During the initial rounds of cloning and screening of the hybridomas, two hybridomas (1A11 and 2B10) were selected for their ability to recognize major proteins from HeV and NiV infected Vero cell lysates (Figure 1A). The 1A11 antibody recognized a protein similar in size to the N protein (~58 kDa) from HeV as well as from both strains of NiV (Malaysia and Bangladesh). The 2B10 antibody detected a protein of slightly less than 100 kDa from 2 NiV strains and HeV. It also weakly reacted with a protein similar in size to the N protein in all infected lysates (Figure 1A). These two hybridomas were subjected to further cloning and screening against cell lysates containing individual NiV Malaysia protein P, V, W, C or N, that were expressed from plasmid DNA. The resulting antibody from 1A11 C1 was specific for the N protein (Figure 1B). Antibody from 2B10 p4 strongly recognized the P protein (migrated with apparent molecular weights between 80 and 98 kDa) and a more weakly product of approximately ~40 kDa, which likely represents a degradation or premature termination product. In addition, 2B10 p4 could also weakly detect the V and W proteins (55 kDa, Figure 1B). The C protein which is translated from an alternative reading frame on the P gene was not recognized by either antibody (Figure 1B). Antibody isotypes of 1A11 C1 and 2B10 p4 were determined as IgG2a and IgG1, respectively.

To determine the linear epitope of monoclonal antibody 1A11 C1, 24 or 25-mers of non-overlapping peptides spanning over the entire N protein sequence of NiV were screened by direct ELISA. Only a C-terminal peptide (a.a. 509-532) yielded specific signals above the 10 ng/ml concentration of plate-coated peptide (Figure 2). The epitope in this C-terminal region was not only detected by NiV hyperimmune mouse ascites fluid (HMAF) but also by HeV HMAF, rabbit anti-HeV serum and a pool of NiV sero-positive swine sera from the 1999 Malaysia outbreak (Figure 3). The overall pattern of shared epitopes among polyclonal antibodies could be observed regardless of the source of immunogen (NiV or HeV) or the mammals from which the antibodies were generated (Figure 3). Interestingly, the rabbit anti-HeV serum bound to a distinct epitope (a.a.101-124) which was not recognized by antibodies generated from other animal species (Figure 3). Several human convalescent sera from the 1999 outbreak did not react to the C terminal peptide of the N protein (Figure 3).

The V and W proteins of henipaviruses are transcribed from the same reading frame as the P protein until reaching the internal mRNA editing site in the P gene [32-34]. We have shown that 2B10 p4 recognized protein products generated from P gene transcripts (NiV P, V, and W proteins) by Western blot (Figure 1B). This suggests that the epitope of 2B10 p4 is located within the common N-terminal sequence (a.a. 1-407) of these proteins. However, linear epitope mapping over this region did not result in any peptide binding even with the inclusion of reported single phosphorylated site at Ser-240 [21] (Data not shown).

Monoclonal antibodies 1A11 C1 and 2B10 p4 were further analyzed to test their abilities to capture native viral proteins on ELISA plates. Antibody 1A11 C1 was able to detect NiV Malaysia, NiV Bangladesh and HeV from infected Vero cell lysates (Figure 4A, B and 4C). Successful detection of N proteins was achieved at dilutions ranging 1:1 to 1:7290, with the cutoff value derived from 3 times the standard deviation of the average OD of the uninfected cell lysate controls being 0.2-0.3. The P/V/W specific antibody 2B10 p4 could only detect NiV proteins at very low dilutions (1:1 to 1:270, Figure 4A and 4B); however, applying infected cell lysate at a dilution less than 1:270 resulted in increased non-specific signals (e.g. Marburg hemorrhage fever (MHF) virus HMAF coated controls in Figure 4A, B and 4C). The 2B10 p4 captured HeV P/V/W protein at cell lysate dilutions of 1:1 to 1:2340 (Figure 4C).

