Detection of influenza viruses by a waveguide-mode sensor

Abstract

Methods are demonstrated for antibody-based detection of influenza viruses using the monolithic-sensing-plate technology of an evanescent-field-coupled waveguide-mode sensor. To facilitate the detection of viruses at lower concentrations, a strategy of complexing the virus with Coomassie Brilliant Blue dye was used to amplify the signal. By applying this method, hemagglutinins from avian influenza virus (H5N1) and human influenza virus (A/Panama/2007/1999) were detected. An intact human influenza virus (A/Brisbane/59/2007) was also detected. The method is suitable for the detection of virulent viruses that may emerge at various stages of evolution.

---

Introduction

Influenza is an infectious disease caused by RNA viruses of the family Orthomyxoviridae that affects birds and mammals. Influenza spreads around the world in seasonal epidemics, occasionally resulting in pandemics that can cause millions of deaths. Influenza A viruses are spherical or filamentous and enveloped; they range in diameter from 80 to 120 nm.¹ Influenza viruses express two major proteins on their surfaces: hemagglutinin (HA) and neuraminidase (NA). HA comprises up to 556 amino acid residues and it is the principal determinant of variation among the major types of influenza virus that cause disease in humans. The influenza A virus subtypes are classified according to their H number (for HA) or N number (for NA). There are 16 different HA antigens (H1–H16) and nine different NA antigens (N1–N9) for influenza A. Because of antigenic drift on HA, several subtypes of the virus have arisen and caused major epidemics around the world. Subtypes of influenza viruses are identified and named according to the isolate that they resemble and are therefore presumed to share a lineage. The HA protein is critical for binding to cellular receptors and for viral fusion, and it mediates the binding of the virus to the cells of the host.^2,3 Various anti-HA probes are available for detecting viral infections and differentiating subtypes of the virus; these include anti-influenza aptamers^4,5 and anti-influenza antibodies.^5,6 Several diagnostic methods that use these probes have been shown to be capable of detecting and characterizing the viruses. These methods include enzyme-linked immunosorbent assay,⁶ immunoblotting,⁷ immunosensor-based methods,⁸ interferometry,⁹ fluoroimmunoassay,¹⁰ methods based on surface plasmon resonance,⁵ and immunochromatography.¹¹ Here, we demonstrate method for detecting the viruses by using the monolithic-sensing-plate technology of evanescent-field-coupled waveguide-mode (EFC-WM) sensors. Because HA is the major determinant for causing epidemics, we used the anti-HA antibody as the capturing molecule that was immobilized on the silicon-based sensor chip.

The principles and design of the waveguide sensor system that we used was similar to that of a surface plasmon resonance (SPR) system, the only difference being that the mode used for measurement is not a surface mode but a waveguide mode (see Suppl. Fig. S1†).^12,13 The advantage of the waveguide sensor over the SPR sensor is that higher sensitivity is easily obtained by perforating the waveguide layer and/or by using color materials, such as dyes or metal nanoparticles, as a label.^12,14,15 In the case of the SPR sensor, the wavelength of an incident light is restricted by the material used in order to induce the surface plasmon, whereas there is no such restriction in the case of the waveguide sensor. As a result, any kind of color materials can be applied as a label, which means that cheaper and more effective labels can be available. In addition, amorphous SiO₂, one of the most popular materials of the waveguide layer of the waveguide sensor, is physically and chemically more stable than Au or Ag, which are typical materials for SPR. Previously, by using silicon chip-based waveguide-mode sensors, we have detected biomolecular interactions of a wide range of molecules, including DNA–DNA, DNA–RNA, and RNA–protein interactions, and interactions with small molecules.^12,16,17 In these cases, we used nanometre-sized holes with a diameter of 50 nm in the surface of the chip to improve the accommodation of biomolecules. However, such nanoscale holes can behave as molecular sieves toward biomolecules such as viruses, leading to a lack of sensitivity in the detection of viruses because of their greater size (around 100 nm).¹⁸ Based on our previous report,¹⁸ increasing the nanoholes on the chips larger than 50 nm will broaden the spectrum and it also leads to lower sensitivity. In addition, using Coomassie Brilliant Blue (CBB) it is expected that sensitivity will be increased at least 100-fold.¹⁵ To achieve an adequate level of detection for viruses, we complexed them with CBB on the sensor chips; because CBB has an absorbance near to 600 nm, the wavelength is suitable for use with our sensor chip thereby increasing the sensitivity of our sensor system. In this way, we were able to avoid having to make nanoscale holes in the chip surface and we achieved a greater sensitivity than achieved with a surface containing nanoscale holes.

Experimental

Chemicals and biomolecules

We purchased 3-(triethoxysilyl)propan-1-amine from Sigma–Aldrich (Tokyo, Japan). The dye Coomassie Brilliant Blue G-250 was obtained from Biorad Laboratories. The monoclonal antibody for avian influenza H5N1 and the HA antigen were purchased from Biodesign International, Meridian Life Science, Inc. (Saco, ME, USA). Anti-HA serum for H1N1 (Brisbane/59/2007) and inactive viruses were purchased from Denka Seiken (Tokyo, Japan). Haemagglutinin from human influenza A H3N2 virus was purified as described previously.^16,19 Human anti-H3N2 monoclonal antibody was purchased from Chemicon, Millipore Corporation (USA).

Preparation of the waveguide-mode sensor chips and measurements

The design of the waveguide-mode sensor with Kretschmann geometry is shown in Suppl. Fig. S1†. All measurements were performed as described previously.¹⁵ For fabrication of the sensor chip, we used a silicon-on-quartz (SOQ) substrate (Shin-Etsu Chemical, Japan) consisting of a 265 nm-thick single crystalline Si (100) layer on a 1.2 mm-thick SiO₂ glass substrate. The substrate was cut into plates measuring 25 × 25 mm and thermally oxidized in an electric furnace in an atmosphere of O₂ containing water vapor at 1000 °C at ambient pressure for 62 min. The oxidation process converted the surface of the Si layer into a SiO₂ waveguide layer (Si = 35 ± 5 nm, SiO₂ = 520 ± 10 nm).¹⁸ The sensor chip was mounted on the base of a SiO₂ glass prism to form an optically contiguous medium. An S-polarized He–Ne laser operating at 632.8 nm was used as the light source for the measurements. The beam was directed onto the base of the prism at an angle that gave total internal reflection. To observe the optical waveguide modes, the reflectivity was measured for various angles of incidence at a constant temperature of 20.8 °C.

Analyses of interactions of antibodies and viruses

To analyze HA antigens and their interactions with various types of HA or with viruses on the monolithic sensing plate of our EFC-WM sensors, amino coupling on the chip followed by immobilization with 2.5% aqueous glutaraldehyde (Glu) was performed as previously described.¹⁷ On this surface, either the monoclonal antibody against avian H5N1 or human H3N2 or anti-HA serum against H1N1 of Brisbane/59/2007 virus was attached to the tail end of the Glu molecules by injecting 100 μl of sample and incubated for 3 h as it is the slow reaction. The surface was blocked with 1 M 2-aminoethanol (ethanolamine) with same sample volume for 30 min and similarly followed in all other steps. Antibody–antigen assembly was induced by injecting the antigen or virus and then measured. The sensitivity toward these complexes was enhanced by injection or complexation with CBB, and changes in the dip of the resonance upon attachment of complexed CBB to the protein or virus were monitored.

The complex of CBB with the HA of the H3N2 virus was prepared by mixing the HA and Bio-Rad Protein Assay Solution (diluted 1 : 4 with buffer and filtered), and rocked gently for one hour at room temperature. Uncomplexed free dye was separated from the complex by filtration through Amicon Ultra (Millipore) with a molecular-weight cutoff of 30,000 Da. The complexed molecules of 100 μl were adjusted to the desired concentration and used for the interactive analyses and incubated for 30 min. The specific binding of anti-HA antibody to HA was analyzed by the SPR method on Biacore 2000 with a CM4 sensor chip from Biacore (Uppsala, Sweden).

