African green monkey kidney cell line (Vero, ATCC® CRL-1586, Manassas, Virginia, USA) was used as host cells in this study. YF virus strain 17D was used. Dulbecco’s modified Eagle’s medium (DMEM, Gibco/BRL, Gaithersburg, Maryland, USA) with 2% FBS (Gibco/BRL, 16000-044) and 0.05 mg/ml gentamycin (Gibco/BRL, 15710-064) was used as culture medium. When the Vero cells grew to 70–80% confluency, the culture was washed with PBS (pH 7.4) and then dengue, zika, chikungunya, mayaro and YF viruses were inoculated at a concentration of 0.5×10⁴ pfu/ml. After inoculating each virus, the flask was shaken intermittently for 1 hr every 10 to 20 min to facilitate adsorption of the virus. After washing with PBS (pH 7.4), the flasks were added with DMEM containing 2% FBS and then cultured at 37°C in a 5% CO₂ incubator. When the cytopathic effect (CPE) of cells was observed, the culture medium was harvested and centrifuged for 10 min at 8,000 rpm to obtain the supernatant.

In order to purify YF NS1 protein, we used an affinity column that was developed in a preceding study [16] in which anti-dengue NS1 polyclonal antibodies are conjugated with cyanogen bromide (CNBr) beads (GE Healthcare, Chicago, Illinois, USA). After equilibrating the affinity column with 20 mM KH₂PO₄ (pH 7.4), the acquired YF virus culture medium was loaded. After washing the unbound proteins using the same buffer, YF NS1 bound to the column was eluted with 0.1 M glycine (pH 2.8). The pH level was adjusted by adding 1 M Tris-HCl (pH 7.9) solution to 1/10 th volume of the eluted solution. The eluted YF NS1 was dialyzed with PBS (pH 7.4) and stored at 4°C. The purified YF NS1 was dried by 1 μl dripping method on nitrocellulose membrane. The nitrocellulose dot strips were blocked for 2 hr with PBS containing 5% skim milk, and then the antigen diluted with dilution solution (1% BSA in PBS) was added to each strip and reacted for 30 min. Each dot strip was washed with PBS-Tween for 1 min 5 times, and then reacted for 30 min at room temperature with gold-conjugated protein G diluted 1:1,000 in dilution solution. Finally, it was washed with distilled water and the reaction stopped. In this study, we obtained YF NS1 5 mg/L from the yellow fever culture media, 1 week post infection.

Procedures to produce mAb against NS1 of YF were followed the methods described earlier [16,17] with some modifications. Briefly, a 6-week SPF female mouse (Samtaco, Suwon, Korea) was immunized primarily with yellow fever NS1 25 μg in 25 μl PBS mixed with Freund’s Complete adjuvant (Sigma Chem Co., St. Louis, Missouri, USA) of 1:1 volume. Two weeks later, secondary immunization was carried out using incomplete adjuvant (Sigma) by the same method. Two weeks later, antibody titration was identified through indirect ELISA with blood collected from the tail. An absorbance of 1: 1,000 diluents of 0.2 or higher seemed successful immunization, and 25 μg of YF NS1 50 μl was injected into the mice’s feet 5 days before cell fusion to boost the immunity.

Myeloma cells (SP2/0 cells, ATCC CRL1581) were cultured in T-75 flasks using DMEM containing 20% FBS. The cells were fused with lymph node cells of the immunized mouse under the polyethylene glycol 1500 (PEG 1500, Sigma). Then, after reacting for 5 min at 37°C, the hybridoma cells were cultured in HAT culture medium containing 20% FBS. When the cells multiplied to 60–70% of the well plate at the bottom, the antibody titer was measured by ELISA using a plate coated with YF NS1 protein and the well supernatant in addition to differentiated with cross-reactions using plates coated with Zika and Dengue viruses. Wells with high antibody titers were moved into 24 well plates and cultured according to limiting dilution so that one cell per well would fit in a 96-well plate, and finally a monoclonal cell line was established.

