Title: Gold−Silver Alloy Nanoparticle-Incorporated Pitaya-Type Silica Nanohybrids for Sensitive Competitive Lateral Flow Immunoassay

Journal: Analytical Chemistry

IF: 7.4

Original link: https://doi.org/10.1021/acs.analchem.3c03569

Reporter: Qiaoying Wu, Master of Grade 2022

Although the competitive lateral flow immunoassay (CLFIA) using gold nanoparticles (AuNPs) as labels has been widely adopted for the rapid detection of small molecules, its sensitivity is often constrained by the insufficient colorimetric signal produced by conventional AuNPs labels. Herein, we introduce a new type of intensified colorimetric label, denoted as SAAS, which is engineered by integrating gold−silver alloy nanoparticles (Au−Ag NPs) within a dendritic silica scaffold.These pitaya-type silica nanohybrids combine the advantages of the amplified molar extinction coefficient of alloy units with the signal collective effect of numerous Au−Ag NPs in a singular label. The SAAS-based CLFIA strips not only achieve qualitative screening of aflatoxin B₁ (AFB₁) at an extraordinarily low concentration of 0.2 ng/mL by the naked eye but also enable precise AFB₁ quantification through a smartphone, with a remarkable limit of detection of 0.00314 ng/mL. Moreover, by leveraging SAAS as a quencher, we have delved into transforming the conventional signal-off mode of competitive immunoassay into a signal-on configuration. This innovation led to the development of a fluorescent LFIA that augments interpretative precision and sensitivity. Our study demonstrates the substantial potential of the proposed nanohybrid labels in enhancing the sensitivity of CLFIA for detecting small molecules.

Immunoassay of hazardous small molecules in a point-of-need (PON) manner is of great importance, given its high accuracy and versatility that fulfill the requirements of on-site disease diagnosis, drug testing, and food safety monitoring. Although the enzyme-linked immunosorbent assay is commonly utilized for the sensitive and specific detection of small compounds, its time-consuming procedure, extensive washing steps, and reliance on skilled operators render it less optimal for rapid testing. In contrast, lateral flow immunoassay (LFIA) offers a simpler procedure that yields a visible signal output, enabling rapid detection in resource-scarce settings due to its straightforward operation and immediate visual interpretation of results. Among the colorimetric labels employed in LFIA, gold nanoparticles (AuNPs) are widely utilized due to their advantageous physical and chemical properties. The high molar extinction coefficient of AuNPs guarantees a robust color signal that can be readily perceived by the naked eye. Moreover, the uncomplicated conjugation of AuNPs with biological receptors, such as antibodies and nucleic acids, through electrostatic interactions or thiol-gold coordination, further enhances their effectiveness for PON applications.

In the immunochemistry, small molecules like mycotoxins are classified as haptens. In the LFIA, effective recognition of these small molecules is achieved through a competitive model. The outcome of a competitive assay relies on the disappearance of the colored test line (T-line), indicating a positive result. However, the application of this “signal-off” approach in responding analytes with low concentrations is often limited due to the insufficient and undistinguishable darkness alternation of colorimetric signal generated by AuNPs. Consequently, various signal amplification strategies have been proposed to enhance the sensitivity of competitive LFIA (CLFIA), such as catalyzing enzyme-mediated color development, increasing the signal intensity of labels, and transforming the detection into a “signal-on” pattern. Therefore, the development of a colorimetric label with potent absorbance attributes is of paramount importance for achieving the highly sensitive detection of small molecules in LFIA. In this study, we developed a highly sensitive colorimetric label of gold−silver alloy nanoparticles (Au−Ag NPs) incorporated silica colloids, referred to as SAAS, to detect small molecules in CLFIA. The combination of amplified molar extinction coefficient of alloy nanoparticles and assembly of multiple Au−Ag NPs into a single label synergistically enhances the colorimetric signal. Recently, Kim et al. reported a two-step method for assembling Au−Ag NPs onto the surface of solid silica microspheres. In contrast to irregular Au−Ag NPs synthesized in an aqueous solution, we utilized hydrophobic Au−Ag NPs with uniform particle sizes as signal units. The chosen carrier, depicted in Scheme 1A, was dendritic mesoporous silica (dSiO₂) decorated with a central radial pore structure. Unlike conventional solid silica templates, dSiO₂ offers an increased specific surface area for loading Au−Ag NPs units; meanwhile, the assembly process driven by metal−ligand coordination effectively preserves the inherent absorption properties of the nanoparticles. Additionally, an organic silica−silica bilayer coating was applied to improve the stability and dispersibility of the microspheres, making them more suitable for immunochromatographic probe conditions. Aflatoxin B₁ (AFB₁), a highly toxic and carcinogenic mycotoxin, is classified as a Class I carcinogen by the International Agency for Research on Cancer. In this work, AFB₁ was chosen as the model molecule to illustrate the detection principle using the developed immunoassay platform (Scheme 1B). In negative tests, the SAAS-mAbs/AFB₁-BSA immune complex formed on the T-line, resulting in a dark brown color. Conversely, in positive tests, the presence of AFB₁ in the sample solution led to the binding of the SAASmAbs probe to the target, hindering its interaction with the predeposited antigen. As a result, the T-line appeared colorless, indicating a positive result. By capitalizing on the signal amplification capability of SAAS-mAbs, the developed CLFIA strips not only enabled visual screening of AFB₁ at an exceptionally low concentration by the naked eye but also allowed for quantitative detection by reading the strip with a smartphone. Moreover, SAAS can be utilized as a quencher, facilitating a highly sensitive FLFIA with a signal-on pattern.

