, the thiol compound generated by the enzymatic reaction with the trapped conjugate was accumulated on a gold thin film located downstream in the microchannel to monitor the real-time SPR angle shift. We achieved a detectable concentration range of 5 pg/mL-100 ng/mL by monitoring the SPR angle shift, which covers the required detection range for the BNP concentrations found in blood. This success resulted from the use of a T-shaped microfluidic device structure, which prevents the sample solution from flowing over the gold film used for SPR detection. We were able to measure trace levels of BNP peptide (15 fg) within 30 min since the procedure with our immunosensor is simpler than a multistep immunoassay through the simultaneous use of a labeled enzymatic reaction and the real-time monitoring of enzymatic product accumulation in the microfluidic device. We employed the procedure to detect serum BNP by using spiked samples in human serum and achieved satisfactory recovery for heat-treated samples to denature the esterase in the serum before the immunoreaction. B-type natriuretic peptide (BNP) was discovered in 1988,1 and its reliable role in the diagnosis of cardiac failure has been studied since 1990.2 Today, BNP is one of the most important cardiac markers for the prediction and diagnosis of heart failure. The blood BNP concentration under normal conditions is ∼20 pg/ mL (6 pM), and it increases to ∼2 ng/mL (600 pM) for patients diagnosed with severe congestive heart failure.3 Generally, radio* Corresponding author. Tel: 029-861-6158. Fax: 029-861-6177. E-mail:
[email protected]. † Present address: University of Hyogo, 3-2-1 Kouto, Kamigori-cho, Ako-gun, Hyogo 678-1297, Japan. (1) Sudoh, T.; Kangawa, K.; Minamino, M.; Matsuo, H. Nature 1988, 332, 7881. (2) Mukoyama, M.; Nakao, K.; Saito, Y.; Ogawa, Y.; Hosoda, K.; Suga, S.; Shirakami, G..; Jougasaki, M.; Imura, H. N. Engl. J. Med, 1990, 323, 757758. 10.1021/ac060480y CCC: $33.50 Published on Web 06/28/2006
© 2006 American Chemical Society
immunoassay or immunochromatography combined with a fluorescence detection system has been used to measure blood BNP concentrations.4,5 This is because the conventional simple immunochromatography technique could not be used to detect BNP owing to its extremely low concentration in blood. Thus, BNP analysis has mainly been undertaken in laboratories located far from the patient because these detection techniques need huge and expensive equipment such an excitation light source or photomultiplier. Recently, one of the authors reported a highly sensitive and low-cost immunoassay method for BNP measurement based on the electrochemical desorption of preconcentrated thiols on an electrode surface.6 However, an assay took ∼2 h to complete because of the long accumulation and immunoreaction time. Furthermore the method is complicated since the electrode potential has to be scanned in a deaerated alkaline solution. In addition, the detection limit (20-40 pg/mL) was still insufficient to detect the basal level of blood BNP. Surface plasmon resonance (SPR) is a surface-sensitive optical technique that is used to study a thin layer on a metal surface. Many researchers have reported immunosensors based on the SPR angle shifts caused by the formation of an immunocomplex on an antibody- or antigen-modified sensing surface. The SPRbased immunosensor has advantages as regards bedside monitoring and point of care because the method is a rapid, simple, safe, and low-power technique that does not use any isotopes or fluorescence labels.7-10 However, the sensitivity for small molecules is insufficient compared with conventional radioimmunoassay or fluorescence-based detection. For example, the detection limit of the SPR immunoassay is reported to be around several (3) Dao, Q.; Krishnaswamy, P.; Kananegra, R.; Harrison, A.; Amirnovin, R.; Lenert, L.; Clopton, P.; Alberto, J.; Hlavin, P.; Maisel, A. S. J. Am. Coll. Cardiol. 2001, 37, 156-166. (4) Knudsen, C. W.; Omland, T.; Clopton, P.; Westheim, A.; Abraham, W. T.; Storrow, A. B.; McCord, J.; Nowak, R. M.; Aumont, M. C.; Duc, P.; Hollander, J. E.; Wu, A. H. B.; McCullough, P. A.; Maisel, A. S. Am. J. Med. 2004, 116, 363-368. (5) Chen, H. H.; Lainchbury, J. G.; Harty, G. J.; Burnett, J. C. Circulation 2002, 105, 999-1003. (6) Matsuura, H.; Sato, Y.; Niwa, O.; Mizutani, F. Anal. Chem. 2005, 77, 42354240. (7) Kanda, V.; Kariuki, J. K.; Harrison, D. J.; McDermott, M. T. Anal. Chem. 2004, 76, 7257-7262. (8) Fitzpatrick, B.; O’Kennedy, R. J. Immunol. Methods, 2004, 291, 11-25. (9) Akimoto, T.; Ikebukuro, K.; Karube, I. Biosens. Bioelectron. 2003, 18, 14471453. (10) Choi, J. W.; Park, K. W.; Lee, D. B.; Lee, W.; Lee, W. H. Biosens. Bioelectron. 2005, 20, 2300-2305.
