Ultrasensitive and Visible Detection of Influenza A Virus Based on Enzymatic Properties of Layered Gold Nanoparticles

Considering the urgent demand for reliable and rapid detection of infectious respiratory viruses during unpredictable pandemics, an innovative ultrasensitive colorimetric immunoassay for influenza A (H1N1) virus detection is developed herein. The proposed approach leverages dual amplification by combining layer‐by‐layer interactions with the nanozyme effect of biotinylated gold nanoparticles (BGNPs). BGNPs assemble around the target via repeated incubation cycles under optimized conditions, resulting in a layered structure that increases optical density, producing a more intense signal proportional to the viral titer. Additionally, the nanozyme effect of the layered BGNPs induces oxidation of 3,3',5,5'‐tetramethylbenzidine, which further enhances the visible signal detectable by the naked eye. This synergetic nanoprobe‐based system demonstrates remarkable sensitivity, with a limit of detection of 101.29 EID50 mL−1, which is 2500‐fold higher than that of commercial rapid kits and conventional enzyme‐linked immunosorbent assays, within a rapid 55 min timeframe. Furthermore, the anti‐interference capability and portability of the developed system reinforce its practicality, making it a promising tool for field diagnostic tests that offers advanced, ultrasensitive, and early detection of respiratory viruses.

highly sensitive, and portable measurement of the target biomolecule. [18,19]Immunodiagnostic methods based on hostguest interactions, such as lateral flow assays (LFAs) based on colloidal gold nanoparticles (GNPs), have been utilized as rapid diagnostic tests, as they offer simplicity, rapid turnaround times, and ease of interpretation. [20]Although LFAs are powerful tools for POCT, they do have limitations, such as low sensitivity and weak detection signal intensity.This is because the detection signal in the test line relies on a single substitutional binding between the colloidal gold-conjugated antibody and the target and is only detected when the signal from that binding exceeds a certain threshold.Therefore, new POCT technologies that identify low-abundance biomarkers in the early stages of infection and improve sensitivity through signal generation and enhancement are required as alternatives to existing diagnostic technologies, which have limited field use, low sensitivity, and a tendency for false negatives that prevent the development of effective control measures. [21,22]he detection of target biomolecules with high-intensity signals is critical to the performance of immunosensors.Researchers have explored various signal amplification strategies using nanomaterial-based signal-generating reporters.For instance, Huang et al. reported an immunosensor utilizing GNPs self-assembled around a target, and Lin et al. developed a gold nanorod probe for electrochemical sensing based on a superstructure immunocomplex. [23,24]These systems achieved high optical densities by converting a single target binding event into a cascade of numerous signaling molecule bindings.Therefore, the assembly of particles resulted in a more pronounced expression for signal, enabling highly sensitive target diagnosis.
Another approach involves the use of nano-biocatalysts to enhance the intensity of the signal generated by the target detection events; this approach primarily involves the use of enzyme complexes and inorganic nanomaterials with enzyme-like activity.Representatively, Yu et al. reported that paper electrode-based flexible pressure sensor for POC of carcinoembryonic antigen using platinum nanozyme which can efficiently catalyze the decomposition of H 2 O 2 to O 2 . [25]Additionally, they presented a strategy for implementing signal amplification and developing ultrasensitive nonenzymatic biosensors based on hollow Prussian blue nanoparticles with their enzyme-like peroxidase activity. [26]These researches have enabled considerable progress in the effective amplification of target substance detection signals by actively utilizing the enzymatic properties of nanomaterials.They also have achieved signal amplification with lower costs and greater stability compared to artificial enzymes.However, major barriers persist in controlling the sensitivity of nanobiocatalysts and in addressing false signal amplification due to nonspecific adsorption.Therefore, meticulous engineering of the structure and surface properties of nanoprobes is necessary to achieve high specificity for target substances and to maximize catalytic activity. [27]o address the need for a highly sensitive point-of-care influenza virus detection system with effective signal amplification, we developed a diagnostic system that leverages the unique properties of GNPs, which are well suited for surface engineering applications and exhibit strong catalytic activity.Especially, the unique surface property of GNPs, localized surface plasmon resonance, generates a strong optical signal, which is advantageous to optimize nanoparticles through sophisticated analysis.Simultaneously, ease of surface modification can be used to stably conjugate the desired biomolecules to the surface of the GNP.Therefore, we chose GNPs as signal probes, as all of their characteristics were beneficial for our intended system. [28,29]Our system comprises a signal amplification mechanism involving layer-by-layer (LBL) assembly of avidin-linked biotinylated gold nanoparticles (BGNPs) that utilizes the strong biotin-avidin binding affinity. [30,31]This LBL assembly strategy mediates target-specific events into a densely stacked configuration of BGNPs, significantly boosting signal output and consequently enhancing the sensitivity of the diagnostic system.Additionally, the layered BGNPs exhibit peroxidase-like activity, leveraging a catalytic property known as the nanozyme effect to further amplify the target-specific detection signal. [32,33]This dual-amplification strategy combines precise control of the surface properties of GNPs, target-specific amplification via avidinbiotin interactions, and the nanozyme effect, resulting in visible signal detection within a 1 h timeframe.Furthermore, our diagnostic system streamlines the process by eliminating the need for lysis pretreatment, and has the advantage of being able to detect intact viruses.By detecting the virus in its intact form, our system mitigates the risk of false-positive signals caused by residual nucleic acids that can remain in the body even after the virus has ceased to cause an infection. [34]This aspect addresses a limitation of conventional PCR-based nucleic acid diagnostic technologies, thereby complementing existing methods to more accurately assess active viral infections with increased precision and reliability.Our diagnostic assay involving dual-signal amplification achieved high sensitivity and selectivity without the need for further specimen processing and demonstrated outstanding sensitivity than other GNP-based and existing commercialized LFAs, underscoring its suitability as a robust POCT for virus detection.

