Ultrasensitive Detection of Bacterial Protein Toxins on Patterned Microarray via Surface Plasmon Resonance Imaging with Signal Amplification by Conjugate Nanoparticle Clusters

Abstract

Sensitive detection and monitoring of biological interactions in a high throughput, multiplexed array format has numerous advantages. We report here a method to enhance detection sensitivity in surface plasmon resonance (SPR) spectroscopy and SPR imaging via the effect of accumulation of conjugated nanoparticles of varying sizes. Bacterial cholera toxin (CT) was chosen for the demonstration of enhanced immunoassay by SPR. After immobilization of CT on a gold surface, specific recognition is achieved by biotinylated anti-CT. The signal is amplified by the attachment of biotinylated 20 nm AuNP via streptavidin bridge, followed by attachment of 5 nm streptavidin-functionalized Fe₃O₄NP to the AuNP–biotin surface. The continuous surface binding of two differently sized conjugated nanoparticles effectively increases their packing density on surface and significantly improves SPR detection sensitivity, allowing quantitative measurement of CT at very low concentration. The dense packing of conjugated nanoparticles on the surface was confirmed by atomic force microscopy characterization. SPR imaging of the immunoassay for high-throughput analysis utilized an Au-well microarray that attenuated the background resonance interference on the resulting images. A calibration curve of conjugated nanoparticle binding signal amplification for CT detection based on surface coverage has been obtained that shows a correlation in a range from 6.31 × 10⁻¹⁶ to 2.51 × 10⁻¹³ mol/cm² with the limit of detection of 5.01 × 10⁻¹⁶ mol/cm². The absolute quantity of detection limit using SPR imaging was 0.25 fmol. The versatile nanoparticles and biotin–streptavidin interaction used here should allow adaptation of this enhancement method to many other systems that include DNA, RNA, peptides, and carbohydrates, opening new avenues for ultrasensitive analysis of biomolecules.

Graphical Abstract

Surface plasmon resonance (SPR) biosensors have become an indispensable tool for studying biomolecular interactions in pharmaceutical and biomedical research. Capable of real-time and end-point detection that can be label-free and highly sensitive, SPR allows for measurement of analyte concentration and binding kinetics in a fast, convenient, and nondestructive fashion. The Kretschmann configuration is the most widely employed physical setup for SPR, which utilizes a thin (approximately 50 nm) layer of a noble metal (usually gold) deposited on a glass substrate that is attached to a prism.¹ Measuring the optical characteristics of light reflected from the prism, including resonance angle, intensity, phase, and polarization, provides information about the biomolecular interactions taking place on the surface.^2–4 Examples of use include protein conformation studies,⁵ biomarker profiling,⁶ aptamer selections,⁷ and antibody selections,⁸ which have produced high-affinity ligands that recognize specific protein targets. While highly useful for determining a variety of important kinetic and affinity parameters of biological interactions, one major drawback of SPR is its low throughput.⁹ This problem has been largely circumvented with the advent of SPR imaging. Rather than the scanning-angle or scanning-wavelength measurements commonly employed in SPR spectroscopy, SPR imaging generally measures at a fixed angle. Differences in reflectivity are monitored over time, which provides spatial capabilities of imaging and allows for multiplexed detection and high-throughput bioanalysis.¹⁰ By using a CCD camera for signal detection, the images of the chip can be recorded allowing simultaneous analysis of many interactions. SPR imaging has been reported for the bioaffinity detection of a wide range of analytes, including proteins,^11,12 nucleic acids,^13–15 carbohydrates,¹⁶ Escherichia coli bacteria,¹⁷ and receptor–guest interactions.¹⁸