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006 5525
tens of nanomolars for small peptides.11,12 Thus, it is impossible to use a conventional SPR immunosensor to detect blood BNP. With a view to overcoming the sensitivity problem, several researchers have reported highly sensitive SPR immunosensors realized by introducing signal amplification using a sandwich immunoassay technique with an antigen or antibody labeled with latex or gold particles.13 Severs and Schasfoort first reported the signal amplification of the SPR angle shift by using a sandwich immunoassay,14 and Lyon et al. achieved the picomolar detection of immunoglobulin using a secondary antibody labeled with gold particles.15 A large SPR angle shift can be obtained even if the analytes are small. However, it is still difficult to measure trace level biomolecules in real samples using the above methods because of the nonspecific adsorption of biomolecules, especially protein fouling, which occurs when the sample solution is directly introduced onto the sensing surface. Since an SPR sensor simply responds to refractive index changes on the sensing surface, it is impossible to distinguish whether the SPR angle shift is caused by specific or nonspecific adsorption. Various modifications of the film on the sensing surface have been used to suppress nonspecific adsorption.16-19 Nevertheless, the complete prevention of nonspecific adsorption is difficult at present. Thus, when monitoring trace level biomolecules, it is inappropriate to introduce the blood sample directly onto the sensing surface during SPR measurement. Recently, many researchers have reported integrated analysis systems called lab-on-a-chip or micro total analysis systems that are small, light, and capable of integrating all sample-handling steps in the microflow channels. These techniques may make it possible to undertake clinical measurements simply and rapidly at the bedside. Some researchers have reported on-chip immunoassay methods that use microchannels combined with fluorescence,20-22 chemiluminescence,23,24 electrochemical,25,26 and thermal lens microscope27 detection techniques. The most commonly (11) Gomes, P.; Andreu, D. J. Immunol. Methods 2002, 259, 217-230. (12) Kuriyama, M.; Wang, M. C.; Papsidero, L. D.; Killian, C. S.; Shimano, T.; Valenzuela, L.; Nishiura, T.; Murphy, G. P.; Chu, T. M. Cancer Res. 1980, 40, 4568-4662. (13) Gobi, K. V.; Miura, N. Sens. Actuators, B 2004, 103, 265-271. (14) Severs, A. H.; Schasfoort, R. B. M. Biosens. Bioelectron. 1993, 8, 365-370. (15) Lyon, L. A.; Musick, M. D.; Natan, M. J. Anal. Chem. 1998, 70, 51775183. (16) Barrett, D. A.; Hartshorne, M. S.; Hussain, M. A.; Shaw, P. N.; Davies, M. C. Anal. Chem. 2001, 73, 5232-5239. (17) Uchida, K.; Otsuka, H.; Kaneko, M.; Kataoka, K.; Nagasaki, Y. Anal. Chem. 2005, 77, 1075-1080. (18) Kurita, R.; Tabei, H.; Iwasaki, Y.; Hayashi, K.; Sunagawa, K.; Niwa, O.; Biosens. Bioelectron. 2004, 20, 518-523. (19) Kurita, R.; Hirata, Y.; Yabuki, S.; Kato, D.; Sato, Y.; Mizutani, F.; Niwa, O. Electrochemistry 2006, 74, 121-124. (20) Cheng, S. B.; Skinner, C. D.; Taylor, J.; Attiya, S.; Lee, W. E.; Picelli, G.; Harrison, D. J. Anal. Chem. 2001, 73, 1472-1479. (21) Endo, T.; Okuyama, A.; Matsubara, Y.; Nishi, K., Kobayashi, M.; Yamamura, S.; Morita, Y.; Takamura, Y.; Mizukami, H.; Tamiya, E. Anal. Chim. Acta 2005, 531, 7-13. (22) Cesaro-Tadic, S.; Dernick, G.; Juncker, D.; Buurman, G.; Kropshofer, H.; Michel, B.; Fattinger, C.; Delamarche, E. Lab Chip 2004, 4, 563-569. (23) Yakovleva, J.; Davidsson, R.; Lobanova, A.; Bengtsson, M.; Eremin, S.; Laurell, T.; Emne´us, J. Anal. Chem. 2002, 74, 2994-3004. (24) Tsukagoshi, K.; Jinno, N.; Nakajima, R. Anal. Chem. 2005, 77, 1684-1688. (25) Farrell, S.; Ronkainen-Matsuno, N. J.; Halsall, H. B.; Heineman, W. R. Anal. Bioanal. Chem. 2004, 379, 358-367. (26) Wang, J.; Ibanez, A.; Chatrathi, M. P.; Escarpa, A. Anal. Chem. 2001, 73, 5323-5327. (27) Sato, K.; Tokeshi, M.; Kimura, H.; Kitamori, T. Anal. Chem. 2001, 73, 12131218.
