Research ArticleFLAVIVIRUSES

A recombinant platform for flavivirus vaccines and diagnostics using chimeras of a new insect-specific virus

See allHide authors and affiliations

Science Translational Medicine  11 Dec 2019:
Vol. 11, Issue 522, eaax7888
DOI: 10.1126/scitranslmed.aax7888

Fighting flavi with flavi

Insect-transmitted flaviviruses can cause hemorrhagic fever in humans and contribute to morbidity and mortality worldwide. Hobson-Peters et al. isolated a new flavivirus from mosquitoes in Australia that can only infect insect cells. Binjari virus permits swapping of structural proteins from other flaviviruses such as dengue or Zika virus. The chimeric viruses grow to high titers in insect cells but do not infect human cells. They can be used to efficiently produce vaccines or antigens for diagnostics, overcoming safety and manufacturing hurdles of other approaches. A chimeric Binjari virus platform could facilitate and possibly accelerate development of much needed diagnostics and interventions for flaviviruses.


Flaviviruses such as dengue, yellow fever, Zika, West Nile, and Japanese encephalitis virus present substantial global health burdens. New vaccines are being sought to address safety and manufacturing issues associated with current live attenuated vaccines. Here, we describe a new insect-specific flavivirus, Binjari virus, which was found to be remarkably tolerant for exchange of its structural protein genes (prME) with those of the aforementioned pathogenic vertebrate-infecting flaviviruses (VIFs). Chimeric BinJ/VIF-prME viruses remained replication defective in vertebrate cells but replicated with high efficiency in mosquito cells. Cryo–electron microscopy and monoclonal antibody binding studies illustrated that the chimeric BinJ/VIF-prME virus particles were structurally and immunologically similar to their parental VIFs. Pilot manufacturing in C6/36 cells suggests that high yields can be reached up to 109.5 cell culture infectious dose/ml or ≈7 mg/liter. BinJ/VIF-prME viruses showed utility in diagnostic (microsphere immunoassays and ELISAs using panels of human and equine sera) and vaccine applications (illustrating protection against Zika virus challenge in murine IFNAR−/− mouse models). BinJ/VIF-prME viruses thus represent a versatile, noninfectious (for vertebrate cells), high-yield technology for generating chimeric flavivirus particles with low biocontainment requirements.


Flaviviruses such as dengue virus (DENV), yellow fever virus (YFV), Zika virus (ZIKV), West Nile virus (WNV), and Japanese encephalitis virus (JEV) are major causes of human morbidity and mortality globally (summarized in table S1). Flaviviruses are transmitted to humans via the bite of an infected arthropod vector, such as mosquitoes or ticks. Commercial vaccines are available for DENV (Dengvaxia), YFV (YF-Vax), JEV (ChimeriVax-JE/IMOJEV and IXIARO), and WNV (equine) and are in development for ZIKV and WNV (human). Dengvaxia and ChimeriVax-JE are live attenuated chimeric vaccines, with the structural protein genes (prME) of DENV serotypes and JEV, respectively, added to the backbone of the YFV vaccine strain 17D (1). Replacements for these vaccines are being sought to address safety concerns associated with live attenuated vaccines and manufacturing issues (summarized in table S1) (1).

Diagnosis of flavivirus infections is achieved by a spectrum of different tests (24). For instance, the most widely used DENV diagnostic test involves detection of anti-DENV immunoglobulin M (IgM) by enzyme-linked immunosorbent assay (ELISA) using DENV envelope (E) protein antigens (3, 5). A new technology that allows simultaneous evaluation of IgM responses to multiple viruses is the multiplex microsphere immunoassay (MIA) (6). Inactivated virus (6) or recombinant viral proteins (7) are coupled to beads, and IgM from serum samples bound to the bead-bound antigen(s) is then detected by flow cytometry. There is also a considerable need for rapid point-of-care tests (8); unfortunately, such tests currently have relatively poor performance profiles (9). Cross-reactivity between different (often cocirculating) flaviviruses remains an issue for many serodiagnostic tests (2). Neutralization tests are often viewed as the gold standard for differential serodiagnoses of flaviviruses but require live infectious virus (7, 10). Because ZIKV, the Kunjin strain of WNV (WNVKUN), and DENV are usually classified as biosafety level (BSL)–2 and WNV, YFV, and JEV are classified as BSL-3, production of live virus for these assays requires appropriate biocontainment infrastructure, training, and safety protocols. Although existing live attenuated or chimeric vaccine strains (e.g., ChimeriVax) have potential use as safer options for producing diagnostic antigens, they remain infectious for humans. Exposure risks via aerosols or needle stick injuries are thus retained for attenuated viruses, especially when high-titer stocks are being used during manufacturing.

Aside from the vertebrate-infecting flaviviruses (VIFs) described above, viruses in the genus Flavivirus also include insect-specific flaviviruses (ISFs). Recently, there has been an increased appreciation of the complexity of the viromes of mosquitoes and other insects (11), and we have isolated and characterized a number of insect-specific viruses (1214) and ISFs (1518). ISFs, in contrast to VIFs, are unable to replicate in vertebrate cells and phylogenetically cluster into two lineages. Lineage I ISFs were the first to be found and are phylogenetically quite distant from VIFs (19). Lineage II ISFs represent a much less studied group of viruses that are more closely related to VIFs and may have evolved from VIFs but, at some point in their evolution, have lost their ability to replicate in vertebrates (19).

Here, we describe the discovery of a new lineage II ISF, named Binjari virus (BinJV). BinJV emerged to be remarkably tolerant for the exchange of its prME genes for those of a range of the medically important VIFs described above. We demonstrate that BinJ/VIF-prME chimeric viruses represent a new platform technology for generation of chimeric antigens for a range of VIFs, with these chimeric particles providing authentic antigens for diagnostic and vaccine applications.


Chimeric viruses of a new lineage II ISF, BinJV

A new ISF, designated BinJV, was first suggested by reverse transcription polymerase chain reaction (RT-PCR) analysis of extracts from Aedes normanensis mosquitoes collected near the indigenous Binjari community (Katherine, Northern Territory, Australia). The virus was subsequently isolated in 2016 from a pool of A. normanensis that were trapped at the Bradshaw Field Training Area (BFTA, Northern Territory, Australia). The complete coding region of this isolate (designated BFTA20) was sequenced (GenBank accession number MG587038). Phylogenetic analysis illustrated that BinJV grouped with the lineage II ISF clade with ≥99% support and also clustered with other ISFs isolated from Aedes mosquito species (Fig. 1A and table S2). Consistent with other lineage II ISFs, BinJV replicated in the Aedes mosquito–derived cell line C6/36 (Fig. 1B, C6/36, green staining), but not in a range of vertebrate cell lines. The latter included cell lines from viper (VSW), chicken (DF-1), human [human embryonic kidney (HEK) 293T], and murine embryo fibroblasts from interferon-α/β receptor–deficient mice (IFNAR−/− MEFs) (Fig. 1B and fig. S1).

Fig. 1 BinJV and BinJ/VIF-prME chimeras.

(A) Dendogram showing phylogenetic relationship between BinJV and other flaviviruses using a maximum likelihood model and complete amino acid sequences. GenBank accession numbers are provided in table S2. (B) BinJV was used to infect C6/36 cells (A. albopictus), VSW cells (Russell’s viper), Vero cells (African green monkey kidney), DF-1 cells (chicken embryo fibroblast), and IFNAR−/− MEFs (murine embryonic fibroblasts from interferon α/β receptor–deficient mice). Cells were immunolabeled with anti-BinJV E (BJ-1C1) mAb (green staining) and cell nuclei stained with Hoechst 33342 (blue staining). (C) Schematic of the modified circular polymerase extension reaction (CPER) strategy to generate infectious DNA of chimeric BinJ/VIF-prME viruses. The prME genes of each indicated VIF (red arrows) were inserted into the BinJV backbone (blue arrows) (replacing the BinJV prME). OpIE2-CA, a modified Orgyia pseudotsugata multicapsid nucleopolyhedrosis virus immediate-early 2 promoter; HDVr-pA, hepatitis delta virus ribozyme–poly A, with the ribozyme autocleavage providing an authentic 3′ untranslated region (UTR). (D) IFA staining of C6/36 cells infected (MOI, 0.1) with diluent (Mock), WNVKUN, or BinJ/WNVKUN-prME virus and immunolabeling with anti-VIF E (4G2) and anti-WNV NS1 (3.1112G) mAbs. (E) Growth kinetics of BinJV, BinJ/WNVKUN-prME, and WNVKUN after infection of C6/36 cells (MOI, 0.1). Three biological replicates, with the lower limit of detection for each individual sample of 2 log10CCID50/ml. (F) As for (D) using ZIKV and BinJ/ZIKV-prME virus and anti-E (4G2) and anti-ZIKV NS1 (Z3H3) mAbs. (G) As for (E) using ZIKV and BinJ/ZIKV-prME. (H) As for (D) using DENV2 and BinJ/DENV2-prME virus and anti-E (DEN-1B7 or 4G2) and anti-DENV NS1 (DEN-1H7) mAbs. (I) As for (E) using DENV2 and BinJ/DENV2-prME virus.

Chimeric BinJ/VIF-prME viruses were constructed using a modified circular polymerase extension reaction (CPER) methodology (2022), with the BinJV prME genes replaced with the prME genes from a series of VIFs (Fig. 1C). Generation of replication-competent BinJ/WNVKUN-prME virus (Fig. 1, D and E), BinJ/ZIKV-prME virus (Fig. 1, F and G), and BinJ/DENV2-prME virus (Fig. 1, H and I) was demonstrated by immunofluorescence assays (IFAs) and viral expansion in C6/36 cells. Deep sequencing confirmed the correct amino acid sequences (table S3). Similar IFA and replication data were obtained for BinJ/DENV1-prME, BinJ/DENV4-prME, BinJ/YFV-prME, and BinJ/JEV-prME (fig. S1).

The BinJ/VIF-prME chimeras often replicated to higher titers in C6/36 cells than their parental VIFs (Fig. 1, E, G, and I, and fig. S1, B and C). For instance, BinJ/ZIKV-prME and BinJ/DENV2-prME virus reached 109.47 and >108.35 cell culture infectious dose (CCID50)/ml, respectively, ≈200- and >800-fold higher than their parental VIFs (Fig. 1, G and I). Together, these data attest to the high level of tolerance BinJV has for substitution of its prME genes with prME genes from a range of VIFs, without loss of replication efficiency, in comparison to wild-type (WT) BinJV.

BinJ/VIF-prME viruses do not replicate in vertebrate cells

To illustrate the host specificity of BinJV/VIF-prME viruses, WNVKUN, BinJV, and BinJ/WNVKUN-prME viruses were used to infect C6/36 cells and a series of mammalian cell lines. All three viruses infected C6/36 cells (at 28° and 34°C) as illustrated by virus-specific immunolabeling (Fig. 2A, green staining, top rows of both composite panels). WNVKUN infected C6/36 cells and all the mammalian cells at all temperatures tested (Fig. 2A, green staining, second column). Neither BinJV nor BinJ/WNVKUN-prME showed any immunolabeling at 37° or 34°C in the mammalian cell lines (including IFNAR−/− MEFs) (Fig. 2A, 37° and 34°C). The lack of immunolabeling at 34°C illustrated that this was not a temperature-dependent phenomenon (23).

Fig. 2 BinJV and BinJ/VIF-prME viruses do not replicate in vertebrate cells.

(A) IFA analysis by confocal microscopy of WNVKUN, BinJV, and BinJ/WNVKUN-prME virus–infected C6/36 cells and mammalian cells. Cells were fixed and immunolabeled 5 days after infection and culture at 37°C (top) or 34°C (bottom). Mammalian cells: BSR (baby hamster kidney), WT MEFs (wild-type murine embryonic fibroblasts), IFNAR−/− MEFs, and Vero cells (African green monkey kidney). Viruses were immunolabeled with mAb 4G4 (green), and cell nuclei were stained with Hoechst 33342 (blue). (B) IFA analysis of C6/36 and Vero cells transfected with RNA derived from WNVKUN, BinJV, and BinJ/WNVKUN-prME viruses. Cells were fixed and immunolabeled 5 days after transfection after culture at the indicated temperatures as described for (A). (C) RNA from BinJ/WNVKUN-prME virus, WNVKUN, or BinJV was used to transfect the indicated cell lines (x axis) and cultured for 5 days. Virus titers in the supernatants (three biological replicates) were then determined by CCID50 assays using C6/36 cells and mAb 4G4. The limit of detection was 2 log10CCID50/ml. (D) BinJ/VIF-prME viruses and WNVKUN were used to infect C6/36 cells and a panel of vertebrate cells. Cells were fixed and immunolabeled as in (A) 5 days after infection and cultured at 28°C (C6/36) or 37°C for the vertebrate cell lines: VSW (viper), 3CPL (crocodile), DF-1 (chicken), Vero (monkey), and IFNAR−/− MEFs (mouse). As a positive control, all the vertebrate cell lines were shown to be replication competent for WNVKUN (bottom row). Note that the IFA images for the infection of Vero cells and IFNAR−/− MEFs with BinJ/WNVKUN-prME appear in (A).

