Identification of tolerated insertion sites in poliovirus non-structural proteins

Abstract

Insertion of nucleotide sequences encoding "tags" that can be expressed in specific viral proteins during an infection is a useful strategy for purifying viral proteins and their functional complexes from infected cells and/or for visualizing the dynamics of their subcellular location over time. To identify regions in the poliovirus polyprotein that could potentially accommodate insertion of tags, transposon-mediated insertion mutagenesis was applied to the entire nonstructural protein-coding region of the poliovirus genome, followed by selection of genomes capable of generating infectious, viable viruses. This procedure allowed us to identify at least one site in each viral nonstructural protein, except protein 2C, in which a minimum of five amino acids could be inserted. The distribution of these sites is analyzed from the perspective of their protein structural context and from the perspective of virus evolution.

Figure 1

Schematic presentation of the steps involved in isolation of PVs with 15-nt insertions. Insertion mutant libraries were generated in three PV sublcones shown with the positions of the corresponding fragments within the PV genome. Each mutant full-length plasmid library was transcribed in vitro and the pooled transcripts were used to transfect HeLa cells for genetic selection. Viable viruses were isolated from individual plaques and their RNA genomes were sequenced to determine the locations of the transposon-mediated insertions. The locations of the 15-nt insertions identified from independent virus isolates are indicated by the arrows at the bottom. The numbers indicate the amino acid residue of the corresponding PV protein after which insertion occurred.

Figure 2

Plaque-forming viruses containing 15-nt insertions. The insertions and corresponding viruses are designated by the name of the protein containing the insertion and the number of the amino acid residue after which 5 additional amino acids were inserted. Insertions after the same amino acid that were encoded by15 nts inserted in different reading frames of the same or adjacent codons are named as “a” and “b” respectively. (A) Nucleotide and amino acid sequences in viruses with 15-nt insertions. For each insertion, the nucleotide (upper line) and the predicted amino acid (lower line) sequences are shown. The nucleotide numbers at the start of each sequence refer to the nucleotide numbers in the Mahoney strain of PV type 1 genome. The nucleotide sequences introduced by transposon insertion are in bold italics. In each case, five nts that were duplicated from the viral genome during transposon insertion are underlined, as are the five amino acid sequences introduced by the insertion. For the C-terminal insertions in protein 3C, the (potential) 3C-3D cleavage site(s) are indicated by arrowheads. The far right column indicates the number of independent isolates of each mutant that was recovered. (B) Plaque phenotypes of viruses with 15-nt insertions. Plaques were stained 48 h after infection with the indicated viruses. Virus plaque phenotypes were analyzed in four separate experiments (a – d), and each included a parental wild-type virus control for comparison. For panel (a), the parental wild-type virus was from pPVM, which forms slightly smaller plaques than that from pXpA, used as the parent for panels (b – d).

Figure 2

Plaque-forming viruses containing 15-nt insertions. The insertions and corresponding viruses are designated by the name of the protein containing the insertion and the number of the amino acid residue after which 5 additional amino acids were inserted. Insertions after the same amino acid that were encoded by15 nts inserted in different reading frames of the same or adjacent codons are named as “a” and “b” respectively. (A) Nucleotide and amino acid sequences in viruses with 15-nt insertions. For each insertion, the nucleotide (upper line) and the predicted amino acid (lower line) sequences are shown. The nucleotide numbers at the start of each sequence refer to the nucleotide numbers in the Mahoney strain of PV type 1 genome. The nucleotide sequences introduced by transposon insertion are in bold italics. In each case, five nts that were duplicated from the viral genome during transposon insertion are underlined, as are the five amino acid sequences introduced by the insertion. For the C-terminal insertions in protein 3C, the (potential) 3C-3D cleavage site(s) are indicated by arrowheads. The far right column indicates the number of independent isolates of each mutant that was recovered. (B) Plaque phenotypes of viruses with 15-nt insertions. Plaques were stained 48 h after infection with the indicated viruses. Virus plaque phenotypes were analyzed in four separate experiments (a – d), and each included a parental wild-type virus control for comparison. For panel (a), the parental wild-type virus was from pPVM, which forms slightly smaller plaques than that from pXpA, used as the parent for panels (b – d).

Figure 2

Plaque-forming viruses containing 15-nt insertions. The insertions and corresponding viruses are designated by the name of the protein containing the insertion and the number of the amino acid residue after which 5 additional amino acids were inserted. Insertions after the same amino acid that were encoded by15 nts inserted in different reading frames of the same or adjacent codons are named as “a” and “b” respectively. (A) Nucleotide and amino acid sequences in viruses with 15-nt insertions. For each insertion, the nucleotide (upper line) and the predicted amino acid (lower line) sequences are shown. The nucleotide numbers at the start of each sequence refer to the nucleotide numbers in the Mahoney strain of PV type 1 genome. The nucleotide sequences introduced by transposon insertion are in bold italics. In each case, five nts that were duplicated from the viral genome during transposon insertion are underlined, as are the five amino acid sequences introduced by the insertion. For the C-terminal insertions in protein 3C, the (potential) 3C-3D cleavage site(s) are indicated by arrowheads. The far right column indicates the number of independent isolates of each mutant that was recovered. (B) Plaque phenotypes of viruses with 15-nt insertions. Plaques were stained 48 h after infection with the indicated viruses. Virus plaque phenotypes were analyzed in four separate experiments (a – d), and each included a parental wild-type virus control for comparison. For panel (a), the parental wild-type virus was from pPVM, which forms slightly smaller plaques than that from pXpA, used as the parent for panels (b – d).

Figure 3

The structures of PV proteins displaying sites of the insertions of 5 amino acids. Amino acid residues preceding the 5-aa insertions are indicated in red. Ribbon diagrams are shown of (a) a putative model of PV 2A protein modeled using the I-TASSER server was used to predict the PV 2A structure from the NMR-derived magnetic resonance structure of Coxsackievirus B4 2A protein (PDB ID: 1Z8R) (Baxter et al., 2006); (b) the predicted homodimer formed by the N-terminal soluble portion (60 amino acids) of PV protein 3A (PDB ID: 1NG7) (Strauss et al., 2003); (c) peptide 3B (PDB ID: 2BBL) (Schein et al., 2006); (d) protein 3C (PDB ID: 1L1N) lacking amino acids 180-182 at the C-terminus) (Mosimann et al., 1997); (e) protein 3D (PDB ID:1RA6 (Thompson and Peersen, 2004); and (f) protein 3CD (PDB ID: 2IJD) (Marcotte et al., 2007). The structure analysis and graphics generation were done using PyMOL Viewer .

Figure 4

Conservation of the non-structural polyprotein part and location of PVM insertions. A plot of the conservation along the non-structural part of the polyprotein alignment is shown for three evolutionary scales which include eleven Coxsackie A virus plus three poliovirus serotypes forming the species Human enterovirus C (A), one representative for each of the four Human enterovirus species (B) and one representative for each of the eleven species of the genus Enterovirus (C). The normalized similarity measure was compiled under the Blosum62 substitution matrix and smoothed using a sliding window with a size of 5 aa and an overlap of 3 aa positions. The mean similarity of single protein alignments is indicated by dashed horizontal lines and their positions are shown using rectangles and names on top. Below, the positions of PVM insertions resulting in viable virus are indicated by vertical bars. Neighboring insertions are grouped together and highlighted with grey background. Each intersection of an insertion with a similarity distribution curve is magnified in a respective inlet that zooms in on the intersection.