Attomole Detection of Hemagglutinin Molecule of Influenza Virus by

Apr 30, 2008 - Optimization of the analytical conditions enabled determination of the hemagglutinin molecules of the influenza virus in the concentrat...
0 downloads 9 Views 1MB Size
Anal. Chem. 2008, 80, 4020–4025

Attomole Detection of Hemagglutinin Molecule of Influenza Virus by Combining an Electrochemiluminescence Sensor with an Immunoliposome That Encapsulates a Ru Complex Naoyoshi Egashira,*,† Shin-ichi Morita,† Emi Hifumi,‡,§ Yoshiharu Mitoma,† and Taizo Uda§,⊥ Department of Environmental Sciences, Faculty of Life and Environmental Sciences, Prefectural University of Hiroshima, Shobara-shi, Hiroshima, Japan, Research Center for Applied Medical Engineering, Oita University, Oita-shi, Oita, 870-1192, Japan, CREST, Japan Science and Technology Agency (JST), Kawaguchi-shi, Saitama 332-0012, Japan, and Department of Applied Chemistry, Faculty of Engineering, Oita University, Oita-shi, Oita, 870-1192, Japan An immunoliposome (80 nm in diameter) encapsulating a Ru complex with two aminobutyl moieties was prepared to detect the presence of hemagglutinin molecules, which play an important role in influenza virus infection. The highly sensitive detection was accomplished by electrochemiluminescence (ECL) from the Ru complex adsorbed onto Au electrodes after competitive immunoreactions. This method clarified that the adsorption of the Ru complex onto the electrode was an important factor in obtaining high sensitivity. Optimization of the analytical conditions enabled determination of the hemagglutinin molecules of the influenza virus in the concentration range of 3 × 10-14 (6 × 10-19 mol/50 µL sample) to 2 × 10-12 g/mL. The sensitivity was far superior to that obtained by conventional ELISA as well as to that obtained by biosensors and reported thus far. In the past 3 decades, many biosensors have been developed. Immunosensors to detect antigen-antibody reactions, surface plasmon resonance (SPR) sensors, and quartz crystal microbalance (QCM) sensors have become widely used for many applications. They have proved to be very convenient tools for the rapid detection of substances such as a variety of antigens1,2 and substrates.3,4 However, they have some drawbacks, in particular, for the detection of minute amounts of antigens (less than nanograms per milliliter). Throughout the world, the appearance of new influenza viruses (especially the H5 type) has been identified as a very serious threat to human life as are other types of viruses such as HIV (human immunodeficiency virus), etc. At present, to prevent a * Corresponding author. Phone: +81-824-74-1796. Fax: +81-824-74-1796. E-mail: [email protected]. † Prefectural University of Hiroshima. ‡ Research Center for Applied Medical Engineering, Oita University. § Japan Science and Technology Agency. ⊥ Faculty of Engineering, Oita University. (1) Baac, H.; Hajos, J.; Lee, J.; Kim, D.; Kim, S.; Shuler, M. Biotechnol. Bioeng. 2006, 94, 815. (2) Pribyl, J.; Skladal, P. Anal. Chim. Acta 2005, 530, 75. (3) Ishihara, T.; Arakawa, T. Sens. Actuators, B 2003, 91, 262. (4) Larsson, A.; Angbrant, J.; Ekeroth, J.; Mansson, P.; Liedberg, B. Sens. Actuators, B 2006, 113, 730.

4020

Analytical Chemistry, Vol. 80, No. 11, June 1, 2008

pandemic of influenza viruses, we must develop new tools such as biosensors that have extremely high detection sensitivities of less than 1 ng/mL. Recently, there have been several reports of studies aiming to fabricate unique sensors using liposomes that carry targetrecognition molecules on their surface and encapsulate reporter molecules in the inner aqueous phase. Antibody-modified liposomes, immunoliposomes, have been utilized to detect E. coli by encapsulating dyes in the inner aqueous phase.5 A liposome polymerase chain reaction (PCR) assay using encapsulated DNA reporters has accomplished highly sensitive detection (0.02 fg/ mL) of a biological toxin. However, this procedure requires complex analytical processes.6 In addition, biotin-modified liposomes have been used to determine the concentration of a protein (100 ng-10 mg/mL) by combining electrogenerated chemiluminescence (ECL) with magnetic microbeads.7 Unfortunately, these new methods using liposomes could not achieve the sensitivity required to detect the very low amounts of the viruses mentioned above. In this decade, we have developed various flow-through detectors using ECL with which we were able to detect many compounds such as oxalates,8 amines,9 amino acids,10,11 and peptides12 with a high level of sensitivity. Furthermore, we have tried to fabricate an ECL sensor for detecting proteins by employing an immunoliposome encapsulating Ru complexes.13 It is well-known that the hemagglutinin (HA) molecule, which situates as a trimer of protein on the surface of the envelope membrane of the virus, plays an important role in influenza virus infection of human cells, followed by cell fusion, entry of the (5) (6) (7) (8) (9) (10) (11) (12) (13)

Ho, J.; Hsu, H. Anal. Chem. 2003, 75, 4330. Mason, J.; Xu, L.; Sheng, Z.; O’Leary, T. Nat. Biotechnol. 2006, 24, 555. Zhan, W.; Bard, A. J. Anal. Chem. 2007, 79, 459. Egashira, N.; Kumasako, H.; Ohga, K. Anal. Sci. 1990, 6, 903. Egashira, N.; Kumasako, H.; Uda, T.; Ohga, K. Electroanalysis 2002, 14, 871. Egashira, N.; Piao, J. E.; Hifumi, E.; Uda, T. Bunseki Kagaku 2000, 49, 1029. Piao, J.; Mitoma, Y.; Zhao, C.; Hifumi, E.; Uda, T.; Egashira, N. ITE Lett. Batteries, New Technol. Med. 2003, 4, 606. Piao, J.; Mitoma, Y.; Uda, T.; Hifumi, E.; Shimizu, K.; Egashira, N. Electroanalysis 2004, 16, 1262. Hirata, T.; Mitoma, Y.; Uda, T.; Egashira, N. Chem. Sens. 2006, 22, 58. 10.1021/ac702625d CCC: $40.75  2008 American Chemical Society Published on Web 04/30/2008

Figure 1. Ru(II) complex.

viruses into the cells, propagation of the viruses, and so on. That is to say, the HA molecule is essential for the activity of the influenza virus. It has been noted that several hundred HA molecules are present on the surface of an influenza virus. Therefore, detection of the molecules with very high sensitivity is an important issue from the perspective of preventing a virus pandemic. To achieve the above goal, in the present study, we set out to fabricate such a tool and successfully constructed a very highly sensitive biosensor by using our unique technique, which combines both ECL and an immunoliposome encapsulating a Ru complex in the inner, aqueous phase. Through the use of this biosensor, we were able to detect the HA molecule with a level of sensitivity on the order of up to 10-14 g/mL (10-19 mol/ sample). Here, we will describe the structure of the sensor, discuss the method of detection, etc. MATERIALS AND METHODS Reagents. The ruthenium complex, bis(2,2′-bipyridine)[4,4′bis(4-aminobutyl)-2,2′-bipyridine] ruthenium perchlorate (Figure 1), was prepared by the Gabriel synthesis of bis(2,2′-bipyridine)[4,4′-bis(4-bromobutyl)-2,2′-bipyridine]ruthenium perchlolate according to a paper previously reported.14 1,2-Dipalmitoyl-racglycero-3-phosphoethanolamine (DPPE), 1,2-dipalmitoyl-rac-glycero-3-phosphacholine (DPPC), and dithiodibutanoic acid were obtained from Sigma-Aldrich Company. N-Succinimidyl-3-(2-pyridyldithio)propionate (SPDP) was obtained from Pierce Chemical Company. 1-Ethyl-3-(dimethylaminopropyl)carbodiimide hydrochloride (WSC) and N-hydroxysuccinimide (NHS) were obtained from Wako Pure Chemical Industries, Ltd. Other reagents were of reagent grade unless otherwise noted. Preparations of HA-Peptide and anti-HA Monoclonal Antibody (JN-1-2 mAb). The HA peptide (TGLRNGITNKVNSVIEKAA; MW, 2105.30) used in this experiment was synthesized by the Fmoc solid-phase method and identified by mass spectrometry. The purity was over 95% by HPLC analysis. Regarding the amino acid sequence of HA peptide, TGLRN and GITNKVNSVIEK are the conserved amino acid residues located in the HA-1 domain and HA-2 domain, respectively, of the hemagglutinin molecule of H1 and H2 types of influenza. (Note that there are H1-H16 types among hemagglutinin molecule in influenza virus at the present time.) In this study, the above two conserved peptide sequences were combined and synthesized as one peptide at whose C-terminal two alanine residues were introduced in order to hold the helix conformation. In the preparation of anti-HA peptide monoclonal antibody, at first, female Balb/c mice were subcutaneously immunized with (14) Egashira, N.; Morita, S.; Tashiro, H.; Mitoma, Y.; Uda, T. ITE Lett. Batteries, New Technol. Med. 2006, 7, 276.