When mammalian cell lines were transfected with viral RNA, the same infection pattern emerged (Fig. 2B), illustrating a post-entry restriction of replication of BinJV and BinJV chimeras. To further illustrate this restriction, a range of cell types (Fig. 2C, x axis) were transfected with RNA from BinJ/WNVKUN-prME, BinJV, or WNVKUN. As expected, high amounts of infectious virus were detected in the supernatants of C6/36 cells for all three viruses (Fig. 2C, C6/36), and WNVKUN RNA produced infectious virus in all the cell lines (Fig. 2C, red squares). No infectious virus was detected after transfection of BSR (baby hamster kidney), IFNAR−/− MEFs, or Vero (African green monkey) cell lines with BinJV or BinJ/WNVKUN-prME viral RNA (Fig. 2C, ND). These cell lines are ordinarily highly permissive for flavivirus infection.

An expanded panel of vertebrate cell lines, which included viper (VSW), crocodile (3CPL), chicken (DF-1), and human (HEK293T) cell lines, was infected with BinJ/ZIKV-prME, BinJ/DENV2-prME, and BinJ/WNVKUN-prME viruses. All viruses replicated in C6/36 cells, and WNVKUN replicated in all cells, but no virus-specific fluorescent signal was seen in vertebrate cell lines infected with BinJ chimeras (Fig. 2D and fig. S1G). Inoculation of high titers of BinJ/ZIKV-prME virus into immunocompromised mice [IFNAR−/− and NRG (NOD.Cg-Rag1tm1Mom Il2rgtm1Wjl/SzJ)] also failed to show any viral genome amplification (fig. S2).

BinJ/ZIKV-prME virus particles are structurally and antigenically authentic

The structure of mature and immature BinJ/ZIKV-prME virions was determined by cryo–electron microscopy (cryo-EM) (Fig. 3, fig. S3, and table S5). The mature structure (Fig. 3, A to D, and fig. S3, A and B) and immature structure (fig. S3, C to H) of the BinJ/ZIKV-prME virions were consistent with those previously reported for mature ZIKV (24) [Protein Data Bank (PDB): 5IRE] and immature ZIKV (25, 26) (PDB: 5UWV). The mature BinJ/ZIKV-prME comprised virions of ≈50 nm in diameter, with a classical flavivirus architecture (Fig. 3A). The three-dimensional (3D) structure of the chimeric particle was resolved using single particle analysis (SPA) to a resolution of 7.1 Å [Fourier shell correlation (FSC) of 0.143] (fig. S3G) and displayed the typical “rafts” of three head-to-tail E dimers that are organized in a herringbone-like pattern across the pseudo-icosahedral virion surface (Fig. 3A). Clear densities attributable to the virion membrane bilayer and transmembrane helices were also observed (Fig. 3B). Rigid body fitting of the icosahedral asymmetric unit derived from the atomic-level structure of the WT ZIKV (PDB: 5IRE) into our resolved map of the chimeric virus revealed a very close fit (Fig. 3, C and D) [fit-in-map cross-correlation of 0.908, calculated in University of California San Francisco (UCSF) Chimera; table S5]. Densities corresponding to the single glycosylation moiety within ZIKV E were visible within the resolved map, matching the position previously reported (24, 26) for ZIKV (Fig. 3D).

Fig. 3 Cryo-EM 3D reconstructions of BinJ/ZIKV-prME chimeras.

(A and B) Surface and central slab representations of the reconstructed BinJ/ZIKV-prME cryo-EM map. E proteins are represented in cyan, blue, and pink corresponding to the three positions within the asymmetric unit. Two-, three-, and fivefold symmetry axes are indicated. (C and D) Fitting of the published high-resolution cryo-EM prM/E raft structure of wild-type ZIKV (PDB: 5IRE) to the 3D reconstruction of mature BinJ/ZIKV-prME is shown from top and side-on positions. The high-resolution cryo-EM model is shown in ribbon representation and colored with the three E domains I, II, and III in red, yellow, and blue, respectively. E stem and M helices are colored light blue and peach, respectively. Glycan densities are indicated. (E to G) Kd values for binding of mAbs to the indicated BinJ/VIF-prME and corresponding VIFs; each dot represents one mAb (mAbs are described in table S6). Black, E-specific mAbs; yellow, E domain II–specific mAbs; purple, E domain III–specific mAbs; cyan, pr-specific mAbs. Statistics were performed using Pearson correlations. Quaternary epitope binding mAbs C8 and C10 are indicated in (F). (H) Cryo-EM micrograph of BinJ/ZIKV-prME. (I) Cryo-EM micrograph of BinJ/ZIKV-prME complexed with C8 Fab (an E dimer–specific monoclonal antibody). (J) Three-dimensional reconstruction of the BinJV/ZIKV-prME particle in complex with C8 Fab, calculated to a resolution beyond 10 Å. The reconstruction is colored to provide radial depth (color key inset in angstroms). Examples of projecting Fab molecules are indicated with arrows. (K) Side view cut through of the complex reconstruction, with densities from the asymmetric unit fitted with ZIKV structural proteins (PDB: 5IRE). BinJV/ZIKV-prME particle colored as in (D). Atomic structure of Fab C8 (PDB: 4UTA) is fitted to the remaining density and shown in pale green. (L) Top-down view of fitted cryo-EM reconstruction colored as in (K).

Cryo-EM of BinJ/ZIKV-prME virions produced in the presence of NH4Cl, which raises the pH of the secretory pathway thus preventing efficient virion processing, revealed a mixed virion population with a majority of “spiky” immature particles (fig. S3C). The immature structure was resolved by SPA to 12.2 Å at an FSC of 0.143 (fig. S3, D to F and H) and closely resembles the published immature ZIKV structure (PDB: 5UWV), with a cross-correlation of fit of 0.936 (table S5).

Flavivirus virus-like particles (VLPs) can display structural irregularities, and quaternary epitopes may not be authentically displayed (27, 28). The antigenic authenticity of three BinJ/VIF-prME viruses was therefore investigated by determining the apparent dissociation constants (Kd) of a panel of monoclonal antibodies (mAbs) (table S6) for (i) the parental VIF and (ii) their respective BinJ/VIF-prME. Pearson correlations illustrated that apparent Kd values for mAb binding to WNVKUN and BinJ/WNVKUN-prME were nearly identical (Fig. 3E). Similar data were obtained for the binding of 12 mAbs to ZIKV and BinJ/ZIKV-prME (Fig. 3F) and the binding of 10 mAbs to DENV2 and BinJ/DENV2-prME (Fig. 3G). These data illustrate a high level of antigenic authenticity for BinJ/VIF-prME chimeric particles as compared to their WT VIF counterparts.

