CSNK2B modulates IRF1 binding to functional DNA elements and promotes basal and agonist-induced antiviral signaling

Abstract

Interferon regulatory factor 1 (IRF1) is a critical component of cell-intrinsic innate immunity that regulates both constitutive and induced antiviral defenses. Due to its short half-life, IRF1 function is generally considered to be regulated by its synthesis. However, how IRF1 activity is controlled post-translationally has remained poorly characterized. Here, we employed a proteomics approach to identify proteins interacting with IRF1, and found that CSNK2B, a regulatory subunit of casein kinase 2, interacts directly with IRF1 and constitutively modulates its transcriptional activity. Genome-wide CUT&RUN analysis of IRF1 binding loci revealed that CSNK2B acts generally to enhance the binding of IRF1 to chromatin, thereby enhancing transcription of key antiviral genes, such as PLAAT4 (also known as RARRES3/RIG1/TIG3). On the other hand, depleting CSNK2B triggered abnormal accumulation of IRF1 at AFAP1 loci, thereby down-regulating transcription of AFAP1, revealing contrary effects of CSNK2B on IRF1 binding at different loci. AFAP1 encodes an actin crosslinking factor that mediates Src activation. Importantly, CSNK2B was also found to mediate phosphorylation-dependent activation of AFAP1-Src signaling and exert suppressive effects against flaviviruses, including dengue virus. These findings reveal a previously unappreciated mode of IRF1 regulation and identify important effector genes mediating multiple cellular functions governed by CSNK2B and IRF1.

Graphical Abstract

CSNK2B interacts directly with IRF1 and modulates the binding of IRF1 to chromatin, thereby promoting expression of antiviral genes.

Figure 1.

IRF1-associated cellular proteins identified by MS and functional validation of CSNK2B as an IRF1 cofactor. (A) Coomassie brilliant blue staining of affinity-purified FLAG-tagged IRF1 proteins on SDS-PAGE gels. The gel was imaged in the 700 nm channel on the Odyssey Infrared Imaging System. (B) Gene ontology analysis and functional annotation of IRF1-associated proteins identified by MS. (C) Validation of cellular proteins affecting the endogenous PLAAT4 transcript level by siRNA knockdown experiments. RNA levels were determined by RT-qPCR assays of RNA extracted from PH5CH8 cells transfected with siRNA pools targeting indicated genes for 72 h. *P < 0.05, **P < 0.01, ***P < 0.0001 versus control (line at mean, n = 3, two-tailed Student's t-test). (D) PH5CH8 cells were transfected with selected siRNA pools and infected 48 h later with hepatitis A virus (HAV) at an m.o.i. of 10. Relative abundance of HAV RNA was determined 4 day post-infection (d.p.i.) by RT-qPCR (upper panel, n = 4). Knockdown efficiency of target genes was determined 6 days post-siRNA transfection (lower panel, n = 3). *P < 0.05, **P < 0.01 versus control (one-way ANOVA with Dunnett's multiple comparisons test). (E) HAV/NLuc (18f/NLuc) infectious titers released from PH5CH8 cells transfected with CSNK2B siRNA or control as in (D) were determined 5 d.p.i. *P < 0.05 (n = 3, two-tailed Student's t-test). (F) HAV RNA levels at 4 d.p.i. in PH5CH8 cells expressing CSNK2B versus control sgRNAs. **P < 0.01, ***P < 0.0001 versus control (n = 3, one-way ANOVA with Dunnett's multiple comparisons test). (G) Schematic representation of secretory Nanoluciferase (NLuc) reporter analysis of 4 × IRF1-NLuc stably transfected in PH5CH8 cells. (H) NLuc activities in cells transfected with IRF1, CSNK2B and control siRNAs. The promoter activity was determined 3 days post-transfection. ***P < 0.0001 versus control (n = 3, one-way ANOVA with Dunnett's multiple comparisons test). (I) NLuc activities and PLAAT4 mRNA levels in PH5CH8 cells expressing CSNK2B versus control sgRNAs. The promoter activity was determined 3 days post-transduction. *P < 0.05, **P < 0.01, ***P < 0.0001 versus control (one-way ANOVA with Dunnett's multiple comparisons test). (J) Stability of IRF1 protein in PCH5CH8 cells transfected with CSNK2B siRNA or control after treatment with 50 μg/ml puromycin. Data were fit to a one-phase decay model (n = 3, R² = 0.9588–0.9832).

Figure 2.

CSNK2B interacts with IRF1 and regulates its antiviral activity independently of phosphorylation. (A) 293FT cell lysates expressing IRF1-FLAG or empty vector was immunoprecipitated with anti-FLAG M2 antibody. Proteins eluted from the precipitates were subjected to an SDS-PAGE followed by western blotting with specific antibodies against CK2 components. (B) Pull-down analysis showing direct interaction between purified IRF1-FLAG and recombinant human CSNK2B (rCSNK2B) proteins. (C) Phos-tag SDS-PAGE of FLAG-tagged IRF1. PH5CH8 cell lysates were subjected to immunoblotting before (−) and after (+) digestion with lambda protein phosphatase (λ PPase). An arrowhead shows phospho-IRF1. (D, E) Phos-tag gel analysis of ectopically expressed IRF1-FLAG in 293FT cells (D) or endogenously expressed IRF1 in PH5CH8 cells (E). Means ± S.D. of values for abundance of phospho-IRF1 relative to non-phosphorylated IRF1 are shown on right (n = 3). (F) PH5CH8 cells were transfected with siRNAs targeting CK components and infected 48 h later with HAV at an m.o.i. of 10. Percentage of HAV RNA levels relative to non-target control siRNA was determined 4 d.p.i. by RT-qPCR. Immunoblots showing depletion of catalytic CK2 subunits are shown on right. **P < 0.01 versus control (n = 3 or 4, one-way ANOVA with Dunnett's multiple comparisons test). (G) Immunoblots showing CSNK2B protein abundance in Huh-7.5 cells treated with 10 μM CX-4945 for 24 h. (H) Effects of CX-4945 on replication of HAV/NLuc (18f/NLuc) in Huh-7.5 cells and cell viability. **P < 0.01, ***P < 0.0001 versus control (n = 3, one-way ANOVA with Bonferroni's multiple comparisons test). (I) Effects of CX-4945 on HAV replication and PLAAT4 expression in IRF1-depleted and control Huh-7.5 cells. **P < 0.01, ***P < 0.0001 (n = 3, two-way ANOVA with Sidak's multiple comparisons test). (J) NLuc reporter analysis of Huh-7.5 cells expressing NLuc reporter (pNL-4×IRF1). Cells were transfected with indicated siRNAs for 48 h, and relative NLuc values secreted at 48–72 h post-transfection are shown. Immunoblots showing depletion of each siRNA target are shown on right. *P < 0.05, **P < 0.01 versus control (n = 4, one-way ANOVA with Dunnett's multiple comparisons test).

Figure 3.

CSNK2B promotes IRF1-dependent expression of PLAAT4 and restricts HAV RNA replication. (A) Scheme showing the CSNK2B/IRF1-mediated restriction of HAV replication via PLAAT4. (B) Relative HAV RNA levels at 5 d.p.i. in PH5CH8 cells transfected with siRNA targeting either CSNK2B or PLAAT4 alone or the combination of 2 siRNAs (bottom). NS, not significant. **P < 0.01 (n = 4, two-tailed Student's t-test). Knockdown efficiency of CSNK2B and PLAAT4 is shown on right. ***P < 0.0001 versus control (n = 3, one-way ANOVA with Dunnett's multiple comparisons test). (C) A549 cells transfected with CSNK2B versus control siRNAs were challenged with HAV/NLuc (18f/NLuc). NLuc activities at indicated time points post-infection are shown. Light units (LU) of noninfected lysates was 28.25 ± 9.83 (mean ± S.D.). Immunoblots showing depletion of CSNK2B are shown on right. ***P < 0.0001 (n = 3, two-way ANOVA with Sidak's multiple comparisons test). (D) Firefly luciferase (FLuc) activity produced from transfected subgenomic HAV-Luc (sgHAV-Luc) RNA or its replication-incompetent mutant (Δ3D) in Huh-7.5 cells depleted of CSNK2B versus control. Immunoblots showing depletion of CSNK2B are shown on right. L.O.D., limit of detection showing LU of non-transfected lysates. ***P < 0.0001 (n = 3, two-way ANOVA with Sidak's multiple comparisons test).

