RNA Editing as a General Trait of Ebolaviruses

Abstract

RNA editing has been discovered as an essential mechanism for the transcription of the glycoprotein (GP) gene of Ebola virus but not Marburg virus. We developed a rapid transcript quantification assay (RTQA) to analyze RNA transcripts generated through RNA editing and used immunoblotting with a pan-ebolavirus monoclonal antibody to confirm different GP gene–derived products. RTQA successfully quantified GP gene transcripts during infection with representative members of 5 ebolavirus species. Immunoblotting verified expression of the soluble GP and the transmembrane GP. Our results defined RNA editing as a general trait of ebolaviruses. The degree of editing, however, varies among ebolaviruses with Reston virus showing the lowest and Bundibugyo virus the highest degree of editing.

RNA editing is an important mechanism to generate protein diversification in different organisms and, thus, plays a critical role in evolution [1, 2]. For example, the human cellular editing enzyme adenosine deaminase acting on RNA (ADAR) plays a key role in antiviral defense and proteome diversity [3]. ADAR-driven RNA editing results in site-specific base substitution and has been described for hepatitis delta virus [4, 5]. In contrast, RNA editing driven by viral RNA-dependent RNA polymerase (RdRP) results in nontemplate nucleotide insertions and has been described for the phosphoprotein (P) gene of paramyxoviruses and the glycoprotein (GP) gene of ebolaviruses [6–10].

In the case of Ebola virus (EBOV), species Zaire ebolavirus, the unedited GP gene transcript encodes a nonstructural soluble glycoprotein (sGP) [8, 9, 11, 12]. It has been postulated that sGP may act as an important target for adaptive immune responses, exhibit anti-inflammatory activities, and serve as a virulence factor [13, 14]. Most recently, it was shown that sGP activates the MAP kinase signaling pathway and, thus, may contribute to EBOV pathogenicity [15]. The structural transmembrane glycoprotein (GP_1,2), which mediates EBOV entry and is the major target for neutralizing antibodies, is expressed by the +1 open reading frame (ORF) of the GP gene accessed through RdRP-driven RNA editing by the predominant insertion of a single nontemplate adenosine residue [8, 9]. The +2 ORF accessed through RdRP-driven RNA editing by insertion of mainly 2 or the deletion of 1 adenosine residue expresses a recently identified nonstructural small soluble glycoprotein (ssGP) of unknown function [7].

RdRP-driven RNA editing has not been experimentally evaluated on the transcript level for other species of the Ebolavirus genus such as Bombali ebolavirus, Bundibugyo ebolavirus, Reston ebolavirus, Sudan ebolavirus, and Taï Forest ebolavirus [12, 16]. Sequence comparison supports the existence of RNA editing with Lloviu virus (species Lloviu cuevavirus, genus Cuevavirus), but experimental data do not exist [17]. In contrast, members of the genus Marburgvirus express GP_1,2 as the primary and sole product of the GP gene with no evidence for RNA editing [18]. Based on sequence data, there is no evidence for RNA editing in the more recently classified genera Dianlovirus, Striavirus, or Thamnovirus [16].

To investigate if RNA editing is an inherent feature of the replication cycle of ebolaviruses, we determined GP gene–specific transcripts and expression products during infection with representative strains of 5 ebolavirus species [16]. For transcript identification, we established a rapid transcript quantification assay (RTQA). Protein detection was analyzed by immunoblotting using a pan-ebolavirus monoclonal antibody. Both methods verified that RdRP-driven RNA editing is a trait of all ebolaviruses.

MATERIALS AND METHODS

Ethics Statements

Whole venous blood was collected from healthy donors in accordance with a protocol approved by the Institutional Review Board for Human Subjects, National Institute of Allergy and Infectious Diseases, National Institutes of Health. The blood and liver samples from an EBOV-infected nonhuman primate (NHP) [19] were derived from a previous animal study protocol, which was approved by the Rocky Mountain Laboratories (RML) Institutional Animal Care and Use Committee and performed according to the guidelines of the American Association of Laboratory Animal Care.

Biosafety Statement

All infectious work using ebolaviruses was approved by the Institutional Biosafety Committee (IBC) of the RML to be conducted under biosafety level 4 (BSL-4) conditions and was performed in the maximum containment laboratories at the Integrated Research Facility of RML (Hamilton, Montana). IBC-approved standard operating procedures were applied for sample inactivation and removal [20, 21].

Cells

Vero E6 cells (African green monkey kidney cells) were maintained in Dulbecco’s modified Eagle medium (Gibco) supplemented with 10% fetal calf serum, 1 mM L-glutamine, 50 U/mL penicillin, and 50 μg/mL streptomycin (all from Gibco). Primary human macrophages were differentiated from monocytes isolated from human peripheral blood mononuclear cells derived from 3 healthy donors as previously described [22]. In brief, a conical tube (50 mL) filled with 15 mL Ficoll-Paque plus (GE Healthcare) was slowly overlaid with 30 mL of ethylenediaminetetraacetic acid blood and cells were separated by centrifugation at 250g. Aspirated cells were washed with phosphate-buffered saline (PBS) twice (total 50 mL, spun at 300g for 10 minutes), diluted in cold (4°C) MACS buffer (Miltenyi Biotec) to 1 × 10⁷cells per 80 μL, and incubated with anti-CD14 magnetic beads (Miltenyi Biotec) (one-fifth of the cell volume) for 15 minutes at 4°C. Following incubation, cells were washed with MACS buffer and applied to a MACS LS column (Miltenyi Biotec). The column was washed 3 times with MACS buffer before elution of the cells according to the manufacturer's instructions. Subsequently, cells were spun at 300g for 10 minutes and resuspended in warm (37°C) Roswell Park Memorial Institute (RPMI) medium (Sigma) without human serum. Cells (3 × 10⁵/well) were seeded in a 24-well plate for 15–30 minutes at 37°C; thereafter, medium was replaced with RPMI containing 10% heat-inactivated AB human serum (Invitrogen). Human primary monocytes were maintained in RPMI with 10% heat-inactivated AB human serum, 1% L-glutamine, 1% pen/strep, and 1% nonessential amino acid (Gibco) under 5% carbon dioxide in a humidified incubator at 37°C and were differentiated into macrophages over several days.

Viruses

The following ebolaviruses were used in this study: Zaire ebolavirus (EBOV, strain Mayinga derived from the Democratic Republic of the Congo in 1976); Sudan ebolavirus (Sudan virus [SUDV], strain Boniface derived from Sudan in 1976); Reston ebolavirus (Reston virus [RESTV], strain Pennsylvania derived from the United States in 1989); Taï Forest ebolavirus (Taï Forest virus [TAFV], strain Côte d’Ivoire derived from Côte d’Ivoire in 1994); and Bundibugyo ebolavirus (Bundibugyo virus [BDBV], strain Bundibugyo derived from Uganda in 2009).

Multiple Sequence Alignment

Clunstal Omega web server (https://www.ebi.ac.uk/Tools/msa/clustalo/) was used to align all ebolavirus GP gene sequences. GenBank accession numbers were as follows: EBOV (AF086833); SUDV (AY729654); RESTV (AF522874); TAFV (FJ217162); BDBV (FJ217161); and Bombali virus (BOMV), Bombali ebolavirus (MF319185).

Ebolavirus Infection

Vero E6 cells (3 × 10⁵ cells) and human monocyte-derived macrophages (3 × 10⁵ cells) were infected with the different ebolaviruses using a multiplicity of infection (MOI) of 0.1. Infected cells and tissue culture supernatants were harvested for RNA extraction and protein analysis at 24, 48, 72, 96, and 144 hours postinfection.

RNA Isolation

Total RNA was extracted from infected cells (3 independent infections) using the RNeasy Kit (Qiagen) according to the protocol provided by the manufacturer. RNA extracted from whole blood and liver of 2 cynomolgus macaques infected with EBOV, strain Kikwit, collected at day 7 postinfection was provided from a previous study [19].

