Phylogenetic analysis of the promoter element 2 of paramyxo- and filoviruses

ABSTRACT Paramyxo- and filovirus genomes are equipped with bipartite promoters at their 3' ends to initiate RNA synthesis. The two elements, the primary promoter element 1 (PE1) and the secondary promoter element 2 (PE2), are separated by a spacer region that must be precisely a multiple of 6 nucleotides (nts), indicating these viruses adhere to the “rule of six.” However, our knowledge of PE2 has been limited to a narrow spectrum of virus species. In this study, a comparative analysis of 1,647 paramyxoviral genomes from a public database revealed that the paramyxovirus PE2 can be clearly categorized into two distinct subcategories: one marked by C repeats at every six bases (exclusive to the subfamily Orthoparamyxovirinae) and another characterized by CG repeats every 6 nts (observed in the subfamilies Avulavirinae and Rubulavirinae). This unique pattern collectively mirrors the evolutionary lineage of these subfamilies. Furthermore, we showed that PE2 of the Rubulavirinae, with the exception of mumps virus, serves as part of the gene-coding region. This may be due to the fact that the Rubulavirinae are the only paramyxoviruses that cannot propagate without RNA editing. Filoviruses have three to eight consecutive uracil repeats every six bases (UN5) in PE2, which is located in the 3' end region of the genome. We obtained PE2 sequences from 2,195 filoviruses in a public database and analyzed the sequence conservation among virus species. Our results indicate that the continuity of UN5 hexamers is consistently maintained with a high degree of conservation across virus species. IMPORTANCE The genomic intricacies of paramyxo- and filoviruses are highlighted by the bipartite promoters—promoter element 1 (PE1) and promoter element 2 (PE2)—at their 3' termini. The spacer region between these elements follows the “rule of six,” crucial for genome replication. By a comprehensive analysis of paramyxoviral genome sequences, we identified distinct subcategories of PE2 based on C and CG repeats that were specific to Orthoparamyxovirinae and Avulavirinae/Rubulavirinae, respectively, mirroring their evolutionary lineages. Notably, the PE2 of Rubulavirinae is integrated into the gene-coding region, a unique trait potentially linked to its strict dependence on RNA editing for virus growth. This study also focused on the PE2 sequences in filovirus genomes. The strict conservation of the continuity of UN5 among virus species emphasizes its crucial role in viral genome replication.