100 µg/mouse of the HA peptide conjugated with human IgG (1.0 mg/mL in phosphate-buffered saline (PBS): pH ) 7.4), which was emulsified with an equal volume of Freund’s complete adjuvant (Difco Laboratories, Detroit, MI). Additional immunizations (second, third, and fourth) were subcutaneously administered by the injection of 100 µg/mouse of the conjugate emulsified in Freund’s incomplete adjuvant (Difco Laboratories) at 20, 36, 50, and 65 days (this was the final booster) after the first immunization. The immunized spleen cells removed from the mice were fused with myeloma SP/NSI/1-Ag4-1 at a ratio of 5:1 using 50% PEG 1500 (Boehringer Mannheim GmbH). The fused cells were placed into the wells of 96-well culture plates and cultivated in hypoxantine-aminopterin-tymidine medium. The fused cells were screened to find the antibody secretion cells by means of a modified sandwich enzyme-linked immunosorbent assay (ELISA) in which HA conjugated with human immunoglobulin G (IgG) and human IgG were coated on the plate. Hybrids that were found to secrete antibodies specific for the peptide were cloned by the limiting dilution method. The isotypes of the resulting monoclonal antibodies were determined by using a mouse monoclonal antibody isotyping kit. The subtype of the heavy chain of JN-1-2 mAb was IgG1, and that of the light chain was κ-type. The ascites fluid was obtained by intraperitoneal injection of the hybridoma cell lines in pristane-primed female Balb/c mice. Expression of HA-2 Domain Protein. cDNA of the HA-2 domain of hemagglutinin from influenza virus (A/Hiroshima/5/ 2001 (H1N1)) was synthesized by using PCR, in which the following primers were employed: forward primer (5′-ACACACATATGGGTTTGTTTGGAGCCAT-3′), which contains an NdeI site (underlined), and the reverse (5′-AAAAAACTCGAGGATGCATATTCTACACT-3′), which contains an Xho I site (underlined). An aliquot of the PCR mixture was analyzed by agarose gel electrophoresis. To construct a plasmid for the expression of recombinant HA-2 domain protein, the PCR-amplified DNA fragment was ligated to the expression vector, pET21a (Novagen, Madison, WI). E. coli JM109 was transformed and induced the HA-2 domain protein by an addition of a final concentration of 1 mM isopropyl-1-thioβ-D-galactoside. The molecular weight of the HA-2 domain protein thus obtained was estimated to be 26 kDa by agarose gel electrophoresis. ELISA. Fifty microliters of HA peptide or the subunit of hemagglutinin (HA-2 domain) dissolved in PBS solution (5 µg/ mL) was fixed on an immunoplate (Nunc, Denmark) at 4 °C overnight. Blocking was performed using 2% gelatin or bovine serum albumin (BSA) for 1 h at room temperature. After the plate was washed, JN-1-2 mAb was immunoreacted along with the dissolved HA peptide, followed by a reaction with antimouse Ig(G + A + M) conjugated with alkaline phosphatase. After the substrate reaction using p-nitrophenyl phosphate, the absorption band at 405 nm was measured three times by use of an immunoplate reader (ImmunoMini, NJ-2300, Nalgen Nunc International K.K., Tokyo, Japan). Each point in the calibration curve showed the mean + and - standard deviation using (n - 1). The detection range was estimated from the value of two standard deviations away from the background. Purification of Monoclonal Antibody (mAb). JN-1-2 mAb was purified according to the purification manual for the Bio-Rad Analytical Chemistry, Vol. 80, No. 11, June 1, 2008

4021

Figure 2. Immunoliposome: part a, Ru(II) complex; part b, lipid bilayer; part c, antibody.

