Smad3 protein is a mediator of transforming growth factor-β (TGF-β) signalling. It is recruited by SARA ('Smad anchor for receptor activation') to the TGF-β-receptor kinase, where it is phosphorylated. Smad3 then dissociates from SARA and the receptor, forms a heterotrimeric complex with cofactor Smad4, and translocates to the nucleus, where the Smad3–Smad4 complex binds to the promoters of target genes. Ski, a co-repressor, can bind directly to Smad3 to cause transcriptional repression. Interestingly, SARA and Ski bind to overlapping sites in a carboxy-terminal domain of Smad3, called MH2. Qin et al. now report in Genes & Development how these different interactions might be regulated.

The authors used analytical ultracentrifugation to show that unbound Smad3 trimerizes in a concentration-dependent way. By contrast, when complexed with SARA, Smad3 does not oligomerize. Might SARA therefore inhibit Smad3-trimer formation? When the authors modelled the Smad3–SARA complex on the trimeric scaffold of a related Smad protein (Smad2), they found that the Smad3–SARA complex makes poor intersubunit contacts. This is because the MH2-domain three-helix bundle — a conserved structure for intersubunit contacts — does not tilt enough towards the neighbouring Smad3 subunit. SARA might therefore be able to inhibit Smad3 trimerization by preventing the three-helix bundle from tilting. On Smad3 phosphorylation, the conformational change caused by tilting of the three-helix bundle represents an allosteric switch that disrupts the interaction of Smad3 with both SARA and the receptor.

In a series of biochemical experiments, Qin et al. found that phosphorylated Smad3 binds to SARA with a low affinity, but that its binding drives phosphorylated Smad3 towards the monomeric state. This suggests that Smad3–SARA and homotrimeric-Smad3 interactions are competing and mutually exclusive.

Kai Lin and co-workers then showed that the interaction between Ski and Smad3 is greatly reduced when conserved residues at the trimer interface are mutated, stabilizing Smad3 in the monomeric form. So, although SARA and Ski bind to overlapping surfaces on Smad3, they recognize the monomeric and trimeric forms, respectively.

In conclusion, phosphorylation-induced Smad3 trimerization functions as a master switch in TGF-β signalling, converting Smad3–receptor interactions into Smad3–transcriptional-comodulator interactions. This initial study might have far wider implications, according to the authors, who say that 'trimerization-dependent interaction between Smad3 and Ski may represent a more general paradigm of Smad–nuclear interactions'.