Infection-specific phosphorylation of glutamyl-prolyl tRNA synthetase induces antiviral immunity

Abstract

The mammalian cytoplasmic multi-tRNA synthetase complex (MSC) is a depot system that regulates non-translational cellular functions. Here we found that the MSC component glutamyl-prolyl-tRNA synthetase (EPRS) switched its function following viral infection and exhibited potent antiviral activity. Infection-specific phosphorylation of EPRS at Ser990 induced its dissociation from the MSC, after which it was guided to the antiviral signaling pathway, where it interacted with PCBP2, a negative regulator of mitochondrial antiviral signaling protein (MAVS) that is critical for antiviral immunity. This interaction blocked PCBP2-mediated ubiquitination of MAVS and ultimately suppressed viral replication. EPRS-haploid (Eprs^+/-) mice showed enhanced viremia and inflammation and delayed viral clearance. This stimulus-inducible activation of MAVS by EPRS suggests an unexpected role for the MSC as a regulator of immune responses to viral infection.

Figure 1. EPRS induces antiviral immune responses to RNA viruses

(a) Expression of genes encoding MSC components (right margin) in bronchial epithelial cells (two replicates; one per column) at various times after infection with PR8 (above columns), showing genes upregulated (yellow) or downregulated (blue) over 1.5-fold relative to their expression before infection (key). (b) Luciferase activity of 293T cells transfected with the control TK-Renilla plasmid and a firefly luciferase reporter plasmid containing the IFNB promoter, plus empty vector (EV) or vector encoding various MSC components (horizontal axis), together with plasmid encoding the amino-terminal domain of RIG-I (N-RIG-I); results are presented relative to those of the renilla luciferase control. Below, immunoblot analysis of Strep-tagged MSC proteins and FLAG-tagged N-RIG-I in total lysates of the cells above. (c,d) Viral replication in RAW264.7 cells transfected with EPRS-specific siRNA (siEPRS) or control (non-targeting) siRNA (siCtrl) (key) and infected with PR8-GFP (multiplicity of infection (MOI) = 1) (top row) or VSV-GFP (MOI = 0.5) (bottom row), assessed by fluorescence microscopy (c) and fluorescence absorbance and plaque assay (d) at 24 h after infection. PFU, plaque-forming units. (e,f) Concentration of IFN-β (e) or IL-6 (f) in supernatants of RAW264.7 cells transfected with siRNA as in c,d (key) and infected with PR8-GFP, VSV-GFP or HSV-GFP (MOI = 1) or treated with poly(I:C) (80 µg) (above plots). (g) Immunoblot analysis of phosphorylated (p-) and total (inactive) IRF3 and STAT1, and of EPRS and actin (loading control), in PR8-GFP-infected RAW264.7 cells expressing EPRS-specific (shEPRS) or non-targeting control (shCtrl) short hairpin RNA. (h–j) Fluorescence microscopy (h), fluorescence absorbance and plaque assay (i), and secretion of IFN-β or IL-6 (j) of RAW264.7 cells transfected with empty vector (Ctrl) or vector encoding EPRS and infected with PR8-GFP. Scale bars (c,h), 100 µm. *P < 0.05, **P < 0.01 and ***P < 0.001 (Student’s t-test). Data are representative of one experiment (a) or three experiments with similar results (b–j), with at least three (b–f,h–j) or two (g) independent biological replicates (mean and s.d. of triplicates in b,d,e,f,i,j).

Figure 2. EPRS is critical for antiviral defense against RNA viruses in mouse BMDMs

(a) Plaque assay of viral titers in Eprs^+/+ and Eprs^+/− BMDMs infected with VSV-GFP (MOI = 5) (top) or PR8-GFP (MOI = 3) (bottom). (b,c) Concentration of IFN-β and IL-6 in culture supernatants of cell as as in a after viral infection as in a (b) or treatment with poly(I:C) (40 µg) (c). (d) Expression of Ifnb, Il6 and other genes encoding interferon-related antiviral products (vertical axes) in Eprs^+/+ and Eprs^+/− mouse-derived BMDMs at 12 h after infection with VSV-GFP. *P < 0.05, **P < 0.01 and ***P < 0.001 (Student’s t-test). Data are representative three experiments with similar results, with three (a–c) or two (d) independent biological replicates (mean and s.d. of triplicates).

Figure 3. EPRS is essential for antiviral immunity in mice

(a) Survival of 6- to 7-week-old female Eprs^+/+ mice (n = 15) and Eprs^+/− mice (n = 14) monitored for 10 d after intravenous injection of VSV Indiana (2 × 10⁸ PFU per mouse). (b) Viral load in brain and spleen tissues of Eprs^+/+ and Eprs^+/− mice (n = 6 per genotype), assessed by plaque assay at day 5 after infection as in a. (c) Viral loads in mice as in b. (n = 5) determined by qPCR of VSV transcripts. (d,e) Plaque assay of viral load (d) and ELISA of IFN-β, IFN-α and IL-6 (e) in serum of Eprs^+/+ and Eprs^+/− mice (n = 8 per genotype) left uninfected (UI) or at 12 h after infection with VSV-GFP (2 × 10⁸ PFU per mouse). (f) Hematoxylin-eosin-stained sections of brain tissue from Eprs^+/+ and Eprs^+/- mice (n = 4 per genotype) left uninfected or on day 5 after mock infection or infection with VSV Indiana (left margin), showing well-preserved neural parenchyma (top and bottom left), or glial nodule formation by reactive microglial cells and mononuclear cells in brain parenchyma (black arrows), perivascular cuffing (arrowhead), and disruption of the ependymal lining (red arrows) at the lateral ventricle (LV), a result of massive infiltration of mononuclear cells (bottom right). (g) Immunohistochemical analysis, with anti-VSV-G, of brain sections from Eprs^+/+ and Eprs^+/− mice (n = 4 per genotype) left uninfected or at days 3 and 5 after infection with VSV Indiana (left margin); nuclei were stained with DAPI; white arrows indicate VSV-positive GFP signals. Scale bars (f,g), 100 µm. *P < 0.05 and **P < 0.01 (log-rank test (a) or Mann-Whitney test (b–e)). Data are representative of one experiment (a) or three experiments with similar results (b–g), with two independent biological replicates (mean and s.e.m. in b–e).

Figure 4. Virus-induced phosphorylation of EPRS induces its release from the MSC

(a) Endogenous immunoprecipitation (IP), with anti-EPRS, followed by immunoblot analysis with anti-KRS and anti-AIMP3 (MSC components), and anti-NSAP1 and anti-GAPDH (GAIT complex components), at various times (above lanes) after infection of U937 cells by PR8-GFP (MOI = 3) or treatment with IFN-γ (500 units/ml). Input (bottom), immunoblot analysis in the cells above, without immunoprecipitation. (b) Confocal microscopy of the colocalization of endogenous EPRS (red) and KRS (green) in cells infected with PR8 (MOI = 5) or treated with IFN-γ (1,000 units/ml). Far right, enlargement of areas outlined at left. Scale bars, 10 µm (left and middle) or 1 µm (far right). (c) Colocalization index of EPRS and KRS in cells as in b. (d) Tandem mass spectrometry of a triply charged EPRS peptide (KDPSKNQGGGLSSSGAGEGQGPK in cells left uninfected (−; left) and of KDPpS*KNQGGGLSSSGAGEGQGPK (*, Ser990-phosphorylation site in cells infected with PR8 (right). (e–g) Immunoblot analysis of EPRS phosphorylated at Ser990 (e), Ser886 (f) or Ser999 (g) in U937 cells infected with PR8-GFP or treated with IFN-γ, for various times (above lanes). (h,i) Immunoblot analysis of 293T cells transfected (above lanes) with empty vector or vector encoding Strep-tagged wild-type EPRS (WT) or various combinations of three phosphomimetic forms of EPRS (S886D, S990D and S999D) (h) or the phosphorylation-resistant EPRS mutant S990A (i), assessed by Strep precipitation (ppt) and detection with anti-KRS, AIMP3 and anti-MRS. Input (bottom), immunoblot analysis without precipitation. *P < 0.01 and **P < 0.001 (Student’s t test). Data are representative of three experiments with similar results, with three independent biological replicates (a,b,e–i) or three experiments (c; mean and s.d.).

Figure 5. EPRS interacts with PCBP2, a negative regulator of MAVS

(a,b) Luciferase assay (as in Fig. 1b) of IFNB promoter activation in 293T cells at 24 h after transfection (above plots) to express N-RIG-I, MDA5, poly(I:C) and MAVS (a) or TRAF3, TBK1 and IRF7 (b), and with increasing concentrations (0, 200 or 800 ng; wedges) of plasmid encoding FLAG-tagged EPRS (horizontal axes). (c) Luciferase assay (as in Fig. 1b) of IFNB promoter activation in Toll-like-receptor-3-expressing 293T cells (293T(TLR3)) transfected for 24 h with increasing concentrations (0, 50, 200 or 800 ng; wedges) of plasmid encoding FLAG-tagged EPRS (horizontal axis), followed by stimulation for 12 h with poly(I:C) (30 µg). (d) Silver staining (top) of Strep-EPRS complexes purified from 293T cells 24 h after transfection with a plasmid encoding Strep-EPRS, followed by infection for 6 h with PR8-GFP (MOI = 5): left margin, size in kilodaltons (kDa); right margin, Strep-tagged full-length EPRS (170 kDa); *, PCBP2 (38 kDa). Below, sequences of peptides identified by mass spectrometry. (e,f) Immunoassay of the interaction between EPRS and PCBP2 in PR8-infected RAW264.7 cells (e) and U937 cells (f), assessed by immunoprecipitation with anti-EPRS, followed by immunoblot analysis with anti-PCBP2 (input, as in Fig. 4a). (g) Confocal microcopy of endogenous EPRS (red) and PCBP2 (green) in HeLa cells at various times (left margin) after infection with PR8 (MOI = 5). Scale bars, 10 µm. NS, not significant (P > 0.05); *P < 0.01 and **P < 0.001 (Student’s t-test). Data are representative of three experiments with similar results, with three independent biological replicates (a–c,e–g; mean and s.d. of triplicates in a–c) or one experiment (d).

Figure 6. Domain mapping required for the interaction between EPRS and PCBP2

(a) Full-length EPRS (top) and constructs of EPRS (below) containing various combinations of the GST-like domain (GST), catalytic domain (CD), tRNA-binding domain (tRNA), WHEP domains 1–3 (W1–W3) and linkers 1–5 (L1–L5), with (+) or without (−) binding to PCBP2 (right margin). (b,d,f) Immunoassay of 293T cells transfected with plasmid encoding GST-tagged PCBP2 and empty vector or plasmids encoding the GFP- or FLAG-tagged EPRS constructs in a (above lanes), assessed by co-immunoprecipitation (with anti-GFP or anti-FLAG) of the EPRS constructs with PCBP2, followed by immunoblot analysis with anti-GST or anti-GFP (b), anti-GST or anti-FLAG (d) or anti-PCBP2 or anti-FLAG (f) (input, as in Fig. 4a). (c,e,g) Luciferase assay (as in Fig. 1b) of IFNB promoter activation in 293T cells transfected with expression plasmids for N-RIG-I and the IFNB promoter, or without N-RIG-I (control (Ctrl)), together with empty vector or plasmids encoding the EPRS constructs in a (horizontal axes). (h) Full-length PCBP2 (top) and constructs containing or lacking (Δ) various combinations of PCBP2 domains (below), with (+) or without (−) binding to EPRS (right margin). (i) Immunoassay of the interaction of EPRS with PCBP2 in lysates of 293T cells expressing Strep-EPRS and various forms of GST-PCBP2, assessed by Strep precipitation followed by immunoblot analysis with anti-GST (input, as in Fig. 4h,i). (j) In vitro precipitation analyzing direct binding between various regions of EPRS (above lanes; amino acids 1–732, 1–196 and 1–168) and the PCBP2 KH1 domain (amino acids 11–82): black arrows, protein fragments derived from EPRS during purification (N-terminal sequences mapped to the MRFDD (amino acids 234–238) and MVTFI (amino acids 565–569) sequences of EPRS); red arrowheads, PCBP2 KH1 domain. *P < 0.001 (Student’s t-test). Data are representative of three experiments with similar results, with three independent biological replicates (mean and s.d. of triplicates in c,e,g).

