Figure 3
FRG2 is activated as a consequence of DUX4 protein activity at its promoter. (A) Genomic snapshot (location indicated at the bottom) of mapped RNA-seq reads at the chromosome 4q FRG2 locus. Overexpression of DUX4 results in the activation of FRG2 in myoblasts and fibroblasts, GFP overexpression was used as a control. (B) Graphical representation of DUX4 binding at the 4q FRG2 promoter (genomic snapshot location indicated at the bottom) as revealed by ChIP-seq analysis (data obtained in myoblasts). (C) Genomic fragments obtained from chromosomes 4 and 10 were cloned upstream of the luciferase gene. Polymorphisms distinguishing both copies are indicated (Variants 1-3) and the three identified DUX4 binding sites are displayed, with nucleotides matching the previously identified core DUX4 binding sequence underlined. All numbers show the relative distance to the TSS of FRG2, in the 10mut1-3 constructs the number indicates the first displayed nucleotide. The 10Δ construct lacks all three DUX4 binding sites, whereas in 10mut1-3 the red nucleotides were mutated to the nucleotides indicated below them, thereby destroying the individual DUX4 binding sites. (D) DUX4 activates the FRG2 promoter in a luciferase reporter assay. Both the 4q and 10q copy of the FRG2 promoter are activated by DUX4. The 10q copy lacking the DUX4 binding sites (10Δ) was not activated by DUX4. Destruction of the three individual binding sites revealed that sites 1 and 2 are mediating the DUX4 dependent activation of FRG2. Counts per second (CPS) are a direct measure of luciferase activity, error bars indicate the SEM of three independent experiments. Asterisks indicate significant differences based on a one-way ANOVA (P <0.0001), followed by pairwise comparison using Bonferroni correction. NS = non-significant.