The identification of 16 different activating mutations in the TSH receptor, found in patients suffering from toxic autonomous adenomas or congenital hyperthyroidism, leads to the concept that this receptor is in a constrained conformation in its wild-type form. We used mild trypsin treatment of CHO-K1 cells or COS-7 cells, stably or transiently transfected with the human TSH receptor, respectively, and measured its consequences on the TSH receptor coupled cascades, i.e. cyclic AMP and inositol-phosphates accumulation. A 2-min, 0.01% trypsin treatment increased stably cyclic AMP but not inositol-phosphates formation. This was not observed after chymotrypsin, thrombin and endoproteinase glu C treatment. The TSH action on cyclic AMP was decreased by only 25%. The effect was also observed in cells expressing the dog TSH receptor. It was not observed in MSH receptor, LH receptor expressing or mock transfected cells (vector alone). It is therefore specific for the TSH receptor, for its action on the Gs/adenylate cyclase cascade, and for the proteolytic cleavage caused by trypsin. Using monoclonal (A. Johnstone and P. Shepherd, personal communication) and polyclonal antibodies directed against the extracellular domain of the TSH receptor, it was shown that treatment by trypsin removes or destroys a VFFEEQ epitope (residues 354–359) from the receptor. The effect mimics the action of TSH as it activates G_sα and enhances the action of forskolin. It is not reversible in 1 h. The results support the concept that activation of the receptor (by hormone, autoantibodies, mutations or mild proteolysis) might involve the relief of a built-in negative constrain. They suggest that the C-terminal portion of the large extracellular domain plays a role in the maintenance of this constrain.