Scheme 1. (A) Process for the Preparation of SAAS-mAbs. (B) Schematic Illustration of Visual and Quantitative Detection of AFB₁ with CLFIA Strip

        1.     Characterizations of SAAS.

      The pitaya-type nanostructure was formed through the assembly of hydrophobic Au−Ag NPs with dSiO₂ templates (see Scheme 1A). Transmission electron microscopy (TEM) images revealed that the Au−Ag NPs have a spherical shape with an average diameter of approximately 10 nm (Figure 1A). The interplanar distance between adjacent lattice fringes was measured by Digital Micrograph 3.9 software to be 0.24 nm, corresponding to the (111) planes of the facecentered cubic (fcc) structure of Au or Ag (Figure 1E). X-ray diffraction (XRD) analysis further confirms the polycrystalline nature of the synthesized Au−Ag NPs (Figure S1). The dSiO₂ templates exhibited a dendritic structure with an average diameter of ∼240 nm (Figure 1B) and an average pore width of ∼35 nm (Figure 1F). Au−Ag NPs were directly assembled with the thiolated dSiO₂ templates in chloroform through thiol-metal coordination.37 The resulting dSiO₂/Au−Ag NPs nanocomposite (Figure 1C,1G) consisted of numerous compactly distributed Au−Ag NPs within the pore channels of dSiO₂. This nanocomposite was then subjected to an organosilica coating and silica shell encapsulation. Figure 1D,1H confirmed the complete filling of radial pores by the silica matrix, forming the spherical and smooth dSiO₂/Au−Ag NPs/SiO₂ (SAAS) structure. High-angle annular dark-field scanning TEM (HAADF-STEM) images (Figure 1I) and elemental mapping by X-ray energy dispersive scattering (EDS; Figure 1J) confirm the even distribution of Au and Ag throughout the SAAS structure.

Figure 1. (A) TEM and (E) HRTEM images of Au−Ag NPs. (B−D) TEM and (F−H) SEM images of dSiO₂ (B, F), dSiO₂/Au−Ag NPs (C, G), and SAAS (D, H). (I) HAADF-STEM image of SAAS. (J) EDS mapping of different elements (J1−J5) for a single SAAS nanosphere.

2.     Optimization of Loading Au−Ag NPs on dSiO₂.

       The loading rate of Au−Ag NPs onto dSiO₂ plays a pivotal role in determining the sensitivity of the nanolabels. The encapsulation efficiency of Au−Ag NPs approached near 100% when the feeding content (defined as the mass percentage of total Au− Ag NPs to dSiO₂) was below 150% (Figure 2B), as indicated by the nearly colorless supernatant (inset of Figure 2B). The loading content (defined as the mass percentage of captured Au−Ag NPs to dSiO₂) of Au−Ag NPs exhibited a proportional increase within the feeding content range of 50−200%, accompanied by higher packing densities of Au−Ag NPs in the nanoassemblies (Figure 2A). The absorbance of the dSiO₂/Au−Ag nanocomposite displayed a continuous rise as the loading content of the Au−Ag NPs increased. In contrast, the plasmonic absorption of SAAS initially increased and then decreased with higher Au−Ag NPs loading (Figure 2C). This phenomenon can be attributed to the effective reduction in interparticle distances between encapsulated Au−Ag NPs through the organosilica/silica shell. The closer proximity of the nanoparticles enhances the electromagnetic coupling between them, resulting in a subsequent decrease in absorbance at 450 nm. Based on these findings, a feeding content of 150% was selected for the assembly process.

