Development of nucleocapsid-specific monoclonal antibodies for SARS-CoV-2 and their ELISA diagnostics on an automatic microfluidic device

Abstract

Coronavirus disease 2019 (COVID-19) pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection has threatened public health globally, and the emergence of viral variants has exacerbated an already precarious situation. To prevent further spread of the virus and determine government action required for virus control, accurate and rapid immunoassays for SARS-CoV-2 diagnosis are urgently needed. In this study, we generated monoclonal antibodies (mAbs) against the SARS-CoV-2 nucleocapsid protein (NP), compared their reactivity using an enzyme-linked immunosorbent assay (ELISA), and selected four mAbs designated 1G6, 3E10, 3F10, and 5B6 which have higher reactivity to NP and viral lysates of SARS-CoV-2 than other mAbs. Using an epitope mapping assay, we identified that 1G6 detected the C-terminal domain of SARS-CoV-2 NP (residues 248–364), while 3E10 and 3F10 bound to the N-terminal domain (residues 47–174) and 3F10 detected the N-arm region (residues 1–46) of SARS-CoV-2 NP. Based on the epitope study and sandwich ELISA, we selected the 1G6 and 3E10 Abs as an optimal Ab pair and applied them for a microfluidics-based point-of-care (POC) ELISA assay to detect the NPs of SARS-CoV-2 and its variants. The integrated and automatic microfluidic system could operate the serial injection of the sample, the washing solution, the HRP-conjugate antibody, and the TMB substrate solution simply by controlling air purge via a single syringe. The proposed Ab pair-equipped microsystem effectively detected the NPs of SARS-CoV-2 variants as well as in clinical samples. Collectively, our proposed platform provides an advanced protein-based diagnostic tool for detecting SARS-CoV-2.

1. Introduction

Since 2019, the coronavirus disease 2019 (COVID-19) pandemic has spread globally and posed an immense threat to public health [1], [2], [3], [4], [5]. As of 18 March 2022, there have been approximately 460 million confirmed cases of COVID-19, including over 6 million COVID-19-related deaths worldwide [6]. COVID-19, a highly pathogenic and transmissible disease, is caused by the infection of newly discovered severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). SARS-CoV-2 is a positive-sense, single-stranded RNA virus belonging to the genus Betacoronavirus, which also includes SARS-CoV and MERS-CoV. Due to the high mutation rate of RNA viruses, multiple SARS-CoV-2 variants have been continually emerging and have become a significant challenge for controlling COVID-19 [7].

Precise identification of SARS-CoV-2 infection is important for the prevention of viral transmission, antiviral treatment, and implementation of government policies [8]. Various diagnostic methods have been developed to detect viral antigens expressed by SARS-CoV-2 in pharyngeal swabs and serum samples of infected individuals [9]. Although viral nucleic acid detection using real-time reverse transcription polymerase chain reaction (RT–PCR) is the gold standard for SARS-CoV-2 detection, this technique has some limitations, such as the sequence mutations in primer binding regions of the viral genome, sample storage, low-quality nucleic acid purification, and cost [10]. A typical RT–PCR requires 4–6 h for completion, which includes sample preparation, gene amplification, and results analysis, and necessitates the use of well-trained technicians. Consequently, despite the outstanding diagnostic odds ratio, the lack of trained human resources, expensive consumables, and need for sophisticated instrumentation have hampered the timely control of the pandemic, making it imperative to develop new and faster diagnostic approaches [11], [12]. To overcome these issues, a variety of types of the POCT platform have been extensively explored such as a lateral flow strip kit (SD Biosensor, Inc., Siemens Healthcare Diagnostics, Inc., Abbott Diagnostics, Ltd., etc), high-throughput 96-well plate processing ELISA platform (Hamilton, Co.) and a QIAstat-Dx (Qiagen, Co.) that is capable of molecular and immunoassay diagnostics for SARS-CoV-2 [13], [14]. In particular, protein-based diagnostic tools have been developed using viral antigen-specific monoclonal antibodies (mAbs). Detection kits are more convenient and reduce the time required to obtain results. Of the various antigens of SARS-CoV-2, the nucleocapsid protein (NP) is the most promising target antigen for diagnostics. NP is one of the predominant proteins and is highly conserved compared to the spike (S) protein, which is a surface antigen responsible for binding to host cells. The amino acid sequence of SARS-CoV-2 NP shares approximately 90% and 50% identity with the SARS-CoV and MERS-CoV homologs, respectively. Previous reports have confirmed that NP may be a marker for the early diagnosis of the SARS-CoV outbreak that occurred in 2003 [15], [16]. A recent study also showed that NP levels in the serum of infected patients may have diagnostic values for the early stage of SARS-CoV-2 infection [17]. To date, some antigen test kits based on Abs have been approved for the diagnosis of COVID-19; however, improvement in these diagnostic platforms using mAbs of certain specificity and affinity or well-matched Ab pairs for specific and effective diagnosis, including sandwich enzyme-linked immunosorbent assay (ELISA) to detect emerging SARS-CoV-2 variants, still remains desired.