Protein A MAPS-II kit (Nippon Bio-Rad, Tokyo, Japan). First, 5 mL of ascites fluid containing JN-1-2 mAb was mixed with the same volume of ammonium sulfate saturated solution. The precipitate was recovered by centrifugation, and then 5 mL of PBS was added to the precipitate. This process was repeated twice, followed by two dialyses against PBS. An aliquot of the PBS solution containing JN-1-2 mAb was mixed with the same volume of the binding buffer of MAPS-II. This mixture was then placed on a bed packed with Affi-Gel (protein A) for elution of the bound mAb. The eluted mAb was dialyzed twice against the buffer, 50 mM Tris/0.15 M NaCl (pH 8.0), at 4 °C and stored at 4 °C or frozen. Preparation of Immunoliposome. DPPE and anti-HA monoclonal antibody were reacted with SPDP to give SPPE-HA mAb and SPDP-DPPE.15 DPPC (10 µmol), SPDP-DPPE (0.15 µmol), and cholesterol (3 µmol) were mixed in CHCl3. After removal of the solvent by evaporation, 10 mM phosphate buffer (PB, pH ) 7.4) containing 5 mM Ru(II) complex solution was added to the remaining mixture. This was vortexed and then extruded through a series of polycarbonate membrane filters with pore sizes of 50, 100, and 400 nm (Lipex Extruder; Northern Lipids, Vancouver, BC, Canada) to form liposomes. The extruded solution was then chromatographed on a Sephadex G50 fine column (mobile phase: 0.1 M PB, pH 7.4). The collected liposome fractions were concentrated, treated with the activated SPPE-HA antibody reduced by dithiothreitol, and chromatographed on a Sepharose 4B column (mobile phase: 0.1 M PB, pH 7.4). A Centriprep YM10 filter unit was used to concentrate the immunoliposome fractions to 1 cm3. The resulting liposomes (Figure 2) were stored under N2 at 4 °C until used. The liposomes thus prepared with the 50 and 100 nm membranes showed diameters of 80.4 and 94.9 nm, respectively, as determined by a Zetasizer Nano, Malvern Instruments Ltd. (U.K.). Destruction of Immunoliposome. The immunoliposome solution (7 µL) was dropped onto a clean Au electrode and allowed to stand for 30 min at room temperature. The electrode was washed with PB (0.1 M, pH 7.4) and dried. Seven microliters of different alcohols or other organic solvents were then added onto the electrode. After the electrode was heated and cooled for 10 min each, the PB solution (0.1 M, pH 7.4) containing 0.1 M triethylamine was added to the electrode, which was applied at 1.3 V versus Au quasi reference electrode to generate ECL after being allowed to stand for 5 min. (15) Ho, R. J. Y.; Rouse, B. T.; Huang, L. Biochemistry 1986, 25, 5500.

4022

Analytical Chemistry, Vol. 80, No. 11, June 1, 2008

Figure 3. Detection procedure: (1) immobilization of hemagglutinin (or antigen peptide) on a gold electrode; (2) binding of immunoliposome with hemagglutinin onto the gold electrode through competitive antigen-antibody reaction; (3) destruction of immunoliposome by addition of ethanol; (4) adsorption of Ru(II) complex by heating at 60 °C for 10 min; (5) ECL measurement on application of potential.

Figure 4. Four Au electrode unit (A) and spacer (B). Part a, counter electrode; part b, working electrode; part c, reference electrode.

