Nonlinear Optical Detection of Proteins Based on Localized Surface

Jun 21, 2008 - The contrast ratio is much greater than those obtained by linear reflectivity ... Current Address: Tokyo Institute of Technology, Gradu...
0 downloads 0 Views 2MB Size
Langmuir 2008, 24, 8367-8372

8367

Nonlinear Optical Detection of Proteins Based on Localized Surface Plasmons in Surface Immobilized Gold Nanospheres Shin-ya Fukuba,† Kazuma Tsuboi,‡,§ Shinya Abe,† and Kotaro Kajikawa*,†,‡ Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology, Nagatsuta, Midori-ku, Yokohama 226-8502, Japan, and PRESTO, Japan Science and Technology Agency (JST), Kawaguchi, Saitama, Japan ReceiVed February 29, 2008. ReVised Manuscript ReceiVed May 5, 2008 A new nonlinear optical method is presented to detect proteins binding to a gold surface without using fluorescentdye labeling. After exposure of the protein-binding surface to a gold nanosphere solution, the nanospheres are immobilized above a gold surface with a nanogap supported by the protein. The gold nanospheres immobilized on the gold surface show strong localized surface plasmon (LSP) resonance, and the formation of this structure results in a marked increase in the optical second harmonic (SH) activity of the gold surface arising from a large enhancement of the electric field localized adjacent to the nanospheres on the LSP resonance. The SH image, therefore, gives a high contrast ratio, 7.0:1, of protein-binding spots to control spots. The contrast ratio is much greater than those obtained by linear reflectivity imaging.

1. Introduction Localized surface plasmons (LSPs) in metallic nanoparticles have received considerable attention in recent years.1–3 While a number of LSP systems have been reported, surface immobilized gold nanospheres (SIGNs) above a gold surface show the following characteristic optical properties. A red-shifted plasmon band emerges at wavelengths longer than the LSP band of isolated gold nanospheres (∼520 nm) as a result of the electromagnetic interaction between the nanospheres and the gold surface.4–7 The amount of the red-shift depends on the ratio D/d, where D is the diameter of the SIGN and d is the gap distance. At the LSP resonance wavelengths, enlarged electric fields are generated both in the nanogap and at the top of the gold nanospheres.6 The field enhancement factors with respect to the field of incident light are calculated to be 70 in the nanogap and 20 at the top of the nanospheres when D/d ) 20, based on the formulas proposed by Wind et al.8 Such a large electric field is effective for the nonlinear optical process,6,7 since the nonlinear polarization is proportional to the power series of the induced electric field. Actually, we have reported an enhancement factor of second harmonic (SH) light of ∼3 × 105 observed from the SIGNs supported by a merocyanine-terminated alkanethiol self-assembled monolayer (SAM).7 Since the resonance condition of LSPs in the SIGN system is sensitive to the ambient dielectrics, this can be used for a label-free biological sensing platform for linear and nonlinear optical detections. Recently, we have reported a sensitive label-free multichannel detection method by use of * To whom correspondence should be addressed. E-mail: kajikawa@ ep.titech.ac.jp. † Tokyo Institute of Technology. ‡ Japan Science and Technology Agency (JST). § Current Address: Tokyo Institute of Technology, Graduate School of Science and Engineering, 2-12-1, Ookayama, Meguro-ku, Tokyo 152-8552. (1) Haes, A. J.; Van Duyne, R. P. J. Am. Chem. Soc. 2002, 124, 10596. (2) Nath, N.; Chilkoti, A. Anal. Chem. 2002, 74, 504. (3) Prodan, E.; Radloff, C.; Halas, N. J.; Nordlander, P. Science 2003, 302, 419. (4) Okamoto, T.; Yamaguchi, I. J. Phys. Chem. B 2003, 107, 10321. (5) Kume, K.; Hayashi, S.; Yamamoto, K. Phys. ReV. B 1997, 55, 4774. (6) Abe, S.; Kajikawa, K. Phys. ReV. B 2006, 74, 035416. (7) Tsuboi, K.; Abe, S.; Fukuba, S.; Shimojo, M.; Tanaka, M.; Furuya, K.; Fujita, K.; Kajikawa, K. J. Chem. Phys. 2006, 125, 174703. (8) Wind, M. M.; Vlieger, J.; Bedeaux, D. Physica A 1987, 141A, 33.

