Proteins linked to cell membranes by a glycosylphosphatidylinositol (GPI) anchor must first undergo cleavage by a putative transamidase between the omega and omega + 1 positions within a proposed small amino acid (SAD) domain in the carboxy terminus of the nascent polypeptide. The requirements for such processing, defined in an engineered placental alkaline phosphatase construct (miniPLAP), suggest the SAD domain functions as an autonomous unit within the context of an otherwise permissive carboxyterminal sequence with only certain amino acids tolerated at the omega, omega + 1, and omega + 2 positions. To test whether this hypothesis could be generalized, we engineered a chimeric molecule containing the extracellular domain of miniPLAP and the carboxy-terminal portion of the urokinase receptor (MP/uPAR) into which various amino acid substitutions were introduced. The variant proteins were translated and metabolically labeled in vitro using a cell-free translation system that contains the enzymatic machinery required for carboxy-terminal processing and GPI anchor addition. The results of this study indicate that the SAD domain functions as an independent, but not an autonomous, unit. The requirements fur processing in miniPLAP and MP/uPAR differed markedly in some respects, in part due to the influence of the amino acid at the omega + 4 position which both modified cleavage between the omega and omega + 1 positions and permitted a second cleavage site to be generated in some cases. In addition, substitution of bulky hydrophobic amino acids in series at the omega + 2 and omega + 3 positions inhibited carboxy-terminal processing in a dose-dependent manner, suggesting the presence of a critical docking site adjacent to the cleavage site. These results suggest the carboxy-terminal transamidase recognizes a more extended structure similar to the mechanism proposed for serine proteases. Further, the data provide a potential means for isolating the transamidase.