Development of a New Reverse Genetics System for Ebola Virus

Ebola virus is among the most dangerous viral pathogens, with a case fatality rate of up to 90%. Since 2013, the two largest and most complex Ebola outbreaks in West Africa have revealed the lack of investigation on this notorious virus.

. Here, we first sought to determine whether Huh7-4P cells support the replication of a Yambuku-Mayinga isolate-based tetracistronic subgenome (SG) consisting of viral gene VP40, VP24, unedited GP, and GLuc-Hygro (Fig. 1A). However, the percentage of VP40-positive cells and EBOV-SG RNA levels (intracellular and extracellular) decreased continuously after transfection of in vitro-transcribed EBOV-SG RNA into the Huh7-4P cells (Fig. S1 in the supplemental material), suggesting that the Huh7-4P cell does not efficiently support the SG replication.
Given that the Huh7-4P cell line is a mixed cell population possessing different stoichiometry of the 4 viral proteins in each cell, which has been shown to be critical for Protocol to establish Huh7-4P-High cells that are more permissive for SG replication and trVLP propagation. (C) Huh7-4P and Huh7-4P-High cells that had been transfected with EBOV EGFP-MG RNA for 48 h were analyzed for the GFP expression by fluorescence microscopy. Scale bar, 100 mm. (D and E) Huh7-4P and Huh7-4P-High cells were transfected with Mayinga-SG RNA and cultured for 27 days. Intracellular SG RNA levels (D) and extracellular trVLP titers (E) were measured at the indicated time points by reverse transcriptase quantitative PCR (RT-qPCR) and titration assay, respectively. The data shown in panels D and E are results from 1 of 2 independent experiments and are presented as the mean 6 SD of n = 2 technical replicates. The data of another experiment are shown in Fig. S2A and B.
EBOV replication (14,17,18), we designed an approach to select a cell from the Huh7-4P pool that efficiently supports EBOV-SG replication. Huh7-4P cells were transfected with an EBOV-MG consisting of an enhanced green fluorescent protein (EGFP-MG) (Fig. 1A). The cells with strong GFP signals were sorted out by flow cytometry (Fig. 1B) Data shown in panels A to C are the results from 2 independent experiments and are presented as the mean 6 SD of n = 2 biological replicates. and then cultured continuously for 1 month, leading to complete loss of the GFP signals and EBOV-MG RNA in the cells (denoted Huh7-4P-High) (data not shown).
To assess the ability of Huh7-4P-High cells to support EBOV replication, EGFP-MG RNA was transfected into Huh7-4P-High or Huh7-4P cells. As shown in Fig. 1C, Huh7-4P-High cells indeed displayed an improved ability to support the EGFP-MG replication. Next, we assessed the ability of Huh7-4P-High cells to support EBOV-SG replication. Transfection of EBOV-SG RNA into Huh7-4P-High cells led to a gradual increase of intracellular SG RNA level ( Fig. 1D and Fig. S2A), extracellular trVLP titers ( Fig. 1E and Fig. S2B), and percentage of VP40-positive cells (Fig. S2C), indicating efficient EBOV-SG replication as well as trVLP spread. Consistently, trVLP infection in Huh7-4P-High cells was more efficient than that in Huh7-4P cells (Fig. S2D).
To further improve the permissiveness for EBOV-trVLP propagation, we subcloned Huh-4P-High cells (Fig. S3A). Seven cell clones were generated, among which clone 1 (denoted Huh7-4PX) showed a strong permissiveness for EBOV-trVLP propagation and EBOV-SG replication ( Fig. S3A and S3B) and thus was selected for the following studies. Of note, we confirmed that no residual EGFP-MG RNA previously introduced to enrich permissive cells was detected in Huh7-4P-high cells and in each of these subclones (Fig. S3C). Next, we assessed the ability of Huh7-4PX cells to support trVLP propagation. Naive Huh7, Huh7-4P, Huh7-4P-High, and Huh7-4PX cells were inoculated with trVLP. As shown in Fig. 2, intracellular SG RNA level ( Fig. 2A), extracellular trVLP titers (Fig. 2B), GLuc activity (Fig. 2C), and VP40 protein expression level ( Fig. 2D and E) increased most rapidly and reached the highest level on day 11 postinfection in Huh7-4PX cells, demonstrating that Huh7-4PX cells indeed possess a superior ability to support EBOV replication and propagation.
Characterization of Huh7-4PX cell-derived trVLP. To characterize the features of EBOV-trVLP propagated in Huh7-4PX cells, we first analyzed trVLP morphologically. Huh7-4PX cells were infected with EBOV-trVLP at a multiplicity of infection (MOI) of 0.1 for 12 days. trVLP from 120 ml of culture supernatants was recovered by ultracentrifugation and then analyzed by transmission electron microscopy (TEM). As shown in Fig. 3A, trVLP displayed filamentous morphology as previously reported for infectious EBOV (19).
Next, we determined whether the produced EBOV-trVLP in Huh7-4PX cells can be inhibited by two previously reported countermeasures. Neutralizing antibody c2G4, a component of ZMapp, efficiently inhibited trVLP infection at 50% inhibitory concentration (IC 50 ) of 0.05 mg/ml (Fig. 3B), which was comparable to its inhibition on infectious virus (20). Remdesivir, a nucleotide analog, also displayed a similar inhibitory effect on trVLP compared to that on EBOV (21) (Fig. 3C). We then evaluated the anti-EBOV effect of type I, II, and III interferons (IFNs), whose effect against EBOV remain controversial from previous studies (22)(23)(24)(25). As shown in Fig. 3D to F, only IFN-g slightly inhibited trVLP infection, whereas neither IFN-a nor IFN-l3 had any effect.
To probe into why the Huh7-4PX cell can more efficiently support EBOV-SG replication than its parental cells, we measured the level of NP, VP35, and VP30 proteins and L mRNA (due to lack of anti-L antibody) in Huh7-4P, Huh7-4P-High, and Huh7-4PX cells. To our surprise, the expression levels of these 4 viral proteins were lowest in Huh7-4PX cells compared to parental Huh7-4P and Huh7-4P-High cells ( Fig. 4A and B). Of note, the VP35 level in Huh7-4PX cells was particularly low, with only 1% of that in the parental Huh7-4P cells. Next, we reconstituted expression of NP, VP35, and VP30 by lentiviral transduction or L by plasmid transfection in Huh7-4PX cells, respectively. As shown in Fig. 4C to G, the elevated expression of NP, VP30, or L in Huh7-4PX cells had either minor or no inhibitory effect on EBOV-trVLP infection, while the increased VP35 expression displayed the most dramatic inhibitory effect. These results suggested that 4 virial proteins, particularly VP35, had been selected to appropriate levels in Huh7-4PX cells, leading to its exceptional permissiveness for EBOV-SG replication and trVLP propagation.
Rescue of infectious virions in Huh7-4PX cells using the full-length EBOV genome. We then proceeded to test whether Huh7-4PX could be employed to rescue the fully infectious virions of EBOV. The in vitro-transcribed full-length genome (vRNA) of EBOV Yambuku-Mayinga isolate (Mayinga-FL-vRNA) was transfected into Huh7-4PX cells. The number of VP40-positive cells increased rapidly after transfection, and nearly all cells became VP40 positive on day 9 posttransfection when massive cytopathic effect (CPE) was observed (Fig. 5A). Consistently, intracellular ( Fig. 5B and Fig. S4A) and extracellular ( Fig. 5C and Fig. S4B) viral RNA levels, extracellular viral titers (Fig. 5D), and intracellular VP40 protein level (Fig. 5E) all increased after transfection. TEM of naive Huh7 cells infected with the rescued virus showed the classical filamentous shape of the budding virus and the cross-section of aligned viral nucleocapsid in the cytosol (Fig. 5F). Moreover, Mayinga-EBOV rescued from the RNA-transfection experiment propagated efficiently in naive Huh7 and Vero cells ( Fig. S5A and B).
It has been reported that anti-genomic RNA (cRNA) of a negative-sense RNA virus is preferred over vRNA as the starting material for rescuing virus, possibly to avoid the antisense RNA effects between the introduced vRNA and the mRNA encoding viral proteins transcribed from the transfected plasmids (26). Thus, we also tested this idea in our system was incubated with neutralizing antibody c2G4 at the indicated concentration for 1 h at 37°C and used to infect Huh7-4PX cells for 2 days. VP40 immunofluorescence foci number of each group was used to calculate the IC 50 with GraphPad software. (C) trVLP-infected Huh7-4PX cells were treated with remdesivir at the indicated concentration for 3 days. Intracellular SG RNA level of each group was used to calculate the IC 50 with GraphPad software. (D to F) Huh7-4PX cells were pretreated with the indicated doses of IFN-a (D), IFN-g (E), or IFN-l3 (F) for 12 h and then infected with trVLP for 60 h. Intracellular SG RNA level was measured by RT-qPCR. Data in panels B to F are presented as the mean 6 SD of n = 2 biological replicates. *, P , 0.05; ns, not significant. by transfection of full-length cRNA of Yambuku-Mayinga isolate (Mayinga-FL-cRNA) into Huh7-4PX cells. As shown in Fig. S6, infectious virions were also successfully recovered.
Next, we determined whether Huh7-4PX cells, which stably express NP, VP35, VP30, and L proteins of EBOV-Mayinga isolate, could be utilized to rescue a different EBOV isolate. For this purpose, we chose the EBOV-Makona isolate that has caused the devastating 2013 to 2016 West Africa outbreak. Similar to that of the EBOV-Mayinga isolate, transfection of vRNA of the Makona isolate led to robust production of infectious virions, and nearly all cells became VP40 positive on day 7 posttransfection (Fig. 6A). Intracellular ( Fig. 6B and Fig. S4C) and extracellular ( Fig. 6C and Fig. S4D) viral RNA levels, extracellular viral titers (Fig. 6D), and intracellular VP40 protein level (Fig. 6E) all rapidly increased after transfection as expected. The rescued Makona-EBOV propagated FIG 4 VP35 expression level is critical for Huh7-4PX to support EBOV-SG replication. (A) NP, VP35, and VP30 protein levels in Huh7, Huh7-4P, Huh7-4P-High, and Huh7-4PX cells were analyzed by Western blotting. (B) L mRNA levels in Huh7, Huh7-4P, Huh7-4P-High, and Huh7-4PX cells were determined by RT-qPCR. (C to D) Huh7-4PX cells that had been transduced with NP-, VP35-, or VP30encoding lentiviruses were infected with trVLP at an MOI of 0.1 for 72 h. NP, VP35, VP30, and VP40 protein levels (C) and intracellular SG RNA levels (D) were analyzed by Western blotting and RT-qPCR, respectively. (E to G) Huh7-4PX cells that had been transfected with L-encoding pcDNA3.1 plasmid were infected with trVLP at an MOI of 0.1 for 72 h. L mRNA (E), VP40 protein levels (F), and intracellular SG RNA levels (G) were analyzed by Western blotting and RT-qPCR, respectively. Data in panels B, D, E, and G are presented as the mean 6 SD of n = 2 biological replicates. **, P , 0.01; *, P , 0.05; ns, not significant. efficiently in naive Huh7 cells (Fig. S5C). Besides, virus rescue with cRNA of the EBOV-Makona isolate was also successful (Fig. S7). Sequencing analysis of the full-length viral genome showed no mutation in any of the rescued EBOV of Mayinga or Makona from the last day sample of transfection (data not shown), and the 7 uridine-stretch RNA editing site in the GP gene (27) remained unchanged.

DISCUSSION
In this study, we developed a new reverse genetics system for studying EBOV under different biosafety levels. This system has several following advantages. First, compared to other current EBOV reverse genetics systems that are dependent upon cotransfection of multiple (often 5 to 6) plasmids expressing viral genome and viral proteins, our new system takes advantages of an engineered cell line (Huh7-4PX) in which the 4 viral proteins essential for EBOV replication have been already permanently introduced and their expression levels and stoichiometry fine-tuned to render permissiveness for EBOV replication. Although the full-length EBOV rescue can be also achieved with a decent efficiency using the multiplasmid transfection-based approach that is more flexible in fine-tuning the components of RNP, our system only requires transfection of a single viral RNA genome, which makes experimentations more convenient. Second, the Huh7-4PX helper cell line can be used to rescue both fully infectious virions and subviral particles. This system is particularly useful to rescue EBOV-trVLP because these subviral particles, once produced, can propagate continuously in the helper cell line. In contrast, propagation of EBOV-trVLP in the previously reported systems requires the repeated cotransfection of the 4 viral proteins essential for EBOV replication. In fact, propagation of EBOV-trVLP in Huh7-4PX cells resembles natural EBOV infection and expansion, rendering a surrogate model to study the entire EBOV life cycle outside BSL-4 conditions. Of note, propagation of EBOV-trVLP in Huh7-4PX cells is considerably less efficient than that of infectious EBOV in Huh7 cells, suggesting that an artificially fixed stoichiometry of RNP components in Huh7-4PX cells may be still less optimal for virion production than the dynamically changing ratio of each RNP component during the natural virus infection.