Figure 1. Scheme of sample preparation for protein detection proposed in this paper. A gold substrate covered with ligand is used as an initial platform. When the substrate is exposed to a sample solution, proteins with affinity to the ligand are immobilized on the substrate surface. Subsequent immersion of the substrate in a gold nanosphere solution brings about immobilization of gold nanospheres by the protein, and SIGN structures are formed with the nanogap supported by the protein-ligand complex.

SH microscopy.9 Here, we report another protein detection method based on the formation of the SIGN structures. Enhanced SH was observed from protein-binding spots with respect to the bare gold surface spots used as the control by a factor of ∼7, which is much larger than the contrast ratio of the linear reflectivity imaging. Figure 1 shows a scheme of the protein detection proposed in this paper. A gold substrate covered with ligand is used as an initial platform for protein detection. When the substrate is exposed to a sample solution, proteins with affinity to the ligand are immobilized on the substrate surface. The linear reflectivity change on the binding of protein is negligibly small at wavelengths (9) Tsuboi, K.; Fukuba, S.; Naraoka, R.; Fujita, K.; Kajikawa, K. Appl. Opt. 2007, 46, 4486.

10.1021/la800643e CCC: $40.75  2008 American Chemical Society Published on Web 06/21/2008

8368 Langmuir, Vol. 24, No. 15, 2008

longer than 550 nm. However, greater than a few percent change can be detected at wavelengths ranging from 380 to 480 nm (blue and purple light). This is called anomalous reflection (AR) of gold.10 The AR originates from the weak metallic character of gold at wavelengths less than 480 nm. Although the AR can detect the adsorption of protein molecules,10,11 it does not give us sufficient sensitivity to detect a small amount of protein. This is because the AR method is approximately 1 order of magnitude less sensitive than the surface plasmon resonance (SPR) method using the attenuated total reflection (ATR) geometry.10 In general, proteins create nonspecific bonding to gold, which is due to surface charge, sulfur atoms, or amino groups in proteins. Hence, immersion of the substrate in a gold nanosphere solution brings about immobilization of gold nanospheres by the protein, and the SIGN structures are formed with the nanogap supported by the protein-ligand complex as shown in Figure 1. Since greatly enhanced SH light is observed from the SIGN structure, this scheme offers sensitive detection of proteins without using fluorescent-dye labeling. Formation of the SIGN structures brings about a marked increase in AR and SH efficiency. A high contrast ratio 0.59:1 was obtained in the AR image, which is much higher than the AR image without amplification by gold nanospheres. The SH image also gives a high contrast ratios of 7.0:1, which is the highest value reported in protein detection without using fluorescent-dye labeling. This method has several advantages over the conventional methods in spatially resolved protein detection: high sensitivity, simple and inexpensive process, and high spatial resolution. Fluorescent-dye labeling gives highly sensitive monitoring for biomolecular interactions, and it has been widely used for DNA detection.12–15 However, it is not necessarily an efficient way for protein detection, since fluorescent-dye labeling is sometimes expensive and problematic. A certain amount of analyte is necessary, and the general labeling process is not established for proteins in contrast to DNA detection. Instead, labeling of the analyte with metallic nanoparticles (NPs) also provides us with sensitive protein detection.16,17 However, an additional process to label the analyte is needed is complicated and inconvenient to perform quick detections. Moreover, binding of the analyte to the ligand is sometimes sterically hindered by the NP labels in solution. SPR detection using ATR geometry is a labelfree method, and it has been widely used these days as a biomolecular detection platform in a microarray format. However, there is difficulty in applying it to a high-density microarray format because surface plasmons propagate along the gold surface with a propagation length of 10-30 µm. Ellipsometry is another label-free optical biomolecular detection platform; however, the spatial resolution is much less than that of the SPR method and the sensitivity is insufficient, so far.18 Among these various biomolecular detection methods, the LSP platform is promising, since it is free from labeling and it has high spatial resolution. While LSP detection has sensitivity similar to that of SPR detection,19 the fluorescent-dye labeling method is more sensitive. Therefore, another highly sensitive label-free method based on LSP is needed for biomolecular detection in a high-density (10) Watanabe, M.; Kajikawa, K. Sens. Actuators, B 2003, 89, 126. (11) Watanabe, S.; Usui, K.; Tomizaki, K.-Y.; Kajikawa, K.; Mihara, H. Mol. BioSystem 2005, 1, 363. (12) Kodadek, T. Chem. Biol. 2001, 8, 105. (13) Mitchell, P. Nat. Biol. 2002, 20, 225. (14) MacBeath, G. Nat. Genetic. Suppl. 2002, 32, 526. (15) Zhu, H.; Snyder, M. Curr. Opin. Chem. Biol. 2003, 7, 55. (16) Templin, M. F.; Stoll, D.; Schrenk, M.; Traub, P. C.; Vohringer, C. F.; Joos, T. O. Trends Biotechnol. 2002, 20, 160. (17) Fritzsche, W.; Taton, T. A. Nanotechnology 2003, 14, R63. (18) Otsuki, S.; Ohta, K.; Tamada, K.; Wakida, S. Appl. Opt. 2005, 44, 5910. (19) Mitsui, K.; Handa, Y.; Kajikawa, K. Appl. Phys. Lett. 2004, 85, 4231.