Working Principle of LBL Signal Amplification using BGNPs
The LBL signal amplification system presented in this study operates based on the principles of antigen-antibody interaction, avidin-biotin binding, and nanozyme activity of BGNPs, as shown in Scheme 1.The assay comprised three consecutive steps, the first of which involved capturing the target virus.In this step, the viral sample and biotinylated polyclonal antibody were introduced sequentially to the plate where the monoclonal capture antibody was immobilized, resulting in the formation of a biotinylated sandwich complex via host-guest interactions.These binding events led to the formation of a substructure, that is, BGNP accumulation, with avidin functioning as a bridge at the first amplification step.The repeating cycle number of sequential avidin and BGNP treatments determined the amplification level in the red signal at a maximum absorbance peak (λ max ) of 524 nm.The intensity of the red hue on the plate showed dose-response signal production, enabling naked-eye detection.This color indicates the assembly of BGNPs around the target as a result of the first amplification, thereby correlating with the viral titer of IAV.The second amplification step, the catalytic signal-increasing stage, was induced by the enzymemimicking properties of the BGNPs.Upon exposure to hydrogen peroxide (H 2 O 2 ), the BGNPs catalyze the oxidation of colorless 3,3',5,5'-TMB to emit an intensified blue color signal.During this amplification stage, a more enhanced intensity of the blue color was observed, surpassing the red color emitted previously.After processing H 2 O 2 and TMB incubation for 15 min at room temperature is completed, a stop solution was used to terminate the enzymatic reaction of the BGNPs.Through this procedure, the oxidation of TMB was stopped and the signal appeared in yellow.These sequential steps generated an amplified signal proportional to the target amount by forming a supramolecular structure of the BGNP probe, acting as a nanozyme, near the target virus.Therefore, as illustrated in Scheme 1b, each stage of the process generates a visible signal that can be measured and discerned with the naked eye.The overall detection process involves capturing the target virus, layering the BGNPs to generate a detection signal, and using the nanoenzymatic properties of the BGNPs to enhance the signal, all of which can be completed within 55 min at room temperature (25 °C).

Nanozyme Activity of BGNPs
For successful signal amplification and sensitive viral detection, the surface properties of the BGNPs were thoroughly examined to determine their avidin-assisted BGNP layering effect and enzymatic activity.BGNPs were prepared by capping the GNPs with biotin PEG thiol (SH-PEG-biotin), which acts as a peroxidase upon the addition of H 2 O 2 and TMB, as shown in Figure 1a.GNP biotinylation is essential for avidin-bridged accumulation, the first signal amplification stage.However, excessive biotinylation may overly cover the surface of GNPs, causing unnecessary impedance of the peroxidase-like characteristics of BGNPs, which are essential for the second signal amplification stage.To perform the proposed assay, we synthesized and characterized GNPs and BGNPs.The PEG brushes on the BGNPs increased their hydrodynamic radii and zeta potentials, imparting colloidal stability and a neutral capping layer (Figure 1b,c). [35]he characteristic peak shift from 520 to 524 nm without any aggregation confirmed the successful synthesis of BGNPs (Figure S1a, Supporting Information). [36,37]During the biotincapping process, the BGNPs maintained their spherical morphology and monodispersity (Figure S1b, Supporting Information).The capping of SH-PEG-biotin in the GNP was confirmed by identifying the characteristic absorption peaks of PEG and biotin using Fourier-transform infrared (FTIR) and nuclear magnetic resonance (NMR) spectroscopy (Figure S2, Supporting Information).Instead of directly conjugating biotin to the GNPs, we utilized PEG chains as linkers to provide colloidal stability and prevent the aggregation of the GNPs, which can be induced by various salts or proteins in the biological environment.The colloidal stability of BGNPs was investigated in NaCl, phosphate-buffered saline (PBS), and 0.3% bovine serum albumin (BSA) buffer conditions.The aggregation level was assessed by calculating the ratio of the absorbance peaks at 520 (OD 520 ) and 630 nm (OD 630 ) based on the properties of the GNPs.Upon aggregation, the absorbance spectrum was found to be accompanied by a redshift (Figure S3, Supporting Information). [38]These results showed that the BGNPs maintained their original red color and absorbance spectrum with a low absorbance ratio (OD 630 /OD 524 ), whereas GNPs exhibited a color change to gray and an increase in the absorbance ratio (OD 630 /OD 520 ) in NaCl and PBS buffer conditions, indicating aggregation. [39]In summary, we synthesized BGNPs with high colloidal stability in a biological environment that successfully enhanced LBL signal amplification.Subsequently, we conducted the meticulous optimization of BGNPs to promote their enzyme-mimicking properties.As BGNP-mediated TMB oxidation occurs on the BGNP surface, the surface area and degree of accessibility of the substrate to the exposed surface are important factors in determining the efficiency of detection signal amplification. [40]In this regard, the diameter of the GNPs and the SH-PEG-biotin capping ratio were considered the main factors affecting the nanozyme effect of BGNPs.We prepared GNPs with diameters of 40 and 60 nm, which showed λ max of 530 and 540 nm respectively, to compare and optimize the nanozyme activity of 12 nm GNPs (Figure S4, Supporting Information).We added TMB and H 2 O 2 to each GNP solution and calculated the degree of oxidation from the signal ratio (optical density [OD] at λ 652nm /OD at λ max of each particle).GNPs showed a higher signal ratio at smaller sizes owing to an increased ratio of volume-to-surface area, with smaller GNPs providing a larger active site while containing the same amount of gold (Figure 1d). [41]GNPs of all three sizes exhibited TMB oxidation, but notably, 12 nm GNPs demonstrated comparatively heightened efficiency with a signal ratio of 3.99.In contrast, GNPs of other sizes failed to attain this level of reactivity.Therefore, we selected 12 nm GNPs that showed the highest TMB oxidation for the target detection experiments.The signal ratio of BGNP-mediated TMB oxidation was then examined at different BGNP and SH-PEG-biotin capping ratios, and the results are shown in Figure 1e.The BGNP capping ratio was calculated based on the mass ratio of gold to PEG-biotin, and BGNP showed the highest TMB oxidation at a 1:0.1 mass ratio of gold to PEG-biotin, which was determined to be the optimal secondary amplification activity condition.Furthermore, we determined the optimal concentration of substrates, including H 2 O 2 and TMB, and the incubation time with BGNPs to maximize the signal (Figure S5, Supporting Information).For H 2 O 2 concentration, a value of 5 M, which exhibited a sharp difference from the previous concentration range, was selected.Additionally, to maximize the signal, the highest concentration of TMB used as a substrate was optimized for incubation at 15 min, corresponding to the saturation time for OD 652 .
Then, we examined the nanozyme activity of BGNPs.As shown in Figure 1f, the BGNPs produced a visible signal at 652 nm under enzymatic reaction conditions requiring both H 2 O 2 and TMB.The BGNP-induced TMB oxidation is highly specific to certain conditions and generates a more intense signal than BGNPs at 524 nm, which is advantageous for detecting trace amounts of targets.As shown in Figure 1g, we confirmed that BGNP-induced TMB oxidation and BGNP concentration showed a strong linear relationship, which implies that the BGNPs accumulated after the target capture and layering stages of BGNPs will accelerate detection signal amplification.