Figure 2. (A) TEM images of dSiO₂/Au−Ag assemblies at different feeding content. (B) Relationships between Au−Ag NPs loading content and Au−Ag NPs feeding content for dSiO₂ hosts. The inset is a photograph of the Au−Ag NPs supernatant (top) and redispersed dSiO₂/Au−Ag assemblies in chloroform (bottom). (C) Relationship of absorbance ratios of dSiO₂/Au−Ag (black line) and SAAS (red line) at 450 nm with different loading content. A0 is the absorbance of dSiO₂/Au−Ag or SAAS with a feeding content of 50%.

3.     Verification of Signal Amplification.

Figure 3A demonstrates that the designed nanolabel offers signal enhancement through two key mechanisms. First, the incorporation of Au− Ag NPs with their higher molar absorption coefficient provides enhanced light absorption compared to conventional AuNPs. Second, the effective assembly of Au−Ag NPs within the dSiO₂ structure amplifies their collective plasmonic response, resulting in a more prominent colorimetric signal. As depicted in Figure 3B, the molar absorption coefficient (ε) of Au−Ag NPs at their maximum absorption wavelength was determined to be 1.31 × 108 M−1 cm−1 , which is ∼1.6 times higher than that of AuNPs. This indicates that the Au−Ag NPs possess a higher efficiency in light absorption and colorimetric signal generation compared with plain AuNPs. To verify the impact of the assembly on signal amplification, we plotted the absorbance of SAAS and Au−Ag NPs against their respective particle concentrations. As shown in Figure 3C, the absorbance of SAAS exhibited a linear enhancement with increasing concentration with a slope of 3.85 × 1011. In contrast, the plot for hydrophobic Au−Ag NPs displayed a slope of 1.98 × 109 , indicating that the absorbance of each SAAS was approximately 194-fold stronger than that of a single Au−Ag NP. The insets in Figure 3D,E displays the linear relationship between the signal intensity and the cumulative number of particles for each nanoparticle. In the conventional turn-off format, we defined the recognition rate (N/N0) as a metric to compare the visual sensitivity between the two types of nanoparticles. Here, N₀ and N represent the cumulative number of particles for negative (discernible by the naked eye) and positive (no detectable signal) results, respectively. Notably, accumulations of approximately 1.93 × 1010 AuNPs (Figure 3D) and 1.76 × 108 SAAS (calculated using the linear equation in Figure 3E) were found to generate the same visible signal of negative result. When the signal completely disappeared, the corresponding number of particles for AuNPs and SAAS were approximately 1.93 × 107 and 3 × 106 , respectively. Through calculations, it was determined that SAAS achieved a 17.04-fold higher recognition rate compared with commercial AuNPs.

Next, we compare the performance of SAAS and AuNPs in the LFIA format.When LFIA strips were inserted into known concentrations of mAbs-labeled SAAS and AuNPs, the nanolabels migrated along the NC membrane and were captured by the capture bioreceptor on the T-lines (Figure 3F,G). Notably, ∼3 × 109 AuNPs and ∼8.92 × 107 SAAS (calculated using the linear equation) were needed to generate the same visible signal of negative result. Conversely, less than ∼1.5 × 107 AuNPs and ∼4.84 × 106 SAAS are not sufficient to produce a detectable colorimetric signal. The 10.85 times higher recognition rate achieved with SAAS-mAbs compared with AuNPs-mAbs in the LFIA format aligns with the findings from the drop-cast test discussed above. These findings provide strong evidence that SAAS can indeed serve as highly sensitive colorimetric nanolabels for the detection of small molecules in a LFIA.

