The product of the X-chromosome's Pig-a gene acts in the first step of glycosylphosphatidylinositol (GPI) anchor biosynthesis, and is thereby essential for attaching certain proteins to the cell surface. The experiments described herein were designed to evaluate whether lack of GPI-anchored proteins could form the basis of an in vivo mutation assay. Specifically, we used a CD59-negative cell surface phenotype to denote Pig-a mutation. Besides anti-CD59-PE, two other fluorescent reagents were used: thiazole orange to differentiate mature erythrocytes, reticulocytes (RETs), and leukocytes; and anti-CD61 to resolve platelets. These experiments were performed with Sprague Dawley rats, and focused on two cell populations, total erythrocytes and RETs. The ability of the analytical method to enumerate CD59-negative erythrocytes was initially assessed with reconstruction experiments whereby mutant-mimicking cells were added to control bloods. Subsequently, female rats were treated on three occasions with the model mutagens ENU (100 mg/kg/day) or DMBA (40 mg/kg/day). Blood specimens were harvested at various intervals, as late as 6 weeks post-exposure. Considering all week 4-6 data, we found that CD59-negative cells ranged from 239 to 855 x 10(-6) and 82 to 405 x 10(-6) for ENU and DMBA, respectively. These values were consistently greater than those observed for negative control rats (18 +/- 19 x 10(-6)). The elevated frequencies observed for the genotoxicant-exposed animals were usually higher for RETs compared to total erythrocytes. These data support the hypothesis that an efficient in vivo mutation assay can be developed around flow cytometric enumeration of erythrocytes and/or RETs that exhibit aberrant GPI-anchored protein expression.