Results and discussion

Influenza is a pandemic virus that spreads on a worldwide scale and can infect large proportions of the human population. HA mediates the binding of the virus to host cells. HA, the major viral antigen, is required for membrane fusion with host cells thereby mediating the early stages of infection by the influenza virus.²⁰ In addition, HA gene is known to induce high levels of macrophage-derived chemokines and cytokines, which lead to infiltration of inflammatory cells and severe hemorrhaging, especially when the HA is derived from a virulent strain.²¹ Currently, several anti-HA antibody-based detection systems are routinely used for detection and characterization of viruses.^6–9,11 Similarly, by using an antibody developed against the HA of the virus, we have developed a method for detecting the HA or the virus on monolithic sensing plates of EFC-WM sensors. In such sensors, the waveguide mode is excited by incidence of p- or s-polarized light under conditions of total internal reflection. The optical setup for surface plasmon resonance, which has been widely used as a powerful tool in biosensing applications, can be used for the excitation of waveguide-mode sensors²² (Suppl. Fig. S1†).

Interactions of anti-HA monoclonal antibody and avian influenza A–H5N1 HA

To detect the antibody–antigen interactions, antigens for fourteen amino acids (synthetic peptide) near the middle of the avian influenza A H5N1 virus HA (Genbank accession no. AAT76166) were purchased. A monoclonal antibody (MAb) specific to the HA protein from the H5N1 strain of avian influenza was used to analyze the interaction. The specific interactions of these compounds were initially checked using the established system Biacore 2000 (Suppl. Fig. S2†). As shown in this figure, the molecules showed specific interactions on Biacore, so we further evaluated them with the waveguide sensor by injecting MAb at a concentration of 60 nM onto the immobilized Glu surface (Fig. 1a). Amine-coupling reactions were performed for Glu as described previously¹⁷ (Suppl. Fig. S3a†). On the Glu surface, the antibody reached saturation level, determined because we could not detect any further increment on injection of 1 M 2-aminoethanol onto a surface with the immobilized antibody. No incremental changes upon injection of 2-aminoethanol, could also be due to its smaller size. 2-Aminoethanol is more appropriate than other blocking agents for our sensor chips. In our waveguide-mode measurement device, the excitation of waveguide modes changes the reflectivity of the incident light over a narrow angular region. The waveguide modes are sensitive to the dielectric environment near the surfaces of the waveguides, and variations in this dielectric environment resulted in changes in the reflectivity. From these changes in the reflectivity, small amounts of adsorbed materials at the surface of a waveguide can be detected.^12,16–18 Upon attaching the MAb to Glu, a significant shift of 0.11 degrees was observed, and this shift was enhanced by injecting CBB, which caused a dip in reflectivity from 0.64 to 0.4, a difference of 0.24. Subsequent injections of CBB with repeated washing showed no further increment in the dip, indicating saturation point on the Glu and antibody surface (Suppl. Fig. S3b†). Upon complexation of the antibody with CBB, the absorbance peak of the dye shifted from 465 nm to 595 nm. This is a favorable condition for our sensor system, because the laser we used is more sensitive to absorbance near 600 nm. No significant changes were found when antigen was injected in the absence of CBB (Fig. 1b; upper panel). We injected antigen prepared at a concentration of 10 nM under appropriate buffering conditions onto the immobilized MAb surface and detected a change in the resonance of 0.12 degree, and a change in the dip of 0.21 (from 0.4 to 0.19) on complexation with CBB, due to significant delay in the incident light by optical absorption (Fig. 1b; lower panel). The sensitivity of these shifts observed at antigen concentration of 10 nM was further evaluated by injecting the antigen at a concentration of 1 nM. As in the previous case, with a 1 nM concentration, we observed changes in reflectivity, with dip of 0.1, which proved that it was possible to detect the interactions of the MAb and the antigen for influenza virus at lower concentrations (Fig. 1c). The difference in changes in reflectivity upon attaching antigen at 10 nM or 1 nM concentrations to the antibody is shown in Suppl. Fig. S3c†. During these injections, we also observed that some dye was removed from the antibody or Glu surface as a result of dissociation of the dye during the injection of the antigen. These changes are considered as background, and changes in the dip could be calculated from previous dye injections (between stages 3 and 5 in Fig. 1b and 1c). By forming complexes of CBB with the immobilized antigen on the chip, it was possible to reduce the detection limit to lower concentrations.

