Rescue and characterization of recombinant cedar virus, a non-pathogenic Henipavirus species

Background Hendra virus and Nipah virus are zoonotic viruses that have caused severe to fatal disease in livestock and human populations. The isolation of Cedar virus, a non-pathogenic virus species in the genus Henipavirus, closely-related to the highly pathogenic Hendra virus and Nipah virus offers an opportunity to investigate differences in pathogenesis and receptor tropism among these viruses. Methods We constructed full-length cDNA clones of Cedar virus from synthetic oligonucleotides and rescued two replication-competent, recombinant Cedar virus variants: a recombinant wild-type Cedar virus and a recombinant Cedar virus that expresses a green fluorescent protein from an open reading frame inserted between the phosphoprotein and matrix genes. Replication kinetics of both viruses and stimulation of the interferon pathway were characterized in vitro. Cellular tropism for ephrin-B type ligands was qualitatively investigated by microscopy and quantitatively by a split-luciferase fusion assay. Results Successful rescue of recombinant Cedar virus expressing a green fluorescent protein did not significantly affect virus replication compared to the recombinant wild-type Cedar virus. We demonstrated that recombinant Cedar virus stimulated the interferon pathway and utilized the established Hendra virus and Nipah virus receptor, ephrin-B2, but not ephrin-B3 to mediate virus entry. We further characterized virus-mediated membrane fusion kinetics of Cedar virus with the known henipavirus receptors ephrin-B2 and ephrin-B3. Conclusions The recombinant Cedar virus platform may be utilized to characterize the determinants of pathogenesis across the henipaviruses, investigate their receptor tropisms, and identify novel pan-henipavirus antivirals. Moreover, these experiments can be conducted safely under BSL-2 conditions.


Background
Hendra virus (HeV) and Nipah virus (NiV) are the prototypical viruses of the genus Henipavirus [1], and are notable highly pathogenic zoonotic paramyxoviruses that have caused numerous severe and often fatal acute respiratory and/or neurologic disease in humans and livestock since their initial recognition in Australia (HeV) and Malaysia, Bangladesh, India and the Philippines (NiV) (reviewed in: [2][3][4]). The identification of HeV following fatal respiratory illness in 17 horses and one human in 1994 [5], was followed by the first recorded spillovers of NiV between 1998 and 1999 in Malaysia, and subsequently Singapore, which caused cases of severe illness in pigs that was also transmitted to farmers and abattoir workers [6][7][8][9][10][11]. A genetically distinct but closely related strain of NiV has been responsible for febrile illness in India and annual outbreaks of acute encephalitis in Bangladesh with notable increased pathogenicity compared to the NiV-Malaysia strain: fatality rates 70-100% [12][13][14]. High pathogenicity and mortality rates associated with HeV and NiV infection have resulted in the classification of both viruses as select agents, and represent the only paramyxoviruses that require biological safety level 4 (BSL-4) containment [15].
The henipaviruses encode two envelope glycoproteins, attachment (G) and fusion (F) glycoproteins, which mediate pH-independent cellular attachment, fusion and virus entry. The functional receptors for HeV and NiV are the highly conserved receptor tyrosine kinase ephrin-B class ligands, ephrin-B2 (EFNB2) and ephrin-B3 (EFNB3) [16][17][18][19][20]. EFNB2 is expressed on vascular endothelial cells and in the brain [21,22] and both HeV and NiV display a tropism for endothelial and neuronal tissue [23][24][25], while EFNB3 is more prominent in the brain and brainstem [22,25,26]. The conserved homology of EFN ligands is thought to facilitate the broad natural and experimental species tropism [10,[27][28][29][30][31], and the physiological distribution of EFNB2 and EFNB3 correlates with pathological outcomes of HeV and NiV infection such as vasculitis, central nervous system involvement and systemic dissemination [23][24][25]. In addition, a more efficient use of EFNB3 by NiV compared to HeV as an alternative receptor has been suggested to contribute to the observed increased likelihood of meningitis and encephalitis with NiV infection [32].
Bats in the genus Pteropus were identified as the natural reservoirs of HeV and NiV [33][34][35][36] and HeV and NiV transmission and spillovers correlated with Pteropus geographical distributions [2,37,38]. Evidence of henipaviruses has been detected in Pteropus lylei populations in Southeast Asia [39,40], Pteropus vampyrus in Indonesia [41] and bat populations endemic to the African continent [42][43][44][45] indicative of a global distribution of henipaviruses. Genetic sequences corresponding to new Henipavirus species have been detected in bats endemic in both Africa and Central America [46,47] and the nearly complete genome of one African henipavirus, Ghanaian bat henipavirus Kumasi Virus (KumPV) has been sequenced [47]. One exception to the preponderance of evidence that Pteropid bats are the natural hosts of henipaviruses, was the detection of Mojiang henipavirus (MojPV) sequences from a rodent host in China [48].
In 2012, Cedar virus (CedPV), a non-pathogenic Henipavirus species was isolated from urine collected under a roost of P. alecto and P. poliocephalus in Australia [49]. Genomic analysis revealed that CedPV was closely related to HeV and NiV, but was distinct in its use of EFNB2, but not EFNB3 for cellular entry, and lack of pathogenicity in animal models of infection [49]. In contrast to HeV and NiV, and nearly all other paramyxoviruses, the CedPV phosphoprotein (P) gene does not undergo RNA editing and does not produce the V or W proteins [49,50]. Both HeV and NiV V and W proteins are potent antagonists of the toll-like receptor signaling and interferon (IFN) pathways [51][52][53][54][55][56], whereas the IFN response was not antagonized by CedPV infection [49,50]. The key role of V protein in pathogenicity was demonstrated by a recombinant NiV, which resulted in a non-lethal, replication competent infection when the V protein was removed [57]. Taken together, the lack of V and W protein expression and inability to utilize EFNB3 could be suggestive of the failure of CedPV to cause clinical disease in animal infection models [49].
Without functional studies the pathogenic potential of novel and related henipaviruses remains obscure, and outside of HeV and NiV, CedPV remains the only documented Henipavirus species isolated. Because CedPV was isolated in a BSL-4 facility it cannot be removed and transferred to a lower containment laboratory. To develop a platform to understand pathogenesis of henipaviruses, we used a reverse genetics approach to rescue replication-competent, recombinant CedPV (rCedPV). Reverse genetic systems have been utilized for the generation of recombinant infectious and replication-competent negative sense RNA viruses with specific mutations and insertions [58,59], particularly NiV and HeV [60][61][62][63][64]. Introduction of reporter genes, such as green fluorescent protein (GFP) or luciferase, provides for an ability to monitor virus replication and spread in real time and/or to perform high-throughput screening [63]. In this study, we describe the rescue of two rCedPV variants, one recombinant wild-type CedPV (rCedPV-wt) and one of which expresses GFP from an additional open reading frame (rCedPV-GFP). We compared the replication kinetics of both rCedPV variants and observed no significant differences between rCedPV-wt and rCedPV-GFP and characterized the ability of rCedPV to activate the IFN pathway. We confirmed that both rCedPV-wt and rCedPV-GFP utilized EFNB2 but not EFNB3 for cellular entry, but also observed rCedPV-GFP replication in cells previously characterized as EFNB2 negative.

Generation of plasmids and rescue of recombinant CedPV
Recombinant CedPV laboratory manipulation guidelines and standard operating procedures under BSL-2 conditions were developed and this work was reviewed and approved by the Uniformed Services University, Institutional Biosafety Committee in accordance with NIH guidelines. To construct the rCedPV antigenome clone, large DNA fragments of CedPV based on the Cedar virus isolate CG1a sequence, NCBI Accession number NC_025351.1 were synthesized (GenScript; NJ, USA). These DNA fragments corresponded to CedPV nucleotide bases: 1-4530, 4531-10,517, and 10,518-18,162. These fragments were sequentially cloned into an expression plasmid, pOLTV5 [65], between the T7 RNA promoter and hepatitis delta virus (HDV) ribozyme. The pOLTV5 vector was similarly used for cloning and expression of recombinant HeV [63]. At the time of the rCedPV cDNA clone design, the nucleotide at position 7 in the CedPV reference genome was a cytosine, which was later revised to an adenine. Our rCedPV cDNA clones have the cytosine at position 7. Furthermore, internal SmaI restriction sites in the CedPV reference genome were removed (C395A and C4816A) in the rCedPV cDNA clone to preserve the cloning strategy. A MluI restriction site was created between the P and M genes at nucleotide position 4531, after the M transcriptional start sequence, to facilitate insertion of a modified turbo Green Fluorescent Protein (GFP) gene (Lonza Inc., Allendale, NJ) [66]. To insert the GFP gene into the rCedPV antigenome plasmid, CedPV untranslated regions: transcriptional P stop, intergenic region, and transcriptional M start sequences (TAAGAAAAAACTTAG-GATCCCAG) were amplified by polymerase chain reaction and cloned into the 3′ terminus of the GFP gene with an additional non-coding 3′ thymine nucleotide to maintain the "rule of six". As noted in the rescue of recombinant HeV [63] and in contrast to the rescue of a recombinant NiV-GFP [60], GFP was inserted between the P and M genes to maintain the level of N and P necessary for proper virus replication. To generate replication helper plasmids, polymerase chain reaction was used to amplify the open reading frames of N, P, and L genes from rCedPV DNA plasmids, which were subsequently cloned into expression vectors that contain a cytomegalovirus promoter (pCMV) [67]. All plasmids were sequenced to obtain at least 2-fold sequence coverage. We adopted reverse genetic methods previously used to rescue recombinant HeV and NiV reporter viruses to generate rCedPV [60,63].
To generate rCedPV-GFP and rCedPV-wt, pCMV-CedPV-N (1.25 μg), pCMV-CedPV-P (0.8 μg), pCMV-CedPV-L (0.4 μg), and pOLTV5 full-length CedPV antigenome plasmid (3.5 μg) were mixed with 12 μL of Lipofectamine LTX (Invitrogen; Carlsbad, CA) in 500 μL OptiMEM (GIBCO; Gaithersburg, MD) and incubated for 30 min at room temperature. This mixture of plasmid DNA and Lipofectamine LTX was used to transfect 5 × 10 5 BSR-T7/5 cells. Four days posttransfection, GFP was observed in BSR-T7/5 cells transfected with the rCedPV-GFP antigenome and N, P, and L helper plasmids. Adherent BSR-T7/5 cells were collected along with culture supernatant and subjected to three rounds of freeze-thaw using − 80°C EtOH and 37°C water baths. Vero cells at a density of 1 × 10 6 cells/well in 6-well cell culture plates (Corning Inc.; Corning, NY, USA) were infected with 500 μL of freeze-thawed cell culture supernatant. Six days post infection, when maximal GFP signal and syncytia were observed, supernatant was clarified by centrifugation (2400 rpm) and 300 μL of virus supernatant was passaged to fresh 1 × 10 6 Vero cells in 6well cell culture plates. Within 24 h post infection (hpi), GFP signal and syncytia were observed. After 3 days, maximal GFP signal was observed and supernatant was again clarified then passaged to fresh Vero cells (75 cm 2 flasks) for amplification of rCedPV-GFP. Virus supernatant was concentrated by ultracentrifugation (28,000 rpm; 2 h) through a sucrose cushioned buffer. Stock rCedPV-GFP was serially diluted and incubated with Vero cells for 72 h to determine titer. Similarly, rCedPV-wt was harvested when maximal syncytia was observed, clarified and passaged onto fresh Vero cells for amplification. Stock titers of rCedPV-wt were determined by CPE-based plaque assay (see below).

Replication kinetics of recombinant CedPV
Vero cells at a density of 5 × 10 4 cells/well in 96-well cell culture plates were infected at a multiplicity of infection (MOI) of 1.0. Supernatants were collected at 0, 8, 24, 48 and 72 hpi and viral titers were determined by plaque assay as described by Weingartl et al. [68]. Briefly, 400 μl/well of virus inoculum was added to Vero cells at a density of 5 × 10 5 cells/well in 12-well cell culture plates and incubated for 1 h at 37°C, 5% CO 2 . Two mL of 2% carboxymethylcellulose sodium salt (medium viscosity) + DMEM-3% CCS was then added to each well and incubated for 5 days at 37°C, 5% CO 2 . The plates were fixed with 4% formaldehyde and then stained with 0.5% crystal violet-80% methanol in phosphate buffered saline (PBS). Plaques were counted and calculated as PFU/mL.

Quantitative reverse transcriptase PCR
HeLa-CCL2 cells were seeded at a density of 1.25 × 10 5 cells/well in a 24-well plate and incubated overnight. Cells were untransfected (mock), transfected with Poly I:C (1 μg/mL) using Lipofectamine 2000 (Thermo Fisher Scientific; MA, USA), or infected with rCedPV-wt (MOI: 0.5, 1.0, 5.0). At 24 hpi, total RNA was extracted using the RNeasy Mini Kit, (Qiagen; MD, USA). An amount of 500 ng of DNase I digested RNA was converted to cDNA using the Superscript III First Strand Synthesis System for RT-PCR (Thermo Fisher Scientific) and oligo(dT) primers. Quantitative reverse transcriptase PCR (q-RT-PCR) was performed using the Power SYBR Green PCR Master Mix (Thermo Fisher Scientific) and the Applied Biosystems 7500 Real-Time PCR System. PCR cycling conditions were: 95°C, 10 min; 40× cycles of 95°C, 15 s; 60°C, 1 min; with a melt curve analysis at the end of each assay. Each sample was analyzed for IFN-α, IFN-β and 18S ribosomal RNA in triplicate and fold changes were calculated relative to 18S ribosomal RNA and normalized to mock samples using the ΔΔCt method. Primer sequences are available upon request.

Ephrin ligand-mediated fusion
HeLa-USU, HeLa-USU-EFNB2, and HeLa-USU-EFNB3 cell lines were seeded at 5 × 10 5 cells/well in 6-well cell culture plates and incubated overnight. The next day, cells were washed with 1× PBS and received fresh cell culture medium with no virus (uninfected) or were inoculated with fresh cell culture media containing rCedPV-GFP or rCedPV-wt at an MOI: 0.1. rCedPV-GFP infected cell cultures were monitored for fluorescence and syncytia and rCedPV-wt infected cells were monitored for syncytia. At 24 and 72 hpi, all cells were fixed with methanol and stained with 0.5% crystal violet-25% methanol. Images were captured with a Zeiss Axio Observer A1 inverted microscope using the 5× objective.

Split-luciferase based cell-cell fusion assay
The quantitative fusion assay was based on a dual-splitreporter assay [69]. HeLa-USU cells (1 × 10 4 in a clear bottom, black wall 96-well plate) were transfected with 60 ng of the expression plasmid for the indicated receptor and 50 ng of the expression plasmid for one half of a split-luciferase reporter protein (DSP1-7, a kind gift of Z. Matsuda). As a control, HeLa-USU cells were only transfected with DSP1-7. Concurrently, HeLa-USU cells (7 × 10 5 in a 6-well plate) were transfected with 500 ng of the other dual-split-reporter expression plasmid (DSP8-11) and 12 h later infected with rCedPV-GFP (MOI: 1.0). As an additional control, HeLa-USU cells were also mock infected. Twenty-four hours postinfection, Versene (0.48 mM EDTA in PBS) (Thermo Fisher Scientific) was used to gently detach the infected HeLa cells from the 6-well plate and 2 × 10 4 cells overlaid on the receptor-expressing HeLa-USU cells in the 96-well plate. EnduRen (Promega; WI, USA) was added as the substrate to the culture medium (DMEM, 10% FBS) according to the manufacturer's instructions. Content mixing between HeLa-USU cells as a result of fusion driven by interactions between virus infected cells and receptor bearing cells was monitored at the indicated times using an Infinite M200 Pro microplate reader (Tecan; Switzerland).

Monoclonal antibody m14F3 neutralization assay
Murine monoclonal antibody (mAb) m14F3, is a CedPV soluble G specific mAb of IgG2 subclass that was developed using standard hybridoma generation techniques by immunization of BalbC mice with purified, recombinant soluble CedPV G glycoprotein produced from a human 293F stably-expressing cell line. HeLa-USU cells were seeded at a density of 5 × 10 4 cells/well in 96-well cell culture plates. The next day, the indicated dilutions of mAb m14F3 were incubated with equal volumes of rCedPV-GFP (MOI: 0.1) for 1 h at 37°C. was removed from the cells and 100 μl of virus or the antibody-virus mixture was added to the wells in triplicate and allowed to incubate for 1 h at 37°C. After the incubation, the antibody-virus mixture and virus only was removed and all cells were washed once with DMEM-10. Fresh DMEM-10 with the varying concentrations of m14F3 were added to the corresponding wells and incubated for 48 h at 37°C. GFP foci were counted with a Zeiss Axio Observer A1 inverted microscope using the 5X objective. The percent neutralization was determined based on the presence of GFP foci in replicate wells at each antibody concentration and calculated based on the reduction in the average number of foci per well to the average number of foci observed in the no antibody control, multiplied by 100 [70].

Statistical analyses
Unless otherwise stated, graphs and images are the average of two independent experiments. All experimental results were expressed as the arithmetic mean. Standard deviations were calculated and represented thusly. All statistical analyses were performed with the unpaired, two-tailed Student T-test using GraphPad's -QuickCalcs software (GraphPad Software Inc.; CA, USA).

Rescue of recombinant viruses
A reverse genetics approach was utilized to produce rCedPV. Two rCedPV antigenome clones were created and used for virus rescue (Fig. 1a, b). The first is a recombinant wild-type CedPV (rCedPV-wt) and the second is a reporter virus that contains a tur-boGFP gene inserted between CedPV P and M genes (rCedPV-GFP). Observation of syncytia in Vero cells infected with cell supernatant passaged from BSR-T7/5 cells transfected with rCedPV antigenome and helper plasmids indicated successful rescue of rCedPV-wt had occurred (Fig. 2a). Rescue of rCedPV-GFP reporter was confirmed by the detection of fluorescent positive syncytia in a monolayer of Vero cells (Fig. 2b).
The replication kinetics of rCedPV in Vero cells infected at a MOI of 1.0 were compared (Fig. 3). At all time points examined, no statistically significant differences between rCedPV-wt and rCedPV-GFP were observed. Although there is approximately a 1 log difference between the viruses at 72 hpi, this was found not to be statistically significant with a p-value of 0.06. This data indicates that introduction of the reporter gene did not interfere with the growth kinetics of these recombinant viruses. Maximal titers of~1-2 × 10 6 PFU/mL were recorded between 48 hpi and 72 hpi.

Recombinant CedPV induces an interferon response
Transfection experiments with CedPV P protein demonstrated that the IFN response in HEK293T or HeLa-   CCL2 cells was less antagonized when to compared to IFN signaling antagonism with HeV P protein [50] and infection with wild-type CedPV stimulated the IFN-β response in infected HeLa-CCL2 cells [49]. To further characterize the phenotype of our rCedPV, we examined the type I IFN response in rCedPV infected human cells.
Poly I:C served as a positive control to demonstrate that the IFN production pathway was functional in the HeLa-CCL2 cells. Total RNA was extracted from uninfected (mock), Poly I:C treated and rCedPV-wt infected HeLa-CCL2 cells at 24 hpi and IFN-α and IFN-β mRNA levels were quantified by q-RT-PCR. As seen in Fig. 4, in contrast to the mock infected samples, rCedPV-wt induced a robust IFN-β response ranging from 33 to 283 fold increase, in a dose-dependent manner. The increase in IFN-β mRNA levels of rCedPV-wt samples was statistically significant at all MOIs tested when compared to the mock infected samples (p < 0.05). In addition, we observed a 1.5-1.7 fold increase of IFNα in rCedPV-wt infected samples when compared to the mock infected samples, which was statistically significant only at the lowest MOI of 0.5 (p < 0.01). This data demonstrates that at 24 hpi, analogous to the wild-type CedPV [49], rCedPV induced a robust IFN-β response, while maintaining IFN-α levels similar to that observed in the mock infected samples. The precise mechanism of the Type-I IFN response in rCedPV infected cells warrants further investigation and is currently under study.

Ephrin receptor tropism of recombinant CedPV
Membrane-bound EFNB2 and EFNB3 facilitate the entry of HeV and NiV into host cells [16][17][18]. Characterization of wild-type CedPV demonstrated that receptor triggered cell fusion was mediated by EFNB2, but not EFNB3 [49]. The HeLa-USU cell line has been used as a henipavirus entry negative cell line and artificial expression of EFNB2 in the HeLa-USU receptor negative background was used to identify EFNB2 as the cellular receptor of both HeV and NiV, and the wild-type CedPV isolate [16,49,71]. To assess whether our rCedPV has similar receptor tropism to the wild-type CedPV isolate, HeLa-USU cells and HeLa-USU cells stably expressing either EFNB2 or EFNB3 were infected with rCedPV-GFP and rCedPV-wt. Unexpectedly, GFP expression was observed in all three cell lines: HeLa-USU, HeLa-USU-EFNB2, and HeLa-USU-EFNB3 (Fig. 5a). However, syncytia following rCedPV-GFP was observed only in HeLa-USU-EFNB2 cells as early as 24 hpi, but not in HeLa-USU-EFNB3 cells at 24 hpi or 72 hpi, which had no observable syncytia and little cytopathogenic effects (CPE) at the later time point. Similarly, the rCedPV-wt infected HeLa-USU exhibited extensive syncytial CPE at 72 hpi when EFNB2 was expressed, and this phenotype was not observed with HeLa-USU cells or HeLa-USU-EFNB3, which remained as healthy monolayers at this post-infection time point (Fig. 5b). Although GFP expression was observed in both HeLa-USU and HeLa-USU-EFNB3 cells infected with rCedPV-GFP, syncytia was not observed and GFP expression appeared to indicate single cell infection. These results further supported the requirement for EFNB2 as a receptor for virus entry and syncytia formation, but also highlighted an uncharacterized, unknown mechanism that facilitated low levels of rCedPV entry that do not result in syncytia and which, without the presence of a reporter gene would have gone unnoticed.
Next, we used a quantitative split-luciferase based reporter assay [69] to compare levels of fusion in standard and EFNB2-positive HeLa-USU cells. In this assay, content mixing between virus-infected effector cells expressing the rCedPV fusion complex and a target cell expressing or not the known henipavirus receptors allows the two halves of Renilla luciferase to functionally interact. As expected, HeLa-USU cells transiently expressing EFNB2 exhibited the highest levels and fastest kinetics of cell-cell fusion (Fig. 6a). On the other hand, the standard HeLa-USU cells also supported cell-cell fusion, although to a 3× lower level and at a 4× slower rate (Fig. 6a, b, respectively). The over-expression of EFNB3 did not increase the level or rate of fusion over that observed for the standard HeLa-USU cells, suggesting that CedPV cannot use this ligand for cell entry. Taken together, the data presented in Figs. 5 and 6 indicated that rCedPV utilized EFNB2 for cellular entry and fusion. Twenty-four hours post-infection, total RNA was extracted, of which 500 ng was used for q-RT-PCR for IFN-α, IFN-β and 18S ribosomal RNA. Each sample was analyzed in triplicate and fold changes were calculated relative to 18S ribosomal RNA and normalized to mock samples using the ΔΔCt method. The graphs are representative of two independent experiments performed in triplicate. Error bars indicate standard deviation.* p < 0.05 and ** p < 0.01

Neutralization of recombinant CedPV-GFP infection
To determine whether the apparent low level of rCedPV infection in HeLa-USU cells observed by GFP reporter signal was CedPV envelope glycoprotein specific entry, we performed a virus neutralization assay targeting the CedPV G attachment glycoprotein with the neutralizing mAb, m14F3. Incubation of rCedPV-GFP with m14F3 completely (100%) neutralized rCedPV-GFP at 10 of the 12 m14F3 concentrations tested. The two lowest concentrations of 0.02 and 0.01 μg/mL decreased rCedPV-GFP entry by 77% and 67% respectively (Fig. 7). These results demonstrated that the low level of rCedPV infection of HeLa-USU cells as detected by the presence of GFP reporter signal is mediated by the G glycoprotein.

Discussion
In this study, we used a reverse genetics approach to rescue rCedPV, and characterized the in vitro functionality of rCedPV. In the first report of CedPV isolation and characterization, Marsh et al. demonstrated that CedPV lacks the coding capacity for the IFN antagonist V protein and that CedPV infection of HeLa-CCL2 cells stimulates the production of IFN-β [49]. Although we cannot directly compare CedPV and rCedPV because the original isolate remains in a secure laboratory, we demonstrate that the rCedPV rescued in this study also stimulates the production of IFN-β. The HeV and NiV P gene encodes accessory proteins, V and W, are potent antagonists of the toll-like receptor signaling and IFN pathways, by sequestering and inhibiting phosphorylation of STAT1 and Fig. 5 Ephrin B2 is recognized as an entry receptor by rCedPV. HeLa-USU, HeLa-USU-EFNB2, and HeLa-USU-EFNB3 cells at a density of 1 × 10 6 cells/well in a 6-well cell culture plates were infected with medium with no virus (uninfected) or with medium containing rCedPV-GFP (a) or rCedPV-wt (b) at an MOI of 0.1. a rCedPV-GFP infected cell cultures were monitored for fluorescence and syncytia by microscopy at 24 and 72 hpi; HeLa-USU-EFNB3 cells 72 hpi are shown. b rCedPV-wt infected cells were monitored for syncytia at 24 and 72 hpi; the latter is shown. Cells were fixed with methanol and stained with 0.5% crystal violet-25% methanol. Images were captured with a Zeiss Axio Observer A1 inverted microscope using a 5X objective. Scale bar at 50 μm and insets show zoom magnified areas of the cell monolayer STAT2 [51][52][53][54][55][56]. This suppression of a type I IFN response may contribute to the high virulence and pathogenicity observed with these viruses [57,62]. In contrast, the CedPV P gene does not undergo RNA editing and therefore does not encode for V or W proteins [49,50]. Furthermore, Lieu et al., performed a sequence alignment of the residues in the P protein that are essential for STAT1 binding and showed that CedPV-P shares only 13% sequence identity with NiV-P and HeV-P in those specified residues and therefore should not interact with STAT1 [50]. Interestingly, in CedPV infected HeLa cells IFN-α mediated STAT1 nuclear trafficking was slightly impaired; however, not to the extent inhibited by HeV [50]. This suggested that CedPV may encode viral factors, such as the C protein produced from an alternate reading frame that may be able to interfere with STAT1 nuclear translocation [50]. In fact, NiV C protein weakly inhibited IFN signaling but the mechanism has yet to be elucidated [62,72]. Henipaviruses have multiple strategies to antagonize the IFN response and therefore to determine the precise relationship between virulence and the inhibition of IFN requires studies with mutantvirus in susceptible animal models.
We confirmed that rCedPV utilizes the HeV and NiV receptor EFNB2 for entry into human cells. Unexpectedly, we observed that rCedPV-GFP infected standard HeLa-USU cells, which are EFNB2-receptornegative, expressed GFP, indicative of rCedPV entry and replication. However, syncytia were only observed in HeLa-USU-EFNB2, but not HeLa-USU or HeLa-USU-EFNB3 cells (Fig. 5a, b). Entry of rCedPV into an EFNB2 negative cell line, HeLa-USU, was corroborated by analysis of virus mediated cell-cell fusion kinetics (Fig. 6a, b), which showed a higher background level of cell-cell fusion than what would be expected for a receptor negative cell line.
Transcriptomic analysis of our parental HeLa-USU cell line demonstrated that this cell line was negative for EFN-B type ligands [16]. Using the formation of giant cells as a marker for CedPV infection, Marsh et al. identified EFNB2 as the receptor for CedPV in the HeLa-USU cell line expressing EFNB2; however, entry into HeLa-USU cells was not noted by this group as only syncytia formation was used as read out of entry [49]. We also did not observe CPE or syncytia in HeLa-USU or HeLa-USU-EFNB3 cell lines, and infection of these cells would not have been noticed if not for the GFP reporter gene (Fig. 5a). In our study, GFP expression in HeLa-USU cells infected with rCedPV-GFP is suggestive that endogenous expression of an uncharacterized cellular receptor is facilitating rCedPV entry. Viral envelope glycoprotein mediated receptor binding and fusion activation facilitates cellular entry of henipaviruses (reviewed in:   [73]) and we demonstrated here that rCedPV entry could be completely blocked by a CedPV G specific mAb. Cell to cell fusion is dependent on both the amount of virus receptor and virus envelope glycoproteins (F and G) present [74,75], which is perhaps why in the presence of an uncharacterized receptor that facilitates either a) low level of entry as seen by GFP expression in single cells or b) inefficient receptor-mediated fusion that does not support giant cell formation. However, at this time we are not able to experimentally rule out the possibility that despite previous transcriptomic analysis of our HeLa-USU cell line, a low-level of EFNB2 expression currently persists and is facilitating the presently observed low-levels of rCedPV-GFP entry.
HeV and NiV do not utilize glycosylphosphatidylinositollinked EFN-A type ligands or EFNB1 for cellular entry [19], and studies of cellular receptors for related henipaviruses, KumPV and MojPV, have revealed some similarities and differences in EFNB2 and EFNB3 recognition. The African henipavirus, KumPV, G glycoprotein mediated receptor binding and triggered fusion with EFNB2 [76,77]. Further analysis confirmed that KumPV G binds to EFNB2, but not EFNB3, and structural analysis demonstrated that despite a low sequence identity between the G glycoprotein of NiV and KumPV, the amino acid residues in the globular domain that bind to EFNB2 are highly conserved [78]. The cellular receptor of MojPV remains elusive, but MojPV does not bind to EFNB2 or EFNB3, and is the first Henipavirus species with this described functional divergence in receptor utility [79]. Despite low sequence identity among the G proteins of CedPV, HeV, NiV, and KumPV, these henipaviruses have highly conserved amino acids in the receptor binding domain of G [22]. Future functional analysis of CedPV F and G and rCedPV tropism will allow us to further investigate the nature of the cellular receptors that are facilitating rCedPV entry into HeLa-USU cells.
In addition to the observed novel receptor tropism, rCedPV can be used as a platform to explore new henipavirus therapeutics. The experimental vaccines and therapeutics currently available for HeV and NiV infection are based on the virus envelope glycoproteins; however, these therapeutics are not presently licensed for human use [80][81][82]. A HeV G glycoprotein subunit vaccine, Equivac® HeV, which breaks the chain of zoonotic transmission from bats to horses to humans, is licensed for horse vaccination in Australia and is also a promising human vaccine antigen candidate that is protective against HeV or NiV challenge in ferret and non-human primate infection models [83][84][85][86]. A human monoclonal antibody (hmAb), m102.4, is cross-protective against HeV and NiV challenge in the ferret and non-human primate models [87][88][89][90] and has been administered by compassionate use to people after natural or lab-based exposure to HeV and lab-based exposure to NiV [81].
Structural studies of the HeV and NiV G glycoprotein in complex with m102.4 revealed that the hmAb recognizes the receptor-binding domain in the G glycoprotein blocking its interaction with EFN receptors [91].
Serological evidence of henipavirus infections has been detected in both bats and humans in Cameroon, suggesting zoonotic transmission of African henipaviruses [45]. The population of Eidolon helvum, a bat natural host of African henipaviruses [43][44][45]92] ranges widely throughout Sub-Saharan Africa, which raises concerns about the potential of zoonotic transmission [92]. Whether therapeutics that are cross-protective against HeV and NiV will be efficacious against African henipaviruses is unknown. Polyclonal serum raised against NiV was demonstrated to less potently cross-neutralize African henipavirus, KumPV [45]. This finding highlighted the possibility that therapeutics based on the G glycoprotein of HeV and NiV might not be effective against less antigenically-related emerging African or novel Asiatic henipaviruses such as KumPV or MojPV. We seek to use rCedPV as a vaccine platform to explore novel pan-henipavirus therapeutics that target either viral proteins more conserved than G or henipavirus life-cycle stages. This strategy has the potential to address emerging henipaviruses, a shortcoming of present therapeutic approaches.

Conclusions
CedPV is the third Henipavirus species to have been isolated from Pteropus fruit bats, and is presently the only non-pathogenic Henipavirus species identified and characterized. Here, we describe the successful rescue of two recombinant CedPV variants using a reverse genetics approach. As the wild-type CedPV isolate, our recombinant viruses induce an IFN-β response in infected human cells and as expected, receptor mediated fusion was triggered by EFNB2 but not EFNB3. Our rCedPV platform can be used to study various aspects of henipavirus cell biology and host cell interactions, as well as an authentic henipavirus platform for antiviral drug discovery or vaccine approaches safely under BSL-2 laboratory containment.