Despite being a high-throughput assay, the ELISA needs to be simplified since the average time-to-result is 2–5 h, and performing the assay requires trained personnel. Hence, incorporating traditional ELISA into a portable system is being actively pursued. Liu et al. described a fully integrated, low-cost, and portable ELISA-based point-of-care (POC) device that combines ELISA and distance readout into a single microfluidic chip [18]. Although this ELISA-Chip provided quantitation of disease biomarkers, the operating time of approximately 2 h remained a limitation. A microfluidic ELISA for rapid detection of food allergens was also proposed [19], but it required the manual procedure.

In the present study, we generated mAbs against SARS-CoV-2 NP and developed a microchip diagnostic platform. Our mAbs, especially 1G6 and 3E10, exhibited higher reactivity to NP and viral lysates of SARS-CoV-2 than other mAbs. The 1G6 and 3E10 mAbs were optimal as an Ab pair for a sandwich ELISA platform to identify SARS-CoV-2 due to the different binding sites on NP. In addition, the Ab pair has the potential to efficiently detect NPs of SARS-CoV-2 variants, including alpha, beta, gamma, and delta. The Ab pair-equipped sandwich ELISA was integrated in a microfluidic platform to automate the entire process for POC testing and the clinical samples were successfully analyzed.

2. Materials and methods

2.1. Reagents

NMC-nCoV02, a SARS-CoV-2 strain (S clade) isolated from a confirmed COVID-19 patient in South Korea in February 2020, was propagated and inactivated as previously described [20], [21]. Briefly, virus was propagated in Vero cells cultured in DMEM (Gibco, NY, USA) supplemented with 1% penicillin/streptomycin (Gibco) and TPCK trypsin (0.5 μg/mL; Worthington Biochemical, NJ, USA) at 37 °C for 72 h in a biosafety level 3 (BSL3) laboratory. Virus-containing culture supernatants were harvested and inactivated by 16 h treatment with 0.5% beta-propiolactone (BPL) at 4 °C, followed by 2 h incubation at 37 °C to hydrolyze the residual BPL. The inactivated virus was purified using sucrose gradient ultracentrifugation at 100,000 × g at 4 °C for 2 h.

The following recombinant viral proteins were purchased from Sino Biological (Shanghai, China): NP, S1, S2, and RBD of SARS-CoV-2 produced in HEK293T mammalian cells harboring a plasmid encoding the codon-optimized protein of SARS-CoV-2 (isolate Wuhan-Hu-1; accession no. YP_009724397.2), NP of SARS-CoV generated in mammalian cells harboring a plasmid encoding the codon-optimized NP of SARS-CoV (isolate Tor2; NP_828858.1), NPs of SARS-CoV-2 variants produced in Escherichia coli (alpha with D3L, R203K, G204R, and S235F mutations; beta with T205I mutation; gamma with P80R mutation; and delta with D63G, R203M, and D377Y mutations), and influenza A H1N1 (A/California/07/2009) NP expressed in baculovirus insect cells. Purified SARS-CoV-2 viral lysates (USA-WA1/2020, accession no. MN985325) were purchased from the Native Antigen Company (Kidlington, United Kingdom). The mAbs specific to SARS-CoV-2 NPs were purchased from Creative Diagnostics (clone 4B21; NY, USA) and Invitrogen (Clone B46F; CA, USA). HRP was conjugated to Ab using an HRP-conjugation kit (Abcam, MA, USA) according to the manufacturer’s instructions.

2.2. Mouse immunization and hybridoma preparation

Six-week-old female BALB/c mice were purchased from Orient Bio (Gyeonggi-do, South Korea) and housed in a specific pathogen-free facility at the Korea Research Institute of Bioscience and Biotechnology (KRIBB). Animal handling and experimental procedures were reviewed and approved by the Institutional Animal Care and Use Committee of KRIBB and performed according to the Guidelines for Animal Experiments of KRIBB (Approval number: KRIBB-AEC-21003). Mice (n = 3) were immunized three times with 10 μg inactivated SARS-CoV-2 (S clade) plus TiterMax Gold adjuvant (Sigma–Aldrich, MO, USA) by footpad injection at 2-week intervals. Two weeks after the final immunization, cells were isolated from the popliteal lymph nodes and fused with myeloma FO cells (ATCC, MD, USA) to generate B-cell hybridoma clones as described previously [22]. A monoclonal SARS-CoV-2 NP-secreting B-cell hybridoma clone was selected, and the Ab was purified from the hybridoma cell culture supernatant using protein G agarose (Cytiva, MA, USA) and characterized.

