Fig 2: Tbx1 dosage affects gene expression and H3K4 monomethylation.(a) Statistically significant gene expression changes in Tbx1 K/D versus control samples. The graph shows the distribution of the number of genes versus log ratios/FC. (b) P19Cl6 cells were transfected with Tbx1 siRNAs (non-targeting siRNA as control) with and without TCP treatment. (top) Acid-extracted histones from P19Cl6 cells run on a 15% SDS–PAGE and stained with Coomassie Blue dye. (bottom) Isolated histones were subjected to SDS–PAGE and monomethylation at histone H3K4 or total H3 as control were detected with specific antibodies. (c) Examples of H3K4me1 ChIP-seq signal distribution in two loci before and after Tbx1 K/D. The shaded area indicate regions with reduced methylation. (d) Venn diagram grouping Tbx1 peaks associated with differentially expressed genes (green) and Tbx1 peaks down-methylated (purple) after Tbx1 K/D. The two groups are clearly distinct with only a small overlap. (e) The graph shows the location relative to the TSS of productive Tbx1 peaks. The peaks associated with differentially expressed genes are clearly closer to the TSS. (f) Canonical pathways by InnateDB analysis of differentially expressed (green) and differentially methylated (purple) genes associated to Tbx1 productive sites. (g) The cartoon illustrates the two types of response to Tbx1 dosage (DE and differential methylation). (h) ChIP–western blot analyses of co-immunoprecipitation experiments with endogenous proteins from P19Cl6 differentiating cells. Immunoprecipitation was carried out using an anti-Tbx1 antibody or rabbit IgG (control). Interacting proteins were eluted from the beads using three consecutive elutions (EL1, EL2 and EL3). Western blots were carried out using anti-Tbx1 (control), anti-Setd7, anti-Kmt2a/Mll1, anti-Kmt2c/Mll3 and anti-Kmt2b/Mll4 antibodies. (i) q-ChIP assay on a subset of Tbx1-binding sites in P19Cl6 cells (with and without Tbx1 K/D) using an anti-Kmt2c antibody, followed by qRT-PCR. Error bars: s.e.m.; n=3; two-tailed Student's t-test. ** P<0.01. FC, fold change; q-ChIP, quantitative ChIP.

Fig 3: Hoxa1 and Hoxb1 genes are required for atrioventricular septation.(A–D) Immunofluorescence on medial sagittal sections showing Tbx1 (red) and Tbx5 (green) protein distribution at E9.5 (23–24 s). (A,B) At E9.5, a boundary is observed in Mef2c-Cre embryos between Tbx1+ cells close to the arterial pole of the heart and Tbx5+ cells in the pSHF. In Hoxb1GoF;Mef2c-Cre embryos the Tbx1+ domain appears reduced although the boundary is maintained. (C) Measurement of the Tbx1+ domain confirmed a decrease of aSHF length in Hoxb1GoF;Mef2c-Cre (GoF) compared to control (Ctr) embryos. (D) Measurement of the Tbx5+ domain revealed a shift of the boundary since the length of the pSHF was increased in Hoxb1GoF;Mef2c-Cre (GoF) compared to control (Ctr) embryos. (E,F) The Tbx5+ domain appears reduced in Hoxb1-/- embryos compared to Hoxb1+/- littermates. (G,H) RNA-FISH on sagittal sections showing the reduction of Tbx5+ domain in the pSHF of Hoxa1-/-;Hoxb1-/- embryos compared to Hoxb1+/- littermates. (I,J) DAPI stained sections of a Hoxb1+/- (I) and a Hoxa1-/-;Hoxb1-/- (J) heart at E14.5 showing the primary atrial septum (pas, arrow). Note the AVSD and absence of the DMP in J, n = 3. at, atria; la, left atrium; lv, left ventricle; ra, right atrium; rv, right ventricle; SHF, second heart field; v, ventricle; la, left atrium; lv, left ventricle; ra, right atrium; rv, right ventricle. Scale bars: 100 µm (E,F); 200 µm (G,H).

Fig 4: TCP partially rescues gene dysregulation caused by Tbx1 K/D in cell culture.(a) Statistically significant gene expression changes in Tbx1 K/D+TCP versus Tbx1 K/D samples. The graph shows the distribution of the number of genes versus log ratios/FC. (b) Venn diagram grouping genes significantly up- or downregulated in Tbx1 K/D versus control samples (blue) and Tbx1 K/D+TCP versus Tbx1 K/D samples. Note the large overlap of 948 genes. (c) Scatter plot depicting the DE of 608 genes ‘rescued' by TCP treatment (discordant) and of 246 genes ‘worsened' by TCP treatment (concordant). (d) Canonical pathways defined in the InnateDB with significant enrichment of genes differentially expressed after Tbx1 KD (red, top) and rescued by TCP treatment (green, bottom). Co-occurrences of rescued pathways with in vivo TCP treatment (Fig. 4n) are indicated with red asterisks. (e) qRT-PCR validation of DE of a subset of rescued genes in P19Cl6 cells (with and without K/D), and WT and Tbx1+/LacZ (HET) E9.5 mice embryos, with and without TCP treatment. Asterisks above bars indicated significant differences compared with the control (*P<0.05; **P<0.01; ***P<0.001). Error bars: s.e.m.; n=3; two-tailed Student's t-test. FC, fold change.

Fig 5: Cartoon showing a working model for Tbx1 interactions with chromatin and TCP rebalancing effect.In the WT state (top row), Tbx1 promotes H3K4me1 deposition (small red dots) through interaction with histone methyltransferases (HMTs). When Tbx1 is absent or of lowered dosage (middle row), H3K4me1 enrichment is lowered because of reduced recruitment of HMTs. TCP treatment (bottom row) re-establishes H3K4me1 by inhibiting the activity of the Lsd1 demethylase. Blue cylinders indicate histone octamers.