A missense mutant of Smad2, Smad2D450E, that was not phosphorylated by transforming growth factor beta (TGF-beta) signaling, was identified in colorectal cancer. Previously, we constructed a mutant Smad3, Smad3D407E, which has an Asp to Glu mutation in the corresponding position of Smad2D450. Smad3D407E was not phosphorylated by the constitutively active form of the TGF-beta type I receptor, and inhibited the phosphorylation of co-expressed wild-type Smad2 and Smad3. In the present study, we examined the inhibitory effects of Smad2D450E on TGF-beta signaling and found that there are considerable differences between Smad2D450E and Smad3D407E. Smad2D450E suppressed the phosphorylation of Smad2, but did not affect the phosphorylation of Smad3, while Smad3D407E blocked the phosphorylation of both Smad2 and Smad3. Consistent with these results, Smad2D450E reduced hetero-oligomer formation of Smad2 with Smad4, but not of Smad3 with Smad4, while Smad3D407E reduced hetero-oligomer formation of both Smad2 and Smad3 with Smad4. However, Smad2D450E reduced the binding of Smad3 to a target DNA element as well as Smad2-binding, and Smad2D450E had inhibitory effects on the transcriptional activity of several targets, as Smad3D407E did. These results suggested that Smad2D450E might block the Smad3-mediated TGF-beta signaling in a hitherto unknown manner after the phosphorylation and hetero-oligomer formation, such as in the process of nuclear translocation or transcriptional regulation, which we could not identify previously by using Smad3D407E.