RSV hijacks cellular protein phosphatase 1 to regulate M2-1 phosphorylation and viral transcription

Abstract

Respiratory syncytial virus (RSV) RNA synthesis occurs in cytoplasmic inclusion bodies (IBs) in which all the components of the viral RNA polymerase are concentrated. In this work, we show that RSV P protein recruits the essential RSV transcription factor M2-1 to IBs independently of the phosphorylation state of M2-1. We also show that M2-1 dephosphorylation is achieved by a complex formed between P and the cellular phosphatase PP1. We identified the PP1 binding site of P, which is an RVxF-like motif located nearby and upstream of the M2-1 binding region. NMR confirmed both P-M2-1 and P-PP1 interaction regions in P. When the P-PP1 interaction was disrupted, M2-1 remained phosphorylated and viral transcription was impaired, showing that M2-1 dephosphorylation is required, in a cyclic manner, for efficient viral transcription. IBs contain substructures called inclusion bodies associated granules (IBAGs), where M2-1 and neo-synthesized viral mRNAs concentrate. Disruption of the P-PP1 interaction was correlated with M2-1 exclusion from IBAGs, indicating that only dephosphorylated M2-1 is competent for viral mRNA binding and hence for a previously proposed post-transcriptional function.

Fig 1. Mapping of the M2-1_core interaction regions on P by NMR.

(A) Superimposed ¹H-¹⁵N HSQC spectra of 35 μM ¹⁵N-labeled P, alone (red contours) and in the presence of 1 molar equivalent of M2-1_core (black contours). Residue-specific assignments are represented by red labels for peaks with an intensity decrease by more than 60% on addition of M2-1_core. The bar diagram represents the intensity ratios (I/I₀) measured for each peak. Straight and broken lines represent the mean and mean ± rmsd values over all signals. The position of the oligomerization domain and regions with transient α-helical or β-sheet secondary structure along the sequence of P, previously determined by NMR [44], is indicated in the cartoon on top. The boundaries and maximal propensity of these secondary structures [44] are indicated below. The red background indicates the region around the transient helix α_N2 where M2-1_core induces >60% intensity decrease. The hatched area corresponds to a P region encompassing the OD and C-terminal helices for which peaks cannot be observed due to severe line broadening without M2-1_core. (B) and (C) Superimposed ¹H-¹⁵N HSQC spectra of 35 μM ¹⁵N-labeled P[1–126] and 75 μM P[1–163], alone (cyan and blue contours) and in the presence of 1 molar equivalent of M2-1_core (black contours). Residue-specific assignments are given for all peaks. Insets on the top left give assignments for the crowded central region of the spectra. These spectra were acquired with less resolution in the ⁵N dimension than the spectrum shown in A. Bar diagrams below represent the intensity ratios. The regions around α_N2 with I/I₀ < (mean-rmsd) are highlighted with light cyan and blue backgrounds for P[1–126] and P[1–163], respectively.

Fig 2. Identification of the M2-1 binding domain on P by GST pull-down.

(A) Schematic illustration of the full-length and truncated forms of GST-P used for M2-1 pulldowns in this study. The oligomerization domain of P is represented as a grey box, and numbers indicate amino acid positions. (B) GST-P proteins were purified on glutathione-Sepharose beads and incubated in the presence of M2-1. After extensive washing, the binding of M2-1 was determined by SDS-PAGE and Coomassie blue staining. For each deletion mutant, the ability to interact with M2-1 is summarized on the right in A.

Fig 3. Identification of residues of P critical for RSV polymerase activity.

(A) Polymerase activity assay in the presence of P mutants. BSRT7/5 cells were transfected with the RSV minigenome composed of plasmids encoding the WT N, M2-1, and L proteins, the pMT/Luc minigenome, and WT or mutated P proteins, together with pCMV-βGal for transfection standardization. Viral RNA synthesis was quantified by measuring the Luc activity after cell lysis 24 h post-transfection. Each luciferase minigenome activity value was normalized based on β-galactosidase expression and is the average of results from three independent experiments performed in triplicate. Error bars represent standard deviations. (B) Western blot analysis showing efficient expression of P variant proteins in BSRT7/5 cells. (C) Helical wheel representation of the putative α-helix located between residues S94 and N111 of P (HeliQuest online program). Residues critical for M2-1-binding are indicated by a star and the phosphorylatable residue T108 is indicated by two stars. Positively charged residues are in blue, negatively charged residues in red, putative phosphorylated residues in purple and hydrophobic residues in yellow.

