Caveats for the use of surface-adsorbed protein antigen to test the specificity of antibodies.

Rabbit antisera against apo-cytochrome c, which was prepared by removal of the covalently bound heme prosthetic group from yeast iso-1 cytochrome c, were tested for reactivity against native yeast iso-1-cytochrome c. When the antigen was adsorbed to a microtiter plate in a conventional enzyme-linked immunosorbent assay (ELISA), the antisera were unable to distinguish between their cognate antigen apo-cytochrome c, a random coil protein, and native cytochrome c, a small globular protein of remarkable conformational stability in solution. However, when the assay was conducted under conditions where antigen and antibody were free to associate in solution, that is in a solution-phase radioimmunoassay (RIA), the antisera were highly specific for apo-cytochrome c. Similarly, antibodies induced by native cytochrome c and discriminating strongly between native and apo-cytochrome c in a solution-phase RIA, did not distinguish between native and apo-cytochrome c in a solid-phase ELISA. This discrepancy of results obtained by different immuno assay procedures clearly indicates that adsorption to plastic alters the antigenