Fig 1.
Intrinsic and innate expression of transcription factor ELF1.
a.–c. Cells were treated with 500 U/ml IFN–beta, and mRNA (left side) or protein (right side) of ELF1 or control ISG IRF1 measured over time. mRNA was determined by RT–qPCR and normalized relative to housekeeping gene RPS–11. Fold increase over pre–treatment control levels from n = 3 independent replicates; data shown as individual replicates; line represents the mean. Paired t–test of each time point compared to control, *p<0.1, **p<0.01, ***p<0.001. Western blots showing ELF1 or IRF1 at ELF1 peak mRNA time point; protein levels quantified by densitometry, normalized to actin control. a. Peripheral blood mononuclear cells (PBMC) from n = 3 donors, BSA as carrier control; b. human airway epithelial cultures (basolaterally stimulated); c. A549 cells. d. Human airway epithelial cultures were treated basolaterally with 500 U/ml of IFN–beta or infected apically with influenza A virus (IAV) or parainfluenzavirus 3 (PIV3). mRNA of ELF1 and IRF1 determined at 24 or / and 72 h post treatment or infection as described in (a). RT–qPCR data shown as mean of n = 3 biological replicates.
Fig 2.
Impact of ELF1 on influenza A virus growth.
a. A549 were transduced to express transgenes and RFP as control. 48 h post transduction, cells were challenged with a low MOI of influenza A virus, and % of virus–infected (NP–positive) cells determined by high content microscopy after one or multiple replication cycles. b. Mean ± SEM of % influenza A/WSN/1933 virus–infected cells by high content microscopy in A549 expressing ELF1, IFITM3 as early (entry) ISG inhibitor control, or empty vector as negative control (n = 4). 12 hpi (one cycle of replication, left y–axis) or 48 hpi (multi–cycle replication, right y–axis). One–way ANOVA and Dunnett’s multiple comparison test versus “empty”. c. A549 were transduced to express transgenes ELF1 or IFITM3, or empty vector control. Endogenous and overexpressed ELF1 or IFITM3 protein and actin control post transduction by western blot. Influenza A/WSN/1933 virus growth kinetics on transduced cells (n = 3). Extracellular virus titers by plaque assay on MDCK cells. Individual t–tests between empty and ELF1, or IFITM3 and ELF1, *p<0.1. d. Influenza A/WSN/1933 virus growth kinetics on primary normal human bronchial epithelial (NHBE) cells expressing ELF1 or empty vector (n = 3). Extracellular virus titers by plaque assay on MDCK cells. Individual t–tests between empty and ELF1, *p<0.1. e. In vivo morpholino oligomers (MO)–mediated ELF1 knockdown in A549. ELF1, ELF1 5’UTR–targeting MO; NTC, 5–base–pair non–targeting mismatch control. Endogenous ELF1 protein and actin control post MO knockdown prior to infection by western blot. % ELF1 protein normalized to actin and mismatch control. Influenza A/WSN/1933 growth kinetics post MO knockdown, n = 3. Mean ± SEM virus titer by plaque assay on MDCK cells. Individual t–tests between PBS control and ELF1 knockdown, or NTC control and ELF1 knockdown, *p<0.1.
Fig 3.
In vivo relevance of Elf1 during influenza A virus challenge.
a. Schematic of PPMO–mediated in vivo knockdown. PPMOs targeting Elf1, a control mismatch PPMO (NTC), or PBS, were administered to BALB/c mice intranasally at two and one days prior to infection. Mice were then challenged with 40 PFU of influenza A/PR8/1934 virus, and monitored for body weight and survival. Lungs were collected at days 0, 3 and 6 post infection for western blot analyses (b) and virus titer determination (e). c. Mean % body weight ± SEM from PPMO–treated or control mice, n = 15 mice per group. Unpaired two–tailed t–test comparing Elf1 to PBS, *p = 0.1, ** p<0.01,***p<0.001, **** p<0.0001. d. % survival (>25% body weight loss) of PPMO–treated or control mice, n = 5 mice per group. Log–rank Mantel–Cox test, *p<0.1. e. Virus titers in mouse lung homogenates were measured by plaque assay on MDCK cells, from n = 5 mice per group. Mann–Whitney test, *p<0.1, **p = 0.01.
Fig 4.
ELF1 impact on multi–cycle replication of diverse RNA and DNA viruses.
a. ELF1 protein domains and mutational strategy. TF, transcription factor domain; NLS, nuclear localization signal; ETS, E26 transformation–specific domain; R, conserved arginine within DNA binding domain; A, alanine substitution of conserved arginine within DNA binding domain. b. Protein expression of mutant ELF1 by western blot. c. Mean ± SEM of % influenza A/WSN/1933 virus–infected (NP–positive) cells by high content microscopy in A549 expressing ELF1 wild type, ELF1 mutants, IFITM3 or empty vector (n = 3). One–way ANOVA and Dunn’s multiple comparison test versus empty vector. *p<0.1, **p<0.01. d. A549 were transduced to express empty vector as negative control, ISG and transcription factor IRF1 as positive control, ELF1 wild type, or ELF1 R8A, a DNA binding domain mutant. 48 h post transduction, cells were challenged with a low MOI of the indicated viruses and % of infected cells determined by high content microscopy. Mean ± SEM of % virus–infected cells (n = 3) at one replication cycle (left y–axes) or multi–cycle viral replication (right y–axes): e. influenza A/WSN/1933 (H1N1), % NP–positive cells, f. human parainfluenzavirus 3–EGFP, g. yellow fever virus–Venus, h. chikungunya–virus–ZsGreen, i. coxsackievirus–EGFP, j. adenovirus–EGFP, k. herpes simplex virus 1–EGFP, or l. vaccinia virus–EGFP. YFV (g) and CHIKV (h) were assayed in the presence of 0.4 μM Ruxolitinib to suppress JAK–STAT signaling and enable virus spread. One–way ANOVA and Dunn’s multiple comparison test versus empty vector of the respective time point, *p<0.1, **p<0.01, ***p<0.001, ****p<0.0001.