ECL Measurement. The ECL measurement was carried out according to the scheme shown in Figure 3. (1) Immobilization of antigen onto Au electrodes: Four Au electrode units were formed on a glass plate (20 mm × 30 mm), as shown in Figure 4. To modify the electrode surface with dithiodibutanoic acid, the clean electrode plate was immersed into ethanol solution containing dithiodibutanoic acid (10 mM) for 12 h and then the plate was washed out with ethanol. The acid-modified Au electrode was activated by the reaction with WSC and NHS for 1 h. Then the hemagglutinin or antigen peptide was fixed onto the electrode surface at 4 °C for 15 h (1 mg/mL 0.1 M PB buffer, pH 7.4). (2) Competitive immunoreaction: after an aliquot (50 µL) of hemagglutinin or antigen peptide solution was added onto the electrode, the immunoliposome solution (10 mL) was further added in order to accomplish competitive immunoreaction for 1 h. (3) Destruction of immunoliposomes: after the electrode was washed with PBS to remove any unbound immunoliposomes, ethanol was dropped onto the electrode to destroy the liposomes bound to the electrode. (4) Adsorption of Ru complex: To enhance the adsorption of the Ru complex on the Au, the electrode was heated to 60 °C and maintained at that temperature for 10 min. After being allowed to stand for 10 min, the electrode was washed with PB solution containing 0.01% Triton X-100. (5) ECL measurement: A PB solution (0.1 M, pH 7.4) containing 0.1 M triethylamine was added onto the electrode, to which a potential was applied at 1.3 V versus Au quasi reference electrode in order to generate ECL. The ECL intensity was measured three times and averaged with a DNA Reader Pro1 (Microtech Nition Company, Ltd., Japan). In the ECL response curves, each point showed the mean + and - standard

Table 1. ECL Intensity for Each Ru(II) Complex Absorbed onto Au Electrodes

Figure 5. Typical ECL measurement for the Ru complex adsorbed onto the Au electrodes. The solid line indicates the current when the potential from 0 to 1.3 V vs quasi Au was applied during 1 s and then maintained. The broken line indicates the generated ECL.

deviation using (n - 1). The detection range was estimated from the value of two standard deviations away from the background. Cyclic Voltammetry (CV). CV was accomplished using a glassy carbon working electrode (3 mm in diameter), a Pt counter electrode, and a Ag/AgCl reference electrode in 0.1 M PB (pH 7.4) in the same way as previously reported.16 The CV apparatus used was a CHI 604 electrochemical analyzer (ALS Company, Japan). RESULTS AND DISCUSSION Cyclic voltammetry for the 0.5 mM Ru complex in PB buffer solution was carried out on a glassy carbon electrode, and the anodic peak potential was +1.1 V versus Ag/AgCl. The potential was almost the same as tris(2,2′-bipyridine)ruthenium complex, which is widely used in ECL research. The redox waves were not clear due to the overlapping of a large anodic current for water oxidation although they gave a well-defined reversible one-electron process in acetonitrile.17 Figure 5 shows a typical ECL curve for the Ru complex adsorbed onto the electrode when the applied potential was from 0 to 1.3 V in 1 s and then kept at 1.3 V. The ECL curve increases steeply in about 1 s and then decays gradually. In contrast, the current decayed steeply after 1 s. To attain the huge enhancement of sensitivity in the ECL measurement, we investigated the following reaction conditions in detail. Because it is important to efficiently destroy the liposome bound onto the electrode, many organic solvents such as methanol, ethanol, 1-propanol, 2-propanol, and 1-butanol were examined to determine how greatly the liposome generates the ECL from the Ru complex. The order of the signal strength by the type of solvent was ethanol (1.5 × 105 counts) > 1-propanol (1.3 × 105) > 2-propanol (0.7 × 105) > 1-butanol (0.5 × 105) > methanol (0.5 × 105). Of all the solvents tested, ethanol gave the strongest ECL. Other organic solvents possessing higher boiling points showed an increase of background in ECL signals, because the solvents have to be heated up in the next step. In the experiment of the destruction of liposome, both process of heating over 60 °C and freezing gave very low efficiency. The addition of surfactants such as Triton X-100 was also poor because of blocking of the adsorption of the Ru complex. (16) Egashira, N.; Minami, T.; Kondo, T.; Hori, F. Electrochim. Acta 1986, 31, 463. (17) Tokel, N.; Bard, A. J. J. Am. Chem. Soc. 1972, 94, 2862.