5526
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
used detection technique employs fluorescence because it has a lower detection limit (around the picomolar level) than a conventional SPR immunosensor. Unfortunately, this is not suitable for bedside monitoring since the system needs relatively high power and an expensive excitation light source and photomultiplier. In contrast, electrochemical detection is advantageous because it is inexpensive and easy to miniaturize while maintaining relatively high sensitivity.28,29 However, the sensitivity and magnitude of the signal depends greatly on the flow rate and potential at a working electrode, and it is difficult to control these parameters in a small microchannel. In addition, the reduction in the response current that accompanies miniaturization requires improved electronics and shielding for the system. In this paper, we propose a new one-chip immunosensor with a small sample volume using a T-shaped microfluidic system that has a B/F separation chamber by immunoreaction and a gold film for the detection of concentrated thiols with a portable SPR system. It is well known that thiols form a self-assembled monolayer on a metal surface, and this has been widely used to modify metal surfaces.30,31 We employed this characteristic for a highly sensitive immunoassay by obtaining the surface preconcentration of thiol molecules formed by the enzymatic reaction of a labeled antibody. A highly sensitive and rapid immunosensor was developed by the real-time monitoring of the SPR angle shift caused by the chemisorption of thiol molecules on the gold surface in the microchannel. EXPERIMENTAL SECTION Reagents. B-Type natriuretic peptide (BNP-32, human) and anti B-type natriuretic peptide antibody (rabbit anti BNP-32) (antiBNP) were purchased from Phoenix Pharmaceuticals (Melmont, CA). Acetylthiocholine, acetylcholine esterase (AChE) (from electric eel, 217 units/mg), and human serum were purchased from Sigma (St. Louis, MO). The 0.1 M phosphate buffer (pH 7.0) was purchased from Nakalai (Kyoto, Japan). N-Hydroxysuccinimide (NHS), N-ethyl-N′-(3-dimethyl aminopropyl) carbodiimide hydrochloride (EDC), sulfosuccinimydyl 4-(N-maleimidiomethyl) cyclohexane-1-carboxylate (sulfo-SMCC), and s-acetylmercaptosuccinic anhydride were purchased from Pierce (Rockford, IL). All solutions were prepared with Millipore water (Millipore Co.). Acetylcholine esterase-labeled anti-BNP (anti-BNP-AChE) was synthesized as follows. The antibody, modified with maleimide moieties, was prepared by adding anti-BNP (0.4 mg/mL) and sulfo-SMCC (0.4 mg/mL) successively to 0.1 M phosphate buffer (pH 7.0) and incubating the solution for 1 h. The incubated solution was filtered to remove excess sulfo-SMCC. Similarly, AChE modified with the thiol group was obtained by reacting the AChE (1 mg/mL) with s-acetylmercaptosuccinic anhydride (0.3 mg/mL) in 0.1 M phosphate buffer for 10 min. The solution was filtered to remove the unreacted s-acetylmercaptosuccinic anhydride. Then the maleimide group-modified anti-BNP and thiol group-modified AChE were mixed with a molar ratio of 1:0.7, and (28) Kurita, R.; Hayashi, K.; Horiuchi, T.; Niwa, O.; Maeyama, K.; Tanizawa, K. Lab Chip 2002, 2, 34-38. (29) Kurita, R.; Hayashi, K.; Fan, X, Yamamoto, K.; Kato, T,; Niwa, O. Sens. Actuators, B 2002, 87, 296-303. (30) Ulman A. Chem. Rev. 1996, 96, 1533-1554. (31) Laibinis, P. E.; Whitesides, G. M.; Allara, D. L.; Tao, Y. T.; Parikh, A. N.; Nuzzo, R. G. J. Am. Chem. Soc. 1991, 113, 7152-7167.
Figure 1. (a) Photograph of a portable immunosensor system with SPR equipment. (b) Schematic representation of a PDMS-based microfluidic immunosensor. The dimensions of the immunosensor and microchannel are shown in the figure.
the mixture was incubated for 1 h with stirring. All syntheses were undertaken at room temperature. Preparation of a Microfluidic Device for BNP Measurement. Figure 1a is a photograph of our immunosensor system with portable SPR equipment (NTT Advanced Technology, Tokyo, Japan). The equipment was developed in collaboration with one of the present authors. The portable SPR system is 16 cm width, 6 cm deep, and 9.5 cm high, and weighs ∼770 g. Figure 1b is a schematic representation of the immunosensor for BNP determination indicated by the circle in Figure 1a. The immunosensor consists of two patterned gold thin films (A and B) formed on a glass plate (n ) 1.525 ( 0.0015, Matsunami Glass Industry, Osaka, Japan) and a poly(dimethylsiloxane) (PDMS) cover with a Tshaped microflow channel. The T-shaped channel consists of microchannel A (1 mm wide, 20 µm deep) and microchannel B (2 mm wide, 20 µm deep) as shown in Figure 1b. The fabrication process is as follows. First, we sealed the glass plate using an adhesive sheet with a 2-mm-diameter round hole and a 1 × 5 mm square hole. We then deposited thin titanium on the glass plates with the adhesive sheet by using rf sputtering equipment (Seed Lab., Tokyo, Japan) and then formed gold film without breaking the vacuum.32,33 The total thickness of the gold and titanium film was 50 ( 2 nm. We placed a 100 µM cystamine solution on gold film A, and a cystamine monolayer was allowed to form on the plate for 2 h. We then modified gold film A with BNP as follows by using the well-known carbodiimide coupling reaction provided by an EDC/NHS system.34,35 We immersed gold film A with the cystamine monolayer in 10 mM phosphate buffer that contained 0.4 mg/mL EDC and 1.1 mg/mL NHS. Then BNP was added to the buffer to give a final concentration of 40 µg/ mL, and the solution was incubated for 2 h at room temperature. (32) Niwa, O.; Kurita, R.; Liu, Z.; Horiuchi, T.; Torimitsu, K. Anal. Chem. 2000, 72, 949-955. (33) Kurita, R.; Tabei, H.; Liu, Z.; Horiuchi, T.; Niwa, O. Sens. Actuators, B 2000, 71, 82-89. (34) Lewis, M. R.; Raubitschek, A.; Shively, J. E. Bioconjugate Chem. 1994, 5, 565-576. (35) La¨nge, K.; Bender, F.; Voigt, A.; Gao, H.; Rapp, M. Anal. Chem. 2003, 75, 5561-5566.
The PDMS cover was formed with previously reported procedures.36 The master pattern was fabricated on a glass wafer by a conventional photolithographic technique with a positive photoresist (PMER P-LA900PM, Tokyo Ohka Kogyo, Japan). Liquid PDMS oligomer (Dow Corning Asia) and a cross-linking agent were mixed and poured onto the master with the flow channel pattern. The PDMS layer was cured for 60 min at 60 °C and then peeled from the master. The flow channel dimensions are shown in Figure 1b. The glass plate with the thin gold films was disposable; however, it was possible to use the PDMS cover with the microchannel repeatedly by rinsing it in ethanol and distilled water after use. BNP Measurement Using the Microfluidic Device with SPR Detection. Before undertaking the measurements, we peeled the adhesive sheet from the glass plate, and then we attached the plate to the PDMS cover with the T-shaped microchannel. We connected three tubes, I′, II′, and III′, to ports I, II, and III, respectively, on the PDMS cover. We also connected the other ends of tubes I′ and II′ to syringes, and each syringe was installed in a syringe pump (CMA102, Stockholm, Sweden). First, the sample BNP and 190 or 2.4 ng/mL anti-BNP-AChE conjugate solution was mixed in a microvial (Simport Corp.) and stirred for a few seconds with a micropipet, and then the mixture solution was immediately introduced from port III to port I through microchannel A at a flow rate of 0.3 µL/min for 10 min by the syringe pump, which was connected to port I, and the other syringe pump was idle. During this process, an immunoreaction takes place between BNP and anti-BNP-AChE while they are flowing through the inlet tube, and then the unreacted anti-BNPAChE is collected on the surface of BNP modified film A as shown in Figure 2b. We rinsed microchannel A by introducing a phosphate buffer solution from port III to port I for 5 min, and then we introduced buffer solution containing 1 mM acetylthiocholine, which is the AChE substrate, through microchannel B from port III to port II by using a syringe pump, which we connected to port II after stopping the other syringe pump. Thiocholine moieties are produced from the acetylthiocholine by (36) Duffy, D. C.; McDonald, J. C.; Schueller, O. J. A.; Whitesides, G. M. Anal. Chem. 1998, 70, 4974-4984.