Fukuba et al.

Figure 2. Samples used in this study. Samples Ia and IIa were used for RA spectroscopy. Samples Ib and IIb are avidin microarrays composed of spots of 100 mm in diameter, and these were used for AR and SH image experiments. Water was spotted at A3, B2, and C1 as a control.

microarray format. The method proposed in this paper has possibility as a promising candidate for such purpose.

2. Experimental Section The samples used in this study are schematically illustrated in Figure 2. The samples were used for reflection absorption (RA) spectroscopy, and they are denoted by IaD, where D (D ) 30 and 80) stands for the diameter of the nanospheres in nanometers. The parameter D is suppressed unless necessary. Samples Ia were prepared using the following procedure. The thin gold film was vacuumevaporated on a glass slide at a pressure lower than 10-4 Pa. The gold shots (99.99%) were purchased from Tanakakikinzoku Kogyo K. K., Japan. The thickness was about 200 nm, and the thin gold film can be regarded as a semi-infinite medium. This substrate was immersed in an aqueous solution of avidin at a concentration of 0.2 mg/mL for 40 min. The avidin was purchased from Wako Chemical, Japan, and was used as received. Avidin was then immobilized on the gold surface, forming nonspecific bonds. It was subsequently immersed in an aqueous solution of gold nanospheres for 2 h, and the SIGN structures supported by the avidin were formed. The gold nanosphere solutions were purchased from Tankakikinzodku Kogyo K. K., Japan, and were used as received. The diameters of the nanospheres used in the present study were 30 and 80 nm. The concentrations of the solutions were approximately 70 ppm. After the deposition of gold nanospheres, the substrates were rinsed with water. Samples IIa were used for RA spectroscopy and were prepared as follows. A 100 nm thick gold substrate was treated with a 1 mM ethanol solution of aminoundecanethiol (AUT) (Dojindo Laboratories, Japan) to form a self-assembled monolayer (SAM). It was immersed in a 1 mM aqueous solution of sulfosuccinimidyl-D-biotin (Dojindo Laboratories, Japan) for 10 min, followed by rinsing with a phosphate buffer solution (pH 7.3). The substrate was then immersed in an aqueous solution of avidin at a concentration of 0.2 mg/mL for 40 min. Avidin was immobilized on the gold surface, forming specific bonding to biotin. To form the SIGN structure, it was immersed in a gold nanosphere solution for 2 h, followed by rinsing with water. For the AR and SH imaging experiments, avidin was spotted on the gold surfaces using a homemade spotting apparatus. The spotting pin was made of a glass capillary, the end of which was covered with poly(dimethylsiloxane) (PDMS). Avidin was dissolved in water at a concentration of 0.2 mg/mL to be used as the ink for the PDMS stamp. The size of the spot was approximately 100 µm. We prepared

Nonlinear Optical Detection of Proteins

Langmuir, Vol. 24, No. 15, 2008 8369

Figure 3. Optical setup for (a) AR imaging and (b) SH imaging.

