Fig 2: PAF1 restricts DENV2 infectious virion production.(A) Immunoblot analysis of PAF1 expression in PAF1 KO and rescue A549 cells. GAPDH is a loading control. (B) DENV2 replication in PAF1 KO and rescue cells, MOI 0.1. Data from three independent biological replicates are plotted as mean values +/- standard deviation. P values were calculated using a paired, one-tailed Student’s t-test. Abbreviations: plaque forming units (pfu), not statistically significant (ns).

Fig 3: NS5 mutants rescue expression of PAF1-dependent genes.Relative gene expression was plotted as log2 fold change for (A) PAF1 KO versus parental A549 cells or (B) STAT2 KO versus parental A549 (same as Fig 2E), and NS5 mutant versus WT. Unsupervised K-means clustering was also performed to identify genes with similar behavior (triangles, circles and squares). P values were adjusted for false discovery rate using the Benjamini Hochberg method. Significant changes in gene expression are plotted for: PAF1 KO (cyan), both (yellow), neither (grey), NS5LGS (orange), NS5GTR (red) and NS52xNLS (purple). Immune response genes (GO:0006955) are highlight with larger markers and opaque coloring.

Fig 4: PAF1 is required for expression of a subset of genes activated by poly(I:C).(A) Parental A549, PAF1 KO/rescue and STAT2 KO/rescue cells were stimulated with poly(I:C) for 3 hours and subjected to RNA-seq and DESeq2 differential gene expression analysis. Results are based on three independent biological replicates. (B) Principal component analysis performed on all samples showed a clear separation between principal components (PC) describing untreated/treated cells and cell genotype. (C) GSEA was performed on genes differentially expressed in KO cells compared to parental A549 cells following poly(I:C) treatment. Up to the top 5 positively and negatively enriched Reactome pathways were plotted for each comparison (p adj < 0.1). A full list of GSEA results is available in S2 Table. (D) Changes in gene expression caused by poly(I:C) treatment are shown for the subset of immune response genes (GO:0006955) significantly upregulated for poly(I:C)-treated parental A549 cells relative to mock-treated A549 cells (log2 fold change > 0.5, padj < 0.05). A Wilcoxon signed rank test with Bonferroni correction was performed to identify significant changes caused by PAF1 or STAT2 KO. (E) All genes significantly upregulated for poly(I:C)-treated parental A549 cells relative to mock-treated A549 cells (log2 fold change > 0.5, padj < 0.05) were plotted based on log2 fold change of PAF1 and STAT2 KO cells relative to parental A549 cells following poly(I:C) treatment. Parental A549 cells were used as a normalization so that it was identical for both comparisons. Unsupervised K-means clustering was also performed to identify genes with similar behavior (triangles, circles and squares). P values were adjusted for false discovery rate using the Benjamini Hochberg method. Significant changes in gene expression are plotted for PAF1 KO (cyan), STAT2 KO (magenta), both (yellow) or neither (grey). Immune response genes (GO:0006955) are highlighted with larger markers and opaque coloring.

Fig 5: Nuclear localization is required for NS5-PAF1C interaction.(A) NLS mutants, including single mutant NS5central (REE396-398AAA) and NS5C-term (RR890-891AA) and the double mutant NS52xNLS (REE396-398AAA and RR890-891AA). (B) Subcellular localization of 2xStrep II tagged NS5s (yellow) and colocalization with PAF1 (cyan) was determined by immunostaining and confocal microscopy. Nuclei were stained with Hoechst (magenta). Scale bar represents 10 µm. (C) The protein interaction between NS52xNLS and PAF1 was tested biochemically. Plasmids encoding NS5s and GFP were transfected and affinity purified in HEK293T cells via a 2xStrep II tag. Immunoblot analysis of lysate and purified (Strep-AP) fractions was performed against PAF1, Strep and GAPDH (loading/negative control). Abbreviations: Strep-affinity purified (Strep-AP).