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
5527
Figure 2. Schematic diagram of the procedure for BNP determination in the immunosensor. (a) Preparation of the immunosensor. Film A is modified with BNP for the collection of unreacted anti-BNP-AChE. Film B is a bare gold film for the detection of concentrated thiocholine. (b) Introduction of a mixture solution containing BNP and anti-BNP-AChE from port III to port I. The unreacted anti-BNP-AChE was collected on BNP-modified film A during the first flow. Then the channel was rinsed with phosphate buffer during the second flow. (c) Acetylthiocholine was introduced from port III to port II. Thiocholine was produced from acetylthiocholine by the collected anti-BNP-AChE on film A, then the thiocholine accumulated on the film B surface located downstream in the microchannel, and was monitored by the SPR angle shift during the third flow.
the enzymatic reaction of anti-BNP-AChE trapped on film A as the following enzymatic reaction, AChE acetylthiocholine 98 (CH3)3N CH2CH2SCOCH3 + H2O +
thiocholine (CH3)3N+CH2CH2SH + CH3COOH Then, thiocholine gradually accumulates and forms a monolayer on the surface of film B located downstream.
(CH3)3N+CH2CH2SH + Au f (CH3)3N+CH2CH2S-Au + 1/2H2 This monolayer formation on the film B surface can be monitored with high sensitivity as an SPR angle shift as shown in Figure 2c. Therefore, it is possible to measure the BNP concentration by monitoring the SPR angle shift because the thiocholine concentration at the film B surface will decrease as the BNP concentration increases due to a reduction in the amount of trapped anti-BNP-AChE on the film A, i.e., a reduction in AChE activity. We optimized the condition under which the generated thiocholine concentration is in the micromolar region since the SPR angle gradually increases over a few minutes in this concentration region. We obtained an SPR curve and measured its minimum reflection angle (θSPR) using the portable SPR equipment shown Figure 1a. Index matching oil (n ) 1.510, Cargille Laboratories, Cedar Grove, NJ) was used to obtain optical contact between the glass plate at the bottom of the immunosensor and the BK7 prism on the SPR equipment. 5528
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
Electrochemical Measurements. We evaluated the surface area of the Au film using electrochemical measurements, which we performed using a three-electrode configuration with a potentiostat (CHI Instruments, model 802). An Ag/AgCl (3 M NaCl inner solution, BAS) and a Pt wire were used as a reference and an auxiliary electrode, respectively. A gold film working electrode (d ) 3 mm) was fabricated using the rf sputtering equipment as mentioned in connection with the SPR measurement. The gold film electrode was immersed in a 1 mM thiocholine solution for 2 h, and then linear sweep voltammograms at the electrode for the reductive desorption of adsorbed thiocholine were monitored in a 0.1 M KOH solution deaerated by bubbled Ar gas. The gold film electrode had a roughness factor of 2.0, which was determined by measuring the ratio of the electrochemical surface area to the geometric area.37 The electrochemical surface area was obtained from the charge required to reduce the surface oxide layer by potential cycling between 0.0 and 1.7 V in a 0.5 M H2SO4 solution. All measurements were carried out at room temperature. RESULTS AND DISCUSSION Measurement of Thiocholine Monolayer Formation. In this work, the process of thiocholine accumulation on a gold surface is important. First, we studied the thiocholine accumulation and monolayer formation on a gold surface by performing electrochemical and SPR measurements using various thiocholine solution concentrations. Figure 3 shows a linear sweep voltammogram for the reductive desorption of thiocholine in 0.1 M KOH solution deoxygenated (37) Gileadi, E.; Eisner, K. E.; Penciner, J. Interfacial Electrochemistry. An Experimental Approach; Addison-Wesley: Reading, MA, 1975.
Figure 3. Linear sweep voltammogram of the reductive desorption of thiocholine on a gold surface in 0.1 M KOH. The gold was immersed in 1 mM thiocholine for 2 h. The dashed line shows a linear sweep voltammogram of an unmodified gold electrode. The sweep rate was 100 mV/s.