Ru(II) complex

ECL intensity, × 105 counts

-H -(CH2)4OH -(CH2)4N(CO)2Ph -(CH2)4NH2 -(CH2)4SH

1.0 5.8 7.2 13.2 16.1

To obtain a strong ECL signal, the Ru complex must be concentrated by the adsorption onto the Au electrodes. Therefore, in this study, several Ru complexes having two side chains in the 4 and 4′ positions of bipyridine, such as -H, -(CH2)4OH, -(CH2)4N(CO)2Ph, -(CH2)4NH2, and -(CH2)4SH, were prepared and then examined for their adsorptive ability on the Au electrode.14 Of these compounds, the Ru complexes having side chains -(CH2)4NH2 and -(CH2)4SH gave the stronger ECL signals (Table 1). In particular, the complex having side chains -(CH2)4NH2 showed a unique characteristic to easily desorb from the Au electrodes rather than did the -(CH2)4SH. The characteristic enabled a convenient and mild washing with an active cationic surfactant. This should be important for practical applications because convenience becomes a big advantage for reusing the Au electrodes. After destruction of the liposome by the addition of ethanol, heating to around 60 °C is required to accelerate the absorption of the Ru complex onto the modified Au electrodes, but temperatures over 70 °C brought an increase of background noise, resulting in a very weak ECL signal. By employing the most suitable reaction conditions and the Ru complex, we conducted the following experiment. The results are shown in Figure 6, which is a calibration curve used to detect the antigen peptide, TGLRNGITNKVNSVIEKAA. The sequence is a crucial part of the hemagglutinin molecule of the influenza virus (in H1 and H2 types) as stated in Materials and Methods (Preparation of HA Peptide). The JN1-2 mAb recognizes the sequence of GITNKVNSVIEKAA but not TGLRN, because the antibody reacted to the HA-2 domain but not to the HA-1 domain of the influenza virus. Anyhow, in the sensing experiment, 0.15 µmol of SPDP-DPPE was used and a 100 nm membrane pore size was employed in preparing the liposome. The detection range was roughly from 1 to 100 ng, which has a higher sensitivity than

Figure 6. Calibration curve for the HA peptide. In this figure, an indication, “PB ”, is the case without HA peptide. The liposome was prepared with 0.15 µmol of SPDP and a membrane with 100 nm pores. Analytical Chemistry, Vol. 80, No. 11, June 1, 2008

4023

Figure 7. Conventional ELISAs for HA peptide and hemagglutinin protein.

Figure 9. Calibration curve for hemagglutinin protein under optimum conditions. The conditions were the same as those in Figure 8.

Figure 10. Control experiments. The ECL responses were obtained with the electrode surface to which BSA protein was bound (A) and the liposome without the antibody (B). Other conditions were the same as those in Figure 8. Figure 8. Calibration curve for HA peptide under optimum conditions. The liposome was prepared with 1.5 × 10-9 mol of SPDP and a membrane with 50 nm pores.

that of ELISA, as shown in Figure 7 where the sensitivity for HA peptide (antigenic peptide) was in the range from 100 ng/mL to several µg/mL. Furthermore, the response for hemagglutinin in the ELISA did not show a clear decrease of the calibration curve, indicating the low sensitivity. It is generally well-known that, in the competitive ELISA system, the rate of antigen to antibody is a crucial factor to enhance the detection limit. Therefore, in order to find the preferable reaction conditions for highly sensitive detection, we investigated several important subjects such as liposome size, amount of antibody fixed onto the liposome surface, and concentration of the immunoliposome in the reaction system. The immunoliposome adjusted by using membranes of different pore sizes, 50, 100 and 400 nm, were prepared. The immunoliposome prepared with a 50 nm pore size membrane gave the strongest ECL intensity. The results were contrary to those expected from simple calculations. In the calculations, we assumed that a spherical liposome can bind with an ordered antigen on a plain electrode and the total volume of bound liposomes is proportional to the number of the Ru complex. Actually, the binding site on the electrode would be more disordered and large liposomes might be at a disadvantage in binding with the antigen because of the disordered arrangement on the electrode. Figure 8 shows the experimental results for detection of the HA peptide, where the liposome was prepared with a 50 nm pore membrane and 1.5 × 10-9 mol of SPDP-DPPE was used. A typical sigmoid curve was obtained, and the detection range was from a concentration of 3 × 10-14 to 2 × 10-12 g/mL (7 × 10-19 to 4 × 4024

