Fig. 4. Human Shn1 contains two Smad interaction domains and can form a complex with Smads on the BRE. (A) Summary of hShn1 Smad-interacting regions. hShn1 fragments 1 (hShn1^1-599), 2 (hShn1^1-702), 5 (hShn1^1496-2213), 6 (hShn1^1756-2544) and 7 (hShn1^1756-2717) (green) co-immunoprecipitate with Mad and Med, whereas polypeptides 3 (hShn1^496-1121) and 4 (hShn1^1002-1635) (black) do not. Presumptive minimal Smad-interaction domains are bracketed. (B) GST pull-down assays using fragments 2 (hShn1^1-702) and 6 (hShn1^1756-2544) confirm the presence of two Smad-interacting regions in hShn1. Equivalent amounts of GST, GST-Smad1 (MH2 domain+linker) and GST-Smad4 (MH2 domain+linker) were co-incubated with in vitro translated [³⁵S]-methionine-labeled hShn1 polypeptides as indicated. Both hShn1 fragments are specifically retained by Smad1 and Smad4, but not by GST alone. Both hShn1 fragments display a higher affinity for Smad1 than Smad4. (C) Gel mobility shifts were performed with radiolabelled Xenopus BRE probe. Co-incubation with Mad and Med results in a low mobility complex (arrow) that is further retarded by the addition of a Drosophila Shn fragment, ShnCT (asterisk, lane 2). Co-incubation of Drosophila Smads with hShn1 polypeptides (hShn1^1496-2213 and hShn1^1-599) also resulted in retardation of the BRE (asterisks, lanes 4 and 6, respectively), indicating that a Mad/Med/hShn1 complex had formed on the DNA. Both of these fragments contain regions of overlap with the Smad-interacting fragments identified in GST pull-downs (see A). Fragments that lack Smad-interaction domains did not alter probe mobility (data not shown).