Evaluation of First and Second Amplification on the Avidin Immobilized Plate
After establishing the optimal BGNP conditions for stable nanozyme activity in a biological detection environment, we explored strategies to amplify the detection signal through effective layering with BGNPs.In the layering process based on the bridging of avidin and BGNP, steric hindrance, which is affected by the length of the PEG chain and SH-PEG-biotin grafting density, was thought to be the main factor determining performance. [37,42]To verify this, BGNPs were fabricated using SH-PEG-biotin polymers with molecular weights of 0.4, 1, 3.4, and 5 kDa.The grafting density of SH-PEG-biotin on the BGNP surface was analyzed via thermogravimetric analysis (TGA), and was found to decrease with increasing polymer chain length (Table S1a, Supporting Information).This result was attributed to an increase in conformational entropy as the chain length of the PEG molecule increased; consequently, the number of SH-PEG-biotin molecules grafted onto the GNPs decreased with increasing PEG molecular weight. [35]As the avidin-BGNP binding event is strongly influenced by the number of biotin molecules, the extent of SH-PEG-biotin capping of BGNP can be a determinant of the layering efficiency.We observed that the lower the PEG molecular weight, the higher the BGNP-derived OD 524 in multiple repeated cycles of BGNP accumulation (Figure S6a and Table S1b, Supporting Information).We further assessed the TMB oxidation efficiency of the accumulated BGNPs at cycle 4 and found that BGNPs capped with lower PEG molecular weights potentiated a higher level of TMB oxidation (OD 652 ) (Figure S6b, Supporting Information).Even though short PEG chains in SH-PEG-biotin showed the highest signal production, BGNPs prepared with 0.4 kDa PEG chain length exhibited a slightly aggregated absorbance spectrum (Figure S7, Supporting Information).This can be explained by the fact that the colloidal stability of BGNPs, provided by PEG-derived steric hindrance, is insufficient in the 0.4 kDa PEG chains, resulting in the aggregation of BGNPs. [43]These results led us to select SH-PEG-biotin with a molecular weight of 1 kDa to maximize both the detection efficiency and the stability of BGNPs.
Following the demonstration of the signal amplification strategy, we assessed the nonspecific binding effect in the process associated with the potential risk of false positives.The BGNP layering process was tested and compared with two control groups for direct comparison: layering of BGNPs without primary avidin immobilization (control 1) and layering of GNPs capped with thiol biotin (SH-PEG) to obtain PGNPs with primary avidin immobilization (control 2) (Figure 2a).For these three cases, the absorbance was monitored at OD 524 of BGNP and OD 652 of BGNP-mediated TMB oxidation in repetitive cycle numbers from 0 to 7. First, to determine the optimal detection time, we tested three layering processes at different incubation times in a cycle, ranging from 1 to 15 min, for avidin and BGNPs.As the incubation time increased, the BGNP layering process produced a higher-intensity signal, with OD 652 signal exposure being particularly strong from 5 min onward (Figure S8, Supporting Information).
To assess the correlation between BGNP accumulation and TMB oxidation, H 2 O 2 and TMB were treated after each cycle and incubated together.In contrast to the two controls, BGNP layers with primary avidin immobilization showed continuous signal amplification and a positive correlation as the number of cycles increased.The nanozyme effect was observed to vary depending on the amount of accumulated BGNPs, resulting in strong optical densities at OD 524 and OD 652 , emitting vivid red and blue colors visible to the naked eye, respectively.This result was attributed to the assembly of BGNPs by avidin-biotin binding, whereas in controls 1 and 2, the cycle failed to initiate the layering process because of the absence of primary avidin immobilization and biotin capping on the GNPs, respectively, resulting in an absence of signal production in either control (Figure 2b,c).This suggests that layering of BGNPs occurs without nonspecific adsorption and that an amplified signal can only be obtained in the presence of the target.This demonstrates that our diagnostic assay can achieve quantitative detection using amplified signals.Notably, in the case of BGNPs layering at cycle 1, the color signal is emitted from the BGNPs monolayer formed by avidin, but from cycle 2 onward, as avidin and BGNPs are processed over the monolayer, a much stronger signal is gradually observed due to the formation of a larger assembly.This multilayer structure mediated by increase in the number of BGNPs induces a more active nanozyme effect and contributes to second signal amplification.

Virus Detection Analysis
This diagnostic system utilizes a sandwich complex that includes target capture and signal amplification based on host-guest interactions to detect IAVs. [44]To evaluate this complex, immunowells were coated with an H1N1-specific monoclonal capture antibody, and BSA was used to block the remaining empty space and prevent nonspecific adsorption.We speculated that upon loading the targeted IAVs into the prepared wells the IAVs were captured and formed sandwich complexes that sequentially bound with a biotinylated detection antibody (dAb).The preparation and sandwich complex formation stages were confirmed by image analysis using atomic force microscopy (AFM) (Figure S9, Supporting Information).The successful formation of the sandwich complex not only plays an important role in target capture but also implies that biotin is present on the surface of the complex for subsequent detection signal amplification stages.Optimization of the conditions that increase the S/N ratio is essential to maximize signal intensity and achieve high sensitivity for IAV detection.Therefore, before evaluating the sensitivity, we selected and tuned several important factors, including the incubation time of the target and dAb, ratio of BGNPs to treat avidin in each cycle, and BGNP concentration (Figure S10, Supporting Information).Short turnaround times are important in point-of-care diagnostics; as sensitivity and turnaround time are generally tradeoffs, it is important to find a balance between the two factors in the development of point-of-care diagnostics. [45]After evaluating the sensitivity of the assay under different incubation time conditions for virus capture, we determined that an incubation time of 10 min was optimal for obtaining high S/N values and did not overly increase the overall detection time (Figure S10a, Supporting Information).Furthermore, we determined that a key factor for maximizing the signal and minimizing nonspecific adsorption was the concentration of BGNP and avidin added in each cycle for a layering effect.We investigated the S/N values of the H1N1 detection results for various molar ratios of avidin to biotin coated on multiple concentrations of BGNPs (Figure S10b and S10c, Supporting Information).We selected a molar ratio of avidin to biotin of 1:2 and a BGNP concentration of 0.15 mg mL À1 as the optimal conditions to achieve high S/N values.Under carefully optimized conditions, the accumulation of BGNPs around the captured virus at each cycle was confirmed by scanning electron microscopy (SEM) at 10 5.7 EID 50 mL À1 H1N1.And it was confirmed that a larger assembly of BGNPs was formed in cycle 2 compared to that in cycle 1 (Figure S11, Supporting Information).
Subsequently, we evaluated the sensitivity of our assay to H1N1 stock solution.The detection results of each amplification stage in repetitive cycles 1-3 were measured and photographed (Figure 3a and S12, Supporting Information).We observed that the values of OD 524 and OD 652 increased with virus titer, indicating that each amplification was successful and provided a dose-response read-out signal, enabling quantitative target virus detection.Notably, the signal in the blank wells showed a significantly lower OD value, indicating that the assay generated a visible detection signal specific for the formation of a complex of the target virus, dAb, and avidin-BGNP bridges.For secondary amplification, the reaction was stopped by adding stop solution (5 M sulfuric acid [H 2 SO 4 ]) after 15 min of TMB oxidation.Subsequently, we calculated the S/N values of the test results from cycles 1 to 3 and obtained a high S/N value in cycle 2, which was selected as the optimal repetitive cycle for H1N1 detection (Figure 3b).Notably, the S/N of cycle 3 is lower than that of cycle 2.This outcome arises from an optimization process that considers the trade-off relationship between reducing the time required for detection and maximizing diagnostic signals.Consequently, the diagnostic signal greatly increased as the number of cycles increased, but it was inevitable that background signal would be also slightly generated.In the dual-amplification step from cycles 1 to 3, the calibration curves determined the LOD for each cycle (Figure S13, Supporting Information).Additionally, we used measurement tools, including ImageJ software 1.47v and the RGB Color Detector application (Android), to analyze the detection results considering visible detection situations (Figure S14-S16, Supporting Information).An LOD of 10 1.29 EID 50 mL À1 was measured by the microplate reader in cycle 2, indicating considerably higher sensitivity compared to other infectious virus diagnostic approaches that employ inorganic materials for colorimetric detection (Table S2, Supporting Information).The linear dependence between OD 450 and the titer of the target virus in Figure 3c also suggests that quantitative detection with good regression is possible over a wide dynamic range (10 5.7 -10 0.7 EID 50 mL À1 ) encompassing preclinical to symptomatic patients.Additionally, we compared the detection performance of different measurement tools (Table S3, Supporting Information) and found that they all had a wide analytical range with high correlation and similar LOD levels.Notably, the photographed test results measured using the smartphone application showed the lowest LOD The specificity analysis of the assay against a range of respiratory viruses was conducted at cycle 2.The virus samples tested had the following titers: H1N1 had a titer of 10 5.7 EID 50 mL À1 ; and H3N2, H9N2, influenza B virus (IBV), Newcastle disease virus (NDV), and human coronavirus (NL63) had titers of 10 7.25 EID 50 mL À1 , 10 5.7 EID 50 mL À1 , 10 4.25 EID 50 mL À1 , 10 7.5 EID 50 mL À1 , and 10 4.1 TCID 50 mL À1 , respectively.e) S/N value of our assay and conventional methods was plotted according to the virus titer.Each value was normalized after subtracting the blank from each readout-signal intensity.f ) Comparison of the results of H1N1 detection using our assay at cycle 2 with those from conventional diagnostic assays (ELISA and a commercial rapid kit) A gray background indicates that it is discernible to the naked eye in our assay.g) Comparison of our assays with conventional diagnostic assays.The limits of detection (LOD) for our assay and ELISA were determined using the OD 450 read-out signal obtained from a microplate.Data (a-e) represent mean AE standard deviation with n = 3.
(10 0.68 EID 50 mL À1 ) in cycle 2, suggesting the applicability of this system for point-of-care diagnosis.
High selectivity to detect only the target virus is essential to address the potential risk of obtaining false-positive results in this diagnostic system.Therefore, we investigated the selective diagnostic ability of our BGNP-layering detection system for IAV (Figure 3d and S17, Supporting Information).Target H1N1 virus and other nontarget viral strains (H3N2, H9N2, IBV, NDV, human coronavirus, and blank (PBS as negative controls)) were evaluated under the same conditions.In cycles 1 and 2, we observed visible detection of the target virus, H1N1, whereas no visible detection signal was generated for the other nontarget viruses, confirming that our system could specifically detect IAV based on the serial amplification of BGNPs.Furthermore, to validate the superiority of the developed diagnostic system over existing approaches, we compared its detection performance with that of a conventional ELISA using a similar detection process and a GNP-based commercial rapid kit that is widely used for viral detection.The titer-dependent signal produced in our assay was an intense visible signal (Figure 3f ).In contrast, rapid kits, which are specifically designed to facilitate naked-eye detection, showed much lower sensitivity for optical signals than this assay.For a more accurate comparison, the intensity obtained using ImageJ software was analyzed based on the S/N ratio, and it was confirmed that our assay had a wider detection range than the commercial kit (Figure S18, Supporting Information).Additionally, our system showed higher sensitivity and a wider analytical range than .Real sample analysis.The detection performance of this assay was evaluated using various biological samples, including saliva and nasal solutions, and was compared with that of commercial rapid kit.a) Schematic illustration of workflow for real sample analysis.b,c) The detection signals of H1N1 in PBS, saliva, and nasal solutions were shown corresponding to their respective virus titers (10 5.7 EID 50 mL À1 , 10 3.7 EID 50 mL À1 , and blank, respectively) and photograph image was analyzed with Image J software.d,e) IAV detection using commercial rapid kits for three separate cases, each prepared using a different biological sample.f ) Signal recovery from environmental samples using our assay and a commercial rapid kit was calculated, with normalization based on the signals obtained from PBS."N/A" indicates that the calculation is not applicable.Data (c,e) represent mean AE standard deviation with n = 3. S19, Supporting Information).The detection performances of these diagnostic methods were compared, as shown in Figure 3e,g.The results suggest that our proposed system enables more sensitive detection within 1 h and with a wider dynamic range than ELISA and commercial rapid kits, underscoring its potential as a naked-eye POCT.

