West Nile virus 5'-cap structure is formed by sequential guanine N-7 and ribose 2'-O methylations by nonstructural protein 5

Abstract

Many flaviviruses are globally important human pathogens. Their plus-strand RNA genome contains a 5'-cap structure that is methylated at the guanine N-7 and the ribose 2'-OH positions of the first transcribed nucleotide, adenine (m(7)GpppAm). Using West Nile virus (WNV), we demonstrate, for the first time, that the nonstructural protein 5 (NS5) mediates both guanine N-7 and ribose 2'-O methylations and therefore is essential for flavivirus 5'-cap formation. We show that a recombinant full-length and a truncated NS5 protein containing the methyltransferase (MTase) domain methylates GpppA-capped and m(7)GpppA-capped RNAs to m(7)GpppAm-RNA, using S-adenosylmethionine as a methyl donor. Furthermore, methylation of GpppA-capped RNA sequentially yielded m(7)GpppA- and m(7)GpppAm-RNA products, indicating that guanine N-7 precedes ribose 2'-O methylation. Mutagenesis of a K(61)-D(146)-K(182)-E(218) tetrad conserved in other cellular and viral MTases suggests that NS5 requires distinct amino acids for its N-7 and 2'-O MTase activities. The entire K(61)-D(146)-K(182)-E(218) motif is essential for 2'-O MTase activity, whereas N-7 MTase activity requires only D(146). The other three amino acids facilitate, but are not essential for, guanine N-7 methylation. Amino acid substitutions within the K(61)-D(146)-K(182)-E(218) motif in a WNV luciferase-reporting replicon significantly reduced or abolished viral replication in cells. Additionally, the mutant MTase-mediated replication defect could not be trans complemented by a wild-type replicase complex. These findings demonstrate a critical role for the flavivirus MTase in viral reproduction and underscore this domain as a potential target for antiviral therapy.

FIG. 1.

Recombinant WNV NS5 and MTase proteins. (A) WNV genome. The 5′ and 3′ termini of the WNV genome contain an m⁷GpppAm cap structure and CU_OH, respectively. The NS5 protein (905 amino acids) consists of the MTase and RdRp domains connected by a bridge domain harboring an NLS. (B) Purification of wild-type MTase and NS5 proteins. Uninduced and IPTG-induced bacteria and purified proteins were analyzed by SDS-PAGE and stained with Coomassie blue. (C) Mutant MTases each containing an alanine substitution within the conserved K₆₁-D₁₄₆-K₁₈₂-E₂₁₈ tetrad. The sizes (kilodaltons) of protein standards (Marker) and recombinant proteins are indicated on the left and right, respectively.

FIG. 2.

Analyses of NS5- and MTase-generated cap structures. (A) TLC analysis of nuclease P1-resistant cap structures released from the WNV m⁷G*pppA-RNA or G*pppA-RNA methylated with full-length NS5, MTase domain, VP39, or no protein (Mock). The asterisk indicates that the following phosphate is ³²P labeled. The positions of the origin and migration of G*pppA, m⁷G*pppA, and m⁷G*pppAm molecules are indicated on the left. The relative conversion from m⁷G*pppA-RNA to m⁷G*pppAm-RNA (lanes 2 to 4) and the relative conversion from G*pppA-RNA to m⁷G*pppAm-RNA (lanes 6 and 7) were quantified with a PhosphorImager and are indicated below the autoradiograph. (B) TLC analysis of tobacco acid pyrophosphatase-digested products from the reactions in panel A. The positions of the origin, m⁷G*p, and G*p molecules are indicated on the left. (C) ³H-methyl incorporation into m⁷GpppA-RNA and GpppA-RNA to form m⁷GpppAm-RNA. The indicated RNA substrates were treated with MTase or NS5 in the presence of [³H-methyl]AdoMet, digested with nuclease P1, and analyzed by TLC. The m⁷GpppAm products were excised from the TLC plate, and ³H-methyl incorporation was measured by scintillation counting. All assays were performed with equal amounts of NS5 and MTase proteins (30 pmol). Average results from three independent experiments are presented.

FIG. 3.