Figure 4.

CSNK2B regulates PLAAT4 expression and down-regulates HAV replication in primary human hepatocytes. (A) Phase contrast microscopy of primary human hepatocytes. Scale bar, 20 μm. (B) Effect of CSNK2B siRNA on CSNK2B mRNA levels at indicated time points after siRNA transfection. **P < 0.01, ***P < 0.0001 versus control (n = 3, two-tailed Student's t-test). (C) Effect of CSNK2B siRNA on PLAAT4 mRNA levels at 7 days after transfection. *P < 0.05 versus control (n = 6, two-tailed Student's t-test). (D) Replication of HAV RNA (p16-18fSP and 18f strains that express NLuc) and infectious virus titers at 6 days p.i. in primary human hepatocytes transfected with indicated siRNAs. GEq, genome equivalents. ***P < 0.0001 versus control (n = 3, two-tailed Student's t-test).

Figure 5.

CSNK2B regulates agonist-induced suppression of HAV via PLAAT4. (A) RT-qPCR determination of PLAAT4 mRNA levels in A549 cells treated with various doses of all-trans retinoic acid (ATRA) for 24 h (n = 2). (B) PLAAT4 mRNA levels in A549 cells transfected with CSNK2B versus control siRNAs for 72 h, followed by treatment with IRF1 agonists, all-trans retinoic acid (ATRA, 3 μM) or IFNγ (1000 U/ml) for 24 h. *P < 0.05, **P < 0.01 (n = 3, two-tailed Student's t-test). (C) Dose-dependent effects of IRF1 agonists on HAV/NLuc (18f/NLuc) in A549 cells. Drugs were added 4 h post-infection with HAV/NLuc at an m.o.i. of 1 and NLuc activity determined 72 h later. (D) Effects of depletion of CSNK2B, IRF1 and PLAAT4 on suppression of HAV/NLuc replication by IRF1 agonists, ATRA (1 μM) and IFNγ (1000 U/ml). LU, light units. *P < 0.05, **P < 0.01 versus control (n = 3, one-way ANOVA with Dunnett's multiple comparisons test). (E) Validation of the NLuc results by quantitation of HAV RNA levels using RT-qPCR. GEq, genome equivalents. *P < 0.05 versus control (n = 3, one-way ANOVA with Dunnett's multiple comparisons test). (F) Immunoblots of CSNK2B, IRF1, PLAAT4 and GAPDH as loading control in lysates of A549 cells transfected with indicated siRNAs. (G) Effects of depletion of either CSNK2B or PLAAT4, or both on suppression of HAV/NLuc replication by an IRF1 agonist, IFNγ (1000 U/ml). Knockdown efficiencies of each gene are shown on right. **P < 0.01, ***P < 0.0001 versus control (one-way ANOVA with Dunnett's multiple comparisons test).

Figure 6.