There are multiple directions of development in the fabrication of the arrays for SPR imaging research. Microfluidics (Glass/PDMA biomolecular screening chip) offer the distinct advantage of possible individual addressability, which enables multiple chemical environments to be explored in a multiplexed fashion.^19,20 However, microfluidics fabrication requires complicated and precise design of flowcells. Contact printing has also been widely used with solid pins, which enables easy deposition of viscous solutions, reproducible and efficient printing, and uses a simple cleaning procedure.^21–23 Effective passivating reagent is necessary after microarray printing to avoid nonspecific adsorption for unarrayed areas.²⁴ SPR imaging analysis is generally displayed through differential sensorgrams²⁵ or rendered false color images with background subtraction/correction.^26–28 During measurement, surface plasmon resonance occurs across the entire surface, giving rise to significant background signals. As the background signal varies in response to surrounding light excitation and solution conditions, the sensitivity and the accuracy of the measurements in the targeted area can be severely compromised with the fixed background analysis. Some efforts have been made to reduce the background resonance by nonchemical approaches such as the use of a patterned SPR-carrying layer to obtain metal spots or islands separated by uncoated glass.^19,29,30 Here, we used a novel SPR imaging microarray fabricated earlier in our group,³¹ wherein a second layer of 100 nm Ti was deposited on gold sublayer via photolithography to generate an Au-well microarray. The spatial variation of the metal thickness restricts the excitation of surface plasmons in the desired patterns and attenuates the evanescent field in the background area, mitigating the possible background interference.

Recent developments of nanomaterials have significantly enhanced the detection sensitivity of biosensors. AuNP has been reported to improve the sensitivity of end-point type of SPR detection for biomolecules such as progesterone,³² DNA, and protein,³³ but the enhancement is limited. Packing density of nanoparticles is important to achieve satisfactory SPR signal amplification. However, the rigid structure of large nanoparticles, such as Au, resists close packing and leaves vacant areas that do not receive the enhancement effect, impairing amplification. To increase the packing density of AuNPs on a surface, Corn et al. has reported the use of a poly(A) polymerase extension reaction coupled with poly(T)-coated AuNP to amplify signals from surface-bound miRNA.³⁴ To sufficiently occupy the space among bound AuNPs, polymer was grown on the AuNP surface to further amplify SPR signal for protein detection.³⁵ Using smaller-sized nanoparticles for signal enhancement in bioassays is less common because it is difficult to achieve enough packing density for the small particles to be effectively sensed by a general technique.³⁶ Here, we combine the approaches and connect two differently sized nanoparticles together to enhance packing density on the surface for SPR signal amplification of cholera toxin (CT) protein detection. For experimental convenience, protein CT was covalently immobilized on plain gold substrate and Au-well microarray to demonstrate the principle. After surface blocking with BSA and immuno-attachment of biotin–antiCT and streptavidin, 20 nm biotin-functionalized AuNPs were applied. The role of AuNP–biotin was 2-fold: to increase SPR detection signal and to act as building blocks to generate space for subsequent Fe₃O₄NP filling. 5 nm streptavidin-functionalized Fe₃O₄NPs were incubated on the surface and attached to biotin-functionalized AuNP via biotin–streptavidin interaction; the Fe₃O₄NPs occupied the gap area between AuNPs and further enhanced the SPR signal. The amplification method relied on the accumulation effect of differently sized conjugated nanoparticles, which efficiently utilize the space and significantly enhance packing density. Given the many and varied types of nanoparticles and wide availability of biotinylated antibodies for specific biomolecules, this method should be broadly applicable to almost any protein of interest.

EXPERIMENTAL SECTION

Materials and Reagents.

Streptavidin and (+)-biotinyl-3,6,9,-trioxaundecanediamine (BA) were obtained from Thermo Scientific (Rockford, IL). N-Hydroxysuccinimide (NHS), 1-(3-(dimethylamino)propyl)-3-ethylcarbodiimide hydrochloride (EDC), 2-(2′-aminoethoxy) ethanol (AEE), polyoxyethylene (20) sorbitan monolaurate (Tween 20), 16-mercaptohexadecanoic acid (16-MHDA), cholera toxin (CT), 11-mercaptoundecanoic acid (MUA), and Bovine serum albumin (BSA) were obtained from Sigma-Aldrich (St. Louis, MO). Biotinylated rabbit anticholera (biotinylated antiCT) serum was from ViroStat (Portland, Maine). All proteins solutions were prepared in 20 mM phosphate buffered saline (PBS) (containing 150 mM NaCl, pH 7.4) unless indicated otherwise. Citrate-stabilized AuNP solution with particle size of about 20 nm was prepared according to the literature³⁷ and stored in brown glass bottles at 4 °C, which is stable for at least 1 month. Poly(acrylic acid) (PAA) wrapped Fe₃O₄NP solution with particle size of about 5 nm was prepared according to the literature,³⁸ which is also stable for at least 1 month.