Real Sample Analysis
A major challenge in the practical application of diagnostic technologies is the presence of many nontarget proteins in analytes collected from test subjects, can interfere with accurate target detection and necessitates additional purification processes. [46]In this regard, we evaluated the detection performance of the assay using realistic test samples to assess its resistance to potential interference from other substances.Because IAVs are mainly transmitted through biofluids in the respiratory tract, we examined its anti-interference ability in saliva and nasal fluid samples.The environmental mock-up samples were prepared by mixing H1N1 virus stock with biofluid and assayed at two virus titers (10 5.7 and 10 3.7 EID 50 mL À1 ) and blank (PBS).We compared the detection signals obtained using our assay with that of the commercial rapid kit through analysis by Image J software as illustrated in Figure 4a.The detection signals obtained by our assay in the saliva and nasal fluid environments were similar to those obtained in PBS buffer, which contained no external proteins, as shown in Figure 4b,c and S20, Supporting Information.The assay results were similar in both the PBS and biological samples.These results confirm the high anti-interference ability of this assay and suggest its potential for field use.In comparison, the test results of the commercial rapid kit showed a lower intensity in biological samples than in PBS, as presented in Figure 4d,e.The difference arises from the fact that, while the rapid kit relies solely on a single host-guest interaction, our assay constitutes a system enabling selective signal amplification through both biotin-avidin binding and host-guest interaction.Hence, even in the presence of interference, signal amplification can proceed seamlessly.In the case of the nasal fluid samples from the commercial rapid kit (Figure 4d), even the C-line was inconsistent, which was presumed to be due to the viscosity of the fluid affecting the capillary force and preventing smooth lateral flow. [47]As indicated by the recovery ratio in Figure 4f, the developed assay demonstrated greater resistance to interference with biological samples than a commercial rapid kit.

Application as a Portable BGNP-Based Diagnostic Platform for On-site Detection
One of the most important features of POCT biosensors in resource-constrained settings is their user-friendliness. [48]onsidering this perspective, we expected that the use of immunoplate in practical applications could pose a complexity for general users, along with the multiple steps.Therefore, we confirmed that our system has the potential to be used with simpler operating tools and is applicable as a portable system for onsite diagnosis.We used a microtube for IAV detection because it is small and transparent, which makes it easy to use in colorimetric detection.Moreover, it has a solid substrate that can immobilize antibodies and layer BGNPs without concerns regarding capillary force loss owing to several washing steps, unlike paper, another useful material. [49]As shown in Figure 5a, a series of amplification processes was conducted in the tube in the same manner as in the immunoplate well, and the signal analysis process was further simplified by taking a picture using a smartphone and analyzing it using the RGB Color Detector smartphone application on the App store.The optical signal was analyzed using the Y intensity of the CMYK colorimeter, which has four channels: cyan (C), magenta (M), yellow (Y), and black (K), as supported by the application.For effective antibody immobilization on a solid substrate, we performed (3-aminopropyl)triethoxysilane (APTES) functionalization to improve the orientation of the monoclonal antibody anchored on the tube. [50]When the detection process was performed in the prepared tube, a significant detection signal was obtained due to the accumulation of BGNPs, as observed in the immunoplate (Figure 5b).The test results observed with the naked eye differed slightly between immunowells and tubes, which we attributed to the structural differences between the curved tubes and relatively flat wells.However, in all analyses conducted using a smartphone, microplate reader, and Image J software, we were able to obtain detection signals corresponding to the virus titer (Figure 5c-e).Notably, the LOD of our portable assay, calculated by image extraction using a colorimetric smartphone application, was 10 1.48 EID 50 mL À1 , showing a similar performance to that on flat wells and a more sensitive detection than the commercial rapid kit (Figure S21, Supporting Information).The results demonstrate that our designed system is applicable as a high sensitivity, user-friendly portable tool that can potentially be used in POCT.