two kinds of arrays of avidin spots. One was prepared on the bare gold surface (samples Ib), and the other was prepared on the biotinylated surface (samples IIb). In both samples, pure water was spotted at A3, B2, and C1, as a control. After deposition of avidin, they were then immersed in a gold nanosphere solution for 2 h, followed by rinsing with water. Polarized RA spectroscopy was performed with a USB2000 spectrometer (Ocean Optics Inc.). The incident light was guided to the sample surface by an optical fiber and was passed through a film polarizer. It was cast upon the surface with a 70° angle of incidence with respect to the surface normal. Light reflected at the gold surface was conveyed to the spectrometer by an optical fiber. Figure 3 shows the sensing systems for AR imaging (panel a) and SH imaging (panel b). In Figure 3a, the light source was a blue light-emitting diode (LED; λ ) 470 nm), which was modulated at a frequency of 107 Hz. The light was focused on a sample surface (spot size φ ) 10 µm), and the reflected light was detected with a photomultiplier tube (PMT; H7827-002, Hamamatsu Co. Ltd., Japan). The sample stage was computer-controlled, and the reflected light intensity was mapped to form a reflectivity image in the computer. The typical pixel size was 10 µm × 10 µm at the sample surface, and the number of pixels was 100 × 100; one image was taken for approximately 15 min. The SH image was taken using the optical setup depicted in Figure 3b. Illumination from a tunable Ti:Sa laser light (Tsunami, Spectra Physics Ltd.) operating at wavelengths in the range λω ) 750-800 nm with a repetition of 82 MHz and a pulse width of 120 fs was used as the fundamental light without focusing. Polarization of the fundamental light was selected to be p-polarization with a half-wave plate and a Glan-Taylor prism. It was incident at a 45° angle through color filters (SCF-60R, Sigma Koki Co. Ltd., Japan) that remove SH light from optical components. The maximum power density at the sample surface was 5.0 × 10-2 W/mm2 at the wavelength of 800 nm. The reflected fundamental light was removed with three color filters (BG-39, CVI Co. Ltd.). We used an analyzer for ultraviolet light for choosing the polarization of the SH light, if necessary. The SH image was then taken with a cooled charge-coupled device (CCD) camera (DU434-BV, ANDOR Technology, U.K.) with an accumulation time of 300 s.

3. Results and Discussion First, we show p-polarized RA spectra of samples Ia and IIa taken at a 70° angle of incidence with respect to the surface normal in Figure 4. The spectra have two RA bands: one excited by the electric field normal to the surface and the other excited by the electric field along the surface. The former is red-shifted from the original band of the isolated nanospheres, as a result of the electromagnetic interaction between the nanosphere and the surface. It emerges only by p-polarized excitation. The latter appears at wavelengths almost the same as the resonance wavelength of the isolated gold nanospheres. It is observed in p- and s-polarization, since both polarizations have the in-plane component. We theoretically calculated the peak wavelengths for the SIGN system at various ambient dielectric constants, ε, by solving the Laplace equation with appropriate

Figure 4. RA spectra of (a) Ia30 and IIa30 and (b) Ia80 and IIa80.

Figure 5. Calculated peak wavelengths as a function of D/d at various ambient dielectric constants ε. Table 1. Peak Wavelength of the Red-Shifted Band and Thickness d of the Supporting Layer Evaluated from the Peak Wavelengtha sample Ia30 IIa30 Ia80 IIa80

peak wavelength (nm)

d (nm)

580 550 720 710

2.0 3.5 1.3 (1.6) 1.4 (1.8)

a For Ia80 and IIa80, d values calculated with the peak wavelengths that were 20 nm blue-shifted are shown in parentheses.

boundary conditions concerning multipolar contributions until the 30th order, based on the quasi-static approximation.8 The peak wavelengths of the red-shifted bands are plotted as a function of D/d in Figure 5. According to previous studies,4,6,7 the calculated wavelengths agree with the experimental results when we use the ambient dielectric constant ε ) 2 even in the air. This is because the contributions of the substrate and the gap-supporting layer are taken into account as an effective dielectric constant of the ambient medium. Table 1 summarizes the peak wavelengths experimentally obtained and the calculated gap distances d at ε ) 2. The gap distance of Ia30, evaluated to be 2 nm, almost agrees with the previous report using atomic force microscopy that a denatured avidin submonolayer in air shows a small thickness of ∼2 nm.20 The gap distance of ∼3.5 nm in IIa30 corresponds to a length of the complex of the AUT SAM, the biotin linker, and the avidin molecules. In the RA spectra of Ia80 and IIa80, the peak wavelengths are found at 720 and 710 nm, respectively. The (20) Misawa, N.; Yamamura, S.; Yoon-Hoon, K.; Tero, R.; Nanogaki, Y.; Urisu, T. Chem. Phys. Lett. 2006, 419, 86.