Serial dilutions of titrated virus stocks (NiV Malaysia prototype and HeV) were prepared in buffer containing nonionic detergent and tested on antigen capture ELISA coated with 1A11 C1 or 2B10 p4 (Figure 5A and 5B). Antibody 1A11 C1 was capable of capturing NiV or HeV at virus titer of log 3.6 pfu/mL (400 pfu per well). The anti-P 2B10 p4 antibody detects HeV at a comparable sensitivity as 1A11 C1 (Figure 5B), but had poor ability to bind NiV (Figure 5A) similar to results obtained with the infected Vero cell lysate (Figure 4A). Dilutions of Lassa virus were included as the background control to calculate the cutoff value of the assay (0.21, Figure 5A and 5B).

In order to evaluate antigen capture ELISA, cell suspensions from γ-irradiated pig tissue specimens from the Malaysian outbreak in 1999 were prepared as target antigens on plates coated with monoclonal antibody 1A11 C1. The results of the antigen capture assay are shown alongside data from RT-PCR, virus isolation, immunohistochemistry (IHC), and antibody detections [10] in Table 1. A low titer of NiV N antigens was detected in the lung of pig 55 using the antigen capture assay (Table 1). Positive RT-PCR, virus isolation and IHC results also confirmed the existence of NiV in the lung of this pig (Table 1). However, the antigen capture assay did not detect viral antigen in the lung and brain of Pig 5, although RT-PCR and virus isolation performed at the time of the outbreak confirmed NiV infections in the lung of Pig 5 (Table 1). No virus was detected in the brain, lung, or kidney of Pig 4 and 59 by capture ELISA, RT-PCR, virus isolation, or IHC (Table 1). No IgM or IgG response was detected during the time of outbreak in the serum of these pigs (Table 1). Due to the shortage of outbreak specimens, we were unable to determine the sensitivity and specificity of antigen capture ELISA on tissue samples.

Hendra virus strain 9409-30-1800 Australia was originated from a horse lung sample from Brisbane, Australia in 1994. Both Nipah virus strain SPB199901924 Malaysia prototype and strain SPB200401617 Bangladesh were isolated from CSF of patients in the 1999 and 2004 outbreaks, respectively. Lassa virus Josiah strain isolated from human in Sierra Leone was used as the control virus stock. Viruses were inoculated in Vero-E6 cell and propagated until CPE reached 3-4+ before harvest. Viruses were titrated by plaque assays on Vero-E6 cells as described [48]. The attached cells were scraped and centrifuged, washed before lysed in borate saline containing 1% Triton X-100, pH 9. The sample was frozen at -70°C and gamma irradiated at 5 × 10⁶ Rad. The resulting material was sonicated and centrifuged to obtain cell lysate. Control Vero cell lysate was made the same way except no virus infection was performed.

For pig tissue samples from 1999 Malaysia outbreak [10], 10% (wt/vol) suspensions of thawed frozen tissue sections (~250 mg) were homogenized on ice in Hank's balance salt solution (HBSS)/5% fetal calf serum with a plastic pestle and 250 mg of alundum (Fischer Scientific, Pittsburgh, PA) in 15 mL conical tubes. The tissue suspensions were clarified by low speed centrifugation before using in antigen capture ELISA as follows.

Twenty five μL of Nipah virus Malaysian prototype stock was i.c. injected into suckling mice in BSL4 laboratory. The brain of the sick suckling mouse was taken out by day 3, γ-irradiated and used for following immunizations. Four weeks old BALB/C mice (Charles River Laboratories, Wilmington, MA) were immunized i.p. with 10% suckling mouse brain, HBSS and 0.05 M Tris buffer, pH 9 emulsified with Ribi adjuvant (Ribi ImmunoChem Research Inc., Hamilton, MT) in 200 μL. Two booster injections of 100 μL were given on days 21 and 35. Mice serum antibody titers were monitored by ELISA and IFA in between boosters. One hundred μL of brain homogenate without adjuvant was injected at day 56 and 4 days later, the splenocytes of the immunized mice were isolated and fused with NS-1 (TIB-18) myeloma cells. Hybridomas were screened for secretion for desired antibodies by ELISA and IFA. Western blot was used for confirmations of monoclonality and specificity of the antibody. Protein G 8 mL or Protein A 36 mL column connected to an ÄKTAprime™ plus system (GE Healthcare, Piscataway, NJ) was used for purification of monoclonal antibody from supernatants from hybridoma cultures. Antibodies were concentrated in Amicon 50 mL Stirred Ultrafiltration Cells (Millipore, Billerica, MA). The isotype of purified monoclonal antibody was determined by IsoStrip Mouse Monoclonal Antibody Isotyping Kit (Roche Diagnostics, Indianapolis, IN).

