The P1N-PISPO trans-Frame Gene of Sweet Potato Feathery Mottle Potyvirus Is Produced during Virus Infection and Functions as an RNA Silencing Suppressor

Abstract

The positive-sense RNA genome of Sweet potato feathery mottle virus (SPFMV) (genus Potyvirus, family Potyviridae) contains a large open reading frame (ORF) of 3,494 codons translatable as a polyprotein and two embedded shorter ORFs in the -1 frame: PISPO, of 230 codons, and PIPO, of 66 codons, located in the P1 and P3 regions, respectively. PISPO is specific to some sweet potato-infecting potyviruses, while PIPO is present in all potyvirids. In SPFMV these two extra ORFs are preceded by conserved G2A6 motifs. We have shown recently that a polymerase slippage mechanism at these sites could produce transcripts bringing these ORFs in frame with the upstream polyprotein, thus leading to P1N-PISPO and P3N-PIPO products (B. Rodamilans, A. Valli, A. Mingot, D. San Leon, D. B. Baulcombe, J. J. Lopez-Moya, and J.A. Garcia, J Virol 89:6965-6967, 2015, doi:10.1128/JVI.00337-15). Here, we demonstrate by liquid chromatography coupled to mass spectrometry that both P1 and P1N-PISPO are produced during viral infection and coexist in SPFMV-infected Ipomoea batatas plants. Interestingly, transient expression of SPFMV gene products coagroinfiltrated with a reporter gene in Nicotiana benthamiana revealed that P1N-PISPO acts as an RNA silencing suppressor, a role normally associated with HCPro in other potyviruses. Moreover, mutation of WG/GW motifs present in P1N-PISPO abolished its silencing suppression activity, suggesting that the function might require interaction with Argonaute components of the silencing machinery, as was shown for other viral suppressors. Altogether, our results reveal a further layer of complexity of the RNA silencing suppression activity within the Potyviridae family.

Importance: Gene products of potyviruses include P1, HCPro, P3, 6K1, CI, 6K2, VPg/NIaPro, NIb, and CP, all derived from the proteolytic processing of a large polyprotein, and an additional P3N-PIPO product, with the PIPO segment encoded in a different frame within the P3 cistron. In sweet potato feathery mottle virus (SPFMV), another out-of-frame element (PISPO) was predicted within the P1 region. We have shown recently that a polymerase slippage mechanism can generate the transcript variants with extra nucleotides that could be translated into P1N-PISPO and P3N-PIPO. Now, we demonstrate by mass spectrometry analysis that P1N-PISPO is indeed produced in SPFMV-infected plants, in addition to P1. Interestingly, while in other potyviruses the suppressor of RNA silencing is HCPro, we show here that P1N-PISPO exhibited this activity in SPFMV, revealing how the complexity of the gene content could contribute to supply this essential function in members of the Potyviridae family.

FIG 1

Synergism between SPFMV and SPCSV. (A) Appearance of AM-MB2 plants propagated vegetatively, with a representative original plant (top) and another superinfected with SPCSV (bottom). Pictures were taken 4 months after whitefly-mediated inoculation of the crinivirus. (B) Normalized number of virus-derived RNA-seq reads in AM-MB2 plants (gray bars) compared to those plants superinfected with SPCSV (salmon bars). Individual viruses are indicated, and for each one the average and standard deviation are plotted. The inset with an extended scale was included to accommodate the large differences between the conditions. (C) Genomic organization of the SPFMV genome assembled from the RNA-seq data. The RNA genome of SPFMV is represented as a solid line flanked by a covalently linked VPg (solid circle) and the poly(A) tail, with the three ORFs corresponding to the polyprotein, PISPO, and PIPO depicted as boxes (details of the conserved G₂A₆ motifs are shown). Boxes with names represent the different gene products. (D) Percentages of RNA-seq reads with insertion of 1 nucleotide at the slippage sites in SPFMV. No other alterations were observed at these points, and the numbers show the values for two independent samples used for the RNA-seq analysis (see Table S7 in the supplemental material).

FIG 2

Taxonomic characterization of potyviruses from the AM-MB2 sweet potato plant. (A) Phylogenetic analysis of the complete nucleotide sequences corresponding to the three potyviruses found in AM-MB2 (highlighted in bold), compared to full-length sequences of SPFMV, SPVC, and SPV2 available in GenBank, using alignments and the maximum-likelihood algorithm to generate a tree with 1,000 replications for the bootstrapping test. (B) Plot of nucleotide identities between SPFMV isolate AM-MB2 and isolates Piu3 (EA strain) and RC-ARg (RC strain), after comparison with a sliding window of 100 nt. The recombination site detected by the SBP and GARD analysis is indicated with an arrowhead and the position number in the viral genome. The dashed horizontal lines correspond to a highly divergent region (between the indicated positions in the viral genome) with abundant gaps, which was eliminated to facilitate the global comparison. A schematic drawing of the expected gene products of SPFMV is shown above the plot to facilitate comparison of the recombination site.