Instrumentation.

A dual channel SPR spectrometer NanoSPR-321 (NanoSPR, Addison, IL) with a GaAs semiconductor laser light source (λ = 670 nm) was used for all SPR spectroscopy measurements. The device comes with a high-refractive index prism (n = 1.61) and 30 μL flow cell. Surface interaction and modification were monitored using the angular scanning mode around the minimum angle.

SPR imaging was conducted using a home-built, previously reported instrument setup.³⁹ Au-well microarray chips were mounted on a home-built optical stage housing a 300 μL flow cell. An equilateral SF2 triangular prism (n = 1.61) was then put in contact with the glass substrate with a matching liquid (n = 1.61). The optical stage was fixed to a rotating goniometer that allows the tuning of the incident angle of a red light (648 nm) emitting diode (LED) that excited the surface plasmons on the metal surface. The reflected images of the microarray were captured by a cooled 12-bit CCD camera (QImaging Retiga 1300) with a resolution of 1.3 MP (1280 × 1024 pixels) and 6.7 μm × 6.7 μm pixel size using p-polarized light. Difference images were obtained by digitally subtracting one image from another. All the experiments were carried out at room temperature (23 °C). Atomic force microscopy images were obtained using a Veeco Dimension 5000 atomic force microscope (Santa Barbara, CA) with manufacturer-provided software. All images were obtained in tapping mode.

Preparation of Functionalized Nanoparticles.

AuNP was modified with hydroxyl and biotin groups according to a literature protocol⁴⁰ with some modification. In brief, a 0.80 nm AuNP dispersion (in citrate) was gently mixed in equal volume with 1.82 mg/mL Tween 20 (in 10 mM PBS), which then stood for 30 min to allow physisorption to occur. A 0.5 mM 16-MHDA solution was added and stood for 3 h to allow chemisorption, which was followed by three sets of centrifuging (30 min at 15700g) and resuspension (in 10 mM PBS with 1.82 mg/mL Tween 20). The resulting 16-MHDA-modified AuNPs were then reacted with 50 mM NHS/200 mM EDC (in 10 mM PBS) for 5 min. After washing with 10 mM PBS (with 1.82 mg/mL Tween 20, 5 min centrifuge), the remaining NHS ester-alkanethiol modified AuNPs were reacted with an aqueous solution of 22 mM AEE and 2.4 mM BA for 20 min. After again washing with PBS, the resulting AuNP–hydroxyl–biotin (hereafter referred to as AuNP–biotin) was stored at 4 °C.

Fe₃O₄NP–streptavidin was prepared first by activating the carboxylate groups of the Fe₃O₄ nanoparticle surface by reacting with a mixture of freshly prepared 100 mM NHS and 400 mM EDC solution (10 mM PBS, pH 7.0) for 5 min. Then, after washing with 10 mM PBS, the Fe₃O₄NP were incubated with 1.6 mg/mL streptavidin solution overnight. After washing with 10 mM PBS, the resulted Fe₃O₄NP–streptavidin was stored at 4 °C.

Au Surface Preparation for Immunoassays.

SPR spectroscopy gold sensor chips were fabricated in a cleanroom via E-beam evaporation, onto cleaned BK-7 glass slides, of a 2 nm thick chromium adhesion layer, followed by deposition of a 46 nm thick gold layer. Au-well microarray was fabricated according to our reported protocol³¹ with modification. As shown in Figure 1, a BK7 glass substrate was used for the fabrication with E-beam evaporation of 2 nm titanium as the adhesion layer and of 48 nm gold as the SPR active layer. A photoresist AZ5214E was then spin-coated and patterned by photolithography. A second E-beam evaporation was performed to deposit 100 nm titanium on the patterned substrate. The photoresist was then removed with acetone. After rinsing with ethanol and water, the obtained Au-well microarray chips were stored under vacuum before use.