The C8 mAb (indicated in Fig. 3F) is a highly neutralizing, cross-reactive mAb that recognizes a complex quaternary epitope present only on the surface of mature E dimers from multiple viruses (29). Incubation of BinJ/ZIKV-prME particles (Fig. 3H) with C8 Fab resulted in clear complex formation, with projections evident on the virions’ surfaces (Fig. 3I). To confirm the binding of C8 to the chimeric particle, the structure of the complex was determined by cryo-EM. The reconstructed map was resolved to 8.2 Å at an FSC of 0.143 (fig. S3I and table S5) and revealed well-defined Fab projections (Fig. 3J, arrows), with a total of 180 Fab molecules bound per particle, representing full occupancy (Fig. 3J).

The atomic resolution of the ZIKV cryo-EM structure (PDB: 5IRE) and the published C8 Fab crystal structure (29) was fitted to the resolved map (Fig. 3, K and L). The independently fitted Fab structure sits directly above the known C8 binding site (Fig. 3, K and L), which spans contact residues across the E dimer surface (30).

BinJ/VIF-prME production and purification

The growth in C6/36 cells in 1 liter of batch cultures of BinJ/WNVKUN-prME was compared with BinJ/WNVKUN-NSW-prME (New South Wales 2011 strain of WNVKUN), which has an E glycan at N154) and BinJ/WNVKUN-prME+gly (F156S substitution to introduce N154 glycosylation site). All these viruses and WNVKUN showed similar growth kinetics in vitro (Fig. 4A). N-linked glycosylation was confirmed by peptide N-glycosidase (PNGase) F digestion (Fig. 4B). Purified viruses were analyzed by SDS–polyacrylamide gel electrophoresis (PAGE), with viral structural proteins clearly evident (Fig. 4C). Yields were higher for viruses with N154 glycans [consistent with previous reports (31)], with a mean of ≈4 to 5 mg/liter of total protein and ≈2 to 3 mg/liter of E protein (Fig. 4D). Yields compare favorably with mammalian cell culture technologies for production of (i) VLPs, 3 mg/liter of E for a DENV VLP (32), and (ii) VIFs, 1.6 × 108 plaque-forming units/ml for ZIKV (33) (compare with fig. S1F). Asymmetrical flow field–flow fractionation with multiangle light scattering (AF4-MALS) analysis of a BinJ/DENV2-prME chimeric virus preparation illustrated a dominant peak, with the expected size range, illustrating a relatively homogeneous preparation of virions (Fig. 4E).

Fig. 4 BinJ/VIF-prME chimeric virus manufacture.

(A) Growth of the indicated viruses at 28°C after infection of C6/36 cells at a MOI of 0.1. BinJV/WNVKUN-NSW-prME contains the prME gene of the WNVKUN-NSW isolate, which is naturally glycosylated. BinJV/WNVKUN-prME+gly contains an F156S substitution creating a glycosylation site on N154. Mean of three biological replicates is shown. Limit of detection is indicated by a dotted line. (B) Western blot of culture supernatants from virus-infected C6/36 cells, treated with and without PNGase F. (C) Examples of chimeric virion preparations (purified by polyethylene glycol precipitation and ultracentrifugation through a sucrose cushion) analyzed by SDS-PAGE. The flavivirus structural proteins (E, envelope; prM, premembrane; C, capsid; M, membrane) are indicated. (D) Total protein and E protein yields from three independent 1-liter C6/36 cultures as determined by BCA (total protein) or densitometry comparison of the E protein band to a BSA standard after SDS-PAGE. (E) Asymmetrical flow field–flow fractionation with multiangle light scattering (AF4-MALS) analysis of BinJ/DENV2-prME chimeric virions (purified as above). UV, ultraviolet.

Use of BinJ/VIF-prME chimeric antigens in diagnostic immunoassays

An example of a MIA for a single flavivirus-positive clinical reference sample is shown (Fig. 5A), illustrating the simultaneous testing for IgM responses to multiple flaviviruses (6). The MIA for this patient indicates a recent infection with DENV2, with DENV2 and BinJ/DENV2-prME giving similar mean fluorescence intensity (MFI) values (Fig. 5A). Seven panels of sera (each with 80 flavivirus antibody-positive clinical reference and 15 negative control samples), each with known IgM reactivity for each of the indicated seven VIFs (Fig. 5B), were used to compare the performance of VIFs and BinJ/VIF-prME viruses. Highly significant correlations and P values were obtained for all seven panels (Fig. 5B), illustrating that BinJ/VIF-prME chimeric antigens are fully functional in MIAs.

Fig. 5 Utility of BinJ/VIF-prME reagents for flavivirus diagnostics.

(A) Example of a MIA performed on the serum of one patient. MFI, mean fluorescence intensity. (B) Seven panels of human serum samples, each containing known positives for the indicated VIFs, were used to compare the performance of BinJ/VIF-prMEs and the parental VIFs. Seven correlations are shown with Pearson’s rho and P values provided; n = 95 serum samples for each panel and graph. (C) Comparison between WNVKUN and BinJ/WNVKUN-prME virus–infected C6/36 cells as antigens in a fixed-cell IgG ELISA. SE of triplicates is shown after background has been subtracted. Horse serum samples were taken before and 21 to 26 days after WNVNY99 infection. (D) As for (C) but using purified virus particles adsorbed onto the ELISA plate as antigens. (E) As for (C) but using viruses captured from culture supernatants onto the ELISA plates using mAb 3.67G. (F) Testing of a panel of flavivirus-seronegative, WNVKUN-seropositive (Australia), and WNV-seropositive (USA) patient serum samples in an IgG capture ELISA. SE of triplicates is shown after background has been subtracted. (G) Testing of a panel of flavivirus-seronegative and DENV2-seropositive patient serum samples in a capture ELISA as in (F) but using mAb 6B6C-1. (H) Testing of a panel of flavivirus-seronegative and ZIKV-seropositive patient serum samples in a capture ELISA as in (F) but using mAb 6B6C-1. Two samples (indicated by arrows) were negative for IgG but were positive for IgM. (I) Microneutralization assay for BinJ/ZIKV-prME and wild-type ZIKV virus using serial dilutions of mAb C8 and C6/36 cells. The 50% inhibitory concentration (IC50) is shown for each virus. The mean of three biological replicates is shown.

Panels of equine and human sera were compared for their reactivities to BinJ/VIF-prME chimeric antigens and their parental VIFs in a series of IgG ELISAs. Similar optical density (OD) values were obtained when either BinJ/WNVKUN-prME or WNVKUN viruses were used as the antigen in (i) fixed-cell ELISAs (Fig. 5C), (ii) antigen-coated ELISAs (Fig. 5D), or (iii) antigen capture ELISAs (Fig. 5E). For panels of human antisera, capture IgG ELISAs with WT or chimeric particles as antigens gave very similar OD values (Fig. 5, F to H). Together with the MIA data, these experiments demonstrate very similar behavior of BinJ/VIF-prME chimeric viruses and parental VIFs in various serodiagnostic settings.

Using C6/36 cells and the neutralizing mAb C8, BinJ/ZIKV-prME and parental ZIKV were shown to be neutralized at similar concentrations of C8 mAb [50% inhibitory concentration (IC50) of 1.66 and 1.98 nM, respectively] (Fig. 5I). These IC50 values are comparable with previous reports for this mAb (29, 30). BinJ/VIF-prME viruses can thus also be used (with C6/36 cells) in flavivirus microneutralization assays.

Utility of BinJ/ZIKV-prME chimeric particles as vaccine antigens

A male IFNAR−/− mouse model of ZIKV infection and testes damage (34, 35) was used to evaluate BinJ/ZIKV-prME chimeric particles as a vaccine antigen. Mice were vaccinated once with 2 or 20 μg of BinJ/ZIKV-prME chimeric virus (Fig. 6A), which resulted in the induction of significant ZIKV-specific IgG1 and IgG2c (P = 0.009 and P = 0.013 for 2 and 20 μg of BinJ/ZIKV-prME, respectively; Fig. 6B), total IgG (P = 0.009 and P = 0.013; Fig. 6C), and neutralizing antibodies (P = 0.045 and P = 0.013; Fig. 6D). For the latter two parameters, the 20-μg dose induced significantly higher responses than 2 μg (P = 0.045; Fig. 6, C and D). Both doses of chimeric virions reduced viremia to below detection after challenge with ZIKVPRVABC59 (Fig. 6E). Post-challenge body weight loss was not seen in mice vaccinated with 20 μg of BinJ/ZIKV-prME chimeric virus and was significantly reduced [compared with the phosphate-buffered saline (PBS) control] for the 2-μg chimeric virus dose (P = 0.028; Fig. 6F). At week 6 after challenge (Fig. 6A, week 13), chimeric virus vaccination was shown to prevent reduction in testes weights (Fig. 6G), as well as testes shrinkage and hemorrhage (Fig. 6H). In the control group, but not the BinJ/ZIKV-prME virus vaccinated groups, hematoxylin and eosin (H&E) staining showed destruction of testes morphology (Fig. 6I), and immunohistochemistry (IHC) illustrated abundant ZIKV antigen (Fig. 6J). Enlargements are shown in Fig. 6K.

Fig. 6 BinJ/ZIKV-prME chimeric virus vaccination and challenge of male IFNAR−/− mice.

(A) Timeline of experiment after a single intramuscular vaccination with 2 or 20 μg of BinJV/ZIKV-prME virus particles, or PBS control. (B) End point anti-ZIKV IgG1 and IgG2c ELISA antibody titers. The limit of detection was 1 in 100 dilution. ND, not detected. Statistics were performed using Kolmogorov-Smirnov tests. (C) As for (B) for total anti-ZIKV IgG responses. (D) ZIKV-specific neutralizing antibody titers. The limit of detection was 1 in 30 dilution. (E) Mean viremias after subcutaneous challenge with 103 CCID50 of ZIKVPRVACB59. The limit of detection was 2 log10CCID50/ml for individual serum samples. n = 5 to 6 mice per group. (F) Percent weight change after challenge. Statistics were performed using repeated-measures ANOVA. n = 5 to 6 mice per group. (G) Weight of testes. Statistics were performed using Kolmogorov-Smirnov tests. n = 5 to 6 mice per group. (H) Photos of selected testes. Arrows indicate overt hemorrhage. (I) H&E staining of testes. (J) IHC of testes using 4G4 (anti-flavivirus NS1) mAb. (K) High-resolution images of testes shown in (I) and (J). Left: H&E/IHC pair from the 2-μg BinJ/ZIKV-prME virus vaccine group showing intact architecture and minimal staining. Right: H&E/IHC pair; PBS control group with dotted ovals show concordance between testes damage by H&E and IHC labeling for ZIKV.

BinJ/ZIKV-prME chimeric virus vaccination also protected female IFNAR−/− mice (fig. S4). Formulation of BinJ/ZIKV-prME virus particles with the adjuvant AddaVax provided no additional benefit in male (fig. S5) or female (fig. S6) vaccinated IFNAR−/− mice.


Here, we describe the discovery of a new lineage II ISF (BinJV) and construction of BinJ/VIF-prME chimeras for a range of medically important flaviviruses. The BinJ/VIF-prME chimeras replicated to high titers in C6/36 cells to produce antigenically authentic flavivirus chimeric particles suitable for diagnostic and vaccine applications. BinJ/VIF-prME was unable to replicate in mammalian cells in vitro (even in IFNAR−/− cells) or in vivo in immunocompromised mice. The biocontainment requirements are thus substantially lower for BinJ/VIF-prME chimeras than for their parental VIFs. Furthermore, in the context of rapid response to emerging outbreaks, authentic chimeric viruses for a new or emerging flavivirus disease can be derived using the BinJV chimeric platform in as little as 2 to 3 weeks, including the import of a gene block coding for the new flavivirus premembrane (prM) and E genes. This represents a unique and powerful pipeline for rapid response to outbreaks of new flavivirus diseases.

Many flavivirus serodiagnostic tests use inactivated virus or recombinant E proteins as antigens (table S7). Inactivation may compromise antigenic authenticity (36), and recombinant proteins are often poor at presenting tertiary and/or quaternary epitopes (3739). BinJ/VIF-prME chimeric virions are correctly folded and do not require inactivation; however, current regulations require all whole virus antigens to be inactivated before use in flow cytometry, so we were unable to assess any potential improvements in MIAs using BinJ/VIF-prME chimeras that have not been inactivated with binary ethyleneimine.