2.3. ELISA

For a direct ELISA, Maxisorp 96-well plates (Thermo Fisher Scientific, Roskilde, Denmark) were coated with recombinant proteins (50 ng/well) and inactivated SARS-CoV-2 lysates (250 ng/well; USA-WA1/2020) in PBS at 4 °C for 16 h. The plates were blocked with 5% BSA in PBS (5% BSA-PBS) at 3 °C for 1 h and washed with 0.05% Tween-20 in PBS (PBST). Serially 3-fold diluted mAbs were prepared in 1% BSA-PBS and added to the plates followed by 2 h of incubation at 37 °C. After washing with PBST, the plates were treated with HRP-conjugated anti-mouse IgG (1:5000 diluted in 1% BSA-PBS; Cell Signaling Technology, MA, USA) for 1 h at 37 °C.

For a sandwich ELISA, Maxisorp plates were coated with mAbs in TBS (100 ng/well) at 4 °C for 12 h, washed with 0.1% Tween-20 in TBS (TBST), and then blocked with 5% BSA in TBST at 37 °C for 2 h. Recombinant NPs (30 ng/well) or SARS-CoV-2 viral lysates (300 ng/well) were subjected to serial 3-fold dilution in 2% BSA-TBST, added to the plates, and incubated at 3 °C for 2 h. After washing with TBST, the plates were incubated with HRP-conjugated anti-NP mAbs (10 ng/well) in 5% BSA-TBST at 37 °C for 2 h. Finally, all plates were washed with PBST or TBST to remove residual Abs, and the color was developed using a 3,3′,5,5′-tetramethylbenzidine substrate reagent kit (BD Biosciences); the reaction was stopped using 2 N H₂SO₄. Absorbance was measured at 450 nm using a VICTOR Nivo™ Multimode Plate Reader (PerkinElmer, MA, USA).

2.4. Isotyping

Isotypes of mAbs were determined using a mouse Ab isotyping kit (Roche, Basel, Switzerland). Briefly, hybridoma cell culture supernatants or the purified mAbs (1 μg/mL) were added to the development tube and incubated for 1 min. The isotyping strips were placed in the development tube, and bands were observed after 10 min to identify the isotypes.

2.5. Epitope mapping assay

DNA templates corresponding to the coding region of the SARS-CoV-2 NP gene were amplified using polymerase chain reaction. The SARS-CoV-2 NP or truncated mutant constructs were cloned into the HindIII and EcoRV sites of the vector pCMV3. All constructs were sequenced by Bioneer Corporation (Daejeon, South Korea) to verify 100% correspondence with the original sequences. The constructs were transfected into HEK293T cells and incubated for 72 h to induce NP fragment expression. Whole cell lysates were prepared using radioimmunoprecipitation assay buffer (Thermo Fisher Scientific) and separated using 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) for WB analysis.

2.6. Design of the integrated microfluidic device

The microfluidic device consists of four chambers to store the sample, HRP-antibody, TMB and washing solution. The serpentine structure of each chamber has a dimension of 1 mm × 1.5 mm × 70 mm [width × height × length] and provides more storage space (up to 105 μL for the sample, HRP-antibody, TMB solution and 550 μL for the washing solution). The end of the four chambers was merged to the passive valve (0.2 mm × 0.2 mm [width × height]) to prevent back-flow and easily control the fluidics during the loading of the solution. Then, the passive valve was linked to the coil micro-reactor (0.5 mm × 0.8 mm × 75 mm [width × height × length]), which served as a reaction chamber of the ELISA assay. The capture antibody was pre-coated in the coil micro-reactor and the ELISA assay proceeded. The advantage of the coil structure is to provide the high surface-to-volume ratio, so that more sensitive ELISA assay is achievable than the conventional well type. In addition, the solution occupied in the coil microchannel can be flushed out completely without residue. Such a low dead volume eliminates the contamination issue in the ELISA processes. All the excess of the sample, HRP-antibody, TMB and washing solution flow to the waste chamber, which also has the serpentine structure with a volume of 1.1 mL. An air vent hole was fabricated at the end of the waste chamber.