viral RNA-dependent RNA polymerase (RdRp) complex composed of the large protein (L) and cofactor(s), such as phosphoprotein (P) or VP35.Although it is common that the viral genomic 3' terminus acts as a replication promoter (termed promoter element 1; PE1) for viral RdRp, viruses in the families Paramyxoviridae and Filoviridae possess bipartite promoters, which require a secondary promoter element (PE2) located in the internal genomic region (2)(3)(4)(5)(6).The paramyxo-and filoviruses employ an RNA-editing mechanism in the P gene, and in the NP, glycoprotein, and L genes, respectively (7)(8)(9).During the mRNA transcription process, viral RdRp recognizes the cis-acting element of the RNA-editing signal within the respective gene.Co-transcriptionally, non-template nucleotides (nts) are appended to the mRNA, enabling the synthesis of multiple proteins from a single gene (10)(11)(12).It is important to note that all viruses possessing bipartite promoters and RNA editing are inherently subject to the "rule of six, " which requires hexamer phasing of nucleotides within a genomic region to facilitate viral propagation (2)(3)(4)(5)(6)13).In both paramyxo-and filoviruses, the genomic RNA is covered by NPs, with each NP monomer binding 6 nts (14)(15)(16)(17).
The bipartite replication promoters play an essential role in that the virus genome is recognized as the multiple of 6 by RdRp of paramyxoviruses.The paramyxoviral PE2 is characterized by the presence of specific nucleotides at every 6 nts.Paramyxovirus nucleocapsids have 13 NP subunits per turn such that PE1 and PE2 are juxtaposed on the same face of the nucleocapsid helix for concerted recognition by the viral RdRp (7,8).In Sendai virus (SeV) in the subfamily Orthoparamyxovirinae, the 14th to 16th hexamers contain 5′-GAAGAC UUGGAC UUGUCC-3′, in which the 6th nt is always a C (Fig. 1A and B, PE2) (2).In parainfluenza virus type 5 (PIV5) in the subfamily Rubulavirinae, the 13th to 15th hexamers contain 5′-CGGGAU CGAUGG CGAGGA-3′, in which the 5′-proximal 2 nts are always CG (Fig. 1A and B, PE2) (4).If a 1-nt insertion occurs upstream in the genome, the repeated C in SeV and the repeated CG in PIV5 would be shifted by 1 nt (Fig. 1B, 1-nt insertion).In this case, viral RdRp cannot recognize the genome as a correct template, resulting it to non-template for further replication, which acts to keep the remaining genome in a multiple of 6 in the infected cells.Although the PE2 is thus essential for viral genome replication, its characteristics have only been studied in limited numbers of virus species.
The family Filoviridae contains several genera, including Ebolavirus, Marburgvirus, and Cuevavirus.The genus Ebolavirus contains at least six viruses, including Zaire ebolavirus (ZEBOV), Sudan ebolavirus (SUDV), Tai forest ebolavirus (TAFV), Reston ebolavirus (RESTV), Bundibugyo ebolavirus (BDBV), and Bombali ebolavirus (BOMBV), which have been shown to have different levels of lethality in humans.The genus Marburgvirus includes Marburgvirus (MARV) and Ravn virus (RAVV).The characteristics of the PE2 of Filoviridae have been studied in detail by using ZEBOV as a representative model system.The ZEBOV genome contains PE2 in the NP mRNA 5' untranslated region (UTR) with uracil (U) every 6 nts (UN 5 ) eight times starting from the 81st nt upstream of the genomic 3'terminus (as shown in Fig. 4B) (5,18).In contrast to the PE2 of ebolaviruses, the genus Marburgvirus, specifically MARV, exhibits a distinctive PE2 characteristic: UN 5 hexamer continuity from the gene start (GS) signal (19).To date, there has been a lack of compre hensive studies on PE2 across strains of Filoviridae, leaving the extent of the diversity within PE2 among these viruses unknown.
In this study, we comprehensively analyzed PE2 sequences using a public database and found that almost all paramyxoviruses have C or CG repeats in PE2; the sequence patterns in PE2 were clearly divided in a subfamily-specific manner.The comparable analysis for the position of PE2 and virological properties of the genome replication indicates a unique characteristic specific for the Rubulavirinae.The filovirus PE2 is composed of consecutive UN 5 hexamers, and we confirmed that the continuity of UN 5 is unique to each virus species (three to eight times) with a high degree of conservation.

Comparison of PE2 sequences in the family Paramyxoviridae
For the comprehensive analysis of PE2 sequences across paramyxoviruses, complete genome sequences annotated as "Paramyxoviridae" were downloaded from the NCBI refseq database, yielding a total of 5,100 sequences.The sequence length ranged from 13,488 nt (Avian orthoavulavirus 1, MK495880.1) to 20,544 nt (Ninove microtus virus, OK623355.1).Within this data set, 1,647 sequences exhibited the conserved paramyxovi ral genome terminus 5′-ACC-GGU-3′, and notably, 99.3% of these sequences conformed to a multiple of six length (Table S1).Further classification of the sequences revealed three subfamilies, Orthoparamyxovirinae, Avulavirinae, and Rubulavirinae, consisting of 675, 789, and 183 sequences, respectively.To assess sequence conservation, we examined the 73 to 98 nts from the 3' end of the genome, which represent the genomic PE2, and from the 3' end of the antigenome, which represent the antigenomic PE2, using the WebLogo sequence logo generator (https://weblogo.berkeley.edu/logo.cgi).Almost all viruses in the subfamily Orthoparamyxovirinae possess a conserved PE2 sequence pattern characterized by a C every 6 nts (N 5 C) in the hexamers number (hex#) 14, 15, and 16 within both the genomic and antigenomic PE2 (Fig. 1C).In contrast, in the majority of the Avulavirinae and Rubulavirinae viruses, CG was found in every 6 nts (CGN 4 ) in hex# 13, 14, and 15 in both the genomic and antigenomic PE2 (Fig. 1C).While some variation emerged within virus genera (Fig. S1), it was clear that PE2 sequence conservation could be definitely categorized into two patterns: N 5 C for hex# 14, 15, and 16 (Orthoparamyxo virinae) and CGN 4 for hex# 13, 14, and 15 (Avulavirinae and Rubulavirinae).

