Mutant cell lines defective in the biosynthesis of glycosylphosphatidylinositol (GPI) described to date were isolated by selecting cells which no longer expressed one or more endogenous GPI-anchored proteins on their surface. In this study, a new mutant in this pathway was isolated from ethylmethane-sulphonate-mutagenized Chinese hamster ovary cells stably transfected with human placental alkaline phosphatase (PLAP) as a marker of GPI-anchored proteins. A three-step protocol was employed. In the first step, cells with decreased surface expression of PLAP were selected by four rounds of complement-mediated lysis with an anti-(alkaline phosphatase) antibody. The surviving cells were cloned by limiting dilution and those with low levels of total alkaline phosphatase activity were selected in the second step. Finally, the ability of each clone to synthesize the first three intermediates in GPI biosynthesis in vitro was assessed to determine which cells with low alkaline phosphatase activity harboured a defect in one of these reactions. Of 230 potential mutants, one was defective in the second step of GPI biosynthesis. Microsomes from this mutant, designated G9PLAP.85, were completely unable to deacetylate either endogenous GlcNAc-phosphatidylinositol (PI) synthesized from UDP[6-3H]GlcNAc or exogenous GlcNAc-PI added directly to the membranes. Complementation analysis with the Thy-1-deficient murine lymphoma cells demonstrated that G9PLAP.85 has a molecular defect distinct from these previously described mutants. Therefore, these results suggest that mutants in GPI biosynthesis could be selected from almost any cell line expressing a GPI-anchored marker protein.