2.7. Fabrication of a microfluidic device

A router CNC machine (Tinyroboto, Korea) was used to manufacture a poly(methyl methacrylate) (PMMA) chip (3 mm thickness). An endmill with a diameter of 1.0 or 0.2 mm was used to create the microfluidic structure on the bottom and top surfaces (JJtools, Korea). The double-sided etched PMMA chip was cleaned using a sonicator with a surfactant solution (deconex® 11 UNIVERSAL, Groesfjeld Diagnostics, Denmark) at 60 °C for 30 min. The microfluidic device was then washed and sonicated in DNase-free water before drying in an oven at 80 °C for 30 min. To immobilize the antibody onto the coil micro-reactor, the coil part was first exposed to O₂ plasma for 30 s at a flow rate of 50 sccm, a power of 60 W, and a frequency of 50 kHz. Subsequently, a 5% w/w of (3-aminopropyl)triethoxysilane (APTES) solution dissolved in DI water was injected to fill the coil structure. After overnight incubation, the APTES solution was withdrawn, and a superhydrophobic reagent (Neverwet Shield, Rust-Oleum, USA) was coated on the other microstructures of the chip. The chip was then heated to 80 °C for 1 h to stabilize the coating layer. The reaction coil was washed three times with DI water. The capture antibody was mixed with EDC/NHS reagents and incubated in the coil micro-reactor for 1 h. Then, the coil part was washed three times with DI water. Next, a SuperBlock (PBS) blocking buffer (Thermo Fisher Scientific, USA) was added to the coil part and incubated for 1 h. Finally, the chip was dried and sealed with pressure-sensitive film (HJ-Bioanalytik GmbH, Germany).

2.8. Development of a portable workstation

To operate the ELISA assay on the microfluidic chip, we set up a small syringe pump (PSD/6 Precision Syringe Pump, Hamilton company, U.S) having an 8-port valve, one syringe, and four tubes to connect the designated ports to the inlets of the four solution chambers on a microfluidic chip. Among the 8-port, four ports (ports #1, #2, #3, and #4) were utilized to connect the inlet hole of the sample-containing microchannel (red), the HRP-antibody conjugate-containing microchannel (green), the TMB substrate-containing microchannel (blue), and the washing solution-containing microchannel (yellow) in Fig. 4B. The other 4 ports were open to air and used to withdraw air to the syringe. Upon withdrawal, the syringe contained air. Then, the air was pushed to the designated microchannel by switching the 8-port valve, so that the fluidic flow of each solution in the microfluidic chip could be tuned. Such an operation of the air in-and-out, the switching of the 8-port valve, and the microfluidic control was performed automatically by an in-house software.

Fig. 4.

(A) Schematic illustration of the microfluidic device. (B) The POC workstation that includes an 8-port valve and a syringe pump. (C) Digital images of the microfluidic device showing the movement of the designated solution at each operation step of ELISA.

2.9. Preparation of COVID-19 infected clinical samples

Oropharyngeal samples were collected from confirmed COVID-19 patients of Kyung Hee University Hospital at Gangdong using a sterile swab and stored in a universal transport medium (UTM). Clinical samples were collected from donors and anonymized to preserve their privacy, in accordance with the Institutional Review Board requirements of Kyung Hee University. The clinical samples were lysed in a PBS (pH 7.4) lysis buffer containing 0.5% w/v Triton X and 0.5 M KCl for 1 h. Then, 40 μL of a lysis mixture was injected into the sample chamber on a chip via the sample hole. The HRP-antibody conjugate (40 μL), the TMB substrate (40 μL), the washing (550 μL) were also loaded into the chip. After loading of each solution, the tubes were connected to the inlets, and the ELISA assay was initiated.

2.10. Statistical analysis

Data are presented as the mean ± standard deviation (SD) and represent three independent experiments. All data, including the EC₅₀ of each mAb, were analyzed using GraphPad PRISM software (GraphPad Software, Inc., CA, USA).

3. Results and discussion

3.1. Generation of anti-SARS-CoV-2 NP mAbs

To produce SARS-CoV-2 NP-specific mAbs, BALB/c mice were immunized with inactivated SARS-CoV-2 mixed with TiterMax Gold adjuvant via footpad injection three times at 2-week intervals. Two weeks after the final immunization, lymphocytes were isolated from the popliteal lymph nodes of the mice and fused with mouse myeloma cell lines to generate Ab-secreting hybridomas. The production of NP-specific Ab was determined using a direct ELISA wherein the culture supernatants of the hybridomas were added to recombinant SARS-CoV-2 NP-coated plates. Of the 503 clones, 118 showed high reactivity to SARS-CoV-2 NP, as evidenced by high optical density (OD) values of > 2 (Fig. 1A). By subsequent subcloning of the hybridomas, we finally selected 4 hybridomas that maintained high reactivity to SARS-CoV-2 NP and designated them 1G6, 3E10, 3F10, and 5B6. Next, we examined the reactivity of mAbs against the SARS-CoV-2 S protein, which consists of S1 and S2 subunits and plays a key role in virus entry into host cells via the receptor-binding domain (RBD) of the S1 subunit. Unlike robust reactivity against SARS-CoV-2 NP, all mAbs showed negligible OD values for S1, S2, and RBD (Fig. 1B). Additionally, we determined the isotypes and complementary-determining regions (CDRs) of the mAbs (Table S1 and Fig. S1). All mAbs were IgG2a, kappa (κ)-chain isotypes except for 3E10, which was an IgG2b, κ-chain isotype.

Fig. 1.