Differences in the antigenomic PE2 location among paramyxovirus subfami lies
The genomic PE2 is presumed to reside within the 3' UTR of the genome, while the antigenomic PE2 is situated in the 5' UTR of the genome.However, our previous investigations have revealed that in the human parainfluenza virus type 2 belonging to the subfamily Rubulavirinae, the antigenomic PE2 is localized within the open reading frame (ORF) of the viral L protein (20).Consequently, paramyxoviruses may be cate gorized into two distinct classes: those harboring PE2 outside of the ORF and those harboring PE2 within the ORF.By examining the length of the 3' UTR and 5' UTR of each viral genome, it is possible to determine whether PE2, which is located at position 73 to 98 from the genomic and antigenomic ends, is situated outside or inside of the ORF.We first comprehensively examined the lengths of the 3' UTRs in 1,647 genome sequences of paramyxoviruses (Fig. S2).The shortest 3' UTR was 98 nts in the Denwin virus in the subfamily Orthoparamyxovirinae (OK623354), indicating that there are no viruses in the database with a genomic PE2 within the ORF.Subsequently, the lengths of the 5' UTRs of the genomes were comprehensively examined in the same database.The findings indicated that all viruses belonging to the subfamily Rubulavirinae, with some exceptions, e.g., mumps virus (MuV), had antigenomic PE2 sequences embedded within the L ORF (Fig. 2A and B).Furthermore, we observed that the antigenomic PE2 sequences in only 15 viruses of the subfamily Orthoparamyxovirinae were partially within the L ORF (Table S1).For all of the other viruses, the antigenomic PE2 sequences were located outside of the ORF.The viruses within the family Paramyxoviridae are categorically delineated, as illustrated in Fig. 3A, based on a phylogenetic tree deduced from the alignment of L proteins (21).The evolutionary trajectory shows an initial divergence of the viruses into an Orthoparamyxovirinae lineage and another lineage that subsequently splits into the Avulavirinae and Rubulavirinae lineages.The initial divergence is strikingly consistent with the divergence of the N 5 C PE2 and CGN 4 PE2 patterns (Fig. 3A).The Rubulavirinae lineage is characterized by the localization of the antigenomic PE2 within the ORF and by a different RNA-editing mechanism (V-mode) (Fig. 3A).P-mode viruses perform RNA editing to generate the accessory V gene, while V-mode viruses perform RNA editing to generate the essential P gene (Fig. 3B) (22)(23)(24)(25)(26)(27).
The genomic PE2 of ebolaviruses has been studied in detail, especially in ZEBOV (5).The 3' UTR of a ZEBOV genome sequence (AF086833) is shown in Fig. 4B as a represen tative sequence.The ZEBOV PE2 is located in the NP mRNA region with U at every 6 nts eight times.We searched for a sequence similar to the PE2 of ZEBOV (AF086833) at the 3' end of the genome in all 2,574 sequences of ZEBOV obtained.Among the 2,574 sequences, PE2 could not be confirmed in 633 sequences because the 3' genomic end containing the PE2 region was not registered in the database.For the remaining 1,941 sequences, the PE2 sequence was identified, and sequence homology was examined The P mode produces an accessory protein via editing, whereas the V mode produces an essential P protein via editing.
using the WebLogo sequence logo generator (Fig. 4C).PE2 homogeneity was confirmed in these sequences, with 1,145 sequences perfectly matching that of AF086833.The other 796 sequences displayed minor mutations in the PE2 region (Fig. 4C).Notably, only 9 of 1,941 sequences exhibited mutations in the U position in the UN 5 hexamers.One sequence had the first U from the 3′ end changed to A (MG572232.1, isolated in Guinea in 2014), while eight other sequences had the fifth U from the 3′ end changed to C (LT630506.1,LT630508.1,LT630510.1,LT630513.1,LT630521.1,LT630541.1,LT630547.1,and LT630586.1,isolated in Guinea in 2015).These eight PE2 sequences are separated into four and three consecutive UN 5 hexamers.
Weik et al. had shown in a minigenome system that three consecutive UN 5 hexam ers located in the proper phase in PE2 were sufficient for ZEBOV minigenome activity (5).They suggested that replication occurs more efficiently when more hexamers are present.We also used the ZEBOV minigenome system to confirm that the deletion of eight consecutive UN 5 hexamers eliminated the minigenome activity (Fig. 4D).Using minigenomes with 6 nts containing a leading U (Uaacuu) added to the deletion region one hexamer at a time, it was shown that three consecutive occurrences of Uaacuu resulted in slight recovery of the minigenome activity, while four consecutive occurren ces of Uaacuu resulted in complete recovery of the activity (Fig. 4D).Consequently, a PE2 sequence divided into four and three consecutive UN 5 hexamers may exhibit no significant impact on viral polymerase activity, as evidenced by the minigenome system.
We subsequently performed a comparative analysis of genomic PE2 sequences in ebolavirus species other than ZEBOV (Fig. 5A).Among 90 SUDV sequences, only two PE2 patterns were identified, each characterized by the consistent presence of five consecutive UN 5 hexamers.TAFV, with only four genome sequences in the database, exhibited uniform PE2 sequences featuring six consecutive UN 5 hexamers.The 32 BDBV genome sequences in the database all exhibited uniform PE2 sequences featuring six consecutive UN 5 hexamers.Within the 22 RESTV sequences, two PE2 patterns were found-one comprised 16 sequences of five consecutive UN 5 hexamers, and the other comprised six sequences of four consecutive UN 5 hexamers.In the eight BOMBV genome sequences in the database, all PE2 sequences exhibited four consecutive UN 5 hexamers.These findings showed the conservation of four or more consecutive UN 5 hexamers in the PE2 sequences across all ebolavirus species.
Seven LLOV sequences within the genus Cuevavirus were found in the database, and they contained three consecutive UN 5 hexamers within PE2 (Fig. 5B).No sequences with more than four consecutive UN 5 hexamers were identified in the 3′ UTR of the LLOV genomes.MARV and RAVV of the genus Marburgvirus exhibit a distinctive PE2 charac teristic: UN 5 hexamers continuity from the GS signal (19).Analysis of the homology between the GS-containing portion and PE2 in MARV and RAVV revealed that all 83 MARV PE2 sequences had 7 consecutive UN 5 hexamers, with no observed mutations in the U (Fig. 5C).All eight RAVV PE2 sequences were identical to one of the MARV PE2 sequences.