Most current flavivirus vaccines for human use are live attenuated and unsuitable for immunocompromised individuals and pregnant women (40). A range of inactivated VIF vaccines have been developed or are in development, and should improve safety profiles, but require BSL-2 or BSL-3 biocontainment during production, inactivation (traditionally formaldehyde fixation) (36), adjuvanting, and multiple vaccinations (table S8) (41, 42). Flavivirus VLP vaccines are also in development and do not require inactivation. However, they are usually manufactured in mammalian tissue culture systems, which are costly, often have poor yields (43), and also usually require adjuvanting and multiple immunizations (table S8). BinJ/VIF-prME chimeric viruses grow to high titers and do not need to be inactivated because they are incapable of replicating in vertebrate cells. Unexpectedly, formulation with adjuvant (AddaVax) did not enhance immunogenicity in mice, suggesting that these chimeric virions have some inherent self-adjuvanting activity. Although we have not established or characterized this potential adjuvant activity, we speculate that this may be associated with the presence of viral RNA (44). Perhaps of note, BinJV-chimeric viruses do not replicate in human (or vertebrate) cells and therefore present a negligible risk of recombination with circulating WT viruses in vaccine recipients.

Hypersensitivity or allergic reactions to egg proteins from vaccines grown in embryonated chicken eggs (e.g., YFV 17D and influenza vaccines) or chicken embryonic fibroblasts (e.g., the measles, mumps, and rubella vaccine) are well described (45). Although a vaccine grown in C6/36 cells might arguably cause similar problems (46), purified vaccine preparations would contain only residual levels of mosquito proteins. Severe reactions to mosquito bites are uncommon (47), and individuals living in high-risk areas for flaviviral infection may also be desensitized to mosquito proteins (48, 49). No swelling or redness was detected at the injection sites, and all body temperatures remained normal, after vaccination of nonhuman primates with an alphavirus vaccine concentrated from C6/36 cell supernatants (50).

Vaccine or biologic production in C6/36 cells has yet to be developed, and as far as we are aware, no good manufacturing practice (GMP)–certified C6/36 cell line has yet been generated. The insect-specific Eilat virus technology being developed for alphavirus diagnostics and vaccines (51, 52) is similarly confronted with this development hurdle. An encouraging recent observation is that C6/36 cells do not contain adventitious viruses (53). Another potential limitation inherent in such technologies is the potential for replication, recombination, and onward transmission. However, these represent poorly explored issues for many attenuated arboviral vaccines, and for flavivirus systems at least, recombination events appear to be extremely rare (54, 55).

In summary, we describe herein a versatile technology for the production of antigenically authentic flavivirus chimeric viruses that are replication defective in vertebrate cells, providing substantial amelioration of biocontainment costs and avoiding issues associated with inactivation. The Australian Office of Gene Technology Regulator (DIR-159) and Australian Pesticides and Veterinary Medicines Authority (permit no. 86845) recently provided approval for an experimental BinJ/WNVKUN-prME vaccine for animals farmed outdoors under minimal biocontainment conditions (essentially BSL-1/PC1) to prevent WNVKUN infections. This represents a very promising ruling for future development of this technology for human applications.


Study design

The aim of this study was to investigate whether the prM and E structural genes of BinJV, a new lineage II ISF, could be efficiently exchanged with those of medically important flaviviruses (e.g., WNV, ZIKV, and DENV) to produce chimeric viruses that could be grown in mosquito cells at high yields but were incapable of replicating in vertebrate cells. We conducted experiments to assess the structural and antigenic authenticity and host-restriction profile of chimeric viruses constructed using the BinJV genetic backbone and the growth kinetics of each chimeric virus assessed in mosquito cells. Host restriction for the BinJ/VIF-prME chimeric virus replication in vertebrate cells was also assessed in flavivirus-permissive cells in vitro and in highly immunocompromised mice (IFNAR−/− and NRG) in vivo. The use of the chimeric particles for diagnostic applications was assessed using assays routinely used for flaviviral disease diagnosis, including MIAs, ELISA, and neutralization assays. An IFNAR−/− mouse model of ZIKV infection was used to test for the protective capability of the BinJ/ZIKV-prME virions as vaccine antigens. Group sizes were selected on the basis of our experience with these systems. All mouse work was conducted in accordance with the “Australian code for the care and use of animals for scientific purposes” as defined by the National Health and Medical Research Council of Australia. Mouse experiments and associated statistical treatments were reviewed and approved by the QIMR Berghofer Medical Research Institute animal ethics committee (P2195). mAb generation was approved by the University of Queensland Animal Ethics Committee (SCMB/AIBN/016/16/QAP, SCMB/329/15/ARC, and MICRO/PARA/487/05/CRC). Human serum samples used in these studies were deidentified. Investigators were not blinded when conducting or evaluating the experiments, and no randomization was necessary. No data were excluded from this study. Primary data are reported in data file S1.

Cell culture

C6/36 cells (Aedes albopictus; ATCC CRL1660) were maintained in RPMI 1640 supplemented with 5% fetal bovine serum (FBS) at 28°C. Vero cells (Cercopithecus aethiops, African green monkey kidney; ATCC CRL1586), BSR cells (Mesocricetus auratus, baby hamster kidney; ATCC CCL-10), and MEFs (Mus musculus, primary mouse embryo fibroblasts) derived from WT and IFNAR−/− mice (21) were grown in Dulbecco’s modified Eagle medium (DMEM) supplemented with 5% FBS. HEK293T cells (HEK containing SV40 T-antigen; ATCC CRL-3216) were cultured in DMEM supplemented with sodium pyruvate and 5% FBS. All cell culture media were supplemented with penicillin (50 U/ml), streptomycin (50 μg/ml), and 2 mM l-glutamine. Vertebrate cells were grown at 37°C (unless indicated) with 5% CO2.

Mosquito collection, detection, and isolation of BinJV

The initial indication of a new flavivirus occurred after trapping and processing of mosquitoes collected near the Binjari community, Northern Territory, Australia in 2010. RT-PCR and sequencing were performed as described previously (17). Subsequent isolation of BinJV occurred from adult mosquitoes that were collected using CO2-baited light traps from the BFTA in the Northern Territory of Australia in 2014, as described previously (56). Screening of mosquitoes for the presence of RNA viruses by mosquito homogenization, inoculation onto C6/36 cell monolayers, and subsequent ELISA assessments were performed as described previously (18, 57). Initial confirmation of the BinJV isolation was achieved using RT-PCR and pan-flavivirus–specific primers (FU2 and cFD3) as described previously (17, 58, 59). BinJV stock was generated using the extracted RNA to transfect C6/36 cells, and the virus in the supernatant was expanded by passage in C6/36 cells with the supernatant harvested on day 7 um after inoculation.

