Four amino acids within a tandem QxVx repeat in a predicted extended α-helix of the Smad-binding domain of Sip1 are necessary for binding to activated Smad proteins.

The zinc finger transcription factor Smad-interacting protein-1 (Sip1; Zeb2, Zfhx1b) plays an important role during vertebrate embryogenesis in various tissues and differentiating cell types, and during tumorigenesis. Previous biochemical analysis suggests that interactions with several partner proteins, including TGFβ family receptor-activated Smads, regulate the activities of Sip1 in the nucleus both as a DNA-binding transcriptional repressor and activator. Using a peptide aptamer approach we mapped in Sip1 its Smad-binding domain (SBD), initially defined as a segment of 51 amino acids, to a shorter stretch of 14 amino acids within this SBD. Modelling suggests that this short SBD stretch is part of an extended α-helix that may fit the binding to a hydrophobic corridor within the MH2 domain of activated Smads. Four amino acids (two polar Q residues and two non-polar V residues) that form the tandem repeat (QxVx)2 in this 14-residue stretch were found to be crucial for binding to both TGFβ/Nodal/Activin-Smads and BMP-Smads. A full-length Sip1 with collective mutation of these Q and V residues (to A) no longer binds to Smads, while it retains its binding activity to its cognate bipartite target DNA sequence. This missense mutant Sip1(AxAx)2 provides a new molecular tool to identify SBD (in)dependent target genes in Sip1-controlled TGFβ and/or BMP (de)regulated cellular, developmental and pathological processes.