WNV MTase requires a viral RNA sequence for N-7 and 2′-O methylations. RNA substrates representing positions A574 to G763 of plasmid pUC19 were prepared, treated with WNV MTase, digested with nuclease P1, and analyzed by TLC as described for WNV RNAs. WNV 5′-RNA substrates were included as positive controls. ³²P-labeled markers m⁷G*pppA and G*pppA are indicated on top. (Right panel) G*pppA-RNA substrates were used to test N-7 MTase activity; the methylation reaction mixtures were incubated at room temperature for 5 min. (Left panel) m⁷G*pppA-RNA substrates were tested for 2′-O MTase activity, and the methylation reaction mixtures were incubated for 1 h. Different incubation times were selected for the N-7 and 2′-O MTase reactions based on the methylation kinetics described in Fig. 4.

FIG. 4.

Time course analyses of the NS5-mediated MTase activities. ³²P-labeled m⁷G*pppA-RNA (A) and G*pppA-RNA (B) were methylated by full-length NS5 for the indicated times (minutes), digested with nuclease P1, and analyzed by TLC. ³²P-labeled markers m⁷G*pppA and G*pppA are indicated on top. For each time point, the relative conversion from m⁷G*pppA to m⁷G*pppAm (A) and the relative conversion from G*pppA to m⁷G*pppA and m⁷G*pppAm (B) are presented below the autoradiograph. The input m⁷G*pppA (0 min for A) or G*pppA (0 min for B) was set at 100%.

FIG. 5.

Methylation activities of mutant MTases. Mutant MTases containing the indicated alanine substitutions were assayed for 2′-O MTase (A) and N-7 MTase activities (B) at 22°C for 1 h and 5 min, respectively. The shorter incubation time (5 min) was chosen for the N-7 MTase assay in panel B to minimize subsequent 2′-O methylation, which would complicate the quantification of the N-7-only methylated product (m⁷GpppA). The experimental details were as described in Fig. 2. ³²P-labeled marker m⁷G*pppA or G*pppA is indicated on top. The relative conversions for 2′-O methylation (m⁷G*pppA to m⁷G*pppAm in panel A) and for N-7 methylation (G*pppA to m⁷G*pppA in panel B) were calculated by comparing the products produced from the mutant MTases with that generated from the wild-type protein (set at 100%).

FIG. 6.

Functional analyses of the effects of the conserved K₆₁-D₁₄₆-K₁₈₂-E₂₁₈ tetrad on WNV translation and replication. (A) A luciferase-expressing WNV replicon. A Renilla luciferase reporter is fused in frame with the open reading frame of the genome at a position where viral structural genes were deleted, resulting in RlucRep. (B and C) RlucRep containing alanine substitutions for residues of the K₆₁-D₁₄₆-K₁₈₂-E₂₁₈ tetrad was transfected into BHK cells (B) or into BHK cells that contained persistently replicating WNV replicon expressing a neomycin phosphotransferase gene (28) (C). At indicated time points, the transfected cells were lysed and quantified for luciferase activities. For each time point, luciferase signals derived from the wild-type RlucRep were set at 100%. For wild-type replicon, the luciferase signals collected at 72 h p.t. were consistently 80- to 100-fold higher than those collected at 2 h p.t., as previously described (34). Average results of three experiments are shown.

FIG. 7.

Effects of RNA cap structures on WNV RNA translation, replication, and stability. (A) Analysis of the effect of 5′-cap structure on viral translation and replication. Wild-type RlucRep RNAs containing indicated cap structures were transfected into BHK cells, and luciferase activities were quantified at indicated time points. For each time point, luciferase signals derived from the m⁷GpppA-RlucRep were set at 100%. (B) Analysis of the effect of 5′-cap structure on WNV RNA stability. A mutant RlucRep that contained a defective WNV polymerase (mt-RlucRep) (21) was used to prepare RNAs with indicated 5′-cap structures. The replicon RNAs (10 μg) were transfected into BHK cells. At 2, 6, 12, and 20 h p.t, a primer/probe set targeting WNV NS5 (27) was used to quantify viral RNA in real-time reverse transcription-PCR assays. For each replicon, the RNA level detected at 2 h p.t. was set at 100%. The degradation rate of RNA is indicated by the decrease in percentage of RNA level at later time points compared with the RNA level obtained at 2 h p.t. Average results of three experiments are shown.