Fig 2: NOG, GREM1, GREM2 have different antagonizing effects on BMP2, TGF-ß1 and BMP8A. For testing the antagonisms to BMP2, BRE-Luc reporter-containing HEK293T cells were treated with BMP2 (1 nM) and graded doses of NOG (A), GREM1 (B) or GREM2 (C) as indicated. For testing the antagonisms to TGF-ß1, CAGA-Luc reporter-containing HEK293T cells were treated with TGF-ß1 (0.3 nM) and graded doses of NOG (E), GREM1 (F) or GREM2 (G) as indicated. For testing the antagonisms to BMP8A, BRE-Luc reporter-containing cells (I–K) or CAGA-Luc reporter-containing cells (M–O) were treated with BMP8A (1 nM) and graded doses of NOG (I and M), GREM1 (J and N) or GREM2 (K and O) as indicated. After treated overnight, reporter activity in the cells was measured. The luciferase data were normalized against the ß-galactosidase levels to correct for transfection efficiency and were expressed as fold changes by normalizing against the values obtained from control cells without treatment. Data are expressed as the mean ± SEM. n = 3. ***P < 0.001. For monitoring the phosphorylated SMAD levels, HEK293T cells were treated with BMP2 (D), TGF-ß1 (H) or BMP8A (L and P) in the absence or presence of indicated antagonists (1 nM) for 30 min. Cell lysates were subjected to immunoblotting to detect the phosphorylated SMADs. Total forms of SMADs and ß-actin served as loading controls