FIG 3

In vitro expression of P1-related products. Wild-type and variant constructs were designed to express selected proteins. Arrowheads and asterisks indicate the translation initiation sites (AUG) and stop codons, respectively. Truncated forms with stop codons introduced by mutagenesis are indicated by asterisks in gray. The expected sizes (in kDa) of the anticipated translated protein products are indicated below each construct, with italics used to indicate the trans-frame products expected in the case of translational frameshifting. Wheat germ extract was programmed with RNA transcripts, including a luciferase mRNA and water as controls. Each construct is connected to the lanes on the SDS-PAGE analysis of the proteins synthesized after in vitro translation and detected by autoradiography. The electrophoretic mobilities of molecular mass markers are shown at the right.

FIG 4

Identification of viral proteins by LC-MS/MS in plant samples. (A) Detection of peptides corresponding to proteins transiently expressed in N. benthamiana leaves agroinfiltrated with a construct that contains the wild-type P1 sequence of SPFMV under control of the appropriate promoter for overexpression. The plasmid construct is schematically drawn inside a leaf outline, with a detail of the expected RNA sequence corresponding to the G₂A₆ motif upstream of the PISPO starting point and the detected protein P1. A question mark besides the P1N-PISPO name indicates that no peptides corresponding to the PISPO domain were found in the analysis. The peptides detected are shown (as gray boxes) distributed along the P1 protein sequence in the right panel, and the percentage of coverage is indicated. The sequences of the detected peptides are shaded in gray on the sequence of the P1 gene product, and the frameshift point is underlined. (B) Detection of peptides as in panel A but in the sample agroinfiltrated with the P1N-PISPO construct, which ensures expression of the trans-frame product. The PISPO region is highlighted and is depicted in italic lettering in the sequence, starting at the underlined frameshift point. Peptides found in the analysis are shown as boxes (gray for P1N and black for PISPO) in the graphic and shaded in the sequence below, using a gray background in the case of peptides corresponding to the P1N or a black background with white letters for the PISPO region. (C) Detection of viral peptides in the AM-MB2 I. batatas plant. The virus SPFMV is represented by a schematic virion inside the outline of the sweet potato leaf, and two variants of viral RNAs with the G₂A₆ or G₂A₇ motifs are indicated (13). Peptides deriving from the common P1N region, from the P1 C-terminal region, and from PISPO are shown in the protein schemes, using gray for the common P1N part and the rest of P1, while the peptides found in the PISPO frame are represented by black boxes. The sequences of peptides are also highlighted in the sequence, with the two variants shown after the frameshift point (underlined), represented in the upper lines for P1 or in the lower (in italic) for PISPO and using as above a gray or black background in the sequence detail, respectively.

FIG 5

RNA silencing suppression activity of SPFMV P1N-PISPO. (A) The constructs used are represented with the same conventions as in the other figures for AUG and stop codons. The distribution of the four WG/GW motifs present in the sequence (positions in parentheses) in the mutant in which the four W residues were replaced by A residues is indicated. (B) Patch design used for coagroinfiltration in N. benthamiana leaves of a GFP-expressing construct together with other constructs expressing SPFMV products (indicated by X), a CVYV P1b positive control (c+), or an empty vector (δ, c−). (C) Pictures of representative agroinfiltrated leaves taken at 3 (top row) or 5 (bottom row) days postagroinfiltration (dpai) under UV light. (D) Northern blot analysis of GFP mRNA and siRNA extracted from agroinfiltrated tissue patches at 3 dpai, comparing the different constructs indicated above each lane. The bottom panels show the ethidium bromide staining of the gels as loading controls. (E) Relative accumulation of GFP mRNAs measured by specific RT-qPCR and normalized against the mean value corresponding to the negative control. The average values ± standard deviations from several experiments, each performed with at least three independent Agrobacterium cultures, are plotted. Significant difference in pairwise comparisons and after applying the Tukey-Kramer test were found only for the positive control CVYV P1b (shown with a broken axis to accommodate the large difference) and for the P1N-PISPO samples.