Conclusions
In summary, we developed a highly sensitive colorimetric detection system using a dual-signal amplification strategy that captures the target virus, creates a layered BGNP structure on its surface, and leverages the enzymatic activity of the BGNP complex.We conducted comprehensive engineering by employing the surface properties of GNPs, aiming to maximize their nanozyme effect while ensuring ample interaction with avidin through biotin modification for layering.By enhancing the target detection signal through repeated incubation cycles of the avidin and BGNP pair and achieving secondary signal amplification via the nanozyme effect of the complex, we effectively addressed the inherent limitation of the weak signal intensity observed in conventional GNP-based diagnostics.Our approach demonstrated a high level of sensitivity, with a low LOD of 10 1.29 EID 50 mL À1 , making it 2500 times more sensitive than conventional diagnostic systems.Furthermore, because this system targets the viral surface proteins, it enables detection of intact viruses without the need for lysis or pretreatment, making it highly convenient (Figure S22, Supporting Information).Additionally, the diagnostic performance was validated in both flat wells and microtubes.This application aimed to enhance user-friendliness by simplifying the multiple steps and improving accessibility, for the general user, to overcome the shortcomings of using immunoplate as a POCT tool.As a result, it demonstrates its potential for adaptation to different hardware setups suitable for field use while retaining high sensitivity.In conclusion, our diagnostic system not only shows high sensitivity and selectivity in practical diagnostic settings, but also has great potential as a POCT that enables on-site diagnosis, as it only requires highly portable hardware and smartphones.Furthermore, its application can be extended in the diagnosis of other infectious disease viruses, and can thus contribute significantly to the effective management of infectious diseases in the field, establishing it as a diagnostic system applicable to future pandemics.
Synthesis of Size-Dependent GNPs: Gold nanoparticle seeds with a diameter of 12 nm were synthesized using the citrate reduction method. [51]In brief, 1 mL of gold (III) chloride solution (100 mM) was added to 100 mL of sodium citrate aqueous solution (5 mM), which was refluxed in an oil bath at 95 °C with vigorous stirring for 20 min.A red color was obtained after refluxing and stabilizing for 80 min.The GNP seed solution was centrifuged 3 times at 7000 g for 10 min and stored at 4 °C after cooling to room temperature (25 °C).Subsequently, GNPs with diameters of 40 and 60 nm were synthesized from 12 nm GNP seed solutions using hydroquinone as a reducing agent. [52]Furthermore, 100 μL of gold (III) chloride solution and GNP seed solution (250 mM) were dispersed in Milli-Q water to obtain a final volume of 10 mL with stirring at 600 rpm.The volume of GNP seed solution was 1.024 and 0.25 mL for 40 and 60 nm GNPs, respectively.Subsequently, 22 μL of sodium citrate was immediately injected after adding 100 μL of hydroquinone (30 mM) and incubated at room temperature for 4 h.Each solution was centrifuged 3 times at 8000 g for 10 min with addition of 0.01% w v À1 PVP solution to stabilize the GNPs.The resulting size-dependent GNP solutions were stored in Milli-Q water at 4 °C for further use.
Fabrication and Characterization of BGNPs: GNPs with a diameter of 12 nm were modified using SH-PEG and SH-PEG-biotin via conventional methods.Each polymer was dissolved in Milli-Q water at a mass ratio 1:0.1 with GNPs.An equal volume of PEG solution was added dropwise to the GNP solution, which was stirred at 450 rpm.The mixture was stirred overnight at room temperature with gentle swirling.The thiol groups of each PEG solution were covalently attached to the gold surface and functionalized the GNPs to PGNPs and BGNPs.The unbound PEG solution was removed by several rounds of centrifugation at 7000 g for 10 min.The final solutions were dispersed in Milli-Q water and remained stable for several months.The size distribution and zeta potential of the BGNPs were compared with those of bare GNPs via dynamic light scattering and zeta potential analysis (ELSZ-2000ZS; Otsuka Electronics, Osaka, Japan).The UV-vis spectrum and OD indicative of successful BGNP synthesis were measured using a Spectra Max i3x microplate-reader (Molecular Devices, CA, USA).The spherical morphology and monodispersity of the BGNPs were visualized using transmission electron microscopy; the images were obtained using a JEM-F200 transmission electron microscope (JEOL, Tokyo, Japan).The chemical bonding of BGNPs due to PEGylation by SH-PEG-biotin was confirmed using FTIR (PerkinElmer, MA, USA) and NMR spectroscopy (Bruker, Bremen, Germany).Characteristic absorption peaks were observed for each measurement.A thermogravimetric analyzer (SDT Q600; TA Instruments, DE, USA) was used to calculate the grafting density of each BGNPs capped with SH-PEG-biotin of different molecular weights.Weight loss of BGNPs synthesized using 0.4, 1, 3.4, and 5 kDa SH-PEG-biotin was measured via TGA under nitrogen gas with a constant heating rate of 10 °C min À1 and temperature range of 50-600 °C.The grafting density (σ) of PEG chains on GNPs was calculated using the following formula: where R is average radius, N A is the Avogadro number, and W is weight fraction.
Layering Test in Avidin Immobilized Plate: To prepare an avidin immobilized well for use in the control and BGNP layering, 100 μL of 10 μg mL À1 avidin was incubated at 37 °C for 1 h.Subsequently, the well was washed 3 times using 200 μL 0.01% Tween 20 w v À1 .PGNPs and BGNPs were diluted to 0.075 mg mL À1 and incubated in 100 μL for 5 min for each experimental case.This was followed by the addition of 10 μg mL À1 avidin solution with the same volume and incubation time.This cycle was repeated 7 times, and the washing step was performed in the same manner for all sample additions.For signal measurement, the OD 524 of each experimental case (the first amplification signal) was measured.Subsequently, 33 μL of TMB and H 2 O 2 at optimized concentrations were added to wells containing 33 μL of Milli-Q water to induce catalysis, and the OD 652 was measured to confirm secondary amplification.