Rapid Transcript Quantification Assay

First-strand DNA synthesis was primed with either an oligo-dT primer to target viral transcript RNA (messenger RNA [mRNA]) or a GP gene–specific primer (5′-AGAGTAGGGGTCGTCAGGTCC) to target viral genomic RNA using SuperScript III reverse transcriptase (Invitrogen) according to the manufacturer's instruction. The complementary DNA (cDNA)–RNA hybrid product was purified using the polymerase chain reaction (PCR) purification kit from Qiagen. DNA amplification for fragment analysis was performed using iProof High Fidelity DNA Polymerase (Bio-Rad) in 20 μL reactions consisting of 1 μL template (purified cDNA-RNA hybrid product), 4 μL 5 × HF reaction buffer, 0.4 μL 10 mM dNTPs, 0.5 μL 6 carboxyfluorescein (FAM)–labeled forward primer (20 μM), 0.5 μL reverse primer (20 μM), 13.4 μL water, and 0.2 μL HF iProof DNA polymerase (2 U/μL). Primers were designed to target GP gene–specific targets covering the editing site region (45 nt upstream and 58 nt downstream). The 6FAM-labeled forward primers and the reverse primers for each ebolavirus are presented in Supplementary Table 1. Thermal cycling conditions were set for an initial denaturation at 98°C for 30 seconds, followed by 30 cycles of 10 seconds of denaturation at 98°C, 15 seconds of annealing at 55°C, and 15 seconds of extension at 72°C, followed by a final extension at 72°C for 7 minutes. All PCR reactions were performed in triplicates unless otherwise stated. The 6FAM-labeled products (110 bp or 111 bp fragments) were diluted 1:400 with water and 1 μL was added to a mix of 10.5 μL HiDi Formamide (Applied Biosystems) and 0.5 uL GeneScan LIZ-120 Size Standard (Applied Biosystems), heat denatured for 3 minutes at 95°C, snap-cooled on ice, and separated with the 3730xL Genetic Analyzer (Applied Biosystems). Fragment analysis was completed using GeneMapper version 4.0 (Applied Biosystems) and peak designations were inferred according to the bin set, which was established using 2 PCR fragments covering the GP gene editing site derived either from a plasmid encoding the sGP ORF (designated 7A-only; 110 nt) or from a plasmid encoding the GP ORF (designated 8A-only; 111 nt). These 2 plasmid-derived PCR fragments were analyzed by fragment analysis using various combinations to cover a broad range of transcript ratios for quantification and establish a standard curve based on 7A and 8A fragment peak heights. Subsequently, the relationship of 7A to 8A genotypes for each virus sample was calculated using the curve fit equation (sixth order polynomial) from the standard curve generated by 7A-only and 8A-only transcripts.

Protein Detection

Tissue culture supernatants from ebolavirus-infected Vero E6 cells were harvested at 24, 96, and 144 hours postinfection and clarified from cell debris by centrifugation at 1000g for 15 minutes. Infected cells were harvested in parallel at the same time points and resuspended in 1 × sodium dodecyl sulfate (SDS)–loading buffer. If needed, protein samples were digested overnight with PNGaseF (NEB) according to the manufacturer's instructions. Protein samples were subjected to 10% SDS polyacrylamide gel electrophoresis (PAGE) and subsequently blotted onto nylon membranes using a semi-dry blotting technique. The membranes were treated overnight with 5% skim milk in PBS/Tween20 to reduce background staining. Following 3 washes with PBS/Tween20, the membranes were incubated for 60 minutes at room temperature with the primary monoclonal antibody MAb 42/3.7 (1:10 000 dilution). This antibody cross-reacts with sGP and GP₁, the large cleavage fragment of GP_1,2, of several ebolavirus species [23, 24]. After 3 washes with PBS/Tween20, membranes were incubated for 60 minutes at room temperature with goat antimouse immunoglobulin G (H + L) (1:10 000 dilution; KPL) and washed 3 times with PBS/Tween20 followed by a single PBS wash. Protein visualization was achieved by using the ECL Plus Western blotting detection kit (GE Healthcare).

Statistical Analysis

All quantification was done in triplicates and data were presented as mean ± standard deviation. No significant analysis was done between samples.