with Ar gas. The thiocholine-modified gold electrode was obtained by soaking the electrode in 1 mM thiocholine for 2 h. The voltammogram (a) shows a sharp cathodic peak at -0.87 V (versus Ag/AgCl) caused by the thiocholine desorption. With an unmodified gold electrode, no cathodic peak was observed in this potential region, which is shown in the same figure by a dashed line (b). Generally, the cathodic peak appears between -0.8 and -1.2 V in a linear sweep voltammogram of alkanethiol in an alkaline solution, and the peak potential increases as the alkyl chain length decreases.38-40 Our result for the cathodic peak potential of thiocholine desorption is equivalent to previous results for the cathodic reduction of short alkenethiols.38-40 The charge from the cathodic peak was 81 µC/cm2. Since our gold surface had a roughness factor of 2.0, the actual density of the thiocholine molecules absorbed on the gold surface was estimated to be 2.5 × 1012 molecules/mm2 (0.49 ng/mm2) because the reductive desorption of thiols occurred through one electroreductive path. This value is consistent with the maximum coverage (2.7 × 1012 molecules/mm2) calculated by assuming the close hexagonal packing of alkanethiol41 when the gold surface was fully covered by alkanethiol. Figure 4a shows typical SPR angle shifts when the thiocholine solutions were introduced from port III to port II at a flow rate of 2 µL/min. After obtaining a stable baseline by using water without thiocholine, we introduced thiocholine aqueous solutions with concentrations ranging from 0.1 to 200 µM. We observed that θSPR gradually increased with time. This is because the refractive index of the film B surface is increased by the formation of the thiocholine monolayer on the film B surface. We also observed that the slope of the SPR sensorgram became steep when the thiocholine concentration increased because the monolayer forms quickly on the surface. However, the θSPR shift value was (38) Widrig, C. A.; Chung, C.; Porter, M. D. J. Electroanal. Chem. 1991, 310, 335-359. (39) Phong, P. H.; Ooi, Y.; Hobara, D.; Nishi, N.; Yamamoto, M.; Kakiuchi, T. Langmuir 2005, 21, 10581-10586. (40) Munakata, H.; Oyamatsu, D.; Kuwabata, S. Langmuir 2004, 20, 1012310128. (41) Chidsey, C. E. D.; Bertozzi, C. R.; Putvinski, T. M.; Mujsce, A. M. J. Am. Chem. Soc. 1990, 112, 4301-4306.
saturated at ∼0.1°, and the saturated value was independent of the thiocholine concentration even if the concentration was high because the surface was fully covered with thiocholine. The θSPR shift value was saturated for 10-min accumulation when the thiocholine concentration was above 100 µM. Figure 4b shows a calibration curve for thiocholine estimated from Figure 4a. The y-axis is the θSPR shift value obtained from Figure 4a 10 min after a thiocholine injection. We are able to estimate the thiocholine concentration from the θSPR shift value 10 min after the injection since we can obtain a sufficient difference between each thiocholine concentration. We would be able to determine very low concentrations of thiocholine after a long accumulation period; however, submicromolar thiocholine was clearly detectable even with an accumulation time of only 10 min. It is reported that protein adsorption of 1 ng/mm2 causes a 0.1° θSPR shift.42,43 In contrast, the results in Figures 3 and 4 reveal that thiocholine adsorption of 0.49 ng/mm2 caused a 0.1° θSPR shift when the gold surface was fully packed with thiocholine. Generally, it is known that SPR sensitivity is poor for small molecules. However, by using the thiocholine monolayer, we obtained almost twice the θSPR shift as that obtained with protein adsorption. This may be because the packing density of thiocholine is higher than that of protein near the gold surface where the sensitivity of the SPR sensor is very high. Accordingly, it would be advantageous for a highly sensitive immunoassay to use an antibody labeled with an enzyme that produces thiocholine because of its accumulation effect and the large θSPR shift. Measurement of B-Type Natriuretic Peptide. Figure 5a shows the calibration curve of our immunosensor when we used 190 ng/mL anti-BNP-AChE conjugate. The y-axis is the θSPR shift value obtained from SPR sensorgrams 15 min after acetylthiocholine solution injection. Although the θSPR shift value remained almost unchanged when the BNP concentration was lower than 1 pg/mL, the θSPR shift value decreased as the BNP concentration increased when the BNP concentration was higher than 5 pg/mL. The concentration of the thiocholine that flows into the film B surface decreases when the BNP concentration increases because the production of thiocholine decreases due to the reduction in the anti-BNP-AChE trapped on film A. The proposed method could be used for determining BNP with high sensitivity over a wide dynamic range from 5 pg/mL to 100 ng/ mL. This indicates that the wide measurable concentration range and low detection limit of our immunosensor is sufficient for monitoring the BNP concentration in patients’ blood since that is reported to range from 20 pg/mL to 2 ng/mL. We also confirmed that it was possible to tune the sensitivity range of our immunosensor by changing the conjugate concentration. The calibration curve in Figure 5a was obtained to achieve a wide dynamic region. However, the small BNP concentration change in the low BNP concentration range could be obtained by reducing the conjugate concentration. Figure 5b shows a calibration curve when using 2.4 ng/mL anti-BNP-AChE conjugate. Although the SPR angle shift decreased by only 5% as the BNP concentration increased from 1 to 10 pg/mL in Figure 5a, a 58% reduction was observed with increasing BNP concentration in Figure 5b because the ratio (42) Fagerstam, L. G.; Frostellkarlsson, A.; Karlsson, R.; Persson, B.; Ronnberg, I. J. Chromatogr. 1992, 597, 397-410. (43) Stenberg, E.; Persson, B.; Roos, H.; Urbaniczky, C. J. Colloid Interface Sci. 1991, 143, 513.