Third, the Huh7-4PX helper cell can be used to rescue EBOV of a heterologous isolate. Although this finding has been reported before (28), it still has an important implication in the application of the Huh7-4PX helper cell in rescuing novel EBOV isolates because it alleviates the needs to construct another Huh7-4PX helper cell that is customized for a novel EBOV isolate, which could be a time-consuming process. Interestingly, we found that the Huh7-4PX cell was able to, albeit at a lower efficiency, support the replication and transcription of the minigenome of two Marburg virus (MARV) isolates (Angola-2005 isolate and Musoke isolate) (Fig. S8). This result raises an exciting possibility that the Huh7-4PX cell may have a broader application in developing reverse genetics systems for other filoviruses of different species or even different genera. More future investigations are warranted to determine the potential expanded application of the Huh7-4PX cell.
There are also some limitations to our system. First, virus rescue and trVLP production are restricted to Huh7 cells. It remains a question whether such a helper cell can be developed using other cell lines. Second, the four viral nucleocapsid protein-coding genes have been stably integrated into the Huh7-4PX cells, making it impossible to use this model to study any mutations within these four viral proteins. In addition, because the expression levels and stoichiometry of the four viral RNP proteins are fixed in the cells, it remains to be determined how it would affect the EBOV life cycle, during which their expressions are believed to be dynamic.
We have explored potential mechanisms why the Huh7-4PX cell is competent in supporting EBOV replication. To our surprise, compared to its parental cells, the Huh7-4PX cell has the lowest expression level of NP, VP35, VP30, and L proteins. Among the four viral proteins, VP35 is most critical. Its expression level in the Huh7-4PX cell has been decreased to a greater degree than others, and its overexpression in the Huh7-4PX cell has the greatest inhibitory effect on EBOV replication. It has been reported that excess VP35 inhibits viral replication by affecting the formation of replication complex (18), which may at least partially explain why VP35 expression level is so critical in Huh7-4PX cells. It would be interesting to further understand the mechanism and to develop new antivirals that could mimic the inhibitory function of VP35. It would also be worth investigating whether the VP35-mediated inhibition could be rescued by overexpression of other viral proteins, which may help understand the regulation of the virus replication complex. As an immune antagonist, VP35 blocks virus-mediated induction of interferon signaling response (29), while VP24 inhibits downstream signaling upon the interferon treatment (30). The potential contribution of VP35-antagonizing interferon signaling to the permissiveness of Huh7-4PX cells needs to be further investigated.
The rescue of negative-sense RNA viruses has lagged much behind positive-sense RNA viruses largely due to the prerequisite to deliver viral ribonucleoproteins alongside the viral genome. We adopted the advantageous method of positive-sense RNA virus rescue and simplified EBOV rescue by directly delivering viral genome into cells. Consistent with a previous study (31), we found that both vRNA and cRNA can be used to rescue EBOV, challenging conventional wisdom in the field that cRNA may be preferred over vRNA in rescuing negative-sense RNA viruses (32)(33)(34)(35). In addition, delivery of vRNA, the viral genome, should recapitulate the early events of EBOV life cycle in a more genuine manner. Further investigations are needed to determine which form of RNA is more efficient in rescuing EBOV.

MATERIALS AND METHODS
Cell culture. Human embryonic kidney cells (HEK293T), African green monkey kidney cells (Vero), human hepatoma cells (Huh7) (obtained from National Collection of Authenticated Cell Cultures, Chinese Academy of Sciences), and their derivative cell lines were maintained in Dulbecco's modified Eagle medium (DMEM) (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS), 10 mM HEPES buffer, 100 U/ml penicillin-streptomycin, 10 mM L-glutamine, and 0.1 mM nonessential amino acids (Thermo Fisher Scientific) at 37°C in a moist environment containing 5% CO 2 . To subclone Huh7-4P-High cells, the cells were trypsinized and resuspended in complete culture medium at 1 Â 10 4 cells/ml as the starting concentration. We added 200 ml of the cell suspension to the first well of a 96well plate, and then 2-fold serial dilutions were conducted across the entire plate. The final medium volume of each well was supplemented to 200 ml. After culturing for weeks, monoclonal cells were trypsinized and seeded back into a 96-well plate for proliferation. The viability, growth kinetics, and morphology of Huh7-4P, Huh7-4P-High, and Huh7-4PX cells were like those of Huh7 cells.
Construction of plasmids. The full-length genome of EBOV Yambuku-Mayinga isolate (NC002549) and Makona-C07 isolate (GenBank accession no. KJ660347) were synthesized in several fragments by GenScript. To construct EBOV MG, SG, and full-length genome (FLG) plasmids, PCR was carried out with Q5 High-Fidelity DNA polymerase (New England BioLabs), and products were assembled with the NEBuilder HiFi DNA Assembly cloning kit (New England BioLabs). DNA was transformed into DH5a or TOP10 competent cells (Tiangen). Plasmid DNA was purified with the TIANprep mini plasmid kit (Tiangen) or the NucleoBond Xtra Midi Plus kit (Macherey-Nagel). All the genome sequences were fused between a T7 promoter and a hepatitis D virus ribozyme followed by a T7 terminator as previously reported (36). All constructs were verified by Sanger sequencing.
Viruses. Viruses used in this study were all recombinants produced by reverse genetics. Rescue of viruses was achieved by electroporation of the corresponding in vitro-transcribed full-length viral genome RNA into Huh7-4PX cells. All experiments involving live EBOV were conducted in the BSL-4 facility of Wuhan National Biosafety Laboratory, Chinese Academy of Sciences.
In vitro transcription. In vitro transcription of EBOV MG, SG, and FLG was performed using the MEGAscript T7 transcription kit (Thermo Fisher Scientific) following the manufacturer's protocol. Following 4.5 h incubation at 37°C, the sample was treated with 1 ml of TURBO DNase for 15 min at 37°C, precipitated with 30 ml of lithium chloride, and chilled for at least 1 h at 220°C. RNA was then pelleted at 13,200 Â g for 10 min and washed once with 1 ml of 75% ethanol. After thorough removal of ethanol, RNA was air-dried at room temperature for 2 min and resolved in 100 ml of RNase-free water.
RNA electroporation. Cells were trypsinized, resuspended in complete growth medium, and pelleted at 260 Â g for 5 min. Cells were then washed once with 5 ml of Opti-MEM (Thermo Fisher Scientific) and resuspended in Opti-MEM at a concentration of 1.5 Â 10 7 cells/ml. We mixed 400 ml of the cell suspension thoroughly with 10 mg of in vitro-transcribed RNA in a 0.4-cm electroporation cuvette (Bio-Rad). Electrical pulse was exerted once at 270 V, 950 mF, and 100 X with the Gene Pulser Xcell electroporation system (Bio-Rad). The electroporated cells were immediately transferred into 10 ml of complete growth medium in a 15-ml Falcon tube. The cell suspension was transferred to a 10-cm dish or T75 flask for normal culturing and was usually passaged at a 1:5 ratio every 3 days. All the supernatant, RNA, protein, and cell samples were harvested during each cell passage with consistent amounts.
Western blotting. Cell lysates were prepared by lysis in 1Â SDS loading buffer (62.5 mM Tris-HCl [pH 6.8], 10% [wt/vol] SDS, 50% glycerol, 50 mM dithiothreitol [DTT], 0.05% [wt/vol] bromophenol blue, and 3.33% b-mercaptoethanol) after removal of cell culture medium. The lysates were incubated at 98°C for 10 min, centrifuged, homogenized, and then loaded on 10% PAGE gels. Transferred polyvinylidene difluoride (PVDF) membrane was probed with primary antibodies for 2 h, followed by incubation with the HRP-conjugated secondary antibodies for 1 h. Chemiluminescence was then developed by incubation of the membrane with the Immobilon Western chemiluminescent HRP substrate (Millipore) and scanned with the Amersham Imager 600 (General Electric). Protein quantification was processed with ImageJ software and normalized against internal actin.
Immunofluorescence assay. Cells seeded in 96-well plates were fixed with 4% of paraformaldehyde for at least 30 min and permeabilized and blocked for 1 h by blocking buffer (0.3% Triton X-100, 3% bovine serum albumin [BSA] [wt/vol], and 10% FBS in phosphate-buffered saline [PBS]) before 1 h of incubation with primary antibody prepared in binding buffer (0.3% Triton X-100, 3% BSA in PBS). Cells were then stained for 1 h with secondary antibody prepared in binding buffer before imaging. Images were acquired and processed with an IX73 inverted microscope (Olympus).
Luciferase assay. Gaussia luciferase activity was measured with the BioLux Gaussia luciferase assay kit (New England BioLabs). A 50-ml volume of cleared culture supernatant was added into a 96-well white plate, and 20 ml of GLuc assay solution was prepared and thoroughly mixed with the sample immediately before luminescence detection with a Synergy H1 reader (Biotek). NanoLuc activity was measured with Nano-Glo luciferase assay system (Promega). A 25-ml volume of cleared culture supernatant was added into a 96-well white plate, and 25 ml of NanoLuc assay solution was prepared and thoroughly mixed with the sample before luminescence detection with a Synergy H1 reader.
Overexpression of EBOV proteins in Huh7-4PX cells. Huh7-4PX cells were infected with NP, VP35, or VP30 lentivirus, respectively, for 24 h. Transduced cells were passaged and cultured for another 72 h before infection with EBOV-trVLP. The cell lysates were harvested at 72 h postinfection for measuring protein and RNA levels. For L overexpression, Huh7-4PX cells were transfected with pcDNA3.1 plasmid encoding L protein. Transfected cells were seeded 24 h posttransfection and then infected with EBOV-trVLP. The cell lysates were harvested at 72 h postinfection for measuring protein and RNA levels.
Flow cytometry. Cells were trypsinized, resuspended in complete growth medium, and pelleted at 260 Â g for 5 min. Cells were washed once with PBS and resuspended in sorting buffer (PBS containing 1% newborn bovine serum, 2.5 mM EDTA, and 10 mg/ml gentamicin). Cell sorting was done with MoFlo Astrios (Beckman Coulter). After debris and nonsinglet events exclusion, a 488-nm laser was used to excite GFP signal, and 0.57% of cells with the highest signal were sorted out. After sorting, cells were pelleted at 260 Â g for 5 min and resuspended with complete medium before seeding into a 96-well plate. Cytometry data were analyzed and processed with FlowJo software (FlowJo LLC).
Titration of EBOV-trVLP. Huh7-4PX cells were seeded into 96-well plates at 1 Â 10 4 cells/well. trVLP samples were serially diluted 10-fold in complete medium and used to infect the cells (3 replicates per dilution). Following 2 days of incubation, the cells were fixed and immunostained as described above. The trVLP titer is expressed as focus-forming units per milliliter of supernatant (FFU/ml).
Purification of EBOV-trVLP. Culture medium containing trVLP was centrifuged for 20 min at 3,000 Â g to remove cell debris and then ultracentrifuged at 100,000 Â g for 2 h. The pellet was completely resuspended, homogenized, and transferred for purification by fractionation with iodixanol. The resuspension was centrifuged at 150,000 Â g for 16 h. Fractions were taken out for RNA quantity and infectivity titer measurements, and the best fraction was used for negative staining with 2% uranyl acetate and processed for transmission electron microscopy (TEM) under an FEI Tecnai G2 20 Twin electron microscope (Thermo Fisher Scientific).
Titration of EBOV. Huh7 cells were seeded into 96-well plates at 4 Â 10 4 cells/well. Virus samples were serially diluted 10-fold in 3% FBS medium and used to infect the cells (8 wells per dilution). Following 6 days of incubation, the cells were fixed and immunostained as described above. The viral titer is expressed as 50% tissue culture infectious dose per milliliter of supernatant (TCID 50 /ml).
Reverse transcriptase PCR and sequencing. Supernatant RNA samples were reverse transcribed with the HiScript 1st Strand cDNA synthesis kit (Vazyme) following the manufacturer's protocol. Nested PCR was then performed with PrimeStar GXL DNA polymerase (TaKaRa) to amplify the cDNA in 10 segments, and DNA products were sent for sequencing. Primer sequences are available upon reasonable request.
TEM of infected cells. Ultrathin-section electron microscopy was performed as described previously (6). Briefly, Huh7 cells were cultured in 0.4-mm-pore-size Transwells (Corning), in which 2 Â 10 5 cells were seeded on polycarbonate membrane for each chamber and infected with EBOV at an MOI of 0.1. At 72 h after infection, cells were fixed with 4% of paraformaldehyde for at least 30 min and stored with the fix buffer at 4°C overnight. Cells and polycarbonate membranes were removed from Transwell chambers as a complete monolayer. The cell monolayer was postfixed with 1% osmium tetroxide in 0.1 M PBS for 30 min and dehydrated with a series of ethanol gradients followed by propylene oxide before being embedded in resin. Thin sections were stained with 2% uranyl acetate and Reynolds lead citrate for 15 min and processed for TEM imaging under an FEI Tecnai G2 20 Twin electron microscope.
Statistical analysis. Prism 7 (GraphPad Software) was used for statistical analysis. One-way analysis of variance (ANOVA) (Kruskal-Wallis test) was used to evaluate the efficacy of IFNs on EBOV-trVLP and to compare the effects of excess viral protein on EBOV-trVLP infection. A P value of ,0.05 was considered to indicate a statistically significant difference.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.