Fig 2: MED12 is hypermethylated by CARM1. (A) Experimental flowchart for detecting post translational modifications (PTMs) of C-terminus of MED12 (1616-2177) in wild type (WT) or CARM1 knockout (KO) MCF7 cells. (B) Cell lysates as described in (A) were subjected to immunoblotting (IB) analysis with antibodies as indicated. Actin was served as a loading control. (C) Schematic representation of the domain architecture of MED12 protein. Leucine-rich (L) domain (light green); Leucine-serine-rich (LS) domain (yellow); Proline-glutamine-leucine (PQL) domain (light blue); Poly-glutamine (Opa) domain (purple). Arginine methylation sites identified in the PQL domain were shown by matchsticks. (D) Methylated arginine residues identified in the C-terminus of MED12 following the protocol as described in (A). me1: mono-methylation; me2: di-methylation. PSM: peptide spectrum match. (E) WT and CARM1 KO cells were infected with lenti-viral vectors expressing Flag-tagged MED12 C-terminus (1616-2177), lysed and subjected to immunoprecipitation using anti-Flag antibody followed by immunoblotting (IB) analysis with antibodies as indicated. Anti-H3R17me2(a) antibody was used to detect methylated MED12.

Fig 3: SCFFBXO9 E3 ubiquitin ligase specifically degrades PRMT4 in lung epithelia.A Screen of F-box (FBXO) family members revealed that overexpression of FBXO9 selectively degrades PRMT4 in BEAS-2B cells. B Ectopically expressed FBXO9 plasmid degrades PRMT4 in a concentration-dependent manner not observed with FBXO24. Densitometric results are presented in the bottom panels. C Co-IP studies showing that PRMT4 interacts with Skp1 and Cullin 1, components of SCF E3 ligase. D Knockdown of FBXO9 stabilizes PRMT4 using protein half-life studies with cellular exposure to cycloheximide (CHX) for various times to inhibit protein synthesis. The relative expression of PRMT4 was plotted in the right panel. *P < 0.05 vs. time 0. Experiments n = 3.

Fig 4: Inhibition of PRMT4 suppresses cell death in a LPS lung injury model.A–D PRMT4 was silenced or overexpressed by i.t. administered lentiviral constructs for 14 days. Mice were then given (i.t.) LPS with or without a PRMT4 inhibitor as indicated for 24 h (n = 8). Lung tissues were stained with hematoxylin–eosin (H&E) (A, upper panels) and TUNEL staining (A, lower panels). B TUNEL-positive cells in lung tissues were quantitated. C PRMT4 overexpression and knockdown in lung tissues from A were analyzed by immunoblotting. The data from C were quantitated and plotted in D (n = 3). *P < 0.05 vs. scrambled control. Scale bar = 100 µm.

Fig 5: CARM1-expressing EOC cells are selectively sensitive to EZH2 inhibitors. a Expression of CARM1, EZH2, and Cyclin A in the indicated EOC cell lines, HOSE, and FTE cells were determined by immunoblot. Expression of ß-actin was used as a loading control. b Expression of CARM1 and a loading control ß-actin in CARM1-high parental and CRISPR-mediated CARM1 knockout (KO) A1847 cells. c Equal number of parental control or CARM1 knockout A1847 cells were plated and treated with each of the 23 individual epigenetic inhibitors for 14 days. The media with the inhibitors was refreshed every 3 days. Cell growth was quantified as integrated density using NIH ImageJ software. Quantification of the average integrated density graphed as a scatter plot. The X-axis indicates the relative growth of treated CARM1 knockout A1847 cells compared with DMSO vehicle controls. Y-axis indicates the relative growth of treated CARM1-high parental A1847 cells compared with DMSO vehicle controls. n = 4; error bars represent SEM. d Representative images of colonies formed by the cells treated with the indicated inhibitors. GSK126 and UNC1999 represent positive hits from the screen. Note that CI994 was used a negative control that showed no difference between parental and CARM1 knockout cells. e Expression of CARM1, H3K27Me3, EZH2, and H3 in parental control and CARM1 knockout A1847 cells. f Relative expression of EZH2 in the TCGA HGSOC cases with (n = 61) or without (n = 467) CARM1 amplification. Relative EZH2 expression levels were transformed to log2 (10+ expression) values. P-values are from two-tailed t-test. Error bars represent SD