8370 Langmuir, Vol. 24, No. 15, 2008

Fukuba et al.

Figure 6. AR images and cross-sectional reflectivity profiles of (a) sample Ib30 and (b) sample Ib80.

corresponding gap distances are 1.3 and 1.4 nm. Isolated nanospheres 80 nm in diameter have an absorption peak at 545 nm in water, which is 20 nm red-shifted with respect to that of nanospheres of 30 nm diameter used in this study (525 nm). This red-shift arises from a retardation effect in the nanospheres, because the approximation D , λ is no longer good for nanospheres 80 nm in diameter. Therefore, we calculated d values with the 20 nm blue-shifted wavelengths to compensate for this retardation effect, and they are shown in parentheses in Table 1 for reference. These small d values suggest that avidin heavily denatured between the nanospheres and the gold surface. The AR images of Ib are shown in Figure 6, and those of IIb are shown in Figure 7. The cross-sectional reflectivity profiles along the x and y directions are also plotted below the AR images. It is noted that the reflectivity of a gold surface is approximately 0.35 at 470 nm, caused by the less metallic character of gold at wavelengths shorter than 500 nm. The SIGN-forming spots have lower reflectivity than the bare gold surface. Hence, we plot the decrease in reflectivity in the upper direction in the AR images, and the avidin spots are accordingly represented as protrusions. The clearly imaged SIGN-forming spots have contrast ratios of 0.59:1 for Ib30 and 0.67:1 for Ib80, which are defined as the ratio of the reflectivity at the SIGN-forming spots to that of the control spots. The contrast ratio is reported to be 0.95:1 in the absence of gold nanoparticles.21 Thus, the introduction of this method to AR imaging gives an 8-fold amplification. (21) Fukuba, S.; Naraoka, R.; Tsuboi, K.; Kajiakwa, K. In preparation.

The AR intensity, that is, the decrease in reflectivity at 470 nm, is mostly influenced by the surface coverage of the gold nanospheres and is independent of the gap distance. This is because the AR is an incoherent optical process and the absorption at 470 nm is scarcely influenced by formation of the SIGN structure. Since a cross-sectional area of each nanosphere 80 nm in diameter is 7.1 times larger than that of each nanosphere of 30 nm, the ratio of the surface number density of the gold nanospheres of Ib30 to Ib80 is evaluated to be 8.8:1 from the contrast of the AR image. This suggests that smaller nanospheres have greater affinity to avidin. In contrast, the SIGN spots are obscurely observed for samples IIb as shown in Figure 7. The contrast ratios are approximately 0.91:1 for IIb30 and 0.99:1 for IIb80. These low contrast values are caused by the SIGN structure being formed at the biotinylated surface around the avidin spots, where the gold nanospheres are immobilized by the residual amino groups in the AUT SAM that have not reacted with sulfosuccinimidyl-D-biotin. It is confirmed that the SIGN structure forms on the biotinylated surface by immersion of the substrate in a nanosphere solution, using RA spectroscopy. The absolute reflectivity could not be obtained, since there was no bare gold surface in Sample IIb. Therefore, we plot the reflected light intensity normalized by the mean reflected-light intensity of the surrounding surface, indicated as “signal intensity” in the vertical axis. The amplification of the contrast ratio owing to the formation of the SIGN structure is about 2 times higher for IIb30. This is much lower than the

Nonlinear Optical Detection of Proteins

Langmuir, Vol. 24, No. 15, 2008 8371

Figure 7. AR images and cross-sectional profiles of (a) sample Ib30 and (b) sample Ib80. The vertical axis is the reflected light intensity normalized by the mean reflected light intensity of the surrounding surface.

Figure 8. SH images and cross-sectional SH intensity profiles of (a) sample Ib30 and (b) sample Ib80 at two fundamental wavelengths, λω. The SH intensity is indicated in gray scale with the count per second (cps) unit. The fundamental wavelength used for the excitation of SHG is written in parentheses.

