Characterization of the catalytic center of the Ebola virus L polymerase

Abstract

Background: Ebola virus (EBOV) causes a severe hemorrhagic fever in humans and non-human primates. While no licensed therapeutics are available, recently there has been tremendous progress in developing antivirals. Targeting the ribonucleoprotein complex (RNP) proteins, which facilitate genome replication and transcription, and particularly the polymerase L, is a promising antiviral approach since these processes are essential for the virus life cycle. However, until now little is known about L in terms of its structure and function, and in particular the catalytic center of the RNA-dependent RNA polymerase (RdRp) of L, which is one of the most promising molecular targets, has never been experimentally characterized.

Methodology/principal findings: Using multiple sequence alignments with other negative sense single-stranded RNA viruses we identified the putative catalytic center of the EBOV RdRp. An L protein with mutations in this center was then generated and characterized using various life cycle modelling systems. These systems are based on minigenomes, i.e. miniature versions of the viral genome, in which the viral genes are exchanged against a reporter gene. When such minigenomes are coexpressed with RNP proteins in mammalian cells, the RNP proteins recognize them as authentic templates for replication and transcription, resulting in reporter activity reflecting these processes. Replication-competent minigenome systems indicated that our L catalytic domain mutant was impaired in genome replication and/or transcription, and by using replication-deficient minigenome systems, as well as a novel RT-qPCR-based genome replication assay, we showed that it indeed no longer supported either of these processes. However, it still showed similar expression to wild-type L, and retained its ability to be incorporated into inclusion bodies, which are the sites of EBOV genome replication.

Conclusions/significance: We have experimentally defined the catalytic center of the EBOV RdRp, and thus a promising antiviral target regulating an essential aspect of the EBOV life cycle.

Fig 1. Multiple sequence alignment of filovirus, RSV, and VSV polymerases.

(A) Pairwise distance of aligned sequences. The percentage of identical (bottom left) or similar (top right) amino acids is shown. (B) Positions 631 to 980 of the multiple sequence alignment. Dots indicate amino acids identical to the consensus sequence. The conserved GDNQ motif is highlighted in yellow.

Fig 2. Effect of mutation of the GDNQ motif on expression and intracellular localization of EBOV L.

(A) Western blot of 293T cell lysates transfected with L-mCherry or L_mut-mCherry. Blots were cut in half, and stained with monoclonal antibodies against mCherry or actin. (B) Quantification of L-mCherry expression. Results from western blotting were quantified, and the expression level of wild-type L set to 100%. Means and standard deviations of 4 biological replicates from 3 independent experiments are shown. (C) Intracellular localization of L-mCherry or L_mut-mCherry. Cells were transfected with expression plasmids for all RNP proteins (including either wild-type L, L with a fluorescent mCherry tag (L-mCherry) or a mutated version of L-mCherry without the GDNQ motif (L_mut-mCherry)), a GFP-expressing replication-competent minigenome, T7-polymerase to facilitate initial transcription of minigenome vRNA, and pmTurquoise2-H2A to label cell nuclei. Live cells were visualized after 48 hours by spinning disc confocal microscopy. The white bar indicates 10 μm.

Fig 3. Functional characterization of the GDNQ motif in EBOV L on viral transcription and replication.

(A) Effects of the GDNQ mutation in L on genome replication and/or transcription. A classical minigenome (MG) assay using a Renilla luciferase reporter (rep) was performed with wild-type L (L) or the mutated L (L_mut). In this assay reporter activity (top graph, shown in relative light units (RLU) on a log scale) reflects genome replication and transcription in combination, but doesn’t distinguish between effects on one or both of these processes. In addition, plasmid-based gene expression from a Firefly (FF) luciferase plasmid was measured as a control and is shown in the bottom graph. Means and standard deviations of 3 biological replicates from 3 independent experiments are shown. (B) Effects of the GDNQ mutation in L on genome transcription. A replication-deficient minigenome assay was performed with wild-type and mutated L. In this assay, cRNAs cannot be copied back into vRNAs, abolishing replication, and thus reporter levels become independent of replication, which normally amplifies the available number of vRNA templates for reporter mRNA transcription. Therefore, Renilla luciferase reporter activity in this system reflects genome transcription alone. Again, plasmid-based gene expression from a Firefly (FF) luciferase plasmid was measured as a control and is shown in the bottom graph. Means and standard deviations of 3 biological replicates from 3 independent experiments are shown. (C) Effects of the GDNQ mutation in L on genome replication. p1 cells were pretransfected with NP, VP35, and L or L_mut, and then infected with trVLPs containing a tetracistronic minigenome. Replication of this minigenome in p1 cells was measured by RT-qPCR targeting the intergenic region located between the GP_1,2 and VP24 ORFs on the minigenome. Means and standard deviations of 4 biological replicates from 2 independent experiments are shown.