Fig. 1 Analysis of the interactions of anti-HA monoclonal antibody and avian H5N1 hemagglutinin on an evanescent-field-coupled waveguide-mode (EFC-WM) sensor. (a) Schematic representation of the assembly of biomolecules for the antibody–antigen interactions. Glutaraldehyde (Glu) was immobilized on the sensor-chip through amino coupling, and the antibody was subsequently attached by a similar reaction. (b) Changes in the resonance upon attachment of HA of H5N1 on the immobilized monoclonal antibody surface in the absence of CBB (upper panel) and in the presence of dye (lower panel). Anti-HA antibody at a 60 nM concentration was attached (red lines) to the Glu surface (black lines), followed by attachment of HA from H5N1 at a concentration of 10 nM (green lines). After the attachment of the antibody and antigen Coomassie Brilliant Blue dye was injected at a 1 : 100 dilution for signal amplification (blue and brown lines). Graphic representation of dip changes are indicted as figure insert. (c) Similar reactions were performed with 1 nM of HA fragment. The reaction steps (1–5, in the X-axis) are followed as in Fig. 1a.

Interactions of anti-HA serum and human influenza A H1N1 virus (A/Brisbane/59/2007)

In the above case, we examined the interactions of the peptide HA from avian influenza and the corresponding antibody. To check further interactions of human viruses, we took the commercially available virus A/Brisbane/59/2007 as an example and we analyzed its interactions with anti-HA serum developed against the virus. According to the manufacturer, the suggested titer for these interactions is 1 : 80 for the agglutination test, and we considered this titer to be specific for our present study. For our analyses, the anti-HA serum was immobilized on the Glu surface and blocked with 2-aminoethanol as described above. The increment in the peak resonance on attaching the antiserum was found to be 0.1 degree. The sensitivity was increased by injecting CBB, and with the complex of CBB and the antiserum, the peak dip changed from 0.65 to 0.47, a difference of 0.18 in reflectivity. Virus at 1 : 80 dilution was injected onto the immobilized antiserum sensor surface, and the dip subsequently changed from 0.47 to 0.43 (0.04 reflectivity) on injecting CBB (Fig. 2). As mentioned above, during the injections some dye was removed from antiserum surface by dissociation, and this was considered as background.

Fig. 2 Analyses of the interactions of anti-HA serum and human influenza A H1N1 virus (A/Brisbane/59/2007) on EFC-WM sensors, showing changes in the resonance upon attachment of A/Brisbane/59/2007 of H1N1 to the immobilized anti-HA serum surface. Anti-HA antibody at a concentration of 60 nM was attached to the Glu surface, followed by attachment of a 1 : 80 dilution of the H1N1 virus. After the attachment of the antibody and antigen, CBB dye was injected at a 1 : 100 dilution for signal amplifications (blue and brown lines). The reaction steps (1–5, in the X-axis) are followed as in the Fig. 1a.

Interactions of Anti-HA monoclonal antibody and HA of human influenza A H3N2 virus (A/Panama/2007/1999)