Genome-wide mapping of CSNK2B-regulated IRF1 binding sites by CUT&RUN. (A) In vitro pull-down experiments using biotinylated DNA probe containing IRF1 binding motifs derived from the PLAAT4 promoter in lysates of CSNK2B-depleted PH5CH8 cells versus control (top). Means ± S.D. of relative values for abundance of DNA-bound IRF1 in CSNK2B-depleted cell lysates versus control (bottom). **P < 0.01 (n = 3, two-tailed Student's t-test). (B) Stacked bar plots showing peak annotation of IRF1-bound sites in PH5CH8 cells transfected with CSNK2B siRNA versus control. The number of peaks is shown on each of the stacked bar. (C) Volcano plots showing changes in relative abundance of the peaks identified in CUT&RUN between PH5CH8 cells transfected with CSNK2B versus control siRNAs. Each symbol represents the log₂ fold-change in CSNK2B siRNA-transfected cells over control from three biological replicates. Combined P-value for the peak was determined using Sime's method. The P-values were adjusted using false discovery rate (FDR) correction. (D) Gene ontology analysis and functional annotation of genes differentially bound by IRF1 in CSNK2B-depleted cells. (E) Logo depiction of HOMER de novo motif analysis of all peaks (top) and peaks significantly decreased in CSNK2B-depleted cells (bottom). (F) Example genomic regions showing binding peaks of IRF1 in control versus CSNK2B-depleted PH5CH8 cells for all three biological replicates. (G) Bar graphs showing RT-qPCR measurements of transcripts from representative CSNK2B-responsive genes in PH5CH8 (upper panels) and A549 (lower panels) cell lines. Cells were stimulated with IFNγ (100 U/ml, + IFNγ) or vehicle control (− IFNγ) for 24 h. *P < 0.05, **P < 0.01, ***P < 0.0001 (n = 3, two-way ANOVA with Sidak's multiple comparisons test). (H) Heatmaps showing the relative abundance of CSNK2B-regulated gene transcripts linked to antiviral defense and innate immune responses as determined by RT-qPCR, with clustering according to CSNK2B regulation of basal expression.

Figure 7.

CSNK2B regulates transcription and phosphorylation of AFAP1 that lowers permissiveness to flavivirus replication. (A) Immunoblots of AFAP1 and GAPDH as loading control in lysates of PH5CH8 cells transfected with indicated siRNAs, with and without IFNγ stimulation. (B) Immunoblots of AFAP1, IRF1 and GAPDH in lysates of PH5CH8 cells depleted of either (or both) CSNK2B and IRF1. Quantitation of AFAP1 protein abundance is shown on the right. **P < 0.01, ***P < 0.0001 (n = 3, one-way ANOVA with Dunnett's multiple comparisons test). (C) Immunoblots of AFAP1 and GAPDH in lysates of PH5CH8 cells treated with either lambda protein phosphatase (λ PPase, left panels) or indicated concentrations of CX-4945 (right panels). (D) Immunoblots of AFAP1 and Src in lysates of PH5CH8 cells treated with 10 μM CX-4945. (E) Immunoblots of PH5CH8 cell lysates transfected with indicated siRNAs targeting AFAP1 (left panels) or CSNK2B (right panels). (F) Scheme of CSNK2B-regulated AFAP1 signaling cascades that lead to Src activation. (G) Immunoblots of PH5CH8 cell lysates transfected with indicated siRNAs (top). DENV RNA levels were determined at 48 h p.i. (bottom). *P < 0.05 (n = 3, two-tailed Student's t-test). (H) Immunoblots of AFAP1 in lysates of PH5CH8 versus Huh-7.5 cells stably transduced with AFAP1 (top). Huh-7.5 cells stably expressing AFAP1 or vector control were challenged with DENV/NLuc (bottom). NLuc activities at the indicated time points post-infection are shown. **P < 0.01 (n = 3, two-way ANOVA with Sidak's multiple comparisons test). (I) PH5CH8 cells were transfected with indicated siRNAs and infected 48 h later with dengue virus (DENV) at an m.o.i. of 0.1. CX-4945 (5 μM) was added 2 h post-infection. Infectious titers were determined 48 h p.i. by focus formation assays. FFU, focus forming units. **P < 0.01, ***P < 0.0001 versus control (n = 3, one-way ANOVA with Dunnett's multiple comparisons test). (J) PH5CH8 cells were infected with DENV or Zika virus (ZIKV) as in (I) and viral RNA levels determined by RT-qPCR. **P < 0.01 (n = 3, two-tailed Student's t-test).