DISCUSSION
We compared the PE2 sequences, which have only been studied in detail in a limi ted number of virus species, using a database containing all available paramyxovirus genome sequences.The results showed that the PE2 of paramyxoviruses can clearly be classified into two groups: those with N 5 C at hex# 14, 15, and 16 (Orthoparamyxovirinae) and those with CGN 4 at hex# 13, 14, and 15 (Avulavirinae and Rubulavirinae) (Fig. 1C).This is consistent with previous findings showing the PE2 sequences of several representative paramyxoviruses (7,8,28).This difference in the PE2 pattern is consistent with the evolutionary phylogenetic tree of the viruses (21); the ancestors first diverged into an Orthoparamyxovirinae lineage and another lineage, which was linked to divergence of PE2 regions into those with three consecutive N 5 C versus three CGN 4 hexamers (Fig. 3A).Using the lengths of the 3' UTR and 5' UTR as a simple indicator of whether PE2 is located inside or outside of the ORF, we found that most viruses in the subfamily Rubulavirinae, with the exception of a few viruses including MuV, have the antigenomic PE2 within the ORF of the L protein (Fig. 2).The Rubulavirinae lineage differs from the Orthoparamyxovirinae and Avulavirinae lineages in that the P mRNA is produced by RNA editing (23,26,27).While the Rubulavirinae are strictly reliant on RNA editing for viral propagation, the Orthoparamyxovirinae and Avulavirinae require it only for accessory protein production (22,24,25).RNA-editing efficiency is regulated by an RNA-editing signal present on the P gene that functions as a cis-acting element (7), and NPs also bind to this cis-acting element region.Notably, the positioning of the cis-acting element on NPs affects the RNA-editing efficiency (29).When NPs bind correctly from the 5' end of the genome, the cis-acting element is placed correctly on the NPs (7,29).The antige nomic PE2 functions as a promoter during genome replication from the antigenome, and the precise placement of the cis-acting element on NP relies on whether the spacer sequence between the antigenomic PE1 and antigenomic PE2 is a multiple of 6 (7,29).Therefore, if there is no mutation in PE2, the insertion/deletion-sensing system of the genome can operate correctly, and the cis-acting element will be placed correctly on the NP.The non-coding regions are more susceptible to mutations than the coding regions.Particularly, the ORF for the essential RdRp L protein has remained remarkably conserved throughout viral evolution.The Rubulavirinae, which have an antigenomic PE2 embed ded within the L ORF, appear to have evolved mechanisms to resist insertions, deletions, and other mutations in PE2 more effectively than viruses of the other subfamilies.Further research is warranted to determine conclusively whether there is any connection between the RNA-editing efficiency and the antigenomic PE2 sequence.
In this study, we systematically searched a public database for all filovirus sequences and investigated the PE2 sequences in viruses within the genera Ebolavirus, Cuevavirus, and Marburgvirus.A total of 1,941 PE2 sequences were acquired for ZEBOV, including those isolated from 1976 to 2020 in various regions, such as the Democratic Republic of the Congo, Gabon, Guinea, Liberia, Mali, Nigeria, Republic of Congo, Sierra Leone, and Uganda (Fig. 4C; Table S2).Despite the diverse geographic origins of the isolates, the ZEBOV PE2 sequences showed remarkable uniformity.The PE2 sequences of viruses of the genus Ebolavirus other than ZEBOV also showed no significant differences in the PE2 sequences among virus strains.The PE2 of viruses in the genus Marburgvirus exhibits a distinct feature with seven consecutive UN 5 hexamers, including the GS signal (19,30); this UN 5 continuity was fully conserved in all 91 sequences including both MARV and RAVV (Fig. 5C).The continuity of UN 5 hexamers is fixed for each virus species in filoviruses with an extremely high degree of conservation.
The PE2 sequences within the genus Ebolavirus exhibit variations in the continuity of UN 5 hexamers: ZEBOV displays eight, TAFV and BDBV display six, SUDV displays five, RESTV displays five or four, and BOMBV displays four consecutive UN 5 hexamers in their PE2.Notably, all ebolavirus species, despite their diversity, harbor four or more consecutive UN 5 hexamers in PE2.Our ZEBOV minigenome experiment showed that a minimum of four consecutive UN 5 hexamers is imperative for optimal minigenome activity (Fig. 4C).The necessity of four or more consecutive UN 5 hexamers for replication appears to be a shared attribute among viruses within the genus Ebolavirus.The majority of the PE2 sequences in ZEBOV are characterized by eight consecutive UN 5 hexamers, although rare instances of altered UN 5 hexamer continuity appear to occur in nature (Fig. 4C).We identified one sequence with a mutation in the U of the first 3′ UN 5 hexamer that resulted in seven consecutive UN 5 hexamers.Additionally, eight sequences displayed a mutation in the U at the midpoint of the UN 5 hexamers, leading to a division of PE2 into four and three consecutive UN 5 hexamers.Viruses with a longer UN 5 hexamer repeat in PE2, such as ZEBOV, may be more resistant to mutations in the essential U since viral genome replication appears to be preserved if at least four consecutive UN 5 hexamers are maintained.
Viruses employing an RNA-editing mechanism during mRNA transcription concur rently have bipartite promoters.Paramyxovirus RdRp recognizes the consecutive C sequence within the RNA-editing site on the P gene as a cis-acting element and inserts G(s) into the mRNA during the transcription process (7,8).Filovirus RdRp recognizes the contiguous U sequence within the RNA-editing site on the NP, glycoprotein, and L genes and inserts A(s) into the mRNA (9).The PE2 of paramyxoviruses encodes C and that of filoviruses U residues every 6 nt.Thus, there appears to be a mecha nism whereby paramyxovirus RdRp recognizes C-based elements, and filovirus RdRp recognizes U-based elements.