HEK 293T cells were transfected with expression plasmids containing P, V, W, C, or N as previously described [34]. Cells were lysed in RIPA buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris Buffer, pH 8). Ten μL of cell lysate was separated on a NuPAGE™ 4-12% Bis-Tris gel (Invitrogen, Carlsbad, CA), then transferred onto a PVDF membrane by iBlot™ Gel Transfer System (Invitrogen). After Blocking with PBS containing 0.05% Tween-20 and 5% skim milk, the membrane was incubated with hybridoma cell supernatant overnight at 4°C. PBS containing 0.05% Tween-20 was used for thorough washing before HRP- conjugated antibody was added to incubation. Signals were developed by using SuperSignal™ West Dura Extended Duration Substrate (Thermo Scientific, Waltham, MA).

Non-overlapped 24, 25 or 26-mer peptides spanning N-terminal half (a.a. 1-407) sequence of Nipah P protein (NCBI Reference Sequence, Accession no: NP_112022) and the entire sequence (a.a. 1-532) of Nipah N protein (NCBI Reference Sequence, Accession no: NP_112021) were synthesized and RP-HPLC purified by the Biotechnology Core Facility in CDC. Peptides were dissolved in water before transferring to microtiter plates, and dried at 37°C overnight as described [49]. The wells were blocked with 100 μL of PBS containing 0.1% Tween-20 and 5% skim milk for 1 hour then washed with PBS containing 0.1% Tween-20. The monoclonal antibody (2B10 p4 or 1A11 C1) or polyclonal antibody (NiV HMAF, HeV HMAF, rabbit anti-HeV serum, NiV infected human convalescent or a pool of NiV seropositive swine sera) was diluted into 1 μg/mL (monoclonal) or 1 to 1000 (polyclonal) with blocking buffer. The antibody was added and incubated for 1 hour at 37°C. The secondary HRP-conjugated goat anti-mouse IgG (Thermo Fisher Scientific, Rockford, IL), goat anti-rabbit IgG (Bio-Rad, Hercules, CA), mouse anti-human IgG Fc (Accurate, Westbury, NY), or goat anti-swine IgG (KPL, Gaithersburg, MD) in 1:8000 was added after washing. After one hour of incubation, color development was measured as described in the following ELISA protocol.

The design and setup of antigen capture ELISA for henipavirus were based on the assay developed for Ebola virus as described previously [43,50]. Five μg/mL of monoclonal antibody (1A11 C1 or 2B10 p4) in 100 μL/well were coated onto each well of microtiter plates (BD Falcon, San Jose, CA) overnight at 4°C. Wells were blocked for an hour at 37°C with PBS containing 0.1% Tween-20 and 5% skim milk then washed with PBS containing 0.1% Tween-20, which also included in subsequent steps. Henipavirus stocks, infected Vero-E6 cell lysate or euthanized pig tissue suspension were serial-diluted with PBS containing 0.1% Tween-20 and 5% skim milk in 100 μL on plate then incubated for one hour at 37°C. Rabbit anti-HeV polyclonal antibody diluted 2000 fold was added and incubated for an hour at 37°C. The wells were incubated with goat anti-rabbit HRP conjugate (Bio-Rad) at a dilution of 1:8000 for 1 hour at 37°C. The peroxidase reaction was developed with ABTS (2, 2'-azino-bis (3-ethylbenzthiazoline-6-sulphonic acid)) substrate system (KPL) for 30 min and optical density (OD) was read at 410 nm (Dynatech MR5000). The OD value subtracted by the background value of control uninfected Vero cell lysate or tissue suspension incubated with coated Marburg virus HMAF (1:1000) was indicated as "Adjusted OD". Test samples were considered positive if their mean OD were greater than the mean OD of uninfected Vero cell lysate or tissue suspension incubated with coated monoclonal antibodies plus 3 times of their standard deviation (indicated as the signal cutoff value).