RESULTS

All Ebolaviruses Possess a GP Gene Editing Site

A comparison of the GP gene sequences (nucleotide positions 997–1042, according to EBOV AF086833) of a representative isolate of 6 ebolavirus species (EBOV, SUDV, RESTV, TAFV, BDBV, and BOMV) revealed a complete conservation of the editing site (7 uridine residues) (Figure 1), indicating that RNA editing may occur with all known ebolaviruses. Another highly conserved region was detected directly upstream of the editing site, whereas the sequences directly downstream of the editing site were less conserved (Figure 1), indicating that the upstream sequences may be functionally involved in RNA editing as has been hypothesized earlier for EBOV [10] and also described for paramyxovirus RNA editing [6].

Figure 1.

Glycoprotein (GP) gene editing sites of different ebolaviruses. Clustal Omega web server was used to generate a multiple alignment of the GP gene editing site sequences for representative members of the different Ebolavirus species (sequences shown in viral genomic RNA sense). The editing site (box) is identical for all ebolaviruses analyzed here. The asterisks identify conserved nucleotides among all viruses that are mainly found upstream of the editing site. For viruses and GenBank accession numbers, please see the Materials and Methods. Abbreviations: BDBV, Bundibugyo virus; BOMV, Bombali virus; EBOV, Ebola virus; GP, glycoprotein; RESTV, Reston virus; SUDV, Sudan virus; TAFV, Taï Forest virus.

Development of a Rapid Transcript Quantification Assay

To quantify RNA transcripts differing by a single or a few nucleotide(s), a PCR-based RTQA was established (Figure 2). This assay does not depend on restriction enzyme digestion and ligation such as used for amplified fragment length polymorphism, the most commonly used technology to determine single-nucleotide polymorphisms (SNPs) [25, 26]. To validate this rapid assay, 2 DNA templates representing unedited and edited transcripts of the EBOV GP gene were generated by amplifying and cloning a region flanking the EBOV GP gene editing site (nucleotide positions 997–1042) (Supplementary Table 1). Plasmid DNA was mixed in defined ratios of 0:100%, 10:90%, 20:80% … 80:20%, 90:10%, 100:0% 7A (unedited) versus 8A (edited) DNA template and subjected to the capillary electrophoresis genetic analyzer. Fluorescence-based quantification of PCR fragments showed the suitability of this method to quantify PCR fragments distinct in length by a single nucleotide (Supplementary Table 2).

Figure 2.

Schematic illustration of rapid transcript quantification assay. Complementary DNA was synthesized from total RNA using an oligo-dT primer followed by primer-specific (6 carboxyfluorescein [★FAM]–labeled forward primer) polymerase chain reaction (PCR) amplification of the editing region. PCR fragments differing by a single nucleotide were differentiated using the capillary electrophoresis–based Genetic Analyzer 3730xl. Internal size standard (GeneScan-120 LIZ) allows detection of PCR fragments. GeneMapper software v4.0 analyzes and quantifies allele-specific signals based on the abundance of FAM-labeled PCR products. Abbreviations: cDNA, complementary DNA; GP_1,2, transmembrane glycoprotein; mRNA, messenger RNA; PCR, polymerase chain reaction; sGP, soluble glycoprotein.

EBOV RNA Editing In Vitro and In Vivo

Using the established and validated method, the GP gene transcripts were quantified from both in vitro and in vivo EBOV infections (Figure 3A). In EBOV-infected Vero E6 cells, 85% of the GP gene transcripts were unedited coding for sGP similar to what has been published previously [7, 8]. This method also quantified GP gene transcripts in 2 main target organs of an EBOV-infected NHP with slightly lower and higher percentages of sGP transcripts for liver (75%) and blood (90%), respectively (Figure 3A). When RTQA quantification was applied to genomic RNA derived from EBOV-infected Vero E6 cells and blood from the EBOV-infected NHP, mainly unedited sequences were found (>97% 7U), indicating that transcript polymorphism is principally a result of RdRP-driven transcription (Figure 3A).

Figure 3.