Production and binding affinity of anti-SARS-CoV-2 NP mAbs. (A) BALB/c mice (n = 3) were immunized three times with 10 μg inactivated SARS-CoV-2 plus TiterMax through footpad infection at two-week intervals. Lymphocytes isolated from popliteal lymph nodes were fused with FO cells to generate hybridomas, and culture supernatants were harvested from the growing hybridomas. Binding affinity to SARS-CoV-2 NPs was determined using the supernatants in an ELISA. (B) Binding affinity for S1, S2, RBD, or NP of SARS-CoV-2 was measured using purified mAbs (1G6, 3E10, 3F10, and 5B6) in an ELISA. (C) ELISA plates were coated with SARS-CoV-2 viral lysates, SARS-CoV-2 NP, or SARS-CoV NP. Influenza A H1N1 NP was used as a negative control. After blocking, the plates were incubated with serially 3-fold diluted mAbs followed by HRP-conjugated anti-mouse IgG.

To examine the sensitivity of mAbs, ELISA was performed by incubating serially diluted mAbs in plates coated with NP or viral lysates of SARS-CoV-2. We also used recombinant NPs of SARS-CoV and influenza H1N1 to evaluate the specificity of the mAbs. The binding affinity was compared to that of two commercial anti-SARS-CoV-2 mAbs, B46F and 4B21. As shown in Fig. 1C, all purified mAbs had a higher reactivity to SARS-CoV-2 viral lysates than the commercial Abs. Notably, 1G6 exhibited robust binding with SARS-CoV-2 NP but not SARS-CoV NP. 3E10 had more an efficient reactivity to SARS-CoV-2 NP than SARS-CoV NP, whereas 3F10 showed the opposite pattern. 5B6 efficiently recognized the NPs of both SARS-CoV-2 and SARS-CoV. The 4 mAbs had marginal reactivity to influenza H1N1 NP as a negative control. The binding reactivity was additionally evaluated by calculating the half-maximal effective concentration (EC₅₀); high reactivity to SARS-CoV-2 NP was observed, in order, by 1G6, 3E10, and 5B6, while low reactivity to SARS-CoV NP was shown in 1G6 only (Table S2). To clarify the promising applicability of Abs in protein-based diagnostics, we carried out various immunoassays and found that our mAbs worked efficiently in WB, immunoprecipitation, and immunofluorescence assays (Fig. S2). These results indicate that the mAbs generated in our study efficiently recognize SARS-CoV-2 viral lysates compared to commercial mAbs and have different reactivities to SARS-CoV-2 NP.

Additionally, we calculated the equilibrium dissociation constant (K_d) values based on the direct ELISA assay using NP of SARS-CoV-2 variants, and the binding affinity values are shown in Table S3. The K_d values are roughly measured but are suitable for relative affinity comparisons among mAbs. 1G6 and 3E10 mAbs exhibit somewhat low K_d values compared with 3F10 and 5B6, indicating that 1G6 and 3E10 used for the on-chip Ab pairs have higher NP-binding affinities than the other two mAbs.

3.2. Selection of an optimal Ab pair for a microfluidic device

Identifying the antigen-binding site (epitope) of an Ab is crucial for the development of sandwich ELISA that requires two Abs with specificity to different epitopes of the same antigen [23]. To determine the SARS-CoV-2 NP epitopes of the purified mAbs using an epitope mapping assay, we designed NP fragments based on its structural characteristic [24] and prepared 7 recombinant NP fragments, including the NP-N-terminal domain (NTD; aa 47–174), NP-C-terminal domain (CTD; aa 248–364), NP-p1 (aa 1–174), NP-p2 (aa 1–248), NP-p3 (aa 1–305), NP-p4 (aa 101–419), and NP-p5 (aa 248–419) (Fig. 2A). Briefly, recombinant NP fragments were inserted into the pCMV3 vector and expressed in HEK293T cells. Then, the individual proteins were separated on 12% SDS–PAGE gels followed by WB analysis using the anti-NP mAbs. As shown in Fig. 2B, all mAbs effectively detected whole NP with aa 1–419 sequences. The IG6 mAb recognized NP-p4 (aa 101–419) and NP-p5 (aa 248–419) but not other truncated NP fragments, implying that IG6 binds to the C-terminal tail (aa 365–419) of SARS-CoV-2 NP. Two mAbs, 3E10 and 5B6, detected NP-NTD, NP-p1, NP-p2, and NP-p3 fragments, indicating that the mAbs bind to part of the NTD (aa 47–100) of SARS-CoV-2 NP. Unlike other mAbs, the 3F10 mAb recognized NP-p1, NP-p2, and NP-p3, showing that the binding site of 3F10 is in the N-arm region (aa 1–46) of SARS-CoV-2 NP. According to the results of the epitope mapping analysis, we confirmed that the 5 Ab pairs (1G6/3E10, 1G6/5B6, 1G6/3F10, 3E10/3F10 or 5B5/3F10) with different binding sites of SARS-CoV-2 NP can constitute a sandwich ELISA for identifying SARS-CoV-2.