Figure 1.

Schematic of the fabrication of Au-well arrays for surface plasmon resonance imaging.

SPR Signal Amplification with AuNP–Biotin and Fe₃O₄NP–Streptavidin for CT Detection.

The gold chip for SPR spectroscopy and Au-well microarray chip for SPR imaging were incubated in 1 mM MUA ethanol solution for 18 h to form a self-assembled monolayer with carboxyl functional group on the surface. After extensive rinsing with copious ethanol and DI water, the chip was dried under a N₂ stream. To activate the carboxyl acid group, EDC (400 mM)/NHS (100 mM) solution was injected into the flow cell and incubated for 30 min followed by 10 min surface rinsing. Each subsequent step in the immunoassay was also followed by a 10 min rinse. Varied concentrations of CT were injected subsequently and incubated for 30 min to allow formation of covalent amide linkages, which was followed by a 10 min rinsing to eliminate any residual CT in solution. Passivation of the unused activated carboxyl groups was preformed by incubation with 1 mg/mL BSA solution for 30 min. The immunoassay binding was carried out with injection of 0.25 mg/mL biotinylated anti-CT and 30 min incubation. Following this, 0.5 mg/mL streptavidin was injected and incubated for 30 min to prepare the surface for conjugated nanoparticle binding.

The amplification with conjugated nanoparticles attachment was conducted by injection of AuNP–biotin solution and 10 min incubation. After 10 min surface rinsing to remove nonspecific and unstable surface bound AuNP–biotin, Fe₃O₄NP–streptavidin was injected subsequently and incubated for 10 min to further enhance SPR signal. The surface was rinsed with water for 10 min to remove unstable surface-bound Fe₃O₄NP–streptavidin.

RESULTS AND DISCUSSION

SPR Signal Amplification with AuNP–Biotin and Fe₃O₄NP–Streptavidin Binding.

AuNP has been widely used to enhance detection signal for SPR biosensors.^41,42 However, the AuNPs have rigid structure that limits its packing density on the surface, so the full extent of amplification is not realized. To overcome this problem, we looked to fill in the space with another type of nanoparticle. Fe₃O₄ has been used previously in a signal enhancing capacity, but not extensively, since it is too small on its own to effectively enhance the signal.^36,43 As a result, surface enzyme polyadenylation reaction was applied to generate poly(A) tail to increase the binding positions of T₃₀-coated AuNPs,³⁴ and atom transfer radical polymerization (ATRP) reaction was initiated from AuNP surface to fill the packing vacancy.³⁵ Magnetic nanoparticles have been frequently used in bioseparation and for analyte purification and enrichment,⁴⁴ but there are few reports about their applicability in bioassays due to their extremely small size so that sensitive detection is difficult to achieve using a general technique.^36,43 Here, 20 nm AuNP–biotin was initially applied on the surface as a building block to create large spatial positions and surface area for binding. Five nanometers of Fe₃O₄NP–streptavidin was then accumulated around AuNP via the strong biotin–streptavidin interaction, causing the Fe₃O₄NPs to occupy the gap areas between the AuNPs. This increased the packing density on the surface for the purpose of amplifying the signal change in SPR detection. The high specificity of the biotin–streptavidin interaction brings Fe₃O₄NPs selectively to the surface-bound AuNP, minimizing effects on the binding signal due to nonspecific adsorption. The SPR signal enhancement was demonstrated with bacterial cholera toxin. As displayed in Figure 2, for simplicity of the experiment, CT is covalently linked to surface via amide bond to demonstrate the feasibility of the overall amplification process. To eliminate nonspecific adsorption, the free activated carboxyl acid groups of MUA on the surface that were not attached to CT were blocked by incubation of 1 mg/mL BSA solution for 30 min. The presence of attached CT on the surface was initially recognized by injection and bonding of 0.25 mg/mL biotinylated anti-CT for 30 min. This was followed by injection of 0.5 mg/mL streptavidin to specifically bind to the exposed biotin tag. The captured streptavidins left additional binding sites available for further binding via incubation of AuNP–biotin. SPR resonance angle shifted 0.59° upon AuNP–biotin binding, and the signal was further increased 1.48° for Fe₃O₄NP–streptavidin incubation (Figure 3a). The combination of AuNP–biotin and Fe₃O₄NP–streptavidin attachment (2.07°) amplified the SPR degree shift to about 40 times more than the degree shift from the initial injection of 0.1 μM CT (0.05°). In the control channel, PBS buffer was injected instead of CT under the same conditions, followed by surface blocking with BSA and surface incubation with biotinylated antiCT, streptavidin, AuNP–biotin, and Fe₃O₄NP–streptavidin sequentially. There was no measurable SPR signal change for biotin–antiCT and streptavidin incubation and little SPR signal increase for AuNP–biotin incubation, demonstrating good specificity of signal enhancement. Unbound Fe₃O₄NP–streptavidin was not completely removed by surface rinsing due to its small size, which resulted in 0.10° from nonspecific adsorption (Figure 3b), but this signal increase is negligible compared with substantial specific binding signal (1.48°, Figure 3a).

