Disodium hydrogenphosphate (Na₂HPO₄, ≥99.0%), sodium phosphate monobasic monohydrate (NaH₂PO₄∙H₂O, 98%–102.0%), sodium carbonate (Na₂CO₃, ≥99.0%), sodium bicarbonate (NaHCO₃, 99.7%–100.3%), 2-(N-morpholino) ethanesulfonic acid (MES), and Tween-20 were purchased from Sangon Biotech (Shanghai, China). Gold (III) chloride tetrahydrate (HAuCL₄∙4H₂O, Au>47.8%) was purchased from Beijing Huawei Ruike Chemical (Beijing, China). Hydrogen hexachloroplatinate (IV) hexahydrate (H₂PtCL₆∙6H₂O, 99%) was purchased from Energy Chemical (Shanghai, China). Polyvinylpyrrolidone (PVP, molecular weight: 10kDa) was purchased from Tokyo Chemical Industry. Streptavidin, sulfo-NHS-LC-biotin, 1-ethyl-3-[3-dimethylaminopropyl] carbodiimide hydrochloride (EDC), and N-hydroxysulfosuccinimide (sulfo-NHS) were purchased from Thermo Fisher Scientific (United States). Sodium citrate tribasic dihydrate (99%), L-ascorbic acid (AA, 99%), bovine serum albumin (BSA), and casein block buffer were purchased from Sigma-Aldrich. Polydimethylsiloxane (PDMS; Sylgard 184) was purchased from Dow Corning (United States); photoresist (SU-8 2050) was purchased from MicroChem Corp. (MA, United States); carboxyl-functionalized magnetic beads (MBs) with an average diameter of 10 µm were purchased from Suzhou Nanomicro Technology (China); one-step TMB substrate solution was purchased from Beijing Makewonderbio (Beijing, China). The respiratory virus proteins and antibodies (mAb IgG) were obtained from the Institute of Biopharmaceutical and Health Engineering, Tsinghua University. All aqueous solutions were prepared using deionized (DI) water with a resistivity of 18.2 MΩ·cm.

Monoclonal antibodies with high specificities against influenza A virus, influenza B virus, respiratory syncytial virus (RSV), SARS-CoV-2, human bocavirus (HBoV), human metapneumovirus (HMPV), adenovirus (AdV), and human parainfluenza viruses (HPIVs) were produced and identified, and their target proteins are shown in Table 1. The antibodies against the nucleocapsid proteins of the HPIVs cannot cross-react with other subtypes from among HPIV-1, HPIV-2, and HPIV-3, so specific antibodies were selected for each subtype and mixed to detect these three subtypes.

Platinum-decorated Au@Pt nanozyme (Au@Pt NPs) was synthesized as described in a previous study (Wu et al., 2022). In brief, 15-nm gold NPs were used as the seed and mixed with 190 mL of purified water, followed by the addition of 4 mL of PVP (20% w/v). This solution was mixed strongly for 2 min so that the polymer could coat and stabilize the AuNP seeds; then, 8 mL of L-ascorbic acid (10% w/v) was added to the solution and mixed for 1 min. Next, approximately 1600 μL of chloroplatinic acid hydrate (0.5 M) was added to the mixture and mixed for 1 min; this solution was immediately heated to 65°C in an oil bath without stirring for 30 min until its color turned brown/black. The synthesized Au@Pt NPs were cooled to room temperature, purified by centrifugation (15 min, 14,000g), and resuspended in distilled water three times. Finally, Au@Pt NPs of size 75 nm were synthesized.

In this study, antibodies were allowed to form coordination bonds with sulfur groups that were attached to the surfaces of the Au@Pt NPs. Briefly, 10 mg of DTT was dissolved per milliliter of water, and this solution was added to each labeling antibody at a concentration of 25 μL/mg. The mixtures were then incubated at 4°C for 30 min, following which the excess DTT was removed by ultrafiltration centrifugation using 20 mM sodium phosphate buffer. The modified L-mAbs should be used immediately in conjugation reactions. Next, 100 μL of 75 nm Au@Pt NPs (2.5 nM) was mixed with 400 μL of 20 mM sodium phosphate buffer, followed by addition of 54 μg of modified L-mAbs; this mixture was incubated for 2 h using gentle rotations at room temperature. The modified particles were subsequently blocked using 200 μL of the blocking solution (phosphate buffer containing 1 wt% casein block and 1 wt% glutathione) for 30 min with gentle rotations at room temperature. The excess reagents were removed through three washing steps by centrifugation (15 min, 14,000g) using 20 mM sodium phosphate buffer. Finally, the Au@Pt NPs-mAbs were resuspended in a storage buffer at a concentration of 500 p.m.

Antibody-functionalized MBs were prepared in accordance with the streptavidin–biotin system, and streptavidin-conjugated MBs were prepared according to the EDC/NHS method. Briefly, 100 mg of carboxyl-functionalized MBs were separated using a magnet for 1 min and washed with MES buffer three times before being activated by 2 mM sulfo-NHS and 5 mM EDC for 30 min at 37°C with gentle rotations. The activated MBs were separated using a magnet, washed with MES buffer three times, and redispersed in 50 mM borate buffer to react with 1.5 mg streptavidin for 3 h at 37°C with gentle rotations. The residual active coupling sites or non-specific binding sites were blocked with a blocking solution for 30 min at 37°C, and the streptavidin-coated MBs were finally washed four times with 20 mM sodium phosphate buffer. Biotinylation of each coated antibody (C-mAb) was performed. Briefly, C-mAbs underwent dialysis against phosphate-buffered saline (PBS) and was diluted to a concentration of 2 mg/mL. Then, sulfo-NHS-LC-biotin was dissolved in DMSO at a concentration of 10 mM, and approximately 13.5 μL of the sulfo-NHS-LC-biotin solution was added to 1 mg of the C-mAbs solution and reacted for 24 h at 4°C. The unreacted biotinylation reagent and byproducts were removed by dialysis against PBS for 24 h at 4°C to purify the biotinylated antibodies. Then, the streptavidin-coated MBs and 0.3 mg of the biotinylated antibodies in PBS were incubated for 1 h at room temperature with gentle rotations. The antibody-coated MBs were separated using a magnet and washed four times using 20 mM sodium phosphate buffer. Finally, the MBs-streptavidin-mAbs were resuspended in phosphate buffer with 1 wt% BSA.

The microfluidic chips were designed using AutoCAD 2019 and fabricated as described in a previous study (Wu et al., 2022). The structure of the multiplexed microfluidic immunoassay chip is shown in Figure 1A and consists of one sample Luer inlet port at the middle (white loop) surrounded by eight reaction Luer ports (yellow loop) and eight Luer outlet ports (red loop). There is a wash chamber (green loop) in each path between the reaction Luer port and outlet port. The inlet and outlet ports are connected to the chamber by channels of 1000 μm width, and a height of 100 μm was considered for the chamber. The structural overview of the microfluidic chip is given in Figure 1B.

For the immunoassay preparation, each respiratory-virus-related nanozyme–antibody conjugate (Au@Pt NPs-mAbs) and antibody-functionalized-MBs (MBs-streptavidin-mAbs) were added to the corresponding reaction Luer ports of the microfluidic chip and freeze-dried overnight in vacuum. Upon removal from the freeze-dryer, the chips were stored in a sealed plastic bag filled with silica bead desiccant at room temperature. The two selected antibodies used in the formulation of the nanozyme–antibody conjugate (Au@Pt NPs-mAb) and antibody-functionalized-MBs (MBs-streptavidin-mAb) were designed to specifically recognize the same antigen protein of each virus. Figure 2 shows the formation of the sandwich immunoassay and its colorimetric formation.

The test procedure begins with placing the chip on the magnet, followed by addition of 400 µL of the standard protein solution or sample lysate into the sample Luer inlet port; then, the T-type silicone rubber plug is placed over the inlet and pressed till it touches the bottom. The samples are equally distributed to the eight reaction Luer ports, where the freeze-dried Au@Pt NPs-mAbs and MBs-streptavidin-mAbs recognize and bind with the antigen proteins in the samples to form a sandwich structure. After incubation for 30 min, T-type silicone rubber plugs are placed over each of the reaction Luer ports and pressed to sufficiently push the reaction mixture toward the wash chamber, where the MBs-complex is isolated from the reaction mixture by the magnet. Next, absorbent paper is placed in the eight Luer outlet ports, and approximately 600 µL of the wash buffer is injected from the sample Luer inlet port using a syringe to wash the MBs-complex, followed by removing the magnet and absorbent paper. Finally, 400 µL of the one-step TMB substrate solution is injected from the sample Luer inlet port using a syringe, and the results are detected after incubation for 1–10 min. Usually, the analytical procedure is completed within 40 min. The TMB substrate solution flushes the MBs-complex in the wash chamber into the Luer outlet port to react with the captured Au@Pt NPs and develop color. These developed colors can be conveniently examined and judged visually, or the mixture can be transferred to a microwell plate for absorbance detection at 630 nm using a microplate reader.

The monoclonal antibodies used in this study were screened, expressed, and purified in the lab. Two antigen-specific antibodies with high affinity were selected for each virus. The binding capacities of these antibodies with their antigens were detected by indirect ELISA, and the results are shown in Figure 3. These antibodies were purified in a high titer and can recognize their antigens even at low concentrations, showing their sensitivity and potential for viral antigen detection.

The Au@Pt NPs were prepared with controlled sizes. The color of the platinum-decorated Au NPs changed from red (Figure 4A) to dark brown (Figure 4B) after heat treatment. Then, these Au@Pt NPs were investigated using transmission electron microscopy (TEM) for their structure and size (Figure 4C). The average size of Au@Pt NPs was further detected by dynamic light scattering (DLS) and was found to be approximately 98.27 nm (Figure 4D).

This multiplexed microfluidic immunoassay chip was designed for the simultaneous detection of eight respiratory viruses, namely, influenza A virus, influenza B virus, RSV, SARS-CoV-2, HBoV, HMPV, AdV, and HPIVs. The sensitivity of this chip was determined by testing recombinant protein standard samples at concentrations of 0.01–1000 pg/mL. The results of these standard samples could be inferred visually and also measured using a microplate reader at an absorbance of 630 nm. With the help of the nanozyme, the differential color intensity between the limit concentration and negative control was easily distinguishable via healthy color vision. As shown in Figure 5, the color intensity of the 0.1 pg/mL protein could be distinguished from the negative control visually. For detection with the microplate reader, the test limit concentration of each protein was as low as 0.1 pg/mL (Figure 6). These data indicate that the multiplexed microfluidic immunoassay chip can conveniently detect the eight viruses with high sensitivity.

The specificity and cross reactivity of the multiplexed microfluidic immunoassay chip was analyzed by detecting the conserved nucleocapsid proteins of the influenza A virus (subtype H1N1), influenza B virus (subtype Yamagata strain), RSV (Long strain), SARS-CoV-2 (B.1.1.529 BA.1 and omicron strains), HMPV, HPIVs (subtypes 1, 2, 3, and 4), VP1 protein of HBoV, and HP protein of AdV. These proteins of each virus were mixed for detection; as shown in Figure 7, the multiplexed microfluidic immunoassay chip could detect each virus specifically without cross reactivity with the other seven viruses. Furthermore, all eight respiratory pathogens were detected successfully, demonstrating that the multiplexed microfluidic immunoassay chip could detect these viruses with high sensitivity and specificity.

The studies involving human samples were reviewed and approved by the Bioethics Committee (BEC) of Tsinghua Shenzhen International Graduate School. In this study, human throat swab samples (n = 315) were collected from Shenzhen Children’s Hospital and were diagnosed by real-time RT-PCR/PCR, whose results are shown in Table 2. These samples were inactivated before testing with the multiplexed microfluidic immunoassay chip. Before detection, approximately 300 μL of the human throat swab sample and 300 µL of the lysis buffer (PBS containing 0.5% NP40, 0.2% Tween 20%, and 0.2% LNAC, pH 7.4) were mixed and incubated for 1 min, and the lysate was directly tested using the microfluidic chip immunoassay. As shown in Table 2, the multiplexed microfluidic immunoassay chip could efficiently recognize all eight respiratory viruses in the throat swab samples. Compared with the results of the real-time RT-PCR assay, the microfluidic chip immunoassay had accuracies of 100%, 96.5%, 100%, 100%, 96%, 96%, 96.5%, and 92% for the detection of influenza A virus, influenza B virus, RSV, SARS-CoV-2, HBoV, HMPV, AdV, and HPIVs, respectively. The positive coincidence rate between the multi respiratory virus microfluidic immunoassay and real-time RT-PCR results was 97.2%, negative coincidence rate was 94.3%, and total coincidence rate was 98.1%.