Fig. 2.

Identification of SARS-CoV-2 NP epitopes recognized by the mAbs. (A) Scheme of SARS-CoV-2 NP constructs (full NP, aa 1–419; NP-NTD, aa 47–174; NP-CTD, aa 248–364; NP-p1, aa 1–174; NP-p2, aa 1–248; NP-p3, aa 1–305; NP-p4, aa 101–419; and NP-p5, aa 248–419). (B) HEK293T cells were transfected with the indicated SARS-CoV-2 NP construct carrying plasmids. Whole cell lysates obtained at 72 h post transfection were subjected to WB analysis using the purified mAbs followed by HRP-conjugated anti-mouse IgG.

To choose the best Ab pair for identifying SARS-CoV-2, we performed sandwich ELISA using 1G6 and 3E10 as capture Abs based on robust reactivity against SARS-CoV-2 NP, as described in Fig. 1C. Either 1G6 or 3E10 was coated on the ELISA plate, and then serially diluted SARS-CoV-2 NP or SARS-CoV-2 viral lysates were applied as target antigens. After washing, HRP-conjugated mAbs (1G6, 3E10, 3F10, or 5B6) were added to the ELISA plate as detection Abs. As shown in Fig. 3A, using IG6 as a capture Ab, the degree of SARS-CoV-2 NP detection was highest after incubation with 3E10-HRP and gradually decreased, in order, with HRP-conjugated 5B6, 1G6, and 3E10. In the coating with 3E10, 1G6-HRP only showed robust detection of SARS-CoV-2 NP (Fig. 3B). Other HRP-conjugated Abs had a weak binding affinity to detection Abs under the same conditions. Next, we evaluated the SARS-CoV-2 viral lysate-specific binding ability of the Ab pairs. Compared to SARS-CoV-2 NP as the target antigen, all the Ab pairs exhibited binding affinities to the viral lysates when using an almost 100-fold higher dose than the NP protein (Fig. 3C, D). Only 4 Ab pairs, including 1G6/3F10-HRP, 1G6/3E10-HRP, 3E10/1G6-HRP, and 3E10/3F10-HRP, had high reactivity to SARS-CoV-2 viral lysates compared to other Ab pairs. Collectively, these results suggest that the 1G6/3E10 Ab pair is optimal for sandwich ELISA-based diagnostics to identify SARS-CoV-2.

Fig. 3.

Comparison of reactivity of Ab pairs in a sandwich ELISA to detect SARS-CoV-2. ELISA plates were coated with the purified 1G6 (A, C) or 3E10 (B, D) mAbs. After blocking, the plates were incubated with recombinant NP (A, B) or viral lysates (C, D) of SARS-CoV-2. HRP-conjugated mAbs were used as detection Abs. Influenza NP-HRP was used as a negative control.

3.3. Operation of the on-chip ELISA assay

The microfluidic device was designed with four solution-loaded chambers, each of which was for the sample, HRP-conjugate antibody, TMB substrate, and washing solution. The antibody-coated coil micro-reactor was located between the solution chambers and the waste chamber. Thus, all solutions were flushed into the antibody-coated coil part and then collected in the waste chamber (Fig. 4A). Fluidic control of the microfluidic chip was mainly mediated by air injected from the syringe pump (Fig. 4B). The overall procedure for the ELISA assay is shown in Fig. 4C. First, the four solutions were loaded into each chamber (step #2). The 8-port valve was switched to port #1, and air previously introduced by to the syringe was transferred to the sample chamber. So, the sample solution was moved to the reaction coil, while the other ports were closed (step #3), preventing the sample solution flowing back into the other chambers. The sample was incubated for 30 min for the NP binding to the capture Ab, and air was pushed from the syringe to deliver the sample solution to the waste chamber and made the coil micro-reactor empty (step #4). Next, port #4 was opened, and the air in the syringe was pushed with a volume of 300 μL to the washing solution chamber. Therefore, only 300 μL among 600 μL of the washing solution could be moved to the coil part (step #5) and the remaining 300 μL of the washing solution can be used for the next washing step. After that, port #1 was opened again and air was introduced through the empty sample chamber to clean the coil part (step #6). The valve was then switched to port #2 and the HRP-antibody conjugate solution was injected into the coil micro-reactor on a chip and incubated for 15 min (step #7). Air pushed again from port #2 to flush the conjugate solution into the waste chamber and emptied the reaction coil (step #8). Then, port #4 was opened, and the remaining 300 μL of the washing solution was injected to wash the coil micro-reactor (step #9). The coil was then emptied by introducing air from port #4 (step #10). Finally, port #3 was opened and the TMB substrate was pushed into the coil structure. The color changed from colorless to blue after 15 min, and the color of the coil part was imaged and analyzed using ImageJ software. Note that the overall procedure was operated automatically by a miniaturized syringe pump and in-house software. Supplementary Video 1 shows the entire operation process.

