Introduction
Filoviruses are responsible for complex hemorrhagic fevers in humans and primates with case fatality rates of 50–90% (1). The filovirus family belongs to the order Mononegavirales and is divided into two genera: Ebolavirus comprising five species, Zaire ebolavirus (EBOV), Taï Forest ebolavirus, Reston ebolavirus, Bundibugyo ebolavirus and Sudan ebolavirus; and Marburgvirus, which contains one species, Marburg marburgvirus. A third genus, Cuevavirus has also been recently proposed (2). Ebolaviruses have a non-segmented negative-sense (NNS) single-stranded RNA genome, a characteristic of the order, Mononegavirales, and most closely resembling the families Rhabdoviridae and Paramyxoviridae in its genetic assembly. The EBOV genome is ~19,000 base pairs in length and contains seven open reading frames that code for seven structural proteins in order, from 3′ to 5′, nucleoprotein (NP), VP35, VP40, glycoprotein (GP), VP30, VP24 and the L protein (3).
The L protein is 2,212 amino acids in length and is the largest protein encoded by the EBOV virus. It does not function independently but is part of an RNA-dependent RNA polymerase (RdRp) complex with the NP and viral transcription factors, VP30 and VP35. This complex is important in the replication and transcription of the viral genome. The L protein is hypothesized to serve as the major catalytic subunit of this complex (4). The majority of the data available on the structure and function of the EBOV L protein is theoretical and inadequate, relying on comparative analysis and data from the L proteins of other well-studied NNS RNA viruses of the order Mononegavirales rather than on experimental observations.
The current 2014 EBOV epidemic in West Africa is the largest ever reported Ebola epidemic across multiple West African countries with unprecedented transmissibility. There are no marketed antiviral therapeutic agents or vaccines against EBOV and the continued search for effective and safe therapeutic agents is required (5). RdRp of other RNA viruses, such as hepatitis C virus, have been successful targets for the development of antiviral therapeutic agents and vaccines due to the presence of various highly conserved motifs and/or residues (6,7). A previous study identified single nucleotide polymorphisms that distinguish the 2014 West African outbreak from previous outbreaks and demonstrated how the virus spread from Guinea to Sierra Leone in a matter of few months (8). The authors have identified 16 novel non-synonymous amino acid substitutions in the L protein that are present in the 2014 outbreak isolates (8).
In the present study, the 2014 EBOV L protein sequences were compared with other NNS RNA viruses, including Paramyxoviridae and Rhabdoviridae. The majority of previously reported L protein specific linear were conserved (9). Novel conserved regions that were not previously documented were also identified in the current study. Notably, all the mutations identified in the 2014 EBOV isolates were tolerant, they were pathogenic with certain examples occurring within previously determined functional conserved motifs, possibly attenuating viral pathogenicity, replication and virulence. Phylogenetic analysis of L protein sequences from past EBOV outbreaks was also conducted and was consistent with previous studies.
Discussion
Sequences of residues ranging in length from four to seven amino acids, clustered into six domains of strong conservation and separated by variable regions have been defined for the families, Rhabdoviridae and Paramyxoviridae (9,16). These six domains were also observed to be conserved in previous EBOV L proteins compared with paramyxoviruses (17). Notably, these conserved regions, are not retained in the current 2014 EBOV isolates, except in the mutated regions. The domains are key for crucial activities conducted by the L proteins.
Three polymerase motifs, motif A (aa 646–665), motif B (aa 892–898) and motif C (aa 2,034–2,063) identified previously in other NNS viruses, including EBOV L proteins were also observed in the 2014 EBOV isolates, but with a number of mutations (17). The invariant tripeptide GHP motif (aa 464–466), recognized previously as a turn structure exposing the histidine residue, was identified to be conserved amongst all the viruses (9). This motif may potentially serve as a target for future antiviral therapeutic agents, and site-specific mutagenesis studies may further elucidate its function.
Another pentapeptide, QGDNQ (aa 894–898), part of motif B is conserved in all NNS viruses, excluding EBOV. This motif is crucial for RNA polymerase function and phosphodiester bond formation and/or template recognition. In EBOV, the first glycine residue is substituted for a methionine residue (18). This substitution is not specific to the 2014 EBOV isolates as it has been demonstrated across all other Ebolavirus species. This substitution was not key in the 2014 outbreak, however, its evolutionary role in the transcription and the replication process requires further elucidation. Furthermore, the GDN part of the sequence corresponds to the strictly conserved residues, GDD in other polymerases (18). It is therefore an ideal target for antiviral therapeutic agents.
