Omics Analyses Uncover Host Networks Defining Virus-Permissive and -Hostile Cellular States

Fig. 1. Host factors and processes underlying different permissiveness between HEK293 and HEK293T cells. A, infection fitness of HIV-1Gag-mCherry (left) and SINVmCherry (right) in HEK293 and HEK293T cells. Fluorescence signals were measured every 15 min in culture conditions using a plate reader (top, n = 3, multiplicity of infection (MOI) = 1). Western blottings were performed using HIV-1Nef-mCherry (left) and SINVmCherry (right) with antibodies of viral capsid proteins (bottom, MOI = 1). B, the result of quantitative proteomic comparison between HEK293 and HEK293T. Bar plot summarised numbers of proteins with FDR <0.1 and FDR <0.01 in each cell line (top). Volcano plot showed log2 fold change and significance (p-value) of each protein between HEK293 and HEK293T cells (bottom, supplemental Table S1). C, differentially enriched Reactome pathways in Gene set enrichment analysis (GSEA). p-values of enrichments were ∗ labelled (∗∗p < 0.01, ∗∗∗p < 0.001). D, a schematic of the innate immune response against the virus with key processes labeled in grey boxes. Differentially enriched genes in each process were labeled red (HEK293-enriched) or blue (HEK293T-enriched). E, enrichments of viral…

Fig. 2. Proteome landscape of HEK293 cells after IFN-α stimulation. A, schematic workflow of proteome and phosphoproteome profiling of IFN-α stimulated HEK293 cells. SILAC samples were mixed with label swap between replicates. B, Volcano plot with text label showing proteins with FDR <0.1 after 20 h IFN-α treatment (supplemental Table S2). Proteins related to IFN-α were drawn in different shapes depending on the source of annotations – triangle (Δ): annotated in microarray (MA) studies in database Interferome; square (□): annotated in GO-BP terms; plus (+): annotated in both MA studies and GO-BP terms; circle (○): no IFN-α related annotation in either source. C, scatter plot comparing log2 fold changes quantified in IFN-α/mock proteome and HEK293/HEK293T proteome for each protein. D, parallel coordinate plots showing protein expression patterns across cellular states. Proteins with FDR <0.1 in both IFN-α/mock and HEK293/HEK293T proteomes were color-labeled. E, as in D but showing proteins in pathways of interest. F, Western blotting with antibodies of STAT1-pTyr701 (pSTAT1) and viral capsid proteins. Cells were treated with 200 U/ml IFN-α for indicated time period, then infected at an…

Fig. 3. Transcriptome analysis of HEK293 cells after IFN-α stimulation. A, MA plots of IFN-α induced transcriptome changes at 4 hpt (left) and 20 hpt (right) in HEK293 cells. Gene expression was measured by CPM (counts per million reads), referring to the number of reads mapped to transcript scaled by the number of sequencing reads (supplemental Table S3). B, scatter plots comparing log2 fold changes in transcriptome and proteome after IFN-α stimulation. Only genes quantified in both datasets were plotted. C, Venn diagrams comparing numbers of IFN-α induced significant changes at 20 hpt at RNA and protein levels. D, Vector plots visualizing temporal expression profiles after IFN-α stimulation at RNA and protein levels.

Fig. 4. Correlation analysis with characterized antiviral functions. A, Proportions of GO-BP annotated ISGs among genes regulated by IFN-α in HEK293 cells at RNA/protein levels. B, proportions of antiviral ISGs verified by FACS screening studies (summarised in Table 1) among genes regulated by IFN-α in HEK293 cells at RNA/protein levels. C, proportions of broad-spectrum antiviral ISGs among genes regulated by IFN-α in HEK293 cells. ISG activities in FACS screening results were grouped by the number of viruses they can inhibit: no inhibition, inhibit 1 virus (virus-specific), and inhibit ≥2 viruses (broad-spectra). D, dot plots of FACS screening results that are summarised in Table 1. Genes with broad-spectrum antiviral activities were color-labeled based on IFN-α response at RNA/protein levels in HEK293 cells. ISGs with Z < −1.5 for each virus were text labeled.

Fig. 5. Phosphoproteome analysis of HEK293 cells after IFN-α stimulation. A, Venn diagrams summarising numbers of phosphosites identified, quantified, and IFN-α regulated in phosphoproteomic result that combined data from two instrument configurations (supplemental Table S4). B, proportions of functional annotations among IFN-α regulated sites (top) and IFN-related GO-BP annotations among phosphoproteins that contain IFN-α regulated phosphosites (bottom). C, proportions of phosphosites that are located inside annotated RNA-binding domains (RBDs; top), or located within 10 amino acids before or after RBDs (bottom). D, Protein-protein interaction network of RNA-binding proteins (RBPs) that contain IFN-α regulated phosphosites. Positions of phosphosites were labeled with small circles adjacent to each RBP. E, enrichments of viral protein interactors and viral RNA interactors among proteins that have IFN-α regulated phosphosites. F, protein-protein interaction network of viral RNA interactors that have IFN-α regulated phosphosites. The virus family and species that each protein interacts with were text-labeled.

Fig. 6. Virus-inhibitory effects against SINV and HIV-1 in stable cell lines expressing candidate genes. A, infection fitness of HIV-1Gag-mCherry in HEK293 inducible stable lines expressing candidate genes with eGFP fusion (top, MOI = 1, n = 3). Fluorescence signals were measured as in Fig 1A. Western blotting analyses were performed with anti-eGFP and anti-p24 (bottom, MOI = 1). B, schematic of fluorescence signal delay analysis that assesses delays in virus gene expression caused by expression of candidate gene (top), and delay analysis results at 42 hpi for all candidates assessed with HIV-1Gag-mCherry (bottom). C, as in A but with SINVmCherry for fluorescence assay and Western blots. D, as in B but with SINVmCherry at 20 hpi. E, Fluorescence signal delay analysis from 4 to 20 hpi for EPSTI1 and UBE2L6 overexpression cells. F, schematic of delay pattern analysis that assesses slope of delays (top), and comparison of delay patterns between 8 and 20 hpi for all candidates assessed with SINVmCherry (bottom).