3.4. Detection of NPs of SARS-CoV-2 variants on the microfluidic device

The emergence of new SARS-CoV-2 variants has raised the need for improved SARS-CoV-2 diagnostic agents that can effectively detect these variants as well. NPs of SARS-CoV-2 variants reportedly harbor various mutations, as shown in Fig. 5A. Thus, a sandwich ELISA was performed coupled with the microchip device using the 3E10/1G6-HRP Ab pair to detect the NPs of SARS-CoV-2 variants, including alpha, beta, gamma, and delta. Additionally, we compared reactivity using NPs of the SARS-CoV-2 Wuhan-Hu-1 strain and influenza H1N1 as positive and negative controls, respectively. Using a direct ELISA, we first checked that all purified mAbs had comparable binding affinity to the NPs of SARS-CoV-2 variants (Fig. S3). Importantly, the sandwich ELISA results confirmed that the 3E10/1G6-HRP Ab pair robustly detected NPs of SARS-CoV-2 variants in the following order: gamma, delta, beta, and alpha (Fig. 5B). These results indicate that immunoassays using the 3E10/1G6-HRP Ab pair may be useful for the specific detection of SARS-CoV-2 variants.

Fig. 5.

Detection of NPs of SARS-CoV-2 variants using the 3E10/1G6-HRP Ab pair. (A) Scheme of NPs of SARS-CoV-2 variants. Mutation sites identified in the sequenced genomes of SARS-CoV-2 variants are indicated in red. (B) ELISA plates were coated with the 3E10 mAb, blocked, and then incubated with recombinant NPs of the SARS-CoV-2 variants. After washing the plates, 1G6-HRP was added as a detection Ab. Influenza H1N1 NP and bovine serum albumin (BSA) were used as negative controls. (C) The color image of the coil micro-reactor on a chip after the treatment of the TMB substrate in the ELISA assay. All the NPs of SARS-CoV-2 variants, including alpha, beta, gamma, and delta produced significant color intensity compared with those of the influenza A H1N1.

In the same context, we applied the optimized 3E10/1G6-HRP Ab pair for the on-chip ELISA assay to detect the SARS-CoV-2 variants. The capture antibody of 3E10 mAb was pre-coated on the coil micro-reactor, and the whole process as described above was followed. The 1G6-HRP Ab was utilized as a detection antibody, and after the treatment of the TMB substrate, the color of the coil part was detected. As shown in Fig. 5C, the NPs of SARS-CoV-2 variants, including alpha, beta, gamma, and delta were successfully diagnosed due to superior binding affinity of the 3E10/1G6-HRP Ab pair to the COVID-19 variants. On the other hand, in case of the influenza A H1N1 NP, the blue color was very weak, demonstrating high specificity of the developed antibody pairs for the COVID-19.

3.5. Limit-of-detection (LOD) test and clinical sample analysis

To evaluate the LOD test of the proposed microfluidic system for the ELISA assay, the purified COVID-19 NPs were obtained from Sino Biological (Beijing, China), and the concentration was serially diluted from 12.5 to 0.02 ng. The resultant color of the coil micro-reactor was analyzed by Hue-Saturation-Brightness color system (HRB) using ImageJ (National Institutes of Health, Bethesda, Maryland, USA). Fig. 6A shows the real images of the coil part depending on the concentration of the NPs as well as the curve for the saturation intensity of the coil versus the amount of the input NPs. The intensity of the saturation percentage was proportional to the amount of the antigen added. Based on the criterion of the signal-to-noise ratio ≥ 3, the LOD was determined as 0.1 ng of NPs. The non-linear regression 5PL sigmoidal fitting model was used to illustrate the standard curve [25].

Fig. 6.

(A) Real image of the coil micro-reactor for the LOD test and the plot of the saturation percentage versus the log of input amount of NPs. (B) The resultant real images of the coil micro-reactors with clinical samples of COVID-19 and the corresponding percentage of saturation for each clinical sample.