Quantification of ebolavirus RNA editing. Transcripts (mRNA) and genomic RNA (vRNA) polymorphism of the ebolavirus glycoprotein (GP) gene were quantified using the rapid transcript quantification assay (RTQA). A, Ebola virus (EBOV) editing in tissue culture and animals. RNA was derived from liver and blood of 2 EBOV-infected nonhuman primates at end-stage disease and Vero E6 cells infected with a multiplicity of infection (MOI) of 0.1 of EBOV (96 h postinfection). mRNA and vRNA GP gene–specific transcripts were quantified by RTQA. B and C, Ebolavirus RNA editing in Vero cells (B) and primary human monocytes/macrophages (C). Cells were infected with representative members of 5 ebolavirus species using an MOI of 0.1. After 96 h, total cellular RNA was isolated and used for ebolavirus mRNA transcript quantification by RTQA. Monocytes/macrophages cultures were used from 3 independent donors, but only 1 representative experiment is shown. Each quantification was done in triplicate and data are shown as mean ± standard deviation. Abbreviations: BDBV, Bundibugyo virus; EBOV, Ebola virus; GP, transmembrane glycoprotein; mRNA, messenger RNA; NHP, nonhuman primate; RESTV, Reston virus; sGP, soluble glycoprotein; SUDV, Sudan virus; TAFV, Taï Forest virus.

RNA Editing Is an Inherent Feature of Ebolaviruses

Vero E6 cells (commonly used for ebolavirus propagation) and primary human macrophages (primary target cell of ebolaviruses) were infected with representative members of 5 of the 6 ebolavirus species (MOI of 0.1); a BOMV isolate was not available at the time of the study. GP gene transcripts were quantified using RTQA at different time points postinfection. For all ebolaviruses, RNA editing did not change significantly over the time period analyzed (24, 96, and 144 hours postinfection); therefore, data only for the 96-hour timepoint are shown here (Figure 3B and 3C) and data on the other timepoints are presented as Supplementary Material (Supplementary Figures 1 and 2). In EBOV-infected Vero E6 cells, unedited transcript (sGP transcripts) accounted for approximately 78% of all EBOV GP gene transcripts (Figure 3B), which is comparable to the previous experiment (Figure 3A) and results previously published (Figure 3A) [7, 8]. Slightly higher amounts of unedited transcript were found for SUDV and RESTV with approximately 87% and approximately 92% of all GP gene transcripts, respectively (Figure 3B). Slightly lower amounts of unedited transcripts were found for TAFV and BDBV with approximately 72% and approximately 74% of all GP gene transcripts, respectively (Figure 3B). In EBOV-infected human macrophages, we detected approximately 85% unedited GP gene transcripts (Figure 3C). Higher amounts of unedited GP transcripts were found for RESTV with approximately 95%, whereas transcripts for SUDV and TAFV were lower with approximately 75% and for BDBV with only approximately 57% of all viral transcripts (Figure 3C). Overall, BDBV and, to a lesser degree, TAFV showed lower amounts of unedited transcripts and thus higher editing frequency compared to EBOV in both cell types (Figure 3B and 3C). Interestingly, the presumably human apathogenic RESTV showed considerably higher amounts of unedited GP gene transcripts in both cell types and thus a considerably lower editing frequency (<10%) compared to the other 4 human pathogenic ebolaviruses studied here (Figure 3B and 3C).

Ebolaviruses Express Soluble and Transmembrane Glycoproteins

To confirm editing on protein level, we analyzed expression of distinct GP gene products in infected Vero E6 cells by SDS-PAGE analysis followed by immunoblotting using the pan-ebolavirus monoclonal antibody MAb 42/3.7 [23]. For infection with the different ebolaviruses representing 5 species in the genus Ebolavirus, secreted sGP (∼50 kDa) and the larger furin cleavage fragment GP₁ of the transmembrane GP_1,2 on virus particles could be demonstrated in the supernatants and cells at 96 hours postinfection (Figure 4A and 4B). The reduced amount of TAFV-secreted sGP likely reflects a lower cross-reactivity of the antibody used. The reduced detection of SUDV and TAFV GP₁ fragments may represent reduced particle formation (reduced virus growth) and in part lower cross-reactivity of the antibody. Unfortunately, MAb 42/3.7 displays variable degrees of cross-reactivity and affinity with the different ebolavirus GPs, making quantification of protein expression impractical. To further characterize sGP, we performed deglycosylation with PNGaseF, which removes almost all N-linked oligosaccharides from glycoproteins (Figure 4C). As previously demonstrated for EBOV [7, 27], sGPs of the 4 other ebolaviruses were sensitive to PNGaseF treatment, demonstrating N-glycosylation as a general hallmark of the secreted ebolavirus sGP (Figure 4C).

Figure 4.

Ebolaviruses express a soluble glycoprotein (sGP) and transmembrane glycoprotein. Vero E6 cells were infected with a multiplicity of infection of 0.1 of different ebolaviruses. Supernatants (A) and cells (B) were collected 96 h postinfection and analyzed by Western blot analysis. C, A portion of the supernatant samples was treated overnight with PNGaseF and analyzed by Western blot analysis. The pan-ebolavirus monoclonal antibody MAb 42/3.7 cross-reacts with the larger furin cleavage fragment of the glycoproteins (GP₁) and the sGP. Abbreviations: BDBV, Bundibugyo virus; EBOV, Ebola virus; GP, transmembrane glycoprotein; RESTV, Reston virus; sGP, soluble glycoprotein; SUDV, Sudan virus; TAFV, Taï Forest virus.

DISCUSSION

RNA editing was discovered in the kinetoplastid protozoa where insertion and deletion of uridylates occur in the mitochondrial pre-mRNA by a posttranscriptional process [28, 29]. RNA editing mechanisms are diverse ranging from nucleoside modification (such as C to U and A to I deamination) to insertion of nontemplate nucleosides. RNA editing has been described in viruses, prokaryotes, and eukaryotes [6, 7, 29, 30]. Although RNA editing may occur posttranscriptionally in prokaryotes [29], in viruses it seems to occur co-transcriptionally as exemplified by paramyxovirus P and EBOV GP gene editing [6–9].

ADAR is the most commonly known and well-studied cellular enzyme-driven RNA editing mechanism, which involves site-specific single-nucleoside modification (base deamination) [2], which has been previously quantified by primer extension [31], hybridization techniques [31–33], or restriction enzyme digest [34]. In addition, several advanced high-throughput methods became available for the detection of edited RNA such as denaturing high-performance liquid chromatography, allele-specific real-time PCR with TaqMan probes, and PCR with allele-specific primers [35]. More recently, a pyrosequencing approach [36] and TaqMan-based reverse-transcription PCR quantitative assays have been reported [37]. However, viral polymerase-driven co-transcriptional RNA editing has so far largely been quantified using cloning and sequencing approaches [7, 8, 38, 39], which are time-consuming and labor-intensive if sample numbers are large. The RTQA established herein provides an alternative and efficient method to existing technologies [40, 41].

For genome mapping and variation analysis, particularly detecting SNPs, the restriction fragment length polymorphism (RFLP) PCR assay has been successfully utilized [42]. While our RTQA assay is similar in principle to RFLP PCR, it is unique in quantifying 2 different transcripts simultaneously. Thus, RTQA was evaluated for its suitability, reproducibility, sensitivity, and specificity and chosen for work presented here. As described in this report, PCR fragments differing by a single nucleotide were mixed in known ratios as a reference standard for accurate quantification (Figure 2). EBOV produces at least 3 transcripts with different ratios expressing 3 distinct proteins (2 nonstructural proteins [sGP and ssGP] and 1 structural protein [GP_1,2]) [7] due to RNA editing [10]. Here, we compared the ratio between ebolavirus sGP (nonedited template) and GP_1,2 (single-nucleoside insertion) transcripts. Once established, GP gene transcripts from both in vitro and in vivo EBOV infections were quantified by RTQA resulting in ratios that were similar to those previously published [7, 8] (Figure 3). Our data demonstrate the value of this new and efficient quantification method. While the RTQA assay successfully quantified 2 major transcripts (sGP and GP_1,2), the quantification of >2 transcripts, in particular those with more than a single-nucleoside insertion or deletion (eg, ssGP), is difficult and likely requires additional extensive optimization and thus was outside the scope of this study.

Subsequently, RTQA was utilized to analyze RNA editing with other ebolaviruses in Vero E6 cells, a cell line commonly used for ebolavirus propagation, and human macrophages, primary target cells of ebolavirus infection [43–45] (Figure 3). Approximately 15%–40% of transcripts were edited during infection with human pathogenic ebolaviruses (BDBV, EBOV, SUDV, and TAFV), whereas RESTV infection led only to 10% of transcripts edited. Whether this lower level of editing may impact pathogenicity with the potentially apathogenic RESTV remains to be determined. The highest level of editing was associated with BDBV, a human pathogenic ebolavirus with a lower case fatality rate compared to EBOV; this observation argues against a correlation between RNA editing and pathogenicity. Overall, RNA editing seems to occur with all ebolaviruses, which clearly distinguishes them from marburgviruses and perhaps other filoviruses. Higher RNA editing frequencies were observed among the henipaviruses, displaying a range of 66%–71% [39]. In contrast, editing frequencies varied more among viruses in the family Paramyxoviridae with a wide range of 31%–82% despite a similar editing site in the P gene [6, 38, 46–48]. Importantly, RNA editing plays an important role in paramyxoviruses such as for Sendai virus pathogenesis [49, 50].

For all ebolavirus species, RNA editing was stable over the course of infection (up to 144 hours postinfection). This result is noticeably different from editing observed during a Nipah virus infection where the editing frequency was low (about 10%) at very early time points, high (up to 90%) around 24 hours postinfection, and decreased thereafter [38]. It should be noted that editing with paramyxoviruses regulates the expression of interferon antagonists [51], whereas ebolavirus editing affects the expression of the glycoproteins, which are not known to antagonize interferon. While ebolavirus RNA editing seems stable during infection, its potential contribution to virus replication or host immune evasion has yet to be determined.

Earlier work using the EBOV infectious clone system has suggested that RNA editing might be needed to regulate GP_1,2 expression and control viral cytotoxicity [52]. Thus, editing could also be associated with pathogenicity of ebolaviruses. This would be supported by the lowest level of editing detected for RESTV, which is only known to cause asymptomatic infections in humans. In contrast, a recombinant EBOV with abolished editing and expression of high levels of GP_1,2 did not show an altered pathogenicity phenotype in guinea pigs [53].

Proteins encoded by the GP transcripts during infection were detected in the supernatants of cells infected with all ebolaviruses studied here. sGP and GP₁, the larger fuirn cleavage fragment of GP_1,2, appeared similar in size for all ebolavirus species even though they were not identical (Figure 4). The third product of the GP gene, ssGP, was not considered in this study due to its low abundance during infection as shown for EBOV even though it likely is hidden under the broad sGP band as the sizes of mature sGP and ssGP are very similar [7]. sGPs from the different ebolaviruses were all found to be N-glycosylated with highly conserved N-glycosylation motifs. The importance of N-glycosylation for sGP is unknown but the same sites also appear in GP_1,2, for which N-glycosylation is an important posttranslational modification. We have previously shown that EBOV sGP is not O-glycosylated in contrast to GP_1,2 [7, 27]. Thus, O-glycosylation was not assessed here.

Whether ebolaviruses are the only filoviruses utilizing RNA editing of their GP gene seems unlikely. Marburgviruses have been shown to not use RNA editing and lack an editing site in their glycoprotein gene [54]. Sequence analysis of cuevaviruses defined the same editing site as found with ebolaviruses [55]. Unfortunately, at the time of the study a cuevavirus isolate did not exist, but this has changed with the recently reported isolation of Lloviu virus [56]. Furthermore, we did not include BOMV, now representing a sixth ebolavirus species, in our experimental study due to lack of a virus isolate; however, a recombinant BOMV has become recently available [57].

In conclusion, we have provided experimental evidence here that RNA editing seems to be a common feature of ebolaviruses. Varying degrees of RNA editing among ebolaviruses may indicate its importance for the virus life cycle and possibly pathogenicity.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments . The authors thank Friederike Feldmann (Rocky Mountain Veterinary Branch, Division of Intramural Research, National Institute of Allergy and Infectious Diseases [NIAID], National Institutes of Health [NIH]), for assistance with high containment work and Anita Mora (Visual Arts, Division of Intramural Research, NIAID, NIH) for help with graphical work. This work was part of the PhD thesis of M. M. performed in the Department of Medical Microbiology, University of Manitoba, Winnipeg, Canada. This work was funded by the Division of Intramural Research, NIAID, NIH.

Financial support . This work was funded by the Division of Intramural Research, NIAID, NIH.

Supplement sponsorship. This article appears as part of the supplement “10^th International Symposium on Filoviruses.”