Fig 2: TRIM33 regulates Ifnb1 gene expression in activated macrophages.(a) Genomic PCR detection of Trim33 in BMDM and PM (top) and western blot analysis of TRIM33 in BMDM from WT and Trim33−/− mice (bottom). (b) Kinetics of Ifnb1 mRNA levels in WT and Trim33−/− BMDM (left panel) or PM (right panel) after activation with LPS. Mean±s.e.m., n=3 to 6. (c) IFN-β protein levels in supernatants of non-activated and LPS-activated WT and Trim33−/− BMDM. Mean±s.e.m., n=4. (d) Relative mRNA levels of Ifnb1 and IFN-β target genes in Trim33−/− BMDM versus WT BMDM treated with LPS (0.1 ng ml−1) for 3 or 24 h. Mean±s.e.m., n=3. (e) Kinetics of Ifna mRNA levels in WT and Trim33−/− BMDM activated with LPS. Mean±s.e.m., n=3. (f) Kinetics of Ifnb1 mRNA levels in NIH3T3 cells expressing a shRNA targeting Trim33 (NIH3T3 TRIM33−) or luciferase (NIH3T3 TRIM33+) and activated with poly(I:C). Mean±s.e.m., n=3. (g) Kinetics of Ifnb1 mRNA levels in WT and Trim33−/− iM treated for the indicated times with LPS. Mean±s.e.m., n=3. (h) Relative expression of Ifnb1 mRNA 0, 2 and 8 h after LPS treatment of Trim33−/− iM transduced with a lentivirus coding for GFP (iM Trim33−/− +GFP), for full-length flag-TRIM33 and GFP (iM Trim33−/− +flag-TRIM33) or for flag-TRIM33 mutants lacking the coiled-coil domain (iM Trim33−/− +flag-ΔCC), the ubiquitin ligase activity (iM Trim33−/− +flag-ΔUb) or the SMAD-interaction domain (iM Trim33−/− +flag-ΔSMAD). Mean±s.e.m., n=5. All qRT-PCR data are normalized to HPRT and, unless otherwise indicated, presented relative to expression of untreated WT cells. *P<0.05; **P<0.01 and ***P<0.001, Mann–Whitney test.

Fig 3: TRIM33 deficiency is associated with high sensitivity to endotoxin challengeA. Survival curves of Control and Trim33−/− mice after intra-peritoneal injection of LPS (0.25mg/20g). (n=8-9 mice for each group). * p<0.02 (Gehan-Breslow-Wilcoxon test). B. Kinetics of serum concentrations of the indicated cytokines after intra-peritoneal injection of LPS. Mean ± SEM, n=5 mice. C. Fold increase of IL6, MCP1 and TNFα in serum of dying Trim33−/− mice versus Control mice (serum collected just before the death of Trim33−/− mice) (n=3 vs 5).

Fig 4: Dynamics of TRIM33 binding in myeloid cell linesA. Venn diagrams of TRIM33 peaks (left) and associated genes (right) in 32D and RAW cells. The number of cell-specific or co-localized peaks and shared genes are shown for each cell line. B. Most enriched sequence motifs in TRIM33 peaks in 32D and RAW cells. E-values on the significance of the motifs are given. The percentages of non-overlapping sites (p< 0.0001) are given in parentheses. C. (Left) Overlap between TRIM33 peaks in 32D and RAW cells and PU.1 peaks in BMDM (data sets from [18]). (Right) Distribution of genes according to their numbers of associated TRIM33 and PU.1 peaks in 32D and RAW cells. D. Occurrence of the indicated transcription factor binding sites in the different TRIM33/PU.1 peak subsets in 32D (green) and RAW (blue) cells compared to their frequency in random PU.1-bound regions in BMDM (red). Transcription factor matrices were extracted from Homer database and analyzed with RSAT Matrix-scan, using a threshold p-value<1e-4. E. MNase profiles in BMDM (data sets from [19]) around the summit of TRIM33 peaks in 32D (blue) or RAW (green) cells and around the summit of PU.1 peaks in BMDM (red). F. (Left) PU.1 binding sites in macrophages were clustered according to their TRIM33 ChIP-seq profiles in 32D and RAW cells. (Right) Gene Ontology (GO) functional annotation analysis of the three clusters of TRIM33/PU.1 target genes.

Fig 5: Nucleosome dynamics and non-coding RNA production at the − 35 kb regulatory region. a TRIM33, SPT16 and RNA Pol II occupancy along with ChromHMM analysis over the Atp1b3 − 35 kb site. b Nascent bidirectional transcripts at the − 35 kb site in WT and Trim33−/− BMDM, measured after 4sU-tagging and purification of newly transcribed RNA. Data are relative to WT BMDM. Mean ± SEM, n = 3. c Nucleosome positioning at the − 35 kb region in MNase-digested chromatin from WT and Trim33−/− BMDM. Data are relative to the undigested chromatin. Mean ± SEM, n = 3. d ChIP-qPCR of SPT16, PU.1 and TRIM33 occupancy at the − 35 kb site in WT BMDM treated with DMSO or CBL for 4 h. Mean ± SEM, n = 3. e Kinetics of non-coding transcripts at the − 35 kb region in WT and Trim33−/− BMDM after CBL treatment. RT-qPCR was performed using total RNA, and data are presented relative to expression of DMSO-treated WT BMDM. Mean ± SEM, n = 3. f SPT16 ChIP-qPCR (left) and nucleosome positioning (right) at the − 35 kb region in Trim33−/− IM (Trim33−/−) and in Trim33−/− IM that contained a lentivirus expressing TRIM33 (Trim33−/− + WT TRIM33). Mean ± SEM, n = 3. g Non-coding transcripts at the − 35 kb region (left), Atp1b3 mRNA (middle) and protein expression (right) in Trim33−/− and Trim33−/− + WT TRIM33 IM. Data are presented relative to expression of Trim33−/− IM. Mean ± SEM, n = 4. h Model for FACT/TRIM33-mediated repression at the Atp1b3 locus