Fig 4. Effects of point mutations targeting the M2-1-binding domain of P on the recruitment of M2-1 to cytoplasmic IBs.

BSRT7/5 cells were transfected with pP (WT and variants), pN, and pM2-1–mCherry plasmids. Cells were fixed 24 h after transfection, labeled with anti-P antibody (green), and colocalization of P and M2-1–mCherry (red) was analyzed by fluorescence microscopy. Nuclei were stained with DAPI. Scale bars, 20 μm.

Fig 5. Effects of P substitutions on P–M2-1 interaction and M2-1 phosphorylation.

(A) GST-P (WT and variants) and M2-1 with a C-terminal 6xHis-tag were expressed in E. coli and purified separately. Beads coated with GST-P (variants and WT), as well as GST-P[1–80] used as a negative control, were saturated with 3% BSA, and incubated alone (-) or in the presence of M2-1 (+), and washed. Complexes were resolved by SDS-PAGE and stained with Coomassie blue. M, molecular weight ladder. (B) Model of a P-M2-1 complex obtained by docking the P D95-F109 helix (green ribbon) onto a M2-1 protomer (grey ribbon, with indication of helix numbers). P residues critical for M2-1 binding are shown in sticks. M2-1 residues critical for P binding are in orange sticks, while residues involved in RNA-binding domain are in red sticks. (C) Analysis of the M2-1–P interaction in mammalian cells by immunoprecipitation. BSRT7/5 cells were transfected with plasmids encoding for N, HA-P (WT and variants) and M2-1. Immunoprecipitations (IP) from cell lysates (L) were performed 24 h post-transfection using an anti-HA antibody. P and M2-1 were revealed by Western blotting using anti-P and anti-M2-1 rabbit antisera. The star indicates the phosphorylated form of M2-1.

Fig 6. Phosphatase PP1 binds to the "RVxF" motif of RSV P.

(A) Sequence alignment of Pneumoviridae P showing the conservation of the "RVxF" motif. Human RSV A strain (HRSV), bovine RSV (BRSV), ovine RSV (ORSV), pneumonia virus of mice (PVM), canine pneumovirus (CPV), swine orthopneumovirus (SOV), human metapneumovirus (HMPV) and avian metapneumovirus (AMPV) (accession codes AAX23990.1, NP_048051.1, Q83956.1, Q5MKM7.1, AHF88957.1, ANO40516.1, YP_012606.1, and AAF05910.1, respectively) P sequences were aligned by Clustal Omega and prepared with ESPript 3. Numbers correspond to RSV P amino acid residues. (B) BSRT7/5 cells were co-transfected with pN and pHA-P or p-HA-P[F87A]. HA-P (WT and variant) was immunoprecipitated from cell lysates and the presence of PP1 in the precipitate was analyzed by Western blot. L, cell lysate; IP, immunoprecipitated products. (C) NMR analysis of GST-PP1α binding to RSV P. Intensity ratios (I/I₀), measured for each peak in ¹H-¹⁵N BEST-TROSY spectra of 25 μM ¹⁵N-labeled P[1–126] or ¹⁵N-labeled full-length P, alone and in the presence of 2 molar equivalents of GST-PP1α, are represented in the bar diagrams. Straight and broken lines indicate mean and mean ± rmsd values. Regions with transient α-helical or β-sheet secondary structure [44] are highlighted by a colored background with the same color code as in Fig 1. The hatched area corresponds to a P region encompassing the OD and C-terminal helices for which peaks cannot be observed in the free form. (D) Proteins GST-PP1, M2-1, P WT and F87A mutant were expressed in E. coli and purified separately. Beads coated with GST or GST-PP1 were saturated with 3% BSA, and incubated with, P WT or F87A mutant and washed. Complexes were pulled down, resolved by SDS-PAGE and stained with Coomassie blue.

