Engineered NS1 for Sensitive, Specific Zika Virus Diagnosis from Patient Serology

Dengue virus (DENV) and Zika virus (ZIKV) belong to the Flaviviridae family of viruses spread by Aedes aegypti mosquitoes in tropical and subtropical areas. Accurate diagnostic tests to differentiate the 2 infections are necessary for patient management and disease control. Using characterized ZIKV and DENV patient plasma in a blind manner, we validated an ELISA and a rapid immunochromatographic test for ZIKV detection. We engineered the ZIKV nonstructural protein 1 (NS1) for sensitive serologic detection with low cross reactivity against dengue and developed monoclonal antibodies specific for the ZIKV NS1 antigen. As expected, the serologic assays performed better with convalescent than acute plasma samples; the sensitivity ranged from 71% to 88%, depending on the performance of individual tests (IgM/IgG/NS1). Although serologic tests were generally less sensitive with acute samples, our ZIKV NS1 antibodies were able to complement the serologic tests to achieve greater sensitivity for detecting early infections.

containing 0.05% Tween and 1% BSA. The strips were then added to polyclonal goat antihuman IgM-AP (Fitzgerald, https://www.fitzgerald-fii.com), followed by 10 µL of washing buffer. Subsequently the strip was dipped in 200 µL of BCIP/NBT (MOSS) for 7.5 min. The reaction was stopped by dipping the membrane strip in 0.3 M NaOH. For the IgG test, 5 µL of diluted sample (1:400) was passed through the membrane, followed by 15 µL of the washing buffer. The strips were then added to goat anti-human IgG Fc HRP (Thermo Fisher), followed by 10 µL washing buffer. The membrane strip was dipped in 200 µL of metal-enhanced DAB solution (Thermo Scientific) for 5 min. The strips were dipped in water before imaging. All the strips were imaged using a Bio-Rad ChemiDoc MP Imaging System (https://www.bio-rad.com) on autoscale setting.
Format 2: To prepare the test strips, 1 µL of anti-human IgG and IgM capture antibodies (1 mg/mL) were immobilized on nitrocellulose membrane strips (3 mm width × 25 mm length) at the downstream and upstream portion, respectively, via vacuum drying. The test strips were then blocked with casein (1% w/v), washed with borate buffer, and vacuum dried before use. To prepare the conjugate pad, the H-zMut1 antigen was first conjugated to 40-nm gold nanoparticles (Au NPs) via covalent binding. The conjugated Au NPs were then diluted to 0.5 optical density (OD) using casein buffer, and then dried on glass fiber strips (3 mm width × 30 mm length). The nitrocellulose test strips were assembled with the glass fiber and an absorbent pad (10 mm width × 20 mm length). Patient plasma samples (neat, 5 µL) were applied to the upstream portion of the nitrocellulose test strip and 60 µL of chasing buffer (1× PBS) was then applied to the glass fiber conjugate pad. As the patient plasma and Au NPs flowed past the IgG/IgM test spots, a visible red signal could be observed by the naked eye within 15 minutes.

ZIKV IgM and IgG ELISA Assay
Polystyrene plates were coated overnight with 1 µg/mL of ZIKV NS1-related antigen in PBS buffer, and blocked with blocking buffer (PBS with 10% nonfat dry milk [Bio-Rad]). To perform the ZIKV IgM ELISA assay, patient serum diluted in blocking buffer was mixed with IgG/Rf stripper (Bio-Rad) (e.g., 0.5 µL of sample with 2 µL of IgG stripper in a total of 60 µL to make a 1:120 sample dilution) and incubated for 30-45 min. After incubation, the sample was transferred to the ZIKV NS1-coated plates and incubated for 25 min at 37°C. Plates were washed, and anti-human IgM-HRP (Abcam, https://www.abcam.com) (1:4300 dilution) was added for an additional 10 min incubation at 37°C. Plates were washed again, and the TMB substrate was added for 10 min before stopping with KPL stop solution (SeraCare, https://www.seracare.com). For the ZIKV IgG ELISA assay, blocking buffer containing patient serum (1:250 dilution) was transferred to ZIKV NS1 antigen-coated plates and incubated for 12 min at 37°C. Upon washing, anti-human IgG-HRP (Thermo Fisher) (1:5500 dilution) was added to the plates for 10 min incubation at room temperature, followed by substrate incubation for 7 min, and stopped by adding stop solution and measuring absorbance at 450 nm. Serum samples' mean ODs were measured from 2 replicates. For the ELISA assays performed using serum samples from the training and validation sets, mean OD of the sera samples (P) was divided by the mean OD of an internal standard (I). P/I ratio of >1.5 was considered as Zika positive for both IgM and IgG assays. The internal standard was built based on a commercial dengue sample that consistently showed minimal cross reactivity in our assays.

Monoclonal Antibody Generation and Production
The anti-ZIKV NS1 antibodies were generated in New Zealand white rabbits according to an approved Institutional Animal Care and Use Committee protocol. Briefly, each animal was immunized with recombinant ZIKV WT NS1 (Native Antigen) and the titer was checked after every round of immunization. Upon achieving a high titer, B cells were isolated and sorted for culture. Supernatants from the B cells were tested using ZIKV NS1 antigen ELISA and variable regions were then recovered from positive clones for subcloning into expression vector containing rabbit constant regions. Monoclonal antibodies were expressed in CHO cells for large-scale production according to manufacturer's protocol (Gibco, https://www.thermofisher.com). Antibodies were purified with Protein A resins (Amintra; Abcam), buffer-exchanged into PBS (Gibco), and concentrated using Amicon Ultra centrifugal filters (Merck Millipore).

ZIKV NS1 ELISA
Polystyrene plates were coated with 1 µg/mL of C12 in PBS and incubated overnight at 4°C. The plates were blocked with 2% bovine serum albumin, fraction V (BSA, Capricorn, https://www.capricorn-scientific.com) in PBS before use. After washing with PBS, 45 µL of serially diluted recombinant ZIKV NS1 antigen (ranging from 0 to 6.4 ng/mL) in normal human serum control or patient sample were co-incubated with 5 µL of 10% BSA and 1% PBST for 1 hr at 37°C; 10 µg/mL of recombinant DENV1 NS1 in normal human serum control was included as negative control. Plates were then washed 5 times with 0.2% PBST and once with PBS. Next, an optimized amount of biotinylated C11 in 1% BSA 0.1% PBST was added for a 1-hour incubation at room temperature. The plates were washed and incubated at room temperature with streptavidin-poly-HRP (SDT) diluted in 1% BSA 0.1% PBST for 30 min. The plates were then developed with 100 µL of TMB for 15 min and terminated with 50 µL of stop solution. The absorbance at 450 nm was taken using a Tecan M200 plate reader (Thermo Fisher). The limit of detection was set at 2× the OD450 of the background (normal human serum). The ZIKV NS1 level in the samples was estimated through interpolation from a standard curve. The cutoff was established by 3 standard deviations (SD) from the mean values of 45 DENV patient samples interpolated from the standard curve. *A total of 35 patient specimens were assayed for detecting ZIKV IgM and IgG. Increase in the normalized OD for both IgM and IgG can be observed in most cases at the second time point, upon disease progression (30/35 cases); 28 of the 35 patients were tracked from acute to convalescent phase (n = 28). Normalized OD >1.5 was highlighted in gray. dpo, days post onset of symptoms; OD, optical diameter.