For the clinical sample analysis, we received 10 different patient samples that were infected by COVID-19 from Gangdong Kyung Hee University hospital (Korea). 20 μL of each clinical samples were mixed with a phosphate-buffered saline (PBS, pH 7.4) containing 0.5% w/v Triton X and 0.5 M KCl (20 μL) for 1 h at room temperature for the lysis reaction. Subsequently, 40 μL of the lysed mixture was injected into the chip and the same ELISA process was followed. Fig. 6B shows that ten clinical samples infected with COVID-19 could be successfully analyzed on the proposed ELISA diagnostic microsystem in 1 h. All the clinical samples showed the percentage of saturation value above 25%, which means positive results based on the criterion of the signal-to-noise ratio ≥ 3. On the other hand, two negative samples (H1N1 infected sample and non-infected sample) revealed almost 0% in the percentage of saturation value. This result demonstrated the robustness of the developed Ab pairs and the fully automatic microfluidic system. When the SARS-CoV-2 sample was injected into the reaction coil, the sample tended to move to the outer ring due to the inertia force, resulting in more intensified signal at the outer edge part in the ring structure as shown in Sample 6. The heterogenous color density in the micro-reactor became more noticeable with low concentration of the SARS-CoV-2 samples due to the insignificant saturation of antigen in the whole surface of the reaction coil (See Sample 1 versus Sample 6). To solve such an unbalanced signal distribution, we averaged color intensity in the whole ring structure to produce the percentage of saturation in the micro-reactor. We summarized the comparative evaluation of our platform with the previously reported systems in Table S4. Compared with the previous results, our system shows similar LOD value with easy colorimetric analysis by naked eye, cost-effective automatic operation, versatile nucleocapsid-specific monoclonal antibodies capable of distinguishing four kinds of SARS-CoV-2 variants [26], [27], [28], [29].

4. Conclusions

In summary, we generated mAbs specific to SARS-CoV-2 NP and selected an optimal Ab pair for detecting SARS-CoV-2 NP. The Ab pair-equipped tools effectively detected NPs of SARS-CoV-2 and the newly emerging variants. We also developed a fully integrated and automatic ELISA microfluidic platform for the identification of COVID-19. The combination of the unique microfluidic device and the syringe pump with an 8-port valve could program the flow of the four solutions in the ELISA assay. This all-in-one type of fluidic control allows for automatic operations, on-site ELISA diagnostics, and reduction of the overall cost and time.

CRediT authorship contribution statement

Jihyun Yang: Methodology, Investigation, Validation, Formal analysis, Writing − original draft. Phan Minh Vu: Methodology, Investigation, Formal analysis, Writing − original draft. Chang-Kyu Heo: Methodology, Investigation, Formal analysis. Hau Van Nguyen: Methodology, Investigation, Formal analysis, Writing − original draft. Won-Hee Lim: Formal analysis. Eun-Wie Cho: Conceptualization. Haryoung Poo: Conceptualization, Validation, Writing – review & editing, Funding acquisition, Supervision. Tae Seok Seo: Conceptualization, Supervision, Writing − review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Biographies

Dr. Jihyun Yang is a researcher in Infectious Disease Research Center, Korea Research Institute of Bioscience and Biotechnology. She obtained the Ph.D. degree from Seoul National University and postdoctoral work at KRIBB. Her research focuses on studying the vaccines and adjuvants and developing a diagnosis of viruses.

Vu Minh Phan received the Bachelor degree in Cell and Molecular Biology from Vietnam National University of Science in 2019. Currently, he is Integrated Master and Ph.D. candidate in Chemical Engineering from Kyung Hee University. His research interests include lab-on-a-chip and microfluidic device for biomolecular analysis.

Dr. Chang-kyu Heo completed his PhD in Bioscience and Biotechnology at Chungnam National University in Daejeon, Korea in 2020. Then he has been a postdoctoral research at the Korea Institute of Bioscience and Biotechnology since 2020.

Hau Van Nguyen received the MSc degree in Chemistry from Thammasat University in 2016. Currently, he is PhD candidate in Chemical Engineering from Kyung Hee University. His research interests include centrifugal microfluidic devices, and lab-on-a-chip for biomolecular analysis a and chemical synthesis.

WonHee Lim graduated from Konkuk University in 2016. He has been in the M.S and Ph.D integration program in Department of Functional Genomics at University of Science and Technology (UST) since 2016.

Dr. Eun-Wie Cho is a senior researcher at Korea Research Institute of Bioscience and Biotechnology in Daejeon. In 2001 she received her Ph.D. in Biological Sciences from KAIST and started her research career at KRIBB. Her research is focused on developing tumor biomarkers, mainly based on tumor-associated autoantibodies.

Dr. Haryoung Poo is a principal investigator in Infectious Disease Research Center, Korea Research Institute of Bioscience and Biotechnology. She received the Ph.D. degree in cellular Immunology from Wayne State University and postdoctoral research in Internal Medicine of University of Michigan. Dr. Poo’s research focuses on studying vaccines, vaccine adjuvant, and diagnosis of influenza virus and corona virus.

Tae Seok Seo is a professor of Department of Chemical Engineering, Kyung Hee University, South Korea. He received his PhD degree from Columbia University and finished his Post-doctoral fellowship at UC Berkeley under the guidance of Prof. Richard A. Mathies. His current research interests are centered on Lab-on-a-chip, microfluidics, genomic technology, single cell assay, and nanobiotechnology.

Appendix A. Supplementary material

Supplementary material

.

Supplementary material

.

Data availability

No data was used for the research described in the article.