The KEKE tetrapeptide (aa 646–649), part of motif A that is responsible for the positioning and binding of the RNA template, was observed to be conserved in all viruses, except vesicular stomatitis and rabies viruses. In Rhabdoviridae, the lysine residue was substituted for glutamic acid, consistent with previous studies indicating that filoviruses are more closely associated with the paramyxoviruses than to rhabdoviruses (9,16,17).
Another pentapeptide motif GG(I/L)EG (aa 860–865) was also conserved in the alignment. EBOV contained an isoleucine rather than a leucine, further suggesting a closer association with the family Paramyxoviridae (9,19). Two invariant dipeptides GG (aa 1,071–1,072) and DP (aa 1,088–1,089) were also conserved as reported previously (9).
The HR motif (aa 1,456–1,457), previously indicated to be important for the PRNTase activity of the L protein at the enzyme-pRNA intermediate formation step, was also observed to be conserved across all the virus sequences (16,20).
Furthermore, the GXGXG pentapeptide (aa 2,057–2,061) part of motif C was demonstrated to be conserved in all members of the Rhabdoviridae family and two members of the Paramyxoviridae family (Sendai and measles) with slight inter-species variations in the X amino acids. This pentapeptide is hypothesized to be involved in mRNA capping via a 2′-O-ribose-methyltransferase action (21).
To the best of our knowledge, this is the first study to demonstrate that the L(N/S)SP(L/I)V motif (aa 41–46) is conserved across the viruses. However, the function of this motif remains to be elucidated. A QK(G/L)W(S/T) motif (aa 867–871) was also observed to be conserved, the exact function of this motif remains to be determined. In addition, the pattern of conservation itself is notable, beyond the 1,540th amino acid, the conservation decreases to the 2,024th amino acid, following which there is another stretch of partially conserved residues to the 2,241th amino acid. Beyond this, there is a highly non-conserved region filled with large gaps. This demonstrates that the conserved residues predominantly lie in the mid-region of the protein, consistent with the results of a previous study (9).
Furthermore, the phylogenetic tree of the EBOV sequences is similar to trees drawn previously on the basis of GP and NP sequences, demonstrating the same pattern of clustering (22). Though the previously drawn trees did not take into account the latest 2014 sequences, the pattern of clustering remains the same, predominantly due to minimal mutations acquired by the viruses in previous outbreaks. The same findings were concluded in a recent study using whole genome analysis (8).
As presented in Table II, it is notable that all the substitutions are tolerated and pathogenic. Of the first four mutations, two are located within domain I that spans aa 313–515 of the EBOV L protein in the alignment in the present study (aa 222–426 without the gaps for improved alignment). As demonstrated previously, this domain contains stretches of highly conserved residues and motifs, including GHP (9). In a previous study, it was demonstrated that the binding domain for the polymerase cofactor VP35, is located within the first 370 amino acids of the L protein towards the N-terminal, while the core binding domain is located between aa 280–370. This binding domain overlaps with part of domain I (23). Previous site-directed mutagenesis studies into the Sendai virus L protein have demonstrated that mutation in this domain leads to a defective replication and transcription mechanism and improper P protein binding (24). The first four mutations presented in Table II are located within this binding domain, with the M337I and Q346H mutations specifically in the core binding region. However, such mutations appear to be advantageous to the 2014 EBOV and may have enhanced genome replication efficiency. The fourth substitution is isolated, it does not lie within any known functional motif, thus, its role, if any, in the virulence of the virus requires elucidation by site-directed mutagenesis studies.
The next five mutations are also notable. All the mutations lie close to each other within a span of 90 amino acids. These mutations occur outside the conserved domains, within a region of high variability also observed within previous filovirus L protein sequences by comparative analysis (17). Further research regarding this region is required prior to drawing any conclusions about these mutations.
Notably, the S1826D mutation lies within the conserved motif C identified earlier (17). The mutation is close to the GXGXG pentapeptide. Earlier mutation studies within this region in the Sendai virus have led to reduced replication and transcription activities, however, it did not completely block the polymerase function (25). This mutation is tolerated and appears to be a gain of function mutation.
The final three mutations lie within a highly unconserved region, for which no functional domains or motifs have been defined or hypothesized. However, the I1951V substitution is notable. The amino acid isoleucine is conserved at position 1,951 in all other Ebolavirus species. It is only substituted to valine in the 2014 EBOV isolates. The impact of this substitution remains to be elucidated and requires further research.
In conclusion, the present study provided an overview of the L protein amino acid sequences of the 2014 EBOV isolates. The sequence and mutation analysis of L polymerase will aid in advancing towards confinement of future outbreaks. The current study may guide future site-directed mutagenesis studies, and design of site-specific direct targeting antiviral therapeutic agents. It may also help with the development of future diagnostic studies.