Fig 2: Subnetwork Exploration Uncovers Positive and Negative Regulators of IFN and ISGs(A) Jupyter widget subnetwork view (see STAR Methods) of predicted upstream regulators of IFNB1 and IFNL1. Arrows indicate transcriptional regulatory events between TFs and targets and do not specify positive or negative effects on target expression. Node size denotes the relative number of edges. Nodes are color coded by network cluster (see Figure 4).(B) Heatmap of gene expression for the indicated TFs during treatment with LPS, poly(I:C), HIV-1-GFP (+Vpx), and LKO-GFP (+Vpx) at 2, 8, 24, and 48 h.(C) Immunoblots of MDDC lysates after treatment with the indicated shRNAs.(D) Flow cytometry of CD86 versus GFP expression in shRNA-modified MDDCs 48 h after mock treatment (+Vpx) or infection with HIV-1-GFP (+Vpx).(E–G) Plots showing %GFP+ (E), %CD86+ (F), and %SIGLEC1+ (G) from MDDCs treated as in (D).(H–I) ELISA of IFNL1 (H) and IP-10 (I) expression in supernatants from shRNA-modified MDDCs that were mock treated or infected with HIV-1-GFP for 48 h.For (E)–(I), plots represent pooled data from 5–8 donors. **p < 0.01, ***p < 0.001, ****p < 0.0001.See also Figure S5.

Fig 3: RARA Is a Negative Regulator of MDDC Maturation(A) Jupyter widget subnetwork view (see STAR Methods) showing edges between RARA, its predicted targets, STAT2, and the remaining top 30 TFs ranked by hypergeometric p value. Arrows point from TFs to targets but do not suggest positive or negative activity.(B) STRING database connections for nodes shown in (A), including NCOR1 and NCOR2. Brackets indicate groups and biological pathways. Nodes connected in the EN-ATAC network with no direct STRING connections are shown to the left. In (A) and (B), nodes are color coded by network cluster.(C) Flow cytometry plots of MDDCs modified by shRNA that were mock infected (+Vpx) or challenged with HIV-GFP (+Vpx) at day 4 for 48 h. Plots show CD86 versus GFP expression and represent 1 of 4 donors (MOI = 0.5, 1.5, and 5).(D) qPCR validation of target knockdown in MDDCs modified by the indicated shRNAs.(E) qPCR of ISG15, CXCL10, and GFP expression in shRNA-modified MDDCs that were either mock treated or infected with HIV-1-GFP (MOI = 5) for 24 h.(F) Flow cytometry of MDDCs treated with vehicle or the RARA agonist Am80 (10 µM) ± infection with HIV-1-GFP for 48 h.(G) Pooled data of CD86 MFI in MDDCs from 4 donors treated as in (F). Am80 concentrations = 10 nM, 100 nM, 1 µM, and 10 µM.(H) Flow cytometry of MDDCs treated with vehicle or the RARA antagonist BMS 195614 (10 µM) ± infection with HIV-1-GFP for 48 h.(I) Pooled data of CD86 MFI in MDDCs from 4 donors treated as in (H). BMS 195614 concentrations = 1 nM, 100 nM, and 10 µM.For (D), (E), (G), and (I), n = 4 donors. *p < 0.05, **p < 0.01, ***p < 0.001. Data represent mean ± SEM. See also Figure S7.

Fig 4: Transient Changes in Chromatin Accessibility Correspond to Activation of the Innate Response during HIV Infection(A) Illustration depicting ATAC-seq sample processing.(B) Flow cytometry plots of MDDCs sorted (red boxes) for ATAC-seq after infection with HIV-1-GFP (+Vpx) at 2, 8, 24, and 48 h (MOI = 5). Plots show CD86 versus GFP expression and are representative data from 1 of 3 donors.(C) Smooth scatter density plots of HIV-1-GFP-infected MDDCs from sorted populations as in (B), showing the genome-wide relationship between transcript levels (RNA-seq average log2 fold change) TSS chromatin accessibility (ATAC-seq average log2 fold change). Blue, density of points across the entire genome; red, hallmark IFN response genes as in Figure 1C. The positions of IFNB1, IFNL1, and select ISGs are labeled for each condition.(D) ATAC-seq gene-associated peak values shown as counts per million (CPM) reads for the indicated time points and conditions. Lines connect samples from independent donors. Statistics were calculated on log2-transformed data using a 2-way ANOVA with multiple comparisons. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.See also Figure S2 and Table S3.