Fig 7. P-PP1 interaction allows recruitment of PP1 to IBs and M2-1 localization in IBAGs.

BSRT7/5 cells were transfected with plasmids encoding the N, L and M2-1-mCherry proteins, the M/Luc RSV minigenome, and either wild type (WT) or F87A mutant P-BFP. Tagged proteins were expressed instead of the corresponding wild type as indicated on the pictures. FISH analyses were performed to detect poly(A) RNA (in red). The expressed tagged proteins are visualized thanks to their spontaneous fluorescence. IBs are delimited by the P fluorescence. Arrows point to IBAGs. Representative images from 3 independent experiments are shown. Images were taken under a Leica SP8 confocal microscope. Scale bars 5μm.

Fig 8. Effect of M2-1 on RSV mRNA poly-adenylation either in the presence or absence of PP1 in IBs.

BSRT7/5 cells were transfected with plasmids encoding the N, L and M2-1-mCherry proteins, the M/Luc RSV minigenome, and either WT or F87A mutant P-BFP. FISH analyses were performed to detect poly(A) or NS1 RNAs (in red). The expressed tagged proteins are visualized thanks to their spontaneous fluorescence. IBs are delimited by the P fluorescence. Arrows point to IBAGs. Representative images from 3 independent experiments are shown. Images were taken under a Leica SP8 confocal microscope. Scale bars 5μm.

Fig 9. Effect of PP1 overexpression on RSV RNA polymerase activity and M2-1 phosphorylation.

(A) Inhibition of RSV RNA polymerase activity by PP1 overexpression. BSRT7/5 cells were transfected with the RSV minigenome, and various amounts of pEGFP-PP1 vector. Viral RNA synthesis was quantified by measuring the Luc activity after cell lysis 24 h post-transfection. Each Luc activity value was normalized based on β-galactosidase expression and is the average of results from three independent experiments performed in triplicate. Error bars represent standard deviations. (B) Expression of phosphorylated (*) or unphosphorylated M2-1, N and P in BSR-T7/5 cells transfected with the RSV minigenome and with increasing quantities of pGFP-PP1vector. Cell extracts were resolved by SDS-PAGE 24 h post-transfection and analyzed by Western-blot using rabbit polyclonal anti-M2-1 or anti-P, and mouse monoclonal anti-α -tubulin antibodies. (C) Ratio of phosphorylated (Phos-M2-1) versus unphosphorylated (Unphos-M2-1) M2-1, normalized to levels of α-tubulin, in the presence of increasing amounts of pGFP-PP1vector. Signals were quantified using a Chemidoc Touch Imaging System (Bio-Rad, France).

Fig 10. Model for phosphorylation turnover of M2-1 taking into account the P–M2-1 and P–PP1 interactions.

(A) The primary sequence of the 80–115 region of P is indicated at the top. Amino acid residues previously identified as critical for M2-1 binding [33, 34] are in green, newly identified residues F98 and I106 are in red. Residue F87 critical for M2-1 dephosphorylation and efficient transcription is highlighted in red font on a yellow background. The left-hand column shows amino acid substitutions. Right-hand columns summarize the impact of P mutations on (i) the polymerase activity, (ii) in vitro interaction between P and M2-1 using recombinant proteins, (iii) interaction between P and M2-1 assessed by co-immunoprecipitation from transfected BSRT7/5 cells, and (iv) on the recruitment of M2-1 to IBs; ND, not determined. (B) Diagram of M2-1 phosphorylation turnover. (1) M2-1 is phosphorylated (at Ser 58 and 61), binds to P in the cytosol and is dephosphorylated by PP1 in the P–PP1 complex. (2) The P–PP1–M2-1 complex is directed to IBs (pink) where transcription takes place. (3) M2-1 binds to neo-synthesized mRNAs at the end of transcription and (4) M2-1-mRNAs complexes are concentrated in IBAGs (blue) before being released in the cytosol (5). M2-1 is then phosphorylated by a cellular kinase and parts with viral mRNAs which will be used for translation of viral proteins (6).