Plasmid construction
The nanoluciferase (Nluc)-expressing ZEBOV minigenome plasmid (ZEBOV-Nluc) was constructed using the pUC57 plasmid backbone.The Nluc gene was amplified by PCR and flanked by the 3' UTR (469 nts) containing the leader sequence and the 5' UTR (740 nts) containing the trailer sequence of the ZEBOV genome.The minigenome is set under the control of the T7 RNA polymerase promoter, and the transcript expressed as a negative-sense RNA is cleaved at both ends by a hammerhead ribozyme and a hepatitis delta virus ribozyme (32).The ZEBOV NP, VP35, VP30, and L genes cloned into a pCAGGS vector were as described previously (33).A pCAGGS-derivative plasmid for the expression of firefly luciferase (Fluc) was also constructed.The deletions of the PE2 in ZEBOV-Nluc were performed by a standard cloning method.Briefly, ZEBOV-Nluc lacking 8, 7, 6, 5, and 4 sets of the eight consecutive UN 5 hexamers (ΔU8, ΔU8 + 1, ΔU8 + 2, ΔU8 + 3, and ΔU8 + 4, respectively) within PE2 were amplified by PCR and cloned into the pUC57 plasmid.The virus sequence used in this study was derived from a ZEBOV strain (GenBank accession number AF086833).