Murine ascites were harvested after 1 week injecting 0.5 ml of incomplete Freud’s adjuvant in the abdomen of a female adult mouse with at least 5×10⁷ hybridoma cells diluted into 0.5 ml PBS (pH 7.4). The collected ascites was centrifuged at 4°C, 8,000 rpm for 10 min, and then the antibody fractions were collected with 20% to 50% ammonium sulfate precipitation. The precipitate was solved in 20 mM KH₂PO₄ (pH 7.4) and dialyzed (Spectrumlabs, 132724) with the same buffer. Then, it was loaded into a protein-G column equilibrated with 20 mM KH₂PO₄ (pH 7.4) to purify antibody by adding 1 M Tris-HCl (pH 7.9) solution. The eluted antibodies were dialyzed with PBS (pH 7.4) and refrigerated. The purified antibodies were quantified at 280 nm, and the protein was confirmed with 12% SDS-PAGE by Coomassie blue staining.

To perform an isotype analysis, an ELISA kit was composed of a certain amount of goat antibody (Sigma) specific for mouse antibody isotypes (IgG1, IgG2a, IgG2b, IgG3, IgM, IgA) to aliquot into immune plates. After 1 hr reaction at 37°C in PBS (pH 7.4), the plate was blocked with 200 μl PBS containing 10% BSA, 5% sucrose, and 0.1% Tween-20 at 37°C for 1 hr, the remaining PBS was removed and dried for 16 hr at room temperature before use. After aliquoting the collected monoclonal culture medium into each well, it was reacted for 1 hr at 37°C, and then washed 3 times with PBS-T (pH 7.4). Peroxidase-conjugated goat anti-mouse IgG (Sigma, A9917) which is specific to mouse IgG is reacted at 37°C for 1 hr, and then substrate containing color fixing agent is used to induce colorization and identify isotypes.

The dissociation constant (Kd) for each monoclonal antibody against antigens was measured by modifying the method of Friquet et al. [18]. After reacting serially diluted antigens and antibodies at a certain concentration for 10 min, dissociation constant was calculated using antigen-coated plates. A dissociation constant value in the range of 10−6 to 10−7 means a low coupling force between antigen-antigen, whereas a value in the range of 10−9 to 10−12 means a high coupling rate.

RDT kit was composed basically of test and control lines which were divided on the nitrocellulose membrane [19]. Membranes were obtained from Advanced Microdevices (Ambala Cantt, Haryana, India) while the absorbant, sample, conjugate, and blood separation pads were obtained from Ahlstrom (Helsinki, Finland). The test line contained antibody at 1.5 mg/ml that worked as a capture site, while the control line contained goat antibody at 2.5 mg/ml. Gold nanoparticle (40 nm in mean diameter) conjugate pads were made by soaking the antibody conjugated gold solution in glass fibers and then drying them, while the sample pads were soaked in cellulose fibers containing 0.5% polyvinyl alcohol and Tween-20 and then drying them. All pads were attached overlapping so that the sample solutions could migrate.

Monkey tissue samples were provided by Bahiafarma (based on the IRB Committee of Bahiafarma 2016-01-02), while negative blood samples were provided by the Catholic University of Korea (based on the IRB committee No. MC18SES10013). Specimens from Bahiafarma and Catholic University were all confirmed by RT-PCR. In brief, we first extracted RNA using QIAamp® Viral RNA Mini Kit (QIAGEN, Hilden, Germany). The reagent and 10 μl of RNA prep sample were then mixed using the RealStar® Yellow Fever Virus RT-PCR Kit 1.0 (Altona, Hamburg, Germany). The final mixture was loaded and run in the RT-PCR machine (CFX96™, Bio-Rad, Hercules, California, USA) after setting the temperature profile. The monkey tissues were crushed with a homogenizer with PBS (pH 7.4) at 1 ml per g tissue, and then used. The samples received were verified 15 min after dropping 100 μl onto the sample pad. The control line was checked in every test, and results without control lines were re-examined. When a band appeared on the test line, it was determined to be positive.