Figure 2.

Sequence of surface modifications for SPR signal amplification: (a) immobilization of cholera toxin, (b) addition of biotin–antiCT and streptavidin, (c) attachment of AuNP–biotin to streptavidin, and (d) binding of Fe₃O₄NP–streptavidin to AuNP–biotin. (e) AFM image of Fe₃O₄NP–streptavidin and AuNP–biotin conjugate nanoparticles binding on sensorchip in the presence of 2.51 × 10⁻¹³ mol/cm² CT.

Figure 3.

SPR sensorgrams for (a) immunoassay of 2.51 × 10⁻¹³ mol/cm² CT with AuNP–biotin and Fe₃O₄NP–streptavidin nanoparticle for signal amplification, (b) the control channel in the absence of CT, and (c) Fe₃O₄NP–streptavidin nanoparticle binding on covalently immobilized biotin–BSA in the presence and absence of AuNP–biotin nanoparticles and streptavidin bridge.

To demonstrate the contribution of AuNP–biotin to Fe₃O₄NP–streptavidin accumulation, biotin–BSA was immobilized on surface and the signal from direct binding of Fe₃O₄NP–streptavidin to biotin–BSA was compared to the binding signal of Fe₃O₄NP–streptavidin in the presence of AuNP–biotin (Figure 3c). In the absence of AuNP, biotin–BSA only provided limited surface area for Fe₃O₄NP–streptavidin attachment, resulting in 0.74° of SPR signal increase. With streptavidin as a connection bridge and AuNP–biotin as building blocks, Fe₃O₄NP–streptavidin had a larger effective surface area to bind and generated 1.39° of SPR signal increase. This result confirms the initial hypothesis that the AuNP–biotin attached on the surface increases spatial positions for Fe₃O₄NP–streptavidin accumulation, which further amplifies the binding signal of Fe₃O₄NP–streptavidin by nearly a factor of 2.

Morphological Characterization of Fe₃O₄NP–Streptavidin and AuNP–biotin Nanoparticle Conjugate.

The physical accumulation of Fe₃O₄NP–streptavidin around AuNP–biotin was characterized by AFM, which can provide useful topographic information for surface immobilized nanoparticles. Figure 2e shows the AFM image of the surface features of the adsorbed Fe₃O₄NP–streptavidin and AuNP–biotin with 0.1 μM CT immobilization. Unlike individual dots isolated on the surface for sole AuNP binding,⁴⁵ here, a densely packed block conjugate was observed on the surface in networks of small particles, resulting in larger islands with thickness above 30 nm and average feature size over 100 nm. The average size of an individual small particle is about 10 nm, which approximates the sum of Fe₃O₄NP (5 nm)³⁰ and streptavidin (3–5 nm).⁴⁶ This confirms that Fe₃O₄NP–streptavidin accumulated around AuNP–biotin, fully occupying the exposed AuNP surface area and significantly enhancing nanoparticle packing density on the surface.

