Temperature-Sensitive Mutants in the Influenza A Virus RNA Polymerase: Alterations in the PA Linker Reduce Nuclear Targeting of the PB1-PA Dimer and Result in Viral Attenuation

Abstract

The influenza virus RNA-dependent RNA polymerase catalyzes genome replication and transcription within the cell nucleus. Efficient nuclear import and assembly of the polymerase subunits PB1, PB2, and PA are critical steps in the virus life cycle. We investigated the structure and function of the PA linker (residues 197 to 256), located between its N-terminal endonuclease domain and its C-terminal structured domain that binds PB1, the polymerase core. Circular dichroism experiments revealed that the PA linker by itself is structurally disordered. A large series of PA linker mutants exhibited a temperature-sensitive (ts) phenotype (reduced viral growth at 39.5°C versus 37°C/33°C), suggesting an alteration of folding kinetic parameters. The ts phenotype was associated with a reduced efficiency of replication/transcription of a pseudoviral reporter RNA in a minireplicon assay. Using a fluorescent-tagged PB1, we observed that ts and lethal PA mutants did not efficiently recruit PB1 to reach the nucleus at 39.5°C. A protein complementation assay using PA mutants, PB1, and β-importin IPO5 tagged with fragments of the Gaussia princeps luciferase showed that increasing the temperature negatively modulated the PA-PB1 and the PA-PB1-IPO5 interactions or complex stability. The selection of revertant viruses allowed the identification of different types of compensatory mutations located in one or the other of the three polymerase subunits. Two ts mutants were shown to be attenuated and able to induce antibodies in mice. Taken together, our results identify a PA domain critical for PB1-PA nuclear import and that is a "hot spot" to engineer ts mutants that could be used to design novel attenuated vaccines.

Importance: By targeting a discrete domain of the PA polymerase subunit of influenza virus, we were able to identify a series of 9 amino acid positions that are appropriate to engineer temperature-sensitive (ts) mutants. This is the first time that a large number of ts mutations were engineered in such a short domain, demonstrating that rational design of ts mutants can be achieved. We were able to associate this phenotype with a defect of transport of the PA-PB1 complex into the nucleus. Reversion substitutions restored the ability of the complex to move to the nucleus. Two of these ts mutants were shown to be attenuated and able to produce antibodies in mice. These results are of high interest for the design of novel attenuated vaccines and to develop new antiviral drugs.

FIG 1

The polypeptides encoded by IAV segment 3. (A) PA and PA-X ORF structures, showing full-length PA in frame 0 and the X-ORF in frame 1, with their codon numbers indicated. PA-X results from a ribosomal frameshift (21). (B) Amino acid sequence alignment of influenza virus A, B, and C PA/P3 linker domains. Conserved residues are indicated in bold. (C, left panel) Ribbon diagram of the PA linker (green) interacting with PB1. Note the presence of three PA helices interacting with PB1. Residues from the PB1 nuclear localization signals are colored magenta. (Right panel) Ribbon diagram of the PA linker indicating the positions at which substitutions inducing a ts phenotype (red) or an absence of rescue (black) were found.

FIG 2

Secondary structure of the PA linker. The far-UV CD spectrum was recorded in 20 mM sodium acetate, pH 6.

FIG 3

Virus recovery and phenotypes of mutants generated in the PA linker domain. (A) The amino acid sequence of the PA linker domain and the conserved residues among influenza A, B, and C viruses are indicated in single-letter codes. Positions that were chosen for the analysis are indicated by squares. Amino acids were replaced with a proline (or an alanine for the P220 and P221 positions), and mutants were recovered via reverse genetics. Substitutions that were found to confer a ts phenotype are labeled in green. White squares indicate efficient virus rescue without the ts phenotype. Black squares indicate no detectable rescued virus. (B) Results of the plaque assay used to characterize temperature sensitivity and for selection of mutants.

