(16). Bovine serum albumin (BSA) was from Pentex, bovine hemoglobin from Sigma, and bovine fetuin from Gibco.The samples to be reduced and alkylated (17) were dissolved at a concentration of about 10 mg/ml in 0.1 M Tris-HCI buffer (pH 8.2) containing 8 M urea. After deaeration with nitrogen, dithiothreitol was added to a final concentration of 10 mM. After 3 hr at room temperature under nitrogen, the reaction mixture was cooled on ice and recrystallized iodoacetamide was added to make the final concentration 20 mM. After another 15 min the sample was dialyzed against distilled water and lyophilized. Phosphate-buffered saline containing 0.05% gelatin, 0.02% sodium azide, and 0.1% Triton X-100 (PBS-GTX) with urea added to it to give a 2 M solution was used to dissolve the dry proteins, some of which were poorly soluble in ordinary buffers. In some cases the reduced and carboxamidomethylated (RC-) protein was diluted directly into this solution for testing without dialysis and lyophilization. HAFP labeled with 1251 was reduced and alkylated similarly and freed of reactants by gel filtration on Sephadex G-15 equilibriated with PBS-GTX.Antisera. Anti-HAFP sera have been described (6,8). Using a similar schedule, we produced one antiserum each to RC-HSA and RC-BSA and two to RC-HAFP in rabbits. The dose used per injection was 5 mg (RC-HSA, RC-BSA) or 0.5-2 mg (RC-HAFP) of protein suspended in PBS and given in complete Freund's adjuvant. A precipitating antiserum to RC-CEA was obtained similarly, using 0.25 mg of RC-CEA per injection.Immunological Tests. Immunodiffusion was performed on plates cast in a conventional buffer and soaked overnight in PBS-GTX prior to use. Radioimmunoassays were performed using the double antibody technique as described (4). The first and second antibody were added in the form of an IgG fraction in assays where native albumin was tested for crossreactivity with AFP. The diluent buffer used throughout was PBS-GTX.
RESULTSLack of Crossreaction between Native AFP and Albumin. Antisera produced against highly purified HAFP do not react with albumin in immunodiffusion (ref. 15, our unpublished results). Radioimmunoassay also failed to reveal any crossreactivity between native HAFP and native HSA or RC-HSA. RC-HAFP was also almost completely devoid of activity in the usual AFP assay (Fig. 1).