Phosphorylation controls RNA binding and transcription by the influenza virus polymerase

Abstract

The influenza virus polymerase transcribes and replicates the viral genome. The proper timing and balance of polymerase activity is important for successful replication. Genome replication is controlled in part by phosphorylation of NP that regulates assembly of the replication machinery. However, it remains unclear whether phosphorylation directly regulated polymerase activity. Here we identified polymerase phosphosites that control its function. Mutating phosphosites in the catalytic subunit PB1 altered polymerase activity and virus replication. Biochemical analyses revealed phosphorylation events that disrupted global polymerase function by blocking the NTP entry channel or preventing RNA binding. We also identified a regulatory site that split polymerase function by specifically suppressing transcription. These experiments show that host kinases phospho-regulate viral RNA synthesis directly by modulating polymerase activity and indirectly by controlling assembly of replication machinery. Further, they suggest polymerase phosphorylation may bias replication versus transcription at discrete times or locations during the infectious cycle.

Fig 1. PKCδ stimulates polymerase phosphorylation and inhibits polymerase activity.

A. The viral polymerase proteins and the mini-gene vNP77 were expressed in cells with a constitutively active form of PKCδ. The polymerase was immunopurified, mock treated or incubated with calf-intestinal phosphatase (CIP), and analyzed by western blot. Long and short exposures of the same blot are shown. A CIP-sensitive species that may indicate phosphorylated PB2 is indicated. B. Primer extension assays were performed on RNA extracted from cells expressing the viral polymerase, vNP77 and constitutively active PKCδ or an empty vector control. Viral replication (vRNA) and transcription (mRNA) were quantified, normalized to the 5S rRNA internal control, and presented relative to polymerase without PKCδ (mean of n = 3 ± sd; * = Student’s t-test P < 0.05). Representative primer extension data are shown with western blots to confirm protein expression.

Fig 2. The catalytic core of the viral polymerase is phosphorylated during infection.

A. Experimental design to detect phosphorylation of an active polymerase in transfected 293T cells or infected A549 cells. Samples were prepared as whole-cell lysate or immuno-purified proteins prior to phospho-peptide mass spectrometry. Most phospho-sites are conserved among circulating human influenza virus strains and highly pathogenic H5N1 viruses. PB1 sequences were aligned to WSN and the percentage where the indicated residue is a serine, threonine, or tyrosine is shown. Note that influenza B virus (IBV) residue numbering shifts relative to WSN. Whether phosphosites were detected during infection (ix), transfection (tx), or both is indicated. See S1 Table for all identified sites. B-C. Phospho-sites on PB1 surround the catalytic core and template entry. The location of PB1 phospho-sites characterized in this study are modeled in yellow on (B) a pre-initiation complex (PDB 4WSB [44]) or (C) a transcriptional elongation complex (PDB 6T0V [36]). The motif C residues D445/D446 in the catalytic site are in orange, 5’ vRNA is light blue, 3’ vRNA is magenta, and transcription product is dark blue. Residue assignment and numbering based on WSN. PA and PB2 are shown as white in the space-fill representation, but are not shown in the close-up structure to increase clarity. Similarly, the 5’ hook vRNA is not shown in C. See additional structures in S3 Fig.

Fig 3. PB1-mutant viruses identify phosphorylation sites that impact polymerase activity and viral replication.

A. PB1 phosphorylation both inhibits and enhances viral transcription. Single-cycle infections with PB1 phospho-mutant viruses were performed in A549 cells (MOI of 0.5 for 8h). RNA from infected cells was subject to qRT-PCR to detect NP and GAPDH mRNA. Fold changes (ΔΔC_T) were determined in triplicate from 2 independent infections. (± SEM; ** = P <0.01 for one-way ANOVA with Dunnett’s post hoc compared to WT). B. Multicycle replication kinetics of phospho-mutant viruses. A549 cells were infected at an MOI of 0.001. Viral titers were measured 12, 24, 48, 72, 96 hpi via plaque assay on MDCK cells. The WT reference curve applies to both the phospho-ablative and phospho-mimetic mutant viruses as all mutants were analyzed in the same experiment, but graphed separately to aid visualization (mean of n = 3 ± sd). P < 0.01 for one-way ANOVA at each time point. Statistics for ANOVA with Dunnett’s post hoc pair-wise comparisons to WT are in S3 Table.

Fig 4. PB1 phospho-mutants are defective in RNA synthesis and RNP formation.

A. Viral RNA synthesis was analyzed in primer extensions assays. RNA extracted from 293T cells expressing the viral polymerase, NP, and segment 6 vRNA was subject to primer extension analysis to detect transcription (mRNA) and replication (cRNA, vRNA) products. Primer extension of 5S rRNA was used as an internal loading control. PB1a, catalytically-dead PB1; PB1 K669A/R670A, transcription-deficient PB1. B. PB1 phospho-mutants form polymerase trimers. FLAG-tagged PB1, PB2, and PA were expressed in 293T cells and cell lysates were subject to PB1-FLAG immunoprecipitation. Immunoprecipitates and input samples were probed for PB1-FLAG, PB2, and PA. C. PB1 phospho-mimetics deficient in RNA synthesis fail to generate productive RNPs. NP immunoprecipitations were performed on 293T lysates generated as in (A). Immunoprecipitates and input samples were western blotted for PB1, PA, and NP.

Fig 5. Phosphorylation at T223 inhibits cRNA stabilization and vRNA binding.

A. PB1 phospho-mutants were tested in a cRNA stabilization assay. WT or mutant PB1 and oligomerization-deficient NP (NP_E339A) were expressed in 293T cells. Cells were treated with actinomyocin D (ActD) prior to infection, RNA extraction and primer extension analysis to detect transcription (mRNA) and replication (cRNA, vRNA) products. A representative primer extension gel is shown. cRNA levels were quantified, normalized to WT, and expressed as mean ± sd. * < 0.05, ** < 0.01, *** < 0.001 = P for one-way ANOVA with Dunnett’s post hoc compared to WT. B. PB1 T223D fails to precipitate vRNA in an RNA-IP (RIP). PB1-FLAG, PB2, PA, and segment 6 vRNA were expressed in 293T cells. Cells were lysed and subject to FLAG immunoprecipitation. RNA extracted from immunoprecipitates and input samples was probed for the presence of segment 6 vRNA via primer extension analysis. Immunoprecipitated PB1 was confirmed via western blot. The ratio of radiolabel quantified for RIP and input is shown at the bottom.