BRD9 is a druggable component of interferon‐stimulated gene expression and antiviral activity

Abstract Interferon (IFN) induction of IFN‐stimulated genes (ISGs) creates a formidable protective antiviral state. However, loss of appropriate control mechanisms can result in constitutive pathogenic ISG upregulation. Here, we used genome‐scale loss‐of‐function screening to establish genes critical for IFN‐induced transcription, identifying all expected members of the JAK‐STAT signaling pathway and a previously unappreciated epigenetic reader, bromodomain‐containing protein 9 (BRD9), the defining subunit of non‐canonical BAF (ncBAF) chromatin‐remodeling complexes. Genetic knockout or small‐molecule‐mediated degradation of BRD9 limits IFN‐induced expression of a subset of ISGs in multiple cell types and prevents IFN from exerting full antiviral activity against several RNA and DNA viruses, including influenza virus, human immunodeficiency virus (HIV1), and herpes simplex virus (HSV1). Mechanistically, BRD9 acts at the level of transcription, and its IFN‐triggered proximal association with the ISG transcriptional activator, STAT2, suggests a functional localization at selected ISG promoters. Furthermore, BRD9 relies on its intact acetyl‐binding bromodomain and unique ncBAF scaffolding interaction with GLTSCR1/1L to promote IFN action. Given its druggability, BRD9 is an attractive target for dampening ISG expression under certain autoinflammatory conditions.