contrast of samples Ib. The AR image of IIb80 shows very low contrast, suggesting that the surface number density of the gold nanospheres at the avidin spots is the same as that at the surrounding surface. The better contrast in IIb30 is consistent with the fact that smaller nanospheres have more affinity to avidin, as concluded above. Figure 8 shows the SH images of samples Ib. The fundamental wavelength used for excitation of the second-harmonic generation (SHG) was chosen between 750 and 800 nm to give a better contrast, and it is written in parentheses. Since the degree of the

resonance enhancement at the fundamental wavelength varies in each case, the excitation wavelength to give a good contrast is different. The absolute intensity is shown in gray scale with the count per second (cps) unit, shown with the bar above the SH image. The cross-sectional profiles are taken at the line indicated with arrows. The avidin spots of Ib are clearly SH-imaged as a bright spots, and the greatest contrast ratio of approximately 7.0:1 was found for both samples, where the contrast ratio is defined as the ratio of SHG intensity at the SIGN-forming spots to that of the control spots. These great contrast values originate from the facts that the SIGN surface has SH activity much greater than that of the bare gold surface and that SH imaging is similar to dark field microscopy, that is, the background signal is negligibly small. The SH intensity for Ib80 is approximately 6 times greater than that for Ib30, as a whole. There are two reasons for this: the image of Ib80 is taken without using any analyzer, while the image of Ib30 is a p-polarized SH image, and the SIGN-forming spots in Ib80 are on resonant to the fundamental light. As shown in Figure 4b, the RA peak is close to the fundamental wavelength (λω )750 nm), whereas the spots in Ib30 are off resonant and the SH intensity is relatively low. The resonance enhancement is a main factor that influences the SH intensity. Figure 9 shows the p-polarized SH images of (a) IIb30 and (b) IIb80 at fundamental wavelengths λω ) 750 and 800 nm. In both samples, not only the avidin spots but also the area surrounding the avidin spots is bright, resulting in the reduction

8372 Langmuir, Vol. 24, No. 15, 2008

Fukuba et al.

than that at the avidin spots because of the smaller gap distance. The competition of the two factors results in the poor contrast in both fundamental wavelengths. Such difference in the resonance condition clearly appears in SH images of IIb80. Although IIb80 spots were scarcely ARimaged, they are negatively SH-imaged at λω ) 800 nm with a contrast ratio of 0.65:1. This interesting result is solely attributed to the difference in the resonance condition, since the surface number density of gold nanospheres is equal in both areas according to the corresponding AR image, as described above. RA spectroscopy revealed that the SIGN-forming biotinylated surface has a more red-shifted RA band at 760 nm. This value is similar to the thickness of an AUT SAM,7 and the corresponding gap distance is approximately 1.4 nm. This indicates that the surrounding area is more resonant to the fundamental wavelength of 800 nm. On the other hand, the poor contrast in the images of IIb80 at λω ) 750 nm is attributed to the fact that both areas are similarly resonant to the fundamental wavelength.

4. Conclusion

Figure 9. SH images and cross-sectional SH intensity profiles of (a) sample IIb30 and (b) Sample IIb80 at two fundamental wavelengths, λω. The SH intensity is indicated in gray scale with the count per second (cps) unit.

of the contrast ratio. The surface around the avidin spots is covered with gold nanospheres with forming the SIGN structures, where the SH activity is similar to the SIGN-forming avidin spots. Although the avidin spots were AR-imaged with a contrast ratio 0.91:1 in IIb30, the corresponding SH contrast is poor. The SH intensity is mostly influenced by two factors: the surface coverage of the nanosphere immobilized at the surface and the degree of resonance enhancement. Since IIb30 is fairly AR-imaged as shown in Figure 7a, the surface number density in the avidin spots is larger than that of the surrounding area. On the other hand, the resonance enhancement at the surrounding area is greater

We have shown an optical method to detect proteins without using fluorescent-dye labeling. The SIGN structures with a nanogap supported by protein are formed by immersion of a protein-binding substrate to a gold nanosphere solution. Formation of the SIGN structures brings about a marked increase in AR and SH efficiency. A high contrast ratio of 0.59:1 was obtained in the AR image of Ib30. This contrast ratio is 8 times higher than that of the AR image without amplification of gold nanospheres. The SH image gives a high contrast ratio of 7.0:1. This value is much larger than the contrast ratio of the AR image, since it is a dark-field detection method. This method has several advantages over conventional methods in spatially resolved protein detection: high sensitivity, simple and inexpensive process, and high spatial resolution. There are still exists a few problems, such as the nonspecific immobilization of gold nanospheres, the shortening of the measurement time, and the reduction of the cost for the SH imaging system. However, it is one of the promising candidates for biomolecular detection in a high-density array format. LA800643E