A model has been proposed in which, in the absence of TSH, the extracellular domain of the TSH receptor would exert a silencing effect on the serpentine domain involved in activation of the G(alpha)(s) protein. Mutation of S281 in the ectodomain is supposed to release this constraint, thereby causing receptor activation. This defines S281 and its neighbors as a segment important in intramolecular signal transduction. The functional importance of this segment was explored by site-directed mutagenesis experiments involving S281, as well as the two cysteine residues (C283, C284) present immediately downstream. S281 was mutated to N, T, G, and A in this study, and the functional characteristics of the mutants were compared. We found that S281N, S281T, and S281G display stronger constitutive activity than S281A mutant, suggesting that increase in constitutive activity is related to the extent of disruption of the local structure of the ectodomain. C283 and C284, the two consecutive cysteines that are highly conserved in glycoprotein hormone receptors, were mutated to serine, either alone (S281HSC or S281HCS) or in combination (S281HSS) and were studied in two different TSH receptor backgrounds. The mutated cysteine ectodomains were either linked to a glycosylphosphatidylinositol anchor or the serpentine domain of the wild-type holoreceptor. Glycosylphosphatidylinositol-anchored ectodomain receptors showed good cell surface expression in CHO cells, but only S281HCS was able to bind TSH specifically, illustrating the importance of C283, or the putative disulphide bond, in maintaining the conformation of the ligand binding site. In contrast, cysteine mutants on an extracellular domain-holoreceptor background displayed severely impaired membrane targeting and were poorly expressed in COS cells. However, basal cAMP production, normalized to expression at the plasma membrane, indicated significant increase in constitutive activity of all three mutants, compared with the wild-type receptor. Altogether, these findings support a model in which the ectodomain would act as a silencer of the basal activity of the serpentine portion of the receptor.