Serine/arginine-rich splicing factor 7 promotes the type I interferon response by activating Irf7 transcription

Abstract

Tight regulation of macrophage immune gene expression is required to fight infection without risking harmful inflammation. The contribution of RNA-binding proteins (RBPs) to shaping the macrophage response to pathogens remains poorly understood. Transcriptomic analysis reveals that a member of the serine/arginine-rich (SR) family of mRNA processing factors, SRSF7, is required for optimal expression of a cohort of interferon-stimulated genes in macrophages. Using genetic and biochemical assays, we discover that in addition to its canonical role in regulating alternative splicing, SRSF7 drives transcription of interferon regulatory transcription factor 7 (IRF7) to promote antiviral immunity. At the Irf7 promoter, SRSF7 maximizes STAT1 transcription factor binding and RNA polymerase II elongation via cooperation with the H4K20me1 histone methyltransferase KMT5a (SET8). These studies define a role for an SR protein in activating transcription and reveal an RBP-chromatin network that orchestrates macrophage antiviral gene expression.

Keywords: CP: Cell biology; CP: Molecular biology; RNA binding proteins; antiviral immunity; chromatin; co-transcriptional splicing; histone methylation; innate immunity; pre-mRNA splicing.

Figure 1.. Macrophage activation state does not influence SRSF7 control of alternative splicing (AS)

(A) RT-qPCR of Srsf7 in Srsf7 KD RAW MΦs relative to SCR controls. (B) Immunoblot of SRSF7 in SCR, shSrsf7 #1, and shSrsf7 #2 RAW MΦs. Right, quantification. (C) Local splicing variations (AS) in Srsf7 RAW MΦs compared to SCR in untreated and +Salmonella (4 h). ΔPSI >0.1; confidence threshold set at 0.90. (D) Genes containing one or more AS event in untreated and Salmonella-infected conditions. (E) Differentially expressed genes (DEGs) (±1.5 fold change) vs. AS genes in untreated and Salmonella-infected conditions. (F) Ingenuity Pathway Analysis of pathways enriched for AS genes in Srsf7 KD RAW MΦs in untreated and Salmonella-infected conditions. (G) MAJIQ PSI quantification and VOILA visualization of Dcun1d2 junctions in SCR (left) and Srsf7 KD (right) in untreated and Salmonella-infected MΦs. (H) Semi-quantitative RT-PCR of Dcun1d2 in SCR, Srsf7 KD #1, and KD #2 untreated and LPS-treated RAW MΦs. Each sample in biological triplicate. Quantification on right. (I) As in (G) but Mphosph9 in Salmonella-infected MΦs. (J) As in (H) but for Mphosph9. For all data, n = 3 ± SEM. Statistical significance was determined using two tailed unpaired Student’s t test. In all figures, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Figure 2.. SRSF7 promotes the type I interferon response

(A) Scatterplot of genes differentially expressed in Srsf7 KD #2 RAW MΦs relative to SCR in each of the conditions queried (x axis = untreated; y axis = +Salmonella). Select interferon-stimulated genes (ISGs) depicted in black. (B) Ingenuity Pathway Analysis of pathways enriched for DEGs in Srsf7 KD #2 RAW MΦs in untreated and Salmonella-infected conditions. (C) Integrated Genomics Viewer track of SCR (black) and Srsf7 KD #2 (teal) reads at Zbp1 at 4 h post-Salmonella infection. (D) As in (C) but for Ifit3. (E) RT-qPCR of Zbp1, Ifit3, and Ifnb1 in SCR and Srsf7 KD RAW MΦs at 4 h post-LPS treatment (100 ng/mL). (F) RT-qPCR of Ifit3 at 4 h post-ISD transfection (1 μg/mL). (G) RT-qPCR of Ifit3 at 4 and 8 h post-Mycobacterium tuberculosis (Mtb) infection (Erdman strain, MOI = 5). (H) Immunoblot of ZBP1 protein levels in SCR and Srsf7 KD #2 RAW MΦs at 0, 2, 4, and 8 h post-LPS. Right, quantification (both bands/protein isoforms are included). (I) RT-qPCR of Srsf7 and Zbp1 in primary MEFs at baseline or 4 h post-LPS. (J) RT-qPCR of Vsvg and Vsvm in SCR and Srsf7 KD RAW MΦs at 4, 8, and 12 h post-infection with VSV (MOI = 0.1). (K) RT-qPCR of Zbp1 and Ifit3 in SCR and Srsf7 KD RAW MΦs at 1, 4, 8, and 12 h post-infection with VSV (MOI = 0.1). For all data, n = 3 ± SEM. Statistical significance was determined using two tailed unpaired Student’s t test.