In a similar manner to the above case, we also examined the interactions of HA for human influenza A virus (A/Panama/2007/1999) with the pure anti-HA MAb rather than the antiserum. The kinetics of the reaction of the monoclonal antibody against the HA of A/Panama (H3N2) virus was previously examined using Biacore 2000, and this analysis revealed that the association (k_on) and dissociation (k_off) constants of the antibody–HA complex were 2.3 × 10⁵ M⁻¹ s⁻¹ and 6.6 × 10⁻⁴ s⁻¹, respectively. The equilibrium dissociation constant (K_d) for the above complex was 2.9 nM,⁵ and this proved the specificity of these samples. For the present analyses, the HAs from A/Panama (H3N2) virus were purified by using a sucrose gradient as described previously.⁵ The interactions of this MAb and the viruses on the EFC-WM were determined after attaching the MAb to Glu (Fig. 1a) as described above. We detected a change in the peak shift of 0.05 degree, and when we injected the CBB, the resonance dip changed from 0.77 to 0.57. Upon injecting the HA as 1 nM of HA complex with CBB, the dip change was from 0.57 to 0.54 (Fig. 3a). When compared to the signal from step 3 to 5, the dip changes were higher in steps 2 to 4. This is due to concentration differences between these steps as 60 and 1 nM, respectively. In all these cases, we were able to improve the sensitivity of detection of HA or viruses, particularly when CBB dye was used. In every case, however, dissociation of CBB was observed upon injection of consecutive samples. In an attempt to overcome this problem, we produced a complex of the antigen with CBB and separated any remaining free unbound CBB by membrane separation with a 30,000 KDa cut-off (as described in the Materials and Methods section) before injection into the cell of the waveguide-mode sensor. In this case, the MAb for HA of A/Panama/2007/1999 was injected as described previously, but the consecutive injection of CBB was omitted. Instead, we injected the HA–CBB complex at a concentration of 1 nM, and we detected changes in the dip of the resonance from 0.78 to 0.74 (Fig. 3b). This change in the dip is nearly same as that for the injection of the noncomplexed dye, showing that CBB interacts specifically with the antigen in both methods. This refinement further increased the sensitivity of our sensor system, permitting the detection of the virus at lower concentrations by ensuring that CBB bound specifically to the virus alone.

Fig. 3 Analyses of interactions of anti-HA monoclonal antibody and HA of human influenza A H3N2 virus (A/Panama/2007/1999) on an EFC-WM sensor. (a) Changes in the dip of resonance upon attachment of 1 nM HA from H3N2 to the immobilized anti-HA antibody surface. Anti-HA antibody at a concentration of 60 nM was attached (red lines) to the Glu surface (black line), followed by attachment of HA from H3N2 virus at a concentration of 1 nM (green line). After attachment of the antibody and antigen, CBB was injected at a 1 : 100 dilution for signal amplification (blue and brown lines). Graphic representation of dip changes are indicted as figure insert. The reaction steps (1–5, on the x-axis) are followed as in Fig. 1a. (b) Changes in the resonance upon attachment of HA from H3N2 Panama virus at a concentration of 1 nM to the immobilized anti-HA antibody surface. Anti-HA antibody at a concentration of 60 nM was attached to the Glu surface, followed by attachment of HA (from H3N2 virus) complexed with CBB at a concentration of 1 nm. A schematic representation of the attachments is shown on the left-hand side of the figure.

Conclusion

Currently, methods such as immunochromatography are in vogue for the rapid identification of influenza viruses and for discrimination between influenza A and influenza B virus types. However, physicians claim that these methods are insufficiently sensitive, particularly in early phases of the disease where insufficient amount of viruses are available in the nasopharyngeal tract. Current tests do not function until at least one day after the onset of fever, and it is preferable to detect the virus at an earlier stage for early diagnosis.¹¹ Bearing this in mind, we have shown in this study that is possible to detect influenza viruses using antibodies immobilized on the monolithic sensing plate of an EFC-WM sensor with higher sensitivity obtained at lower concentrations. This kind of sensor development should pave the way for the development of a method to diagnose the presence of viruses at earlier stages, and therefore will be useful in the monitoring and forecasting of influenza epidemics.

Acknowledgements

This study was supported by Industrial Technology Research Grant Program 2009 from the New Energy and Industrial Technology Development Organization (NEDO), Japan. Part of this work was conducted at the AIST Nano-Processing Facility, which is supported by the “Nanotechnology Support Project” of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. The authors would also like to thank the Advanced Functional Materials Research Center of Shin-Etsu Chemical Co., Ltd. for supplying the samples and Drs. N. Fukuda, K. Kawasaki and P.K.R. Kumar of AIST for experimental support.