The analysis of purified YF NS1 with SDS-PAGE showed a band of approximately 45 to 50 kDa (Fig. 1A). To confirm the antigenicity of the purified YF NS1, a dot blot assay was performed. As a result, a distinct positive reaction was observed only in the general sample in which the anti-flavivirus NS1 polyclonal antibodies were spiked, and not in normal serum (Fig. 1B). YF NS1 with confirmed purity and antigenicity was used for immunization.

Immunization was performed using YF NS1 with confirmed antigenicity. The fused cells were subjected to a screening process to confirm the secretion of antibodies specifically binding to YF NS1. Screening was performed by indirect ELISA and lateral flow assay (LFA). Using indirect ELISA, 10 candidate clones were selected, in which the antibody titer was high. In addition, screening was carried out using plates coated with Zika NS1 and Dengue NS1 to confirm cross-reaction with other flaviviruses. Among the 10 highly reactive candidate clones, 7 clones were selected, except for those (1A7, 5G8, 5G9) that cross-reacted with Zika or Dengue. In order to select clones that recognize the tertiary structure, LFA was used to confirm that the candidate clones (2B10, 5A6), which did not show cross-reaction in indirect ELISA, showed cross-reaction between Zika and Dengue. Therefore, 3A11, 4H5, 3B7, 6C11, and 3C4, which are specific to yellow fever, and do not cross-react with Dengue and Zika, were selected finally. The cloning procedure was then performed twice to establish a monoclonal cell line (Fig. 2).

The antibody was isolated and purified from the ascites using ammonium sulfate precipitation method and protein-G column affinity chromatography. As a result of SDS-PAGE, bands were observed near 50 kDa and 25 kDa (data not shown), that correspond to the heavy chain and light chain in a usual antibody. Isotype and dissociation constants were determined to characterize the purified monoclonal antibodies (3A11, 3B7, 3C4, 4H5, 6C11). All monoclonal antibodies belonged in the IgG class and were confirmed to be isotypes IgG1 and IgG2a. The dissociation constants were determined 10⁻⁸–10⁻⁹, and the antibodies with the highest affinity were 3A11 and 3B7 (Table 1).

1.5 mg/ml of the 5 candidate clone antibodies (3A11, 3B7, 3C4, 4H5, 6C11) were applied on nitrocellulose membranes and gold nanoparticle conjugate solution was prepared for the pair test. Antibodies for capture and antibodies for gold nanoparticle conjugate were tested in 25 cases. Antibodies 3B7 for capture, and 3A11 for gold nanoparticle conjugate showed the best sensitivity (Fig. 3). These 2 monoclonal antibodies were selected as the best pair to apply in developing kits for the detection of YF NS1.

Detection limit was analyzed by the recombinant protein and the virus culture were diluted by 1/5, which resulted in the positive up to 0.8 ng/ml of recombinant NS1 protein and 1.6 pfu/ml of supernatant from virus culture. (Fig. 4).

Both positive and negative monkey tissues from Bahiafarma, and negative blood samples from the Catholic University of Korea were used for clinical trials. Specificity refers to the value that a person without disease is judged to be normal through a diagnostic test. A total of 100 sera and 50 whole bloods were tested for specificity with the YF NS1 kit. No false positives were observed, the background was clean, whereas control lines were definite. Sensitivity refers to the value that a person with a disease is defined as being positive through a diagnostic test. Positive monkey tissue was used to investigate the sensitivity of the kits produced in this study.

Tests were carried out with virus cultures to determine whether cross-reaction with other epidemic Flaviviruses and Alphaviruses occurred. No cross-reaction was observed with Dengue virus type 2 (DENV2), Dengue virus type 3 (DENV3), ZIKV, Chikungunya (CHIKV) or Mayaro virus (MAYV) (Fig. 5).