Activation of non-canonical NF-κB enriches for chromatin-modifying enzymes and regulatory factors.

(A) Schematic highlighting the non-canonical NF-κB pathway and its role in HIV latency reversal. The antagonist of inhibitor of apoptosis proteins (IAPs), AZD5582 – a known HIV latent reversing agent (LRA) – is shown in red. Factors containing question marks surrounding RelB/p52 are used to emphasize that our understanding of how AZD5582 mechanistically functions to activate HIV transcription is not known. (B) Schematic of the doxycycline (dox)-inducible RelB-BirA fusion and the experimental approach of the proximity biotinylation method. (C) Western blot analysis showing the exogenous expression of the RelB-BirA in respect to endogenous RelB. The integrated RelB-BirA fusion, although dox inducible, was observed to express the fusion in the absence of dox in normal FBS-containing medium to levels comparable to endogenous RelB (also see Supplementary Figure S1A). (D) Volcano plot of AZD5582-activated versus untreated control proteomics comparison, with AZD5582-activated hits (p-value > 0.05 and log2 fold change < 0.5) highlighted. The x-axis shows enrichment (Log2FC) of proteins in AZD5582-activated condition compared to the control condition and the y-axis shows -log10(p-value) from students t-test. As shown in the plot, RelB is found in both activated and untreated samples (bottom left), whereas p52 is only identified in AZD5582-activated conditions (far right). Histone-modifying enzymes and chaperones not previously known to associate with RelB/p52 are highlighted. (E) Gene Ontology (GO) analysis of the proteins that were not significantly different between the AZD5582-activated and untreated conditions (log2 fold change -1 < x > 1) reveal enrichment of enrichment of immune response, chromatin, and nucleosome related terms. (F) Gene Ontology (GO) analysis of the proteins that were enriched in the AZD5582-activated condition compared to control (p-value > 0.05 and log2 fold change < 0.5; highlighted proteins in panel (D) reveal enrichment of chromatin and nucleosome related terms.

CRISPR screen identifies chromatin regulators that enhance or impair HIV reactivation downstream of AZD5582.

(A) CRISPR knockout screen employiog a guide library of factors identified in the Bio-ID screen (see Figure 1). Upon reactivation with AZD5582, guides are packaged inside of virions and the ratio of the guides detected in the supernatant vs in the cell can be used to calculate if a guide is depleted or enriched shown as the -Log MAGeCK gene score. Gene knockouts that are enriched are predicted to improve reactivation with AZD5582, and gene knockouts that are depleted are predicted to prevent reactivation with AZD5582. Bolded genes were highlighted in Figure 1. (B-C) The top 20 depleted (B) or enriched (C) genes in the AZD5582 screen. Bolded genes correspond to those highlighted previously in A. Red or green bars highlight genes chosen for validation. (D-E) 24 hour pre-treatment with (D) KAT7 inhibitor WM-3835 inhibitors or (E) p300 inhibitor A-485, followed by 100 nM AZD5582 or an equivalent volume of DMSO in combination with previous treatment for an additional 24 hours. Inhibitor concentrations for KAT7 inhibitor WM-3835: 0, 10, 100, 1000, 10000 nM. Inhibitor concentrations for p300 inhibitor A-485: 0, 100, 1000, 10000, 25000 nM. Statistical significance for one-way ANOVA comparing 0 nM inhibitor concentration to all other concentrations in a respective dose curve (i.e. AZD5582 or DMSO) are indicated as follows: *P < 0.05; **P < 0.001; ***P < 0.0001.

Comparative HIV-CRISPR screening distinguishes canonical and noncanonical NF-kB-associated regulators of HIV reactivation.

The HIV-CRISPR screen shown in Figure 2 was compared with a parallel screen performed using the same guide RNA library and TNF-α stimulation to activate canonical NF-kB signaling. This comparison was used to distinguish genes preferentially affecting AZD5582-mediated ncNF-κB activation from those more broadly affecting NF-κB-dependent HIV expression. (A-B) Comparison of depleted (A) or enriched (B) genes in the AZD5582 and TNF-α screens. Select genes of interest are highlighted. In (A), genes above the diagonal show stronger depletion in the AZD5582 screen, genes along the diagonal show similar depletion in both screens, and genes below the diagonal show stronger depletion in the TNF-α screen. In (B), genes above the diagonal show stronger enrichment in the AZD5582 screen, genes along the diagonal show similar enrichment in both screens, and genes below the diagonal show stronger enrichment in the TNF-α screen. (C–D) J-Lat 10.6 cells were pre-treated for 24 hours with increasing concentrations of the KAT7/HBO1 inhibitor WM-3835 (C) or the p300 inhibitor A-485 (D), followed by treatment with 10 ng/mL TNF-α or an equivalent volume of DMSO in combination with the same inhibitor concentration for an additional 24 hours. HIV reactivation was measured by flow cytometry as the percentage of GFP-positive cells. WM-3835 did not decrease TNF-α-induced HIV reactivation and produced a slight increase at higher concentrations (C). A-485 increased TNF-α-induced HIV reactivation, although this effect was less pronounced than observed with AZD5582 stimulation. WM-3835 was tested at 0, 10, 100, 1,000, and 10,000 nM. A-485 was tested at 0, 100, 1,000, 10,000, and 25,000 nM. Statistical significance was determined by one-way ANOVA comparing each inhibitor concentration with the 0 nM condition within the respective dose curve. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.