Preparation of Viruses: The IAV H1N1 (A/California/04/2009) with its subtypes H3N2 (A/Philippine/2/82) and H9N2, and NDV were propagated in the allantoic cavity of 11 d old embryonated chicken eggs.Each virus stock (100 μL) was inoculated into the cavities of chicken eggs.After incubation for 72 h at 37 °C, the eggs were chilled overnight at 4 °C.The allantoic fluid containing the propagated viruses was harvested and purified by centrifugation at 4000 rpm for 20 min.The supernatants were frozen at À80 °C for long-term storage and future use in the assay.Human coronaviruses NL-63 (HCoV-NL63/CN0601/14) and Flu B (influenza B/Seoul/32/2011[Yamagata lineage]) were propagated in Madin-Darby canine kidney cells.After incubation for 48 h at 37 °C, the cells were frozen and thawed at À80 °C once.The supernatant containing the propagated viruses was collected and purified by centrifugation at 4000 rpm for 20 min, transferred to new test tubes, and frozen at À80 °C for long-term storage.The prepared viral stocks were used directly in the assay.
Virus Detection Analysis: The H1N1 virus was detected by BGNP layering under optimized conditions.To anchor the capture antibody to the immunoplate, 100 μL of 0.1 μg mL À1 monoclonal antibody was incubated at 37 °C for 1 h and washed 3 times with 200 μL DPBS.Subsequently, to prevent nonspecific adsorption of avidin and BGNPs, 200 μL of 0.3% BSA solution was incubated for 1 h at 37 °C and washed 3 times in the same manner.For the preamplification step, 100 μL of serially diluted H1N1 virus (10 5.5 -10 À0.3 EID 50 mL À1 ) was added to the capture antibody immobilized plate.Additionally, 100 μL of 0.02 μg mL À1 polyclonal antibody was added for detection.The preamplification mixture was incubated for 10 min at room temperature and washed 3 times, as described above.Avidin and BGNPs were prepared and added at a molar ratio of 3:1 on a preamplified plate for first amplification.Each treatment was performed for 5 min at room temperature, and cycle 1 was completed by washing 3 times with 0.01% Tween 20 washing buffer at each step.When proceeding to cycle 2, 100 μL of Milli-Q water was added to the plate and the first amplification value was measured at OD 524 .After the measurement of first amplification, the plate was emptied, and Milli-Q water, H 2 O 2 , and TMB were added sequentially at 33 μL.After incubation for 15 min, the second amplification value was measured at OD 652 .Subsequently, 100 μL of 5 M H 2 SO 4 was added to the plate as stop solution, and the final signal was confirmed at OD 450 .The OD was measured at each step using a microplate reader and photographed using a smartphone.The LOD for each amplification and cycle was evaluated as previously described.This procedure was used for several viruses to evaluate the selectivity of the BGNPlayering detection system.
Detection of H1N1 in Biological Solutions: To confirm the target virus detection and anti-interference ability of the biological solution, saliva and nasal solutions were prepared.The H1N1 virus was diluted in PBS, saliva, and nasal solution stocks to achieve concentrations of 10 5.5 and 10 3.5 EID 50 mL À1 .The 100 μL of prepared target virus solution was applied to an immunoplate with an anchored capture antibody.After incubation for 1 h at 37 °C, the immunoplate washed 3 times with 200 μL DPBS and detected in cycle 2 by the BGNP-layering detection system.For comparison with the commercial rapid kit, the H1N1 virus, diluted in the same biological solution, was mixed with 100 μL of the elution buffer included in the commercial rapid kit.At this time, the virus titer of the resulting solution was 10 5.5 and 10 3.5 EID 50 mL À1 , and 150 μL of the mixed solution was applied onto the sample pad.The commercial rapid kit was incubated within 30 min at room temperature (25 °C) for analysis.
Application of the Virus Detection Process to Another Portable Device: To immobilize the capture antibody in the microtube, we used a wellestablished method for fixing antibodies to multiple substrates, with slight modifications. [53]The tube was first incubated with 100 μL of ethanol for 5 min and washed with Milli-Q water 5 times.Subsequently, 100 μL of KOH 1% w v À1 was added to the tube for 10 min and washed.This tube was treated with O 2 plasma and silanized by incubation with 100 μL APTES 2% v v À1 at 37 °C for 1 h.The antibodies prepared via the EDC-NHS reaction were then effectively immobilized by incubation at 37 °C in the APTES tube for 1 h.The subsequent detection steps followed the optimized conditions and methods described above.
Imaging Analysis: AFM was used to confirm the preamplification step and its change in topography.Images were obtained using an NX-10 microscope (Park Systems, Suwon, Republic of Korea).A microscope cover glass was used as bare glass, and the preamplification step described above was conducted in the same manner.Similarly, SEM was used to confirm BGNP assembly according to incubation cycle.The imaging was performed using a JSM-7610F-Plus microscope (JEOL, Tokyo, Japan).A silicon wafer was prepared as bare glass and subjected to the preamplification and first amplification steps described above.Additionally, the detection signal of each cycle was photographed using a smartphone (Samsung Galaxy S10).The immunoplates and commercial rapid kits used for diagnosis at each stage were all photographed at the same location.In all diagnostic tests, negative cases were measured alongside positive cases to prevent a decrease in the reliability of detection results caused by environmental factors such as ambient light.The intensities were measured using ImageJ software 1.47v (National Institutes of Health, Bethesda, MD, USA) and RGB Color Detector application 3.0v (Android).Images of the immunoplate and microtube, which completed the detection process, were uploaded and analyzed using each software.
Statistical Analysis: Each experiment was performed in triplicate (n = 3) and presented as mean AE standard deviation.All analyses were conducted using GraphPad Prism 8.0 (GraphPad, Boston, MA, USA).Logarithmic regression was performed to obtain the LOD using the following formula: LOD = mean blank þ 3 (SD blank ).Significant differences were determined using unpaired t-tests from GraphPad Prism 8.0 and are indicated by n.s.(not significant) and asterisks (*, **, ***, and ****).