Highly Sensitive Detection of Protein Toxin.

CT is responsible for the deleterious effects of cholera infection, one of the most severe illnesses in developing countries. The lethal dose for cholera toxin in human is relatively low (LD₅₀ = 250 μg kg⁻¹), so there is a clinical need for accurate detection of CT with high sensitivity,⁴⁷ and ultrasensitive detection of CT has drawn considerable interest in recent years.⁴⁸ To demonstrate the overall sensitivity of the Fe₃O₄NP–streptavidin and AuNP–biotin SPR signal enhancement, CT was covalently immobilized on the surface in varying amounts. Figure 4 shows a plot of the SPR signal increase after AuNP/Fe₃O₄NP enhancement as a function of CT surface coverage, where the surface coverage of CT was calculated based on our formerly reported method.⁴⁹ The conjugated nanoparticle amplified signal shows two segments of linear correlation to CT concentration. When surface immobilized CT amount is between 2.51 × 10⁻¹³ and 7.94 × 10⁻¹³ mol/cm², the large protein amount generated substantial signal increase based on conjugated nanoparticle binding, resulting in a much steeper calibration curve than at lower concentrations. This type of segmental calibration had been observed by our previously reported method for CT detection⁴⁹ and by others for IgG fluorescence detection.⁵⁰ The other linear range is from 6.31 × 10⁻¹⁶ mol/cm² to 2.51 × 10⁻¹³ mol/cm² (R² = 0.932), which is more relevant for the desired range of ultrasensitive detection. Even at 6.31 × 10⁻¹⁶ mol/cm² (corresponding to CT concentration of 4.52 pM), the immobilized CT still generated discernible SPR signal change after conjugated nanoparticle binding. This concentration is much lower than most previously reported CT detection methods, including SPR signal amplification with ATRP polymer growth,⁴⁹ colorimetric method,⁵¹ fluorescently labeled microarray,⁵² and electrochemical biosensors that rely on lipid bilayer membranes supported on glass fiber⁵³ or with a microfluidic technique.⁵⁴ The detection limit at 3σ was 5.01 × 10⁻¹⁶ mol/cm², giving an absolute quantity detection limit of 0.25 fmol, which is comparable to a highly sensitive enzyme-catalyzed enhanced chemiluminescence reaction.⁵⁵

Figure 4.

SPR response as a function of CT surface coverage after AuNP–biotin and Fe₃O₄NP–streptavidin conjugate nanoparticle amplification with a streptavidin bridge.

SPR Imaging Characterization of Au-well Array and Signal Amplification Measurement.

The SPR imaging experiments were carried out using a home-built instrument arranged in the Kretschmann configuration.⁴⁰ The reflectance from the array was imaged with a CCD camera, and the change in reflectance was recorded in real-time. Contact printing,^48,56 microfluidic cross-patterning with two PDMS chips,⁵⁷ and automatic arrayers⁵⁸ have been used to generate protein arrays on flat gold substrates for SPR imaging. However, an effective blocking method is required to deactivate the unbound area to SPR and prevent background resonance and nonspecific adsorption from interfering with the images. Recently, we developed a microarray of 800 μm diameter gold wells surrounded by raised titanium platforms that attenuate the background evanescent field in the areas in between the wells.³¹ As demonstrated in Figure 5a, surface plasmons only resonated and absorbed light in the circular Au well areas, darkening those areas, whereas the boundary titanium areas do not show any effect from the s-polarized light. With the immobilization of 1.29 × 10⁻¹² mol/cm² CT, the discussed immunoassay binding of biotin–antiCT, streptavidin, and conjugated nanoparticles significantly increased the surface refractive index in the Au wells (Figure 5b). For clearer results and easier analysis, a difference image is taken between pre- and postassay (Figures 5a,b, respectively), and the high reflection intensity in Figure 5d from conjugate nanoparticle amplification clearly demonstrates CT immobilization. By comparison, a difference SPR image was captured for 1.29 × 10⁻¹² mol/cm² CT without antibody binding and conjugate nanoparticle signal enhancement, and the CT immobilization signal was barely observed (Figure 5c). The corresponding 3D (Figure 5e) and 2D (Figure 5f) profiles were generated for the array elements from Figure 5d, demonstrating the high reproducibility across the array and clean background.