Analytical Chemistry, Vol. 80, No. 11, June 1, 2008

10-17 mol/50µL/sample). Note that the amount of SPDP-DPPE, which was the antibody-binding site on the liposome, decreased by 100. The sensitivity was superior to that of the conventional ELISA system by a factor of over 107, because the lower limit of the detection by the ELISA system was estimated to be 5 × 10-7 g/mL. Such high sensitivity was attained due to the establishment of an effective competition reaction that brought about a remarkable decrease in the size of the binding site on the liposome. Similarly, dilution of the liposome solution is effective in increasing sensitivity, but it generally resulted in lowering the ECL intensity, leading to a large error. To achieve the same level of sensitivity, the liposome concentration was adjusted with PB buffer based on the fluorescence intensity from the Ru complex in the liposome. Under the same measurement conditions, the sensor was applied for detecting a hemagglutinin molecule (HA-2 domain). Figure 9 shows the results in which the protein molecule was measured in a concentration range from 3 × 10-13 to 4 × 10-11 g/mL (6 × 10-19 to 8 × 10-17 mol/50 mL sample). To confirm the accuracy of the above results, we performed the following two control experiments. First, the liposome on whose surface no antibody was fixed did not show any change in ECL through the wide range of hemagglutinin concentrations, as shown in Figure 10B. Second, the use of the electrode surfaces coated with BSA protein (shown in Figure 10A) and without protein (only ethanolamine-modified acid; the data was not shown) also gave no change in ECL. The results of the control experiments thus show that the appearance of ECL response is due to the immunoreaction on the electrode. Note that we could only detect 4 × 105 molecules of the proteins by using this method. The enhanced sensitivity was attributable to both a decrease in the amount of antibody on the surface and

the dilution of the liposome solution. Zhan and Bard reported that an ECL system using biotin-modified liposomes and magnetic beads could determine the C-protein in a concentration range of 100 ng/mL to 10 µg/mL.7 To our knowledge, our method exhibits the highest level of sensitivity in the rapid detection methods using liposomes that have been proposed thus far. Conclusively, we have successfully fabricated a new biosensor using both immunoliposomes and an encapsulated Ru complex under the appropriate reaction conditions. The biosensor enables rapid attomole sensing of hemagglutinin proteins. It is very fascinating to see how the low content of influenza viruses can be detected by using the biosensor because the virus possesses many hemagglutinin spikes on its surface (about 500 spikes/virus; each spike is a trimer of the protein). In the near future, we will report the rapid detection of hundreds of virus particles and the adsorption mechanism of the Ru complex onto Au electrodes. We believe that the present sensor should become to a powerful tool in the rapid diagnosis of influenza and further for detecting trace amounts of proteins related to many diseases.

CONCLUSIONS A rapid and highly sensitive detection method has been developed by combining ECL with an immunoliposome encapsulating a Ru complex. Under optimum measurement conditions, hemagglutinin molecules of influenza virus were determined in a concentration range from 3 × 10-13 to 4 × 10-11 g/mL. This level of sensitivity suggests that a detectable lower concentration could be 6 × 10-19 mol/50 µL. The present method with high sensitivity at the attomole level would be applicable for detecting trace amounts of various proteins containing the influenza virus. ACKNOWLEDGMENT The present research was supported in part by a Grant-in-Aid for Scientific Research (No. 18510105) from the Japan Society for the Promotion of Science, and the Japan Science and Technology Agency (Creation of Biodevices and Biosystems with Chemical and Biological Molecules for Medicinal Use). Received for review December 28, 2007. Accepted April 2, 2008. AC702625D

Analytical Chemistry, Vol. 80, No. 11, June 1, 2008

4025