Figure 3.. SRSF7 is required for IRF7 expression

(A) Secreted IFN-α/β in SCR and Srsf7 KD #2 RAW MΦs, expressed as relative light units generated by ISRE reporter cells over a time course of LPS treament. (B) RT-qPCR of Zbp1 in SCR and Srsf7 KD RAW MΦs 4 h post-IFN-β treatment (200 IU). (C) Immunoblot of ZBP1 and ACTIN over a time course of IFN-β treatment in SCR and Srsf7 KD #2 RAW MΦs. Quantification, right. (D) Immunoblot of pY701 STAT1, total STAT1, pS396 IRF3, total IRF3, and ACTIN in SCR and Srsf7 KD RAW MΦs over a time course of LPS treatment. (E) Differential expression of genes encoding innate transcription factor in untreated SCR vs. Srsf7 KD RAW MΦs. (F) As in (E) but for +Salmonella samples. (G) Integrated Genomics Viewer with Irf7 reads from SCR (black) and Srsf7 KD #2 (teal) RAW MΦs, untreated (top) and +Salmonella (bottom). (H) RT-qPCR of Irf7 over a time course of LPS stimulation in SCR and Srsf7 KD RAW MΦs. (I) Immunoblot of IRF7 and ACTIN protein levels over a time course of LPS stimulation in SCR and Srsf7 KD RAW MΦs. Quantification, right. (J) RT-qPCR of Irf7 over a time course of Mtb infection in SCR and Srsf7 KD #2 MΦs. (K) As in (J) but during VSV infection. (L) RT-qPCR of Irf7 in siNC or siSrsf7 MEFs 4 h post-LPS. (M) Immunoblot of IRF7 protein levels in GFP gRNA, Irf7 KO1, and Irf7 KO2 RAW MΦs at baseline. ACTIN loading control. (N) RT-qPCR of Zbp1 and Ifnb1 at baseline in Irf7 KO1 (blue) and GFP gRNA control (black) RAW MΦs. For all data, n = 3 ± SEM. Statistical significance was determined using two tailed unpaired Student’s t test.

Figure 4.. SRSF7 overexpression is sufficient to drive basal expression of IRF7 and ISGs

(A) Immunoblot of 3×FLAG-GFP or 3×FLAG-SRSF7 doxycycline (DOX)-inducible RAW MΦs +DOX for time indicated (1 μg/mL). ACTIN shown as loading control. (B) RT-qPCR of Ifit3 and Irf7 in DOX-treated 3×FLAG-GFP and 3×FLAG-SRSF7 RAW MΦs. (C) Immunoblot of IRF7 in DOX-treated 3×FLAG-GFP and 3×FLAG-SRSF7 RAW MΦs. Quantification on the right. (D) As in (B) but for Xaf, Irak2, and Isg20. (E) RT-qPCR of Irf7 and Ifit3 at 0, 4, and 8 h post-infection with Mtb (Erdman; MOI = 5). For all data, n = 3 ± SEM. Statistical significance was determined using two tailed unpaired Student’s t test.

Figure 5.. SRSF7 recruitment to the Irf7 promoter maximizes STAT1 association and promotes RNA polymerase II elongation

(A) Diagram of the Irf7, Ifit3, and Zbp1 genes with tiling primers designed around the transcriptional start site (TSS). (B) pSer2 RNA polymerase II ChIP-qPCR at the Irf7 promoter in SCR and Srsf7 KD RAW MΦs 4 h post-LPS stimulation. All ChIP data represented as elution/input normalized to that of an intergenic control primer set and upstream primer set (−754). (C) As in (B) but at Ifit3. (D) pY701 STAT1 ChIP-qPCR at the Irf7 promoter in SCR or Srsf7 KD RAW MΦs 4 h post-LPS. (E) As in (D) but at Zbp1. (F) pY701 STAT1 ChIP-qPCR at the Irf7 promoter in 3×FLAG-GFP and 3×FLAG-SRSF7 DOX-inducible RAW MΦs. (G) Anti-FLAG ChIP-qPCR of 3×FLAG-GFP and 3×FLAG-SRSF7 in DOX-inducible RAW MΦs. (H) 3×FLAG ChIP-qPCR of 3×FLAG-GFP, 3×FLAG-SRSF7, and 3×FLAG-SRSF7 +RNase (lysate treated with 20 μg/mL RNase prior to IP) in DOX-inducible RAW MΦs 4 h post-LPS. (I) As in (H) but at Ifit3. (J) As in (H) but at Zbp1. (K) CLIP-RT-qPCR of 3×FLAG-GFP and 3×FLAG-SRSF7 along Irf7 exon 1, intron 1, and exon 2. Data represented as elution/input. (L) Semi-quantitative RT-PCR of Irf7 introns 1, 2, and 3 in SCR, Srsf7 KD #1, and KD #2 RAW MΦs in untreated and LPS-treated samples. For all data, n = 3 ± SEM. Statistical significance was determined using two tailed unpaired Student’s t test.

Figure 6.. SRSF7 interacts with the H4K20me1 histone methyltransferase KMT5a/SET8 to promote Irf7 transcription

(A) Proteins with peptides that immunoprecipitated with 3×FLAG-SRSF7, but not 3×FLAG-GFP, -SRSF1, -SRSF3, -SRSF6, or -SRSF9, manually assigned to different functional groups. (B) IP of 3×FL-SRSF7 and 3×FL-GFP probed for endogenous KMT5a/SET8. (C) IP of 3×FL-SRSF7 and 3×FL-GFP probed for endogenous histone H4 or H4K20me1, ±RNase treatment of lysates prior to pull-down. (D) RT-qPCR and immunoblot of siNC- and siKmt5a-transfected RAW MΦs. (E) RT-qPCR of Irf7 and Ifit3b in siNC and siKmt5a RAW MΦs. (F) RT-qPCR of Kmt5a and Irf7 in siNC- and siKmt5a-transfected BMDMs. (G) As in (F) but at 4 h post-LPS (100 ng/mL). (H) pY701 STAT1 ChIP-qPCR at the Irf7 promoter in siNC and siKmt5a RAW MΦs. (I) H4K20me1 ChIP-qPCR in RAW MΦs, untreated and treated with 4 h LPS at Irf7. Data are represented as elution/input normalized to IgG and upstream primer set (−754). For all data, n = 3 ± SEM. Statistical significance was determined using two tailed unpaired Student’s t test.