Figure 5.

Raw SPR images of Au-well microarray (a) before CT immobilization and (b) after nanoparticles amplification of 1.29 × 10⁻¹² mol/cm² CT; difference SPR images of 1.29 × 10⁻¹² mol/cm² CT (c) before and (d) after nanoparticle amplification; (e) 3D and (f) 2D intensity profiles for the array elements of (c).

SPR imaging quantitative analysis of different concentrations of CT is shown in Figure 6. The displayed images are the difference images, and each image was a 2 × 2 array for CT concentration ranging from 3.16 × 10⁻¹⁵ to 1.29 × 10⁻¹² mol/cm². Reflection intensities after conjugated nanoparticle signal enhancement are listed. Figure 6a is a control array with no CT where the surface was blocked by BSA and subjected to the same experimental conditions for immunoassay binding and conjugate nanoparticle signal amplification. The control only showed a small signal (RI = 803 ± 52.3 au), suggesting the conjugated nanoparticle binding signal enhancement has high specificity to CT. The reflection intensity difference also exhibited pronounced response to increasing CT concentration. The highest signal amplification is obtained with 1.29 × 10⁻¹² mol/cm² CT immobilization (RI = 16540 ± 216.6), while 3.16 × 10⁻¹⁵ mol/cm² CT immobilization generated a signal increase of 1667 ± 107.2. Reflection intensity plotted against CT surface coverage demonstrates a linear relationship from 3.16 × 10⁻¹⁵ to 1.29 × 10⁻¹² mol/cm², and using 3 S/N cutoff, the detection limit was determined to be 2.0 × 10⁻¹⁵ mol/cm². Taking into account the individual Au-well surface area, the absolute quantity of detection limit is 3.2 amol, which is much lower than ultrasensitive electrochemical detection⁵⁸ and fluorescence detection with signal amplification by fluorescent nanoparticles,⁵⁹ and 10⁵ times more sensitive than direct CT detection on Au-well microarray via SPR imaging.⁶⁰ With the choice of appropriate biotinylated antibodies, this method can be applied to high-throughput detection of various proteins.

Figure 6.

Reflectivity as a function of CT surface coverage on microarray after Fe₃O₄NP–streptavidin and AuNP–biotin conjugate nanoparticle amplification. Inset: SPR images of corresponding coverages.

CONCLUSION

We report here a novel SPR signal amplification strategy for highly sensitive detection of proteins by stacking of differently sized nanoparticles. Immunoassay binding of biotin–antibody to the target protein and subsequent attachment of streptavidin is initially accomplished on gold substrate. This is followed by SPR signal amplification by two consecutive nanoparticle bindings of 20 nm AuNP–biotin and 5 nm Fe₃O₄NP–streptavidin, which greatly increases packing density on the surface and thus enhances SPR signal substantially. The immobilized CT has been detected in the range of 6.31 × 10⁻¹⁶ to 2.51 × 10⁻¹³ mol/cm² and has a detection limit as low as 6.31 × 10⁻¹⁶ mol/cm² (0.25 fmol). Additionally, SPR imaging detection of the amplified immunoassay was carried out using a novel Au-well microarray of 800 μm diameter gold wells surrounded by raised titanium platforms for attenuation of background evanescent field interference. The evanescent field is attenuated in the background area, and SPR resonance is restricted in gold wells, therefore avoiding background resonance interference and nonspecific adsorption. CT detection from the Au-well microarray demonstrated a linear relationship from 3.16 × 10⁻¹⁵ to 1.29 × 10⁻¹² mol/cm² with a detection limit of 2.0 × 10⁻¹⁵ mol/cm² and an absolute quantity of detection limit of 3.2 amol. With various nanomaterials and biotin antibodies, this method can be well applied to high throughput detection of a wide range of proteins.