FIG 4

H1N1pdm09 D216P virus phenotype. Serial dilutions (10³ to 10⁵) of the wild type of the D216P mutant H1N1pdm09 virus were used to infect MDCK monolayers at the indicated temperatures, and a plaque assay was performed. The D216P mutant exhibited a ts phenotype.

FIG 5

Kinetics of replication of the mutant viruses at different temperatures (33°C, 37°C, and 39.5°C) upon seeding on MDCK cells at a multiplicity of infection of 0.01. Titers were determined via plaque assays on MDCK cells.

FIG 6

Transcription/replication activity of the polymerase complex with wild type or PA mutants. Plasmids expressing NP, PB1, PB2, and wild type or PA mutants were cotransfected in 293T cells together with the reporter plasmid WSN-NA-firefly luciferase, allowing the quantification of the polymerase activity at different temperatures (33°C, 37°C, and 39.5°C). A plasmid harboring the β-galactosidase (pRSV-β-Gal) gene was cotransfected to control DNA uptake. Luciferase activity was measured in cell lysates 48 h posttransfection. Data are expressed as the mean luciferase activity ± the standard error of the mean of three replicates normalized to β-galactosidase activity. (Upper panel) Amino acids replaced with a proline; (lower panel) amino acids replaced with an alanine.

FIG 7

Subcellular localization of the influenza virus polymerase PB1 subunit transiently expressed in 293T cells with PA mutants. (A) Subcellular localization of a PB1-GFP fusion protein when coexpressed with wild type or L214P PA at 33°C and 39.5°C. (B) Percentage of PB1-GFP-expressing cells with nuclear versus cytoplasmic localization of PB1-GFP. PB1-GFP was expressed alone or with different forms of PA: wild type, mutant forms L214P, D219P, and F223P, or revertant forms P216S and P223H (see Fig. 8 for additional data on PA revertant forms). The means and standard deviations from three experiments are shown, in each of which a mean of 200 cells were scored for each condition.

FIG 8

Normalized luminescence ratios as determined in a split Gaussia luciferase-based complementation assay for the indicated protein complexes and at the indicated temperature. (A) Wild-type or mutant PA proteins fused to Gluc1 were coexpressed with PB1-Gluc2. (B) Wild-type or mutant PA proteins fused to Gluc1 were coexpressed with PB1 and with an IPO5-Gluc2 fusion protein. Luciferase activity was measured in cell lysates 24 h posttransfection. Data are expressed as the mean luciferase activity ± the standard error of the mean of triplicates.

FIG 9

Kinetics of replication of the mutant viruses PA T210P, PB1 R287M, and PA T210P plus PB1 R287M at different temperatures (33°C, 37°C, and 39.5°C) upon seeding on MDCK cells at a multiplicity of infection of 0.01. Titers were determined by plaque assays on MDCK cells.

FIG 10

Characterization of reverse mutations. (A) Kinetics of replication of the mutant viruses D216P and F223P and their revertants P216S and P223H at different temperatures (33°C, 37°C, and 39.5°C) upon seeding on MDCK cells at a multiplicity of infection of 0.01. Titers were determined by plaque assays on MDCK cells. (B) Subcellular localization of the PB1-GFP fusion protein coexpressed with PA reverse mutants D216S and F223H at permissive (33°C) and nonpermissive (39.5°C) temperatures.

FIG 11

Morbidity induced by WSN and the two temperature-sensitive D216P and L219P mutants. Mice (n = 10) received intranasally 10-fold dilutions of viruses (from 10⁶ to 10² PFU/mouse) and were weighed daily. Mice with a weight loss of 25% were considered dead and euthanized. The LD₅₀ was estimated to be 3.5 (on a log PFU scale) for the wt, >6 for L219P, and 5.5 for D216P.

FIG 12

Anti-influenza virus antibody titers induced by infection with PA mutant viruses. The antibody titers were measured 2 weeks postinfection via conventional ELISAs.