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
5529
Figure 4. (a) SPR sensorgrams of thiocholine accumulation when the thiocholine concentration was varied from 0.1 to 200 µM at a flow rate of 2 µL/min. (b) Variation in the θSPR shift value obtained from (a) 10 min after thiocholine injection.
Figure 5. Calibration curves for BNP obtained with the immunosensor when the anti-BNP-AChE conjugate concentration was (a) 190 and (b) 2.4 ng/mL.
of the conjugate concentration to the sample BNP concentration was lower than that in Figure 5a. We previously reported an electrochemical immunoassay of BNP where we used an electrochemical desorption technique with a relatively large volume batch system.6,44 Although the method showed high sensitivity, it took ∼2 h to complete an assay because of the long accumulation and immunoreaction time. Furthermore, the method is complicated since we need to change various buffer solutions. This is because the electrode potential has to be scanned in a strong deaerated alkaline solution to measure the thiocholine reductive desorption current accurately by separating it from the oxygen reduction current after the accumulation process in a neutral solution. In this article, we determined the BNP simply and rapidly because the SPR method can be successfully employed to monitor the thiocholine accumulation in a neutral solution with high sensitivity. The detection limit was also improved compared with that of our previous report. This is because the thiocholine concentration increases efficiently in the microchannel due to the high surface-to-volume ratio, and this also increases the accumulation efficiency. In our previous studies, we reported that thiocholine was produced with 0.01 unit/L AChE activity in 10 mL of phosphate buffer solution for 30 min. This indicated that the thiocholine concentration was 0.3 µM (3 nmol in 10 mL), and we could accumulate and measure only 0.3% of the produced thiocholine molecules since we observed a reductive peak of ∼1 µC (44) Mizutani, F.; Matsuura, H. Curr. Appl. Phys. 2005, 5, 98-101.
5530
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
by thiocholine desorption. With our current method, we estimated that the thiocholine concentration in the microchannel increased ∼1000 times compared with that of the batch-style measurement. This is because the inner volume of the microchannel is more than 3 orders of magnitude smaller although there is only a 2.5fold difference between the trapping areas for anti-BNP-AChE. Thus, this makes it possible to determine very low concentrations of BNP with improved sensitivity. In addition, thiocholine would be accumulated effectively in a thin microchannel (20 µm deep). For example, it needed ∼6 h to reach a steady SPR angle when a 1 µM thiocholine standard solution was introduced into the sensor at a flow rate of 0.3 µL/ min. The number of thiocholine molecules introduced in 6 h is 108 pmol, and the full coverage of thiocholine on the sensing surface is calculated at 13 pmol. Thus, 8.3% of the introduced thiocholine was accumulated on the sensing surface, which is 28 times greater than that of the previous batch-style measurement (0.3%). We were able to complete the assay in 30 min/sample using our immunosensor. This is not only because the immunoreactions finished rapidly in the microspace, but also because of the simultaneous use of a labeled enzymatic reaction and the realtime SPR monitoring of the enzymatic product accumulation in a microfluidic device. Some microchannel-based immunoassay techniques with a small sample volume have been reported; however, these methods often require repeated washing and the
Table 1. Comparison of θSPR Shift Value in Human Serum with That in Buffer for Various Concentrations of BNP Spiked Samplesa run no.
BNP added value
calibration curve from Figure 5/ mdeg
serum measurement/ mdeg
ratio/ %
1 2 3 4 5 6
0 pg/mL 10 pg/mL 100 pg/mL 1 ng/mL 10 ng/mL 100 ng/mL
99 91 75 58 38 26
96 91 76 59 32 23
97 100 101 102 84 88
a The θSPR shift values in buffer solution were estimated from the calibration curve in Figure 5.