Figure 7. EPRS blocks PCBP2-mediated ubiquitination and degradation of MAVS

(a) Immunoassay of the interaction of MAVS with PCBP2 in lysates of 293T cells transfected to express FLAG-tagged MAVS and plasmids encoding various GST-tagged constructs of PCBP2 (above lanes), assessed by co-immunoprecipitation with anti-FLAG and immunoblot analysis with anti-GST (input, as in Fig. 4a throughout). (b) In vitro precipitation assay (as in Fig. 6j) of cells as in a: arrowhead, MAVS. (c) GST-precipitation and immunoblot analysis of the interaction of PCBP2 with MAVS (top) or EPRS (bottom) in 293T cells transfected to express various combinations (above lanes) of FLAG-tagged MAVS and GST-tagged PCBP2 plus increasing amounts (wedges) of Strep-tagged EPRS, probed with anti-FLAG or anti-Strep. (d) GST-precipitation assay (as in c) of interactions between PCBP2 and Itch (top) and EPRS (bottom). (e) Immunoassay of the endogenous interactions between PCBP2 and EPRS or MAVS in lysates of PR8-infected RAW264.7 cells, assessed by immunoprecipitation with anti-PCBP2 (for endogenous PCBP2) and immunoblot analysis with anti-EPRS (top) or anti-MAVS (bottom). (f,g) Immunoblot analysis of exogenous MAVS (f) or endogenous MAVS (g) immunoprecipitated from lysates of 293T cells treated with the proteasome inhibitor MG-132 and then transfected to express various combinations (above lanes) of Strep-tagged EPRS, GST-tagged PCBP2, V5-tagged Itch, FLAG-tagged MAVS (f only) and hemagglutinin (HA)-tagged ubiquitin, probed with antibody to Lys48 (K48)-linked ubiquitin (K48-Ub) and other antibodies (right margin). (h,i) Immunoblot analysis (top) of exogenous MAVS (h) or endogenous MAVS (i) in 293T cells transfected to express FLAG-tagged PCBP2, FLAG-tagged MAVS (h only) and increasing amounts of Strep-tagged EPRS. Below, MAVS band intensity, normalized to that of actin (numbers above bars indicate specific intensity). (j) In vitro assay of the ubiquitination of purified Strep-tagged MAVS after incubation with ubiquitin, E1, E2, and a combination of purified Strep-tagged EPRS, Strep-tagged PCBP2 and V5-tagged Itch, assessed by immunoblot analysis with anti-ubiquitin. Data are representative of three experiments with similar results, with three independent biological replicates.

Figure 8. An EPRS-derived L1 peptide has antiviral activity

(a) Immunoassay (top) of 293T cells transfected with various combinations (above lanes) plasmids encoding GST-tagged PCBP2, V5-tagged Itch and hemagglutinin-tagged ubiquitin and increasing amounts (wedges) of Tat-Epep (20, 50 and 100 µM), assessed by immunoprecipitation with anti-MAVS (to precipitate endogenous MAVS), followed by immunoblot analysis with antibody to K48 ubiquitin (input, as in Fig. 4a). Below, sequence of Tat and Tat-Epep. (b) Immunoblot analysis (top) of endogenous MAVS in 293T cells transfected to express FLAG-tagged PCBP2 and increasing amounts of Tat-Epep (as in a). Below, MAVS band intensity (presented as in Fig. 7h,i). (c) Secretion of IFN-β and IL-6 by VSV-infected RAW264.7 cells treated with PBS (negative control) or various concentrations (horizontal axis) of Tat or Tat-Epep. (d) Fluorescence microscopy of RAW264.7 cells infected with VSV-GFP and treated with Tat or Tat-Epep (above images). (e) Viral titers in infected RAW264.7 cells as in c. (f) Immunoblot analysis of MAVS in RAW264.7 cells transfected with non-targeting control siRNA (siCtrl) or MAVS-specific siRNA (siMAVS) and treated as in c. (g–i) Secretion of IFN-β and IL-6 (g) and microscopy (h) viral titers (i) of MAVS-deficient RAW264.7 cells transfected with siRNA as in f and infected and treated as in c. Scale bars (d,h), 125 µm. *P < 0.05, **P < 0.01 and ***P < 0.001 (Student’s t-test). Data are representative of three experiments with similar results, with least three independent biological replicates (mean and s.d. of triplicates in c,e,g,i).