. Independent validation of BRD9 as a hit in the genome-scale screen for factors important for interferon-stimulated gene expression.
A A549/pr(ISRE).eGFP.A1 cells were transduced for at least 10 days with lentiviruses expressing Cas9 and individual sgRNAs derived from the GeCKOv2 CRISPR-Cas9 library targeting BRD9, STAT1, or IFNLR1. eGFP levels following 16 h of IFN-a2 treatment (1,000 IU/ml), or mock, were determined by flow cytometry. MFI = mean fluorescence intensity. Data represent means and standard deviations from n = 3 biological experiments (individual data points shown). Statistical significance was determined relative to the parental cells stimulated with IFN-a2 using single-tailed ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001; n.s. not significant). B Schematic representations of the known components of the canonical BRG1-or BRM-associated factors (BAF) complex, the Polybromo-containing BAF (PBAF) complex, and the non-canonical BAF (ncBAF) complex. Factors reported to be unique to each complex are indicated with colored shapes. BRD9 is unique to the ncBAF complex. C A549-2D8 cells were transfected for 32 h with the indicated siRNA SMARTpools, or siSCR control, prior to treatment with 0, 10, or 100 IU/ml of IFN-a2 for 16 h. Cells were then infected with VSV-GFP at an MOI of 1 PFU/cell and total integrated green fluorescent intensities were collected using the Incucyte live-cell analysis system over the course of 24 h. Area under the curve (AUC) values for GFP-Intensity during the 10-24 h period post-infection were determined. Statistical significance was determined by 1-way ANOVA on log-transformed values comparing the siSCR + 100 IU/ml IFN-a2 condition to each of the other + 100 IU/ml IFN-a2 conditions (****P < 0.0001; n.s. not significant). Dotted line is a visual guide for minimum virus replication in siSCR cells in the presence of 100 IU/ml IFN-a2. Numbers above IFN-a2-treated bars indicate their approximate difference to the respective untreated conditions. Data represent means and standard deviations from n = 3 biological experiments (individual data points shown).
A B D C E F G Figure EV2. Data generated using an independent knockout cell clone to validate BRD9 as important for interferon-stimulated antiviral activity.
A An A549-derived BRD9-KO cell clone (KO#2) was generated using a crRNA targeting exon 1 of BRD9. The target sequence of the crRNA (termed crRNA2), and the resulting 1nt homozygous genomic insertion determined by NGS, are shown in comparison to an unedited control clone (CTRL#2). The generated insertion leads to a premature termination (PMT) codon in the following exon. Encoded amino-acids are shown below the CTRL nucleotide sequence. B Western blot analysis of lysates from CTRL#2 or BRD9-KO#2 cells. BRD9 and b-actin were detected with specific antibodies. Data are representative of at least two biological replicates. C Western blot analysis of CTRL#2 or BRD9-KO#2 lysates from cells treated for 16 h with 1,000 IU/ ml of IFN-a2. The indicated proteins were detected with specific antibodies. Data are representative of at least two biological replicates. D CTRL#2 or BRD9-KO#2 cells were treated, or not, with 1,000 IU/ml of IFN-a2 for 16 h prior to infection with IAV (WSN/33) at an MOI of 0.01 PFU/cell. Viral titers were determined after 24 h by plaque assay. E CTRL#2 or BRD9-KO#2 cells were stably transduced with BRD9-expressing, or control (EV, empty vector), lentiviruses and treated with a range of IFN-a2 concentrations (0, 10, 100, 1,000 IU/ml) for 16 h prior to lysis and analysis for the indicated proteins by Western blot. Data are representative of at least two biological replicates. F CTRL#2 or BRD9-KO#2 cells were stably transduced with BRD9-expressing, or control (EV, empty vector), lentiviruses and treated with 1,000 IU/ml of IFN-a2 for 16 h prior to infection with IAV (WSN/33) at an MOI of 0.01 PFU/cell. Viral titers were determined after 24 h by plaque assay. G RT-qPCR analysis of NFKBIA levels in CTRL#3 or BRD9-KO#3 cells following treatment, or not, with 10 ng/ml of TNF-a for 2 h. GAPDH transcript levels were used for normalization. Data represent means and standard deviations of fold expression changes (relative to CTRL#3 without TNF-a treatment) from n = 3 biological experiments (individual data points shown). Statistical significance was determined by oneway ANOVA on DC t values (n.s. not significant).
Data information: For (D) and (F), data represent means and standard deviations from n = 3 biological experiments (individual data points shown). Statistical significance was determined by 1-way ANOVA on log-transformed plaque counts (*P < 0.05; **P < 0.01; ***P < 0.001). Dotted lines are a visual guide for maximum and minimum virus replication in control cells in the absence and presence of IFN-a2, respectively. Numbers above IFN-a2-treated bars indicate their approximate difference to the respective untreated conditions. Source data are available online for this figure. A B C D Figure EV3. Targeted degradation of BRD9 is non-cytotoxic and reveals its cell type-independent contribution to interferon-stimulated antiviral activity.
A A549 cells were treated for 22 h with either DMSO, 125 nM dBRD9-A, or 1 µg/ml cycloheximide (CHX). CellTiterGlo was used to determine cell viability relative to untreated cells. Data represent means and standard deviations from n = 3 biological experiments (individual data points shown). B-D Calu-3 (B), U87MG (C), or 3T3 (D) cells were treated with 125 nM dBRD9-A or DMSO for 6 h prior stimulation with 1,000 IU/ml of IFN-a2 (B-C) or 400 IU/ml of universal type I IFN (D) for 16 h. Cells were then infected with VSV-GFP at an MOI of 0.6 PFU/cell and total integrated green fluorescent intensities were determined using the Incucyte live-cell analysis system at 24 h post-infection (B-C). Data represent means and standard deviations from n = 3 biological experiments (individual data points shown). Statistical significance was determined by 1-way ANOVA on log-transformed intensity values (*P < 0.05). For (D), cells were infected with IAV (WSN/33) at an MOI of 0.001 PFU/cell. Viral titers were determined after 52 h by plaque assay. Data represent means and standard deviations from n = 3 biological experiments (individual data points shown). Statistical significance was determined by 1-way ANOVA on log-transformed plaque counts (**P-value < 0.01). Numbers above IFN-treated bars indicate their approximate difference to the respective untreated conditions. A B C Figure EV4. Cell-type specific interferon-stimulated gene subsets are affected by dBRD9-A pretreatment.

A B
A An A549-2D8-derived BRD9-KO cell clone (KO#3) was generated using a crRNA targeting exon 1 of BRD9. The target sequence of the crRNA (termed crRNA2), and the resulting 8nt and 2nt heterozygous genomic deletions determined by NGS, is shown in comparison to an unedited control clone (CTRL#3), also derived from A549-2D8. The generated deletions lead to premature termination codons indicated with an asterisk (*). Encoded amino acids are shown below the CTRL nucleotide sequence. B Western blot analysis of lysates from CTRL#3 or BRD9-KO#3 cells. BRD9 and b-actin were detected with specific antibodies. Data are representative of at least two biological replicates.
Source data are available online for this figure.