introduction of a secondary antibody. We simply determine the trace level of BNP (15 fg in 3 µL) in three solution injection steps. The first step consists of introducing a sample solution containing anti-BNP-AChE. The second involves introducing a buffer solution for rinsing, and the final step is to introduce an acetylthiocholine solution. Despite this simple assay system, we still need to improve the assay time. Since all the steps in our assay technique were carried out manually, we will be able to reduce the required time considerably by automating the operation and optimizing the fluidic channel structure, for example by increasing the BNP immobilizing area. The relative standard deviation (RSD) for five measurements of 100 pg/mL BNP was 33.1%. Although this value is relatively high compared with that of a commercially available immunosensing technique such as a 96-well microtiter plate or an immunochromatography combined with a fluorescence detection system, we consider that the value would be acceptable for on-site monitoring for clinical diagnosis because it was reported that the blood BNP concentration significantly increases in heart failure patients.45 The RSD would be improved by enlarging the film A surface or by modifying film A using a micro-spotter machine because the high RSD would be caused by a fluctuation in the amount of BNP on the film A surface caused by the manual modification. Recovery Tests of Spiked Samples with BNP. Recovery tests were carried out with BNP spiked in human serum. The human serum was a 1:10 diluted sample that was heat-inactivated at 72 °C for 30 min. Table 1 compares θSPR shift values obtained from BNP spiked in human serum with those obtained from BNP in a buffer solution. The θSPR shift values in the buffer solution were estimated from the calibration curve in Figure 5. We obtained satisfactory recovery test values over a wide concentration range of 10 pg/mL to 100 ng/mL. However, we could not obtain good agreement between BNP spiked in serum and BNP in the buffer (45) Maisel, A. S.; Krishnaswamy, P.; Nowak, R. M.; McCord, J.; Hollander, J. E.; Duc, P.; Omland, T.; Storrow, A. B.; Abraham, W. T.; Wu, A. H. B.; Clopton, P.; Stef, P. G.; Westheim, A.; Knudsen, C. W.; Perez, A.; Kazanegra, R.; Herrmann, H. C.; McCullogh, P. A. N. Engl. J. Med. 2002, 347, 161167 (46) Ellman, G. L.; Courtney, K. D.; Andres, V.; Featherstone, R. M. Biochem. Pharmacol. 1961, 7, 88-95. (47) Kurita R.; Yabumoto, N,; Niwa, O. Biosens. Bioelectron. 2006, 21, 16491653.
samples when using unheated human serum. This is because the human serum contained a large amount of endogenous pseudocholinesterase, and excessive thiocholine moieties were produced from acetylthiocholine by the nonspecific adsorbed pseudocholinesterase in the microchannel and the inlet tube. The activity of the pseudocholinesterase in human serum was estimated to be 2685 units/L from spectrometric determination results with 5, 5′dithiobis(nitrobenzoic acid).46 The pseudocholinesterase activity fell to 952 units/L after pretreatment at 57 °C for 30 min, and the activity was not detected when the human serum was heated at more than 67 °C for 30 min. This indicates that temperature is more important than treatment time, and a much shorter treatment time could be employed after optimization. At present, although blood sample measurements need heat inactivation treatment to suppress pseudocholinesterase activity, we may achieve the assay without the heat treatment step by optimizing the rinsing conditions, coating the inner wall of the microchannel to suppress nonspecific adsorption, or filtering the serum sample before adding the anti-BNP-AChE for the immunoreaction. CONCLUSION We have developed a novel miniaturized enzyme immunosensor using microfluidics combined with a portable SPR system. The immunosensor can determine BNP with high sensitivity because it is based on the highly sensitive measurement of AChE activity by the real-time SPR monitoring of the thiocholine accumulation on a metal surface in a T-shaped microfluidic device. The detection limit and assay time are greatly improved compared with our previously reported electrochemical method. The proposed BNP sensing method is one of the most sensitive methods yet reported based on an SPR detection system. As regards frequent on-site BNP monitoring, our system is superior to commercially available immunoassay systems based on fluorescence or isotope detection since our immunosensor is portable and safe and employs a lowpower technique with a high sensitivity equivalent to that of conventional methods. The assay time must be improved if our system is to be used as an on-site cardiac sensor. This can be achieved by adopting automated operation for each step and optimizing the fluidic channel structure, for example, by increasing the BNP immobilizing area. For actual sample measurement, we can simplify our assay system by using the membrane that we previously reported47 to remove large molecules such as proteins including pseudocholinesterase. This work is now in progress. ACKNOWLEDGMENT The authors thank Dr. Hiroaki Matsuura (Riken Keiki Corp.) for helping in the initial stage of this study. The authors also thank Mr. Tatsuya Tobita (NTT-AT Corp.) for helping with device fabrication. Part of this project was supported financially by the New Energy and Industrial Technology Development Organization in Japan. Received for review March 16, 2006. Accepted May 25, 2006. AC060480Y
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
5531