Fig 5.
ELF1 downstream impact on interferon production and signaling.
a. Canonical interferon (IFN) signaling (solid arrows) and proposed second wave of ISG expression (dashed arrows). b. STAT1 protein expression in A549 STAT1–/–or control cells by Western blot. c. Mean ± SEM of % influenza A/WSN/1933 virus–infected (NP–positive) cells by high content microscopy in STAT1–/–A549 expressing ELF1, ISG and transcription factor IRF1 as positive control, or empty vector (n = 3). ANOVA with Dunnett's multiple comparison to empty. ****p<0.0001, ***p<0.001 *p<0.1. d. A549 transduced to express ELF1 wild type (wt), ELF1 loss–of function mutant R8A, IRF1 as positive control, or empty vector control. mRNA of type I and II interferons by RT–qPCR. Data is shown normalized to empty control, as mean ± SEM. ****p<0.0001, **p<0.01. e. Reporter assay for ISRE–driven transcription. 293T encoding firefly luciferase under the control of a promoter carrying the ISRE motif and stably expressing renilla luciferase as control was transfected to express GFP as negative control, MDA5 as positive control, or ELF1, ELF2, ELF3, ELF4 and ELF5, respectively. Data as mean ± SEM from n = 3 independent experiments. f. 293T luciferase reporter cells were treated as in (e), and subsequently stimulated by transfection of polyI:C. Data as mean ± SEM from n = 3 independent experiments.
Fig 6.
a. Schematic of RNA–Seq approach. A549 cells were mock–transduced or transduced to express empty vector, ELF1 wild type (WT), or ELF1 loss–of–function mutant R8A, respectively. At 48 h post transduction, two empty–transduced samples were treated with 500 U/ml of IFN–beta for 6 or 48 h, respectively. RNA was harvested and analyzed by RNA–Seq. All samples were generated from n = 3 independent biological experiments (one ELF1–R8A replicate was excluded from analysis due to technical problems). b. Principal component analysis of 1000 most variable genes across all samples (regularized log transformed counts, after correction for replicate batch effects). c. Heatmap plotting z–scaled expression values for genes differentially expressed (adjusted p–value < 0.05, log2 fold–change >2 or <–2) in either of the interferon–beta contrasts (compared to untreated, empty–vector transduced control) or both of the ELF1 contrasts (ELF1–WT vs ELF1–R8A, and ELF1–WT vs empty vector). Hierarchical clustering (complete method) was performed on Euclidean distance values. Sidebar annotation labels indicate differential expression (up regulation, red /down regulation, blue) for listed contrasts. d. Scatter plots of moderated log2 fold–change values for all expressed genes evaluated in differential expression analyses. Points (genes) are colored according to differential expression (adjusted p–value < 0.05, log2 fold–change >2 or <–2) in the indicated pairwise contrasts. e. Enriched GO terms (Molecular function ontology, adjusted p value < 1.0 x 10−6) and constituent genes in ELF1–WT differentially expressed genes.
Fig 7.
ELF1 genomic occupancy and validation of ELF1 target gene expression.
a. A549 cells were subjected to ChIP–seq using anti–ELF1 antibody. Pie chart visualizes the distribution of ELF1 peaks over promoters, untranslated regions (UTR), introns, exons, transcription termination sites (TTS), and intergenic regions. b. Representative Meta–gene plot of average log2 enrichment of ELF1 ChIP signal (reads per million mapped reads) over input control signal (reads per million mapped reads). Genes were normalized to a length of 5 kb, then the average log2 ChIP signal over input signal was plotted over the meta gene including 3 kb upstream and downstream of the transcription start site (TSS) and transcription termination site (TTS). Blue line shows average log2 ChIP enrichments over positive–strand genes and green line shows enrichment over negative–strand genes. c. Venn diagram depicting the overlap of ELF1 differentially expressed genes (ELF1 WT vs. (ELF1 R8A + empty + mock–transduced), Fig 6, RNA–seq) and genes with at least one ELF1 binding site in regions depicted in panel (a). Number of genes shown in brackets. d. Representative ChIP–seq track of ELF1–enrichment in the promoter region of one of the 27 genes noted in (c).Y–axis represents reads per million mapped reads spanning a genomic position. Transcription start site (TSS) for PTPRB (transcript variant 2) is highlighted in red and genomic coordinates ± 5 kb around TSS are noted above. e. A549 cells were treated with 50 μM of MOs for 24 h and knockdown efficiency determined by Western blot analysis. Protein levels quantified by densitometry, normalized to actin control and A549 untreated control, shown as mean ± SEM from n = 3 independent experiments. One–way ANOVA and Dunnett’s multiple comparison test versus NTC, **p = 0.023. ELF1, cells treated with ELF1–targeting MO, NTC, cells treated with non–targeting MO control, A549, mock–treated cells. f. A549 cells were treated with 50 μM of MOs for 24 h, and mRNA expression of candidate ELF1 target genes or GAPDH as negative control determined by RT–qPCR normalized to housekeeping gene RPS11.–ddCT (normalized CT value of ELF–targeting MO treated cells–normalized CT value of non–targeting MO treated cells). Data shown as n = 3 independent biological replicates; box–and–whiskers bars show range and mean.