Genome sequencing and phylogenetic analysis

RNA from BinJV virions was extracted (QIAamp viral RNA Extraction Kit, QIAGEN) without carrier RNA, deoxyribonuclease (DNase)–treated (Heat&Run DNase, ArcticZymes), and complementary DNA–generated (Protoscript II, New England Biolabs), and was converted to double-stranded DNA using Escherichia coli DNA ligase, DNA polymerase I, and ribonuclease H (New England Biolabs). A library was constructed using the Nextera XT library kit (Illumina) with barcoded primers. The library was sequenced on a NextSeg 500 generating 2 × 151–base pair paired reads. The viral genome was assembled using Geneious R8 software.

Amino acid multiple sequence alignments were performed with MAFFT v7.017 algorithm, using a scoring matrix of BLOSUM62, a gap open penalty of 1.53, and an offset value of 0.123. FastTree 2.1.5 was used to construct a tree that uses the maximum likelihood approximation method, with optimization for Gamma20 likelihood selected. Branch support values were calculated using a Shimodaira-Hasegawa test. These analyses were undertaken within the Geneious R8 package.

Equine and human serum samples

Equine samples were obtained from a previously conducted WNV experimental infection study (60). In that study, the horses were infected with the New York 1999 strain of WNV (WNVNY99) and bled before infection and 21 to 26 days after infection. Pooled sera from WNV-naïve horses from that study were used as an additional negative control.

Panels of deidentified human sera samples were provided by the Public Health Virology Laboratory, Forensic and Scientific Services, Department of Health, Queensland, Brisbane, Australia; diagnosed by various National Association of Testing Authorities, Australia–accredited in-house assays; and confirmed positive for ZIKV, JEV, YFV, DENV, or WNVKUN using a combination of RT-PCR, virus neutralization, and/or other virus-specific serological assays. These patients’ serum samples were referred to the aforementioned Public Health Virology Laboratory from public hospitals or private diagnostic laboratories specifically for flavivirus serology. Serum samples with no known reactivity to flaviviruses were used as negative controls. The deidentified WNV-positive human reference samples were provided by the Centers for Disease Control and Prevention in Atlanta, USA.

Use of human samples in the MIA was approved by the Queensland Health Forensic and Scientific Services Human Ethics Committee (project RSS18-018). Deidentified data were provided for this publication.

ZIKV-positive sample Z11 was collected from an individual (blood collected under approval from the University of Queensland Human Ethics Committee) about 12 months after contracting ZIKV in Ribeirao Preto, Sao Paulo State, Brazil, in March 2016. ZIKV infection was confirmed by RT-PCR and plaque reduction neutralization test (PRNT) (20).

ELISAs with equine and human serum samples

Fixed-cell ELISA. Matched mock–, BinJ/WNVKUN-prME–, and WNVKUN-infected C6/36 cells were fixed 5 days after infection with 20% acetone and 0.02% bovine serum albumin (BSA) in PBS. After blocking with ELISA blocking buffer [0.05 M tris-HCl (pH 8.0), 1 mM EDTA, 0.15 M NaCl, 0.05% (v/v) Tween 20, and 0.2% (w/v) casein], equine serum (diluted 1:200 in blocking buffer) was added in triplicate wells and incubated at 37°C for 1 hour. Wells were washed with PBS containing 0.05% Tween 20 (PBS/T) and horseradish peroxidase (HRP)–conjugated rabbit anti-horse antibody (Sigma-Aldrich) added for 1 hour at 37°C. After washing, substrate solution comprising 1 mM 2,2-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) and 3 mM H2O2 in a buffer prepared by mixing 0.1 M citric acid with 0.2 M Na2HPO4 to give a pH of 4.2 was added (100 μl per well), and plates were incubated in the dark at room temperature for 1 hour. The OD was measured at 405 nm. Averaged mock values for each serum sample were subtracted from the values obtained for BinJ/WNVKUN-prME and WT WNVKUN.

Antigen-coated ELISA. BinJ/WNVKUN-prME and WNVKUN viruses were purified as described for cryo-EM (see the Supplementary Materials). Protein concentrations were determined by bicinchoninic acid (BCA) assays (Pierce), and 80 ng per well of virion antigen in 50 mM NaHCO3 and 50 mM Na2CO3 (pH 9.6) was added to the wells and incubated overnight at 4°C. After washing with PBS/T, the ELISA was performed as described for the fixed-cell ELISA.

Capture ELISA. ELISA plates (high binding, Greiner) were coated with 100 ng per well of purified mAb 3.67G (for WNV assays) or mAb 6B6C-1 (for ZIKV and DENV2 assays) at 4°C overnight or at 37°C for 1 hour in PBS. After washing with PBS/T, the plates were blocked with ELISA blocking buffer, and a predetermined optimal dilution of BinJ/VIF-prME or WT VIF virions from C6/36 culture supernatants was added and incubated for 1 hour at 37°C. The culture supernatant from mock-infected C6/36 cells was used as negative control. After washing with PBS/T, the ELISA was performed as above using equine (1:200 dilution) or human (1:100 dilution) sera assessed in triplicate. For human serum ELISAs, HRP-conjugated rabbit anti-human IgG (Dako) was used. The average binding to the mock C6/36 cell culture supernatant for each serum tested (in triplicate) was subtracted from the BinJ/VIF-prME and WT VIF values.

Microsphere immunoassay

The flavivirus IgM MIA was performed as described previously (6) at the Public Health Virology Laboratory, Forensic and Scientific Services, Department of Health, Queensland, Brisbane, Australia.

Microneutralization assay using mAb C8

PRNT and immune plaque assays were performed essentially as previously described (61). Briefly, twofold serial dilutions of mAb C8 (starting at 50 μg/ml) were mixed 1:1 with ≈50 focus-forming units (FFUs) of WT ZIKV or BinJ/ZIKV-prME and incubated for 10 min at room temperature and then at 37°C for 1 hour, before being added to C6/36 cells for 2 hours at 28°C. The inoculum was removed and overlay medium was added. After incubation for 3 days at 28°C, the overlay medium was removed, and the cells were fixed with cold acetone (80% in PBS). FFU was determined by immune staining as detailed previously (61).

Generation of ZIKVPRVABC59

An infectious clone ZIKVPRVABC59 (GenBank: MH158237.1) was constructed as previously described (62) and was provided by S. Tajima (Department of Virology I, National Institute of Infectious Diseases, Tokyo, Japan). Infectious virus was recovered by transfection of Vero E6 cells. Stock virus for use in neutralization and challenge experiments was generated in C6/36 cells.

Preclinical evaluation of a BinJ/ZIKV-prME chimeric virus vaccine

IFNAR−/− mice on a C57BL/6 J background (63) were vaccinated intramuscularly (50 μl into each quadriceps femoris) (34). Mice were aged 5 to 27 weeks; for each group within each experiment, the same age distribution of mice ±1 week was used. Where indicated, chimeric virions were formulated (1:1) with AddaVax adjuvant (InvivoGen). Antibody responses to ZIKV were determined as previously described (34), except that ZIKVPRVABC59 was used in the neutralization assays and goat anti-mouse IgG1 and IgG2 secondary antibodies (Invitrogen) were also used.

Mice were challenged subcutaneously in the base of the tail with 100 μl of 103 CCID50 ZIKVPRVABC59 with viremia assessed by CCID50 assay as described previously (34). Virus preparations were tested for mycoplasma contamination, and FBS used for propagation was tested for endotoxin. Histology and IHC were performed as described previously (34, 64), with the latter undertaken using the mAb 4G4 (anti-NS1) and the Warp Red Chromogen kit (Biocare Medical). Quantitative RT-PCR analyses (to detect ZIKVPRVABC59 E) were undertaken as previously described (34) but using ZIKV E primers (forward, 5′-CCGCTGCCCAACACAAG-3′; reverse, 5′-CCACTAACGTTCTTTTGCAGACAT-3′).

Statistical analyses

Statistical analysis was performed using SPSS for Windows (version 15.0, 2007; SPSS). For comparison of mouse antibody titers, Kolmogorov-Smirnov tests were used because the differences in variances were >4. For comparisons of weight loss, repeated-measures analyses of variance (ANOVAs) were used with datasets showing differences in variance <4, skewness >−2, and kurtosis <2. For comparison of antibody binding Kd values, the Pearson correlation coefficient (r) was determined with an associated P value; the datasets were paired, normally distributed, and linear.


Materials and Methods

Fig. S1. BinJ/VIF-prME chimeric viruses.

Fig. S2. Inoculation of immunodeficient mice with high titers of BinJ/ZIKV-prME virus.

Fig. S3. Cryo-EM 3D reconstructions and data for BinJ/ZIKV-prME chimeras.

Fig. S4. BinJ/ZIKV-prME vaccination of female IFNAR−/− mice.

Fig. S5. BinJ/ZIKV-prME formulated with and without AddaVax (male mice).

Fig. S6. BinJ/ZIKV-prME formulated with and without AddaVax (female mice).

Table S1. Summary of flaviviruses of medical importance and their respective vaccine status for flaviviruses used in this study.

Table S2. Source of flavivirus sequences used in Fig. 1A.

Table S3. Summary of deep sequencing analyses of indicated chimeric viruses.

Table S4. Previously described and new mAbs used in this study.

Table S5. Refinement statistic details for mature, immature, and C8-complexed BinJ/ZIKV-prME virions to accompany the data in Fig. 3 and fig. S3.

Table S6. Kd values for binding of mAbs to the indicated BinJ/VIF-prME and corresponding VIFs.

Table S7. Examples of detection assays used for diagnostics of flavivirus infection.

Table S8. Examples of inactivated and VLP-based flavivirus vaccines.

Data file S1. Primary data.

References (65145)


Acknowledgments: We acknowledge the facilities and the scientific and technical assistance of the Australian Microscopy and Microanalysis Research Facility at the Centre for Microscopy and Microanalysis and the UQ Protein Expression Facility (PEF), University of Queensland. We thank J. Choo and D. Muller (University of Queensland) for help with the AF4-MALS. We are also grateful to F. Moore (Public Health Virology, Queensland Health) for providing deidentified clinical reference samples and D. Bowen (Colorado State University) for providing equine serum samples. We thank N. Petrovsky (Vaxine, Australia) for provision of Advax adjuvant used for mAb production. We thank D. Williams (CSIRO, Australian Animal Health Laboratory) for supply of VSW cell line and Steven Davis (Berrimah Veterinary Laboratories) for the provision of the 3CPL crocodile cell line. We are also grateful to B. Doms and M. Sanchez (University of Pennsylvania) for provision of anti-WNV antibodies. We thank the animal house and histology services staff at QIMR Berghofer for invaluable assistance. We thank C. O’Brien (the University of Queensland) for technical assistance and V. Lutzky (QIMR Berghofer) for help with the manuscript submission. Funding: This project was funded by the Australian Government National Health and Medical Research Council (NHMRC) project numbers APP1138611, APP1144950, and APP1164216; the Australian Research Council (DP120103994); the University of Queensland strategic funding grant (612023); and an intramural seed grant from the Australian Infectious Disease Research Center. A.A.K. and A.S. are NHMRC Research Fellows. J.J.H., J.E.H., N.D.N., L.J.V., A.M.G.C., and T.B.H.P. were supported by Australian Postgraduate Awards. E.N. was supported in part by the Daiichi Sankyo Foundation of Life Science, Japan. The opinions expressed herein are those of the authors and do not necessarily reflect those of the Australian Defence Force and/or Defence Force Policy. Author contributions: J.H.-P., D. Watterson, H.B.-O., C.T., A.S., and R.A.H. designed the research. J.H.-P., J.J.H., J.E.H., D. Watterson, R.A.H., N.D.N., L.J.V., A.M.G.C., C.T., D. Warrilow, B.H., A.A.A., H.B.-O., E.N., K.Y., B.T., S.W., P.R.M., and M.F. performed experiments. W.K.C., N.K., C.T., Y.X.S., A.A.K., P.R.Y., T.B.H.P., A.M.G.C., J.H.-P., D. Watterson, R.A.H., E.N., and N.M. contributed key reagents or analytical tools. J.H.-P., J.J.H., J.E.H., D. Watterson, L.J.V., A.M.G.C., C.T., D. Warrilow, B.H., A.A.A., A.A.K., A.S., and R.A.H. analyzed the data. J.J.H., D. Watterson, A.A.K., R.A.H., J.H.-P., and A.S. wrote the manuscript. Competing interests: R.A.H. and J.H.-P. are inventors on a patent WO/2018/176075, which includes data reported herein. All other authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials. The GenBank accession number for BinJV coding sequence is MG587038. The EM Data Bank IDs for the cryo-EM reconstructions of the BinJ/ZIKV-prME virions are EMD-20417, EMD-20438, and EMD-20439. Select monoclonal antibodies described in this study are available from Mozzy mAbs ( Other monoclonal antibodies can be provided under a material transfer agreement. These chimeric viruses are available for collaborative research projects under appropriate material transfer agreements.

Stay Connected to Science Translational Medicine

Navigate This Article