Analysis of the PE2 sequences and UTR lengths of the viruses in the family Paramyxoviridae
Sequences annotated as Paramyxoviridae were downloaded from the NCBI refseq database on 3 August 2022.There were 64,057 sequences annotated as Paramyxoviridae in the database.Based on the information that the length of paramyxovirus sequences ranged from 13,488 nt (Avian orthoavulavirus 1, MK495880.1) to 20,544 nt (Ninove microtus virus, OK623355.1),sequences shorter than 10,000 nt and longer than 30,000 nt (n = 58,957) were excluded from the analysis.Within the obtained data set, 1,647 sequences exhibited the conserved paramyxoviral genome terminus 5′-ACC-GGU-3′, and 99.3% of these sequences conformed to a multiple of six length.For each sequence, we detected the genome length, 3′ UTR and 5' UTR lengths, and extracted the sequen ces of the antigenomic PE2 (73 to 96 nts from the 3' terminus of the antigenome) and genomic PE2 (73 to 96 nts from the 3' terminus of the genome) (Table S1).The codes used for this analysis are available on GitHub (https://github.com/shohei-kojima/Filo_Paramyxo_2023).

Search for PE2 sequences in the genome of viruses in the family Filoviridae
Sequences annotated as Filoviridae were downloaded from the NCBI refseq database on 15 February 2024.The codes used for this analysis are available on GitHub (https:// github.com/shohei-kojima/Filo_Paramyxo_2023).There were 4,664 sequences annotated as Filoviridae in the database.The presence of the PE2 sequence was searched for through alignment with the PE2 sequence of each virus as previously reported (6,30).

Statistical analysis
Statistical analyses were performed with the Prism software (version 9.1.2;GraphPad, San Diego, CA, USA).Statistical significance was assigned when P values were <0.05.Inferential statistical analysis was performed by one-way analysis of variance followed by Tukey's test.

FIG 1 (
FIG 1 (Continued) the number of the NP hexamer from the 3' terminus.The line indicates RNA.(B) Schematic of the replication promoters for the genomic RNA of SeV and PIV5.The NPs (gray squares) bind to a sequence of 6 nts.The numbers in the circles indicate the positional number of the NP hexamer from the 3' terminus.The normal states are shown on the left, while the nucleotide positions when a single-nucleotide insertion has occurred upstream are shown on the right.The 24 nts covered by the 13th to the 16th NP monomer collectively constitute the PE2.The conserved nucleotides in PE2 are highlighted in red.(C) Conserved nucleotides within the genomic and antigenomic promoters of viruses belonging to the subfamilies Orthoparamyxovirinae, Avulavirinae, and Rubulavirinae.The numbers in the circles below the diagrams indicate the number of the hexamer from the 3' terminus.

FIG 2
FIG 2 Analysis of the genomic 5' UTR lengths of viruses in the family Paramyxoviridae.(A) The nucleotide length of the genomic 5' UTRs of viruses belonging to the subfamilies Orthoparamyxovirinae, Avulavirinae, and Rubulavirinae.Each thin vertical line represents a virus sequence.The antigenomic PE2 (agPE2) region (nts 73 to 98) is shown in yellow.(B) Viruses are divided into agPE2 Out-or In-ORF types.Tr indicates trailer sequence.

FIG 3
FIG 3 Relationship between PE2 and the virological properties of paramyxoviruses.(A) Relationships within the evolutionary phylogenetic tree, the sequence and location of PE2, and the RNA-editing pattern.(B) Differences in the RNA-editing modes.

FIG 4
FIG 4 Comprehensive analysis of filovirus PE2 sequences.(A) Sequence counts used in the analysis of PE2 sequences.The virus genera are shown in italic letters.(B) The PE2 sequence of ZEBOV.The 3′ UTR sequence from a representative ZEBOV genome (AF086833) is shown.(C) The PE2 sequences of ZEBOV.All sequences found in the data set are shown below.(D) The minigenome study of ZEBOV.The relative values of Nluc expression are shown with the values of the normal ZEBOV minigenome set to 1. NP (−) indicates the value from the assay using an empty plasmid instead of the plasmid encoding ZEBOV NP.Bars represent the means and standard deviations (n = 3 from three independent experiments).