Extended Data Fig. 4: ATPase-competent cBAF promotes ACE2 gene expression.

(a) Venn diagrams depicting overlap of BAF subunit (SMARCA4, SMARCC1, ARID1A) or H3K27ac histone mark occupancy (CUT&Tag) and ATAC-Seq peaks in rescued SMARCA4 KO Vero E6 cells. (b) Metaplots of BAF subunit occupancy (SMARCA4, SMARCC1, ARID1A), H3K27ac histone mark, as measured by CUT&Tag, and ATAC-Seq peaks at all sites in rescued SMARCA4 KO Vero E6 cells. (c) PCA plots of merged SMARCC1 occupancy (CUT&Tag) (right) and ATAC-Seq (left) in rescued SMARCA4 KO Vero E6 cells (d) Scatter plots of ATAC-Seq replicates in SMARCA4 KO Vero E6 cells expressing Empty control, SMARCA4 WT and SMARCA4 K785R. The correlation coefficient ‘r’ is indicated (top left). (e) Metaplots of H3K4me1 and H3K27ac histone mark occupancy (CUT&Tag) at SMARCA4 WT-dependent sites (Cluster 3 from Fig. 2b). (f) Cumulative distribution function plots reflecting genes nearest to SMARCA4 and ARID1A gained sites in SMARCA4 KO cells rescued with SMARCA4 WT versus empty vector in Cluster 3 from Fig. 2b; top 1/10th or 1/20th fraction reflect genes associated with top-changed sites, sites highlighted in red indicate genes that scored as pro-viral determinants in CRISPR screen. (g) CUT&Tag and ATAC-Seq tracks at SLC4A4 and HNF1A enhancers in rescued SMARCA4 KO Vero E6 cells. (h) Correlation between BAF subunit occupancies as profiled by CUT&Tag. (i) Venn diagram of overlap between ATPase-competent BAF-dependent transcripts (adj. p-value <0.01, calculated using EdgeR quasi-likelihood negative binomial generalized log-linear model, followed by Benjamini-Hochberg corrections) and proviral genes in the SARS-CoV-2 CRISPR screen (Z-score > 2). (j) RPKM levels of ACE2, HNF1A, and SLC4A4 in Vero E6 cells across indicated conditions (n = 2 biological replicates). (k) Heatmap of top differentially expressed genes (DEGs) in Vero E6 WT and SMARCA4 KO cells rescued with empty vector, SMARCA4 WT or SMARCA4 K785R. Red rectangle highlights ACE2.

Source data