Scheme 1 .
Scheme 1. Working principle of LBL signal amplification using BGNPs.a) Schematic representation of the two-step colorimetric signal amplification for the detection of IAV.Dual-signal amplification comprised LBL accumulation of BGNPs via avidin-biotin interaction, and enzyme mimicking 3,3',5, 5'-tetramethylbenzidine (TMB) oxidation of the accumulated BGNPs.b) Workflow for the quantitative detection and serial color change within 55 min at room temperature (25 °C).

Figure 1 .
Figure 1.Nanozyme activity of BGNPs.a) Schematic representation of BGNP synthesis and their nanozyme activity.Characterization of GNPs and BGNPs are presented with b) size distribution and c) zeta potential.Optical density ratio (signal ratio) of BGNPs at 524 (OD 524 ) and 652 nm (OD 652 ) for nanozyme activity evaluation is shown as a function of d) the diameter of the GNPs and e) the conjugated mass ratio of the biotin PEG thiol (SH-PEG-biotin) and gold ions comprising the GNPs.f ) UV-vis spectrum of BGNPs for evaluating the nanozyme activity of the BGNPs, BGNP alone (red) or containing H 2 O 2 (orange), 3,3',5,5'-TMB (green), H 2 O 2 þ TMB (blue), and H 2 O 2 þ TMB (gray).g) The calibration curves for OD 524 and OD 652 are presented in relation to the concentration of gold ions (C gold ) in BGNP.Significant differences are denoted by the asterisks with p-values (****p < 0.0001).Data (c-g) represent mean AE standard deviation with n = 3.

Figure 2 .
Figure 2. Evaluation of first and second amplification on the avidin immobilized plate.a) Schematic representation of the BGNP-based LBL primary amplification process involving specific recognition of avidin-biotin.Control 1: avidin nonimmobilized well and BGNP pair; Control 2: avidin immobilized well and PEGylated GNP (PGNP) pair; BGNP layering: avidin immobilized well and BGNP pair.The lower table presents the predicted results of first and second amplification in the presence or absence of avidin and biotin.b) First amplification at three cases after each cycle from 0 to 7 (upper).Plots of the OD 524 recorded from upper image (lower).c) Second amplification at three cases after each cycle from 0 to 7 (upper).Plots of the OD 652 recorded from upper image (lower).Data (b,c) represent mean AE standard deviation with n = 3.

Figure 3 .
Figure 3. Virus detection analysis.a) Detection signal at the first and second amplification under cycle 2 presented according to the virus titer.b) Signal-tonoise (S/N) value for each cycle from 1 to 3 was calculated and plotted by dividing the signal from each titer condition by the signal from the blank.c) The IAV detection signal at cycle 2 showed a linear relationship with viral titer over the analytical range.d)The specificity analysis of the assay against a range of respiratory viruses was conducted at cycle 2.The virus samples tested had the following titers: H1N1 had a titer of 10 5.7 EID 50 mL À1 ; and H3N2, H9N2, influenza B virus (IBV), Newcastle disease virus (NDV), and human coronavirus (NL63) had titers of 10 7.25 EID 50 mL À1 , 10 5.7 EID 50 mL À1 , 10 4.25 EID 50 mL À1 , 10 7.5 EID 50 mL À1 , and 10 4.1 TCID 50 mL À1 , respectively.e) S/N value of our assay and conventional methods was plotted according to the virus titer.Each value was normalized after subtracting the blank from each readout-signal intensity.f ) Comparison of the results of H1N1 detection using our assay at cycle 2 with those from conventional diagnostic assays (ELISA and a commercial rapid kit) A gray background indicates that it is discernible to the naked eye in our assay.g) Comparison of our assays with conventional diagnostic assays.The limits of detection (LOD) for our assay and ELISA were determined using the OD 450 read-out signal obtained from a microplate.Data (a-e) represent mean AE standard deviation with n = 3.

Figure 4
Figure 4. Real sample analysis.The detection performance of this assay was evaluated using various biological samples, including saliva and nasal solutions, and was compared with that of commercial rapid kit.a) Schematic illustration of workflow for real sample analysis.b,c) The detection signals of H1N1 in PBS, saliva, and nasal solutions were shown corresponding to their respective virus titers (10 5.7 EID 50 mL À1 , 10 3.7 EID 50 mL À1 , and blank, respectively) and photograph image was analyzed with Image J software.d,e) IAV detection using commercial rapid kits for three separate cases, each prepared using a different biological sample.f ) Signal recovery from environmental samples using our assay and a commercial rapid kit was calculated, with normalization based on the signals obtained from PBS."N/A" indicates that the calculation is not applicable.Data (c,e) represent mean AE standard deviation with n = 3.

Figure 5 .
Figure 5. Application as a portable tool for improving the potential of POCT.a) Schematic illustration of the LBL detection process within a tube and an image depicting the optical signal during cycle 2. b) Optical images of signal detection in the tube during the first and second amplification cycles, as well as the final signal during cycle 2. c) The detection signal, measured using the smartphone application, was extracted via the CMYK colorimeter in the application, and expressed as the yellow component among the four channels (cyan, magenta, yellow, and black).d) The detection signal measured using the microplate reader.e) The detection signal measured using the Image J software.Data (c-e) represent mean AE standard deviation with n = 3.