References

1. T. Noda, H. Sagara, A. Yen, A. Takada, H. Kida, H. Cheng and Y. Kawaoka, Nature, 2006, 439, 490.
2. K. Subbarao and J. Katz, Cell. Mol. Life Sci., 2000, 57, 1770.
3. G. Neumann, T. Noda and Y. Kawaoka, Nature, 2009, 459, 931.
4. S. H. Jeon, B. Kayhan, T. Ben-Yedidia and R. Arnon, J. Biol. Chem., 2004, 279, 48410.
5. S. C. B. Gopinath, T. Misono, T. Mizuno, K. Kawasaki and P. K. R. Kumar, J. Gen. Virol., 2006, 87, 479.
6. K. Watcharatanyatip, S. Boonmoh, K. Chaichoun, T. Songserm, M. Woratanti and T. Dharakul, J. Virol. Methods, 2010, 163, 238.
7. M. M. Bahgat, M. A. Kutkat, M. H. Nasraa, A. Mostafa, R. Webby, I. M. Bahgat and M. A. Ali, J. Virol. Methods, 2009, 159, 244.
8. R. Wang, Y. Wang, K. Lassiter, Y. Li, B. Hargis, S. Tung, L. Berghman and W. Bottje, Talanta, 2009, 79, 159.
9. S. C. B. Gopinath, K. Awazu, P. K. R. Kumar and J. Tominaga, ACS Nano, 2008, 2, 1885.
10. L. Chen, Z. Sheng, A. Zhang, X. Guo, J. Li, H. Han and M. Jin, Luminescence, 2010 DOI:10.1002/bio.1167.
11. M. Watanabe, N. Nakagawa, M. Ito and T. Ihara, Pediatr. Int., 2009, 51, 211.
12. K. Awazu, C. Rockstuhl, M. Fujimaki, N. Fukuda, J. Tominaga, T. Komatsubara, T. Ikeda and Y. Ohki, Opt. Express, 2007, 15, 2592.
13. W. Knoll, MRS Bulletin, 1991, 7, 29.
14. M. Fujimaki, C. Rockstuhl, X. Wang, K. Awazu, J. Tominaga, Y. Koganezawa, Y. Ohki and T. Komatsubara, Opt. Express, 2008, 16, 6408.
15. M. Fujimaki, K. Nomura, K. Sato, T. Kato, S. C. B. Gopinath, X. Wang, K. Awazu, T. Komatsubara and Y. Ohki, Opt. Express, 2010, 18, 15732.
16. S. C. B. Gopinath, K. Awazu, M. Fujimaki, K. Sugimoto, Y. Ohki, T. Komatsubara, J. Tominaga and P. K. R. Kumar, Anal. Bioanal. Chem., 2009, 394, 481.
17. M. Fujimaki, C. Rockstuhl, X. Wang, K. Awazu, J. Tominaga, T. Ikeda, T. Komatsubara and Y. Ohki, J. Microsc., 2008, 229, 320.
18. S. C. B. Gopinath, K. Awazu, M. Fujimaki, K. Sugimoto, K. Ohki, T. Komatsubara, J. Tominaga, K. C. Gupta and P. K. R. Kumar, Anal. Chem., 2008, 80, 6602.
19. S. C. B. Gopinath, Y. Sakamaki, K. Kawasaki and P. K. R. Kumar, J. Biochem., 2006, 139, 837.
20. J. J. Skehel and D. C. Wiley, Annu. Rev. Biochem., 2000, 69, 531.
21. D. Kobasa, A. Takada, K. Shinya, M. Hatta, P. Halfmann, S. Theriault, H. Suzuki, H. Nishimura, K. Mitamura, N. Sugaya, T. Usui, T. Murata, Y. Maeda, S. Watanabe, M. Suresh, T. Suzuki, Y. Suzuki, H. Feldmann and Y. Kawaoka, Nature, 2004, 431, 703.
22. K. H. A. Lau, L. S. Tan, K. Tamada, M. S. Sander and W. Knoll, J. Phys. Chem. B, 2004, 108, 10812.

---

Footnote

† Electronic supplementary information (ESI) available: Supplementary figures. See DOI: 10.1039/c0ay00491j

---