Figure 6: Mirnome analysis revealed that GAS5 competes miR-2467/-3200/Let7e to sustain NODAL expression in hESCs.

(a) A scheme of miRNA target network showing the relationship of GAS5 downregulated miRNAs and their predicted target mRNAs. (b) A list of top 30 downregulated candidate miRNAs on GAS5 overexpression analysed by high-throughput miRNA sequencing. The right panel shows predicted candidate miRNAs that bind to both GAS5 and NODAL. (c) The analysis of NODAL and GAS5 expression in Dicer deficient hESCs. ShDicer represents lentiviral-mediated Dicer knockdown. NC represents scramble siRNA control. **P<0.01, t-test, n=3. (d) The effect of candidate miRNAs on the mRNA levels of GAS5 and NODAL in hESCs. MiRNA mimics are transfected into hESCs at the concentration of 20 pmol per ml. **P<0.01, t-test, n=3. (e) A graph illustrating the binding sites of candidate miRNAs to GAS5 and NODAL transcript, and the scheme of constructing wild-type and mutant NODAL luciferase miRNA reporters and GAS5 MS2 constructs. The red cross indicates mutated sites, see also Supplementary Data 4. (f) Luciferase activities of relative reporter in e with different miRNAs transfection in HEK-293T cells. *P<0.05, **P<0.01, t-test, n=3. (g) qPCR analysis of MS2-mediated candidate miRNAs pulldown assay. Data were normalized to input, and the enrichment ratios compared with input level were shown. MS2-RL serve as a negative control, and MS2-OCT4 are served as positive control in detecting miR-145-5p. **P<0.01, t-test, n=2. (h) AGO2 competing assay using AGO2 antibody-mediated RNA-IP in the indicated groups. RNA levels of precipitated NODAL, linc-ROR and H19 are tested using qPCR. NS, not significant, **P<0.01, t-test, n=2. (i) MS2-mediated miRNA pulldown in the indicated transcript overexpression groups. MS2-RL serves as control. NS, not significant, **P<0.01, t-test, n=2. Error bars represent s.d. of the indicated experiment replicates. RNA level of β-actin served as internal reference for qPCR. See also Supplementary Fig. 6.