Figure 3. (A) Schematic illustration of signal amplification by SAAS. (B) Molar absorption coefficients of AuNPs and Au−Ag NPs at different wavelengths. (C) The linear relationships between the absorbance and the particle concentration of Au−Ag NPs and SAAS. Inset: the photographs of particles at different concentrations. (D, E) The 8-bit ImageJ-processed image (left) and signal intensity (right) obtained from NC membrane drop-casted with different particle numbers of AuNPs (D) and SAAS (E). Inset: the linear relationship between the signal intensity and the cumulative number of particles. (F, G) Schematic illustration of LFIA format (left) and signal intensity obtained from T-lines (right), with AFB₁- BSA used as capture ligand, after exposure to different particle numbers of AuNPs-mAbs (F) and SAAS-mAbs (G). Inset: the photograph of the LFIA strips (bottom) and the linear relationship between the signal intensity and the cumulative number of particles (top). The round circles in (D−G) denote the number ranges of AuNPs or SAAS needed to give positive results.

4.     Visual and Quantitative Assay for AFB₁.

Under optimal conditions, the visualization capability of the SAAS-based CLFIA strips was examined. Remarkably, as the concentration of AFB₁ increased, the brown color of the T-line gradually diminished until it completely disappeared (Figure 4A). The visible limit of detection (LOD) of the test strip was found to be 0.2 ng/mL. SEM images (Figure 4B) confirmed the migration of SAAS-mAbs on the T-line of the test strip at different AFB₁ concentrations. In a negative assay, numerous SAAS nanospheres were captured in the T-line, forming SAASmAbs-AFB₁ immunocomplexes, resulting in a dark brown color on the T-line under daylight (Figure 4B(i)). In a positive test where the AFB₁ concentration was 0.05 ng/mL, the analyte and precoated antigen competitively coupled with SAAS-mAbs, leading to fewer SAAS nanospheres being captured by the T-line, accompanied by a light brown color under daylight (Figure 4B(ii)). When the AFB₁ concentration increased to 0.2 ng/mL, no SAAS spheres were observed, and the T-line appeared colorless under daylight (Figure 4B(iii)), which was consistent with the visual observations made by the naked eye (Figure 4A).

For quantitative analysis, the signal intensities of the T-line and C-line were extracted from each image using ImageJ software (Figure 4C). With the increase of AFB₁ concentration, the signal intensity at the T-line gradually decreases until it disappears, while the signal at the C-line remains relatively stable. The ratio of signal intensity (IT/IC) showed an inverse correlation with the AFB₁ concentration, and a linear relationship was established in the range of 0.01 to 0.2 ng/mL, with a LOD of 0.00314 ng/mL.

We further compared the visual and quantitative performance of the SAAS-based CLFIA strip with those of the existing LFIA strips for AFB₁ detection. As depicted in Figure 5A, the visible LOD of the CLFIA strip was significantly lower than that of colorimetric and fluorometric LFIA. Moreover, in comparison with the conventional AuNPsbased LFIA for AFB₁ quantification, the SAAS-based CLFIA strip exhibited higher sensitivity. Although some fluorometricand surface-enhanced Raman spectroscopy (SERS)-based LFIA can detect lower concentrations of AFB₁, the current colorimetric CLFIA strip provides rapid, sensitive, and quantitative results in a more convenient manner.

Figure 4. (A) Photographs of CLFIA strips at different concentrations of AFB₁. The black arrow indicates the T-line identifying the visible LOD interpreted by 12 independent users. (B) Photographs (left) and SEM images (right) of the CLFIA strip membrane at the AFB₁ concentrations of (i) 0 ng/mL, (ii) 0.05 ng/mL, and (iii) 0.2 ng/mL. The blue arrows annotate SAAS spheres captured in the T-line. (C) Signal intensity of the Tline and C-line in CLFIA strips at different concentrations of AFB₁. (D) Relationship of the IT/IC ratio of the T-line and C-line of SAAS-based CLFIA strips with AFB₁ concentration. (E) The concentration dependence of the IT/IC ratio is in the low concentration range.

5.     Real Sample Analysis.

The specificity and stability of the SAAS-based CLFIA were thoroughly investigated to assess its feasibility for practical applications. To examine the specificity of the CLFIA strip for AFB₁ detection, various common aflatoxin analogues, including AFG₁, AFM₁, AFB₂, and AFG₂, were selected as negative controls. As depicted in Figure 5B, the value of IT/IC for AFB₁ was significantly lower than that of any of the negative controls and the blank control, even when the concentration of interferences was 100-fold higher than that of AFB₁. Furthermore, the presence of nontargets (AFG₁, AFM₁, AFB₂, and AFG₂) alongside AFB₁ had a minimal impact on the original signal of the strip. These findings affirm the high specificity of the SAAS-based CLFIA. To further assess the reliability of the SAAS-based CLFIA, we conducted analyses on two types of real samples, namely maize and soy sauce, that were randomly and artificially contaminated with AFB₁. Commercial AuNPs-based LFIA strips were also utilized for comparison. As depicted in Figure 5C, the visible LOD obtained for maize and soy sauce samples using the commercial AuNPs-based LFIA strips was 5 ng/mL. In contrast, the visible LOD determined by SAASbased CLFIA strips were 0.2 and 0.5 ng/mL, respectively, which was at least 10 times lower than that of commercial LFIA strips.

Figure 5. (A) Comparison of visualization and quantification performance of different LFIA for AFB₁ analysis (black circle: colorimetry; green circle: fluorometry; blue circle: SERS). Yellow region: only visualization; blue region: only quantification; pink region: both visualization and quantification. The threshold value (5 ng/mL) represents the maximum limit concentration of AFB₁ in the European Union dairy products. (B) Photographs and corresponding histograms for the specificity of CLFIA strips toward AFB₁ detection. The concentration of AFB₁ and other interferences are 0.1 and 10 ng/mL, respectively. (C) Photographs captured for commercial AuNPs-based LFIA (top) and SAAS-based CLFIA (bottom) strips at various concentrations of AFB₁ in maize and soy sauce. The black arrows indicate the T-line identifying the visible LOD interpreted by 12 independent users.

6.     AFB₁ Detection Using SAAS-Based FLFIA Strips.

       To further verify the versatility of SAAS, an FLFIA strip with a signal-on pattern using SAAS as a quencher was also prepared, in which the T-line was immobilized with a mixture of AFB₁- BSA and blue fluorescent microspheres (BFM), instead of only AFB₁-BSA as in the CLFIA strip. In the case of a negative assay, where AFB₁ is absent in the sample, SAAS-mAbs are captured by AFB₁-BSA preimmobilized on the T-line (top panel, Figure 6A). On the other hand, in a positive assay, little or no SAAS is captured on the T-line, resulting in a blue appearance under UV irradiation (bottom panel, Figure 6A). As the concentration of AFB₁ in the sample increases, the T-line appears bluer. The visible LOD of FLFIA was determined to be 0.1 ng/mL (Figure 6B), which is much lower than that of the colorimetric CLFIA strip with a signal-off pattern. Additionally, the B value at the T-line is proportional to the AFB₁ concentration (Figure 6C), and a linear relationship was established with a LOD of 0.457 pg/mL. This breakthrough in sensitivity makes SAAS-based FLFIA a powerful tool for the detection of lower concentrations of AFB₁.

Figure 6. (A) Schematic illustration by SAAS-based FLFIA strips. (B) Photographs of test strips at different concentrations of AFB₁ under daylight (left) and UV lamps (right). The white arrow indicates the T-line identifying the visible LOD interpreted by 12 independent users. (C) Standard curve for AFB₁ quantitative analysis by measuring the B value on the T-line.

In summary, the incorporation of SAAS as a novel colorimetric label has revolutionized the sensitivity and performance of the CLFIA for detecting AFB₁. The exceptional properties of SAAS, such as its amplified molar extinction coefficients and the assembly of Au−Ag NPs, have significantly enhanced the colorimetric signals in CLFIA strips. The assay demonstrates the capability to screen for AFB₁ at an extraordinarily low concentration of 0.2 ng/mL, surpassing the sensitivity of commercial AuNPs-based LFIA strips. Moreover, the SAAS labels enable precise and sensitive quantification of AFB₁ at concentrations as low as 0.00314 ng/mL by using a smartphone. Furthermore, this study demonstrates the versatility of SAAS as a quencher in FLFIA, facilitating a signal-on pattern for even more sensitive AFB₁ detection. The successful application of the SAAS-based detection system in real samples, such as maize and soy sauce, validates its potential for practical use in diverse environments, including resource-constrained settings. Overall, the SAAS-based CLFIA and FLFIA platforms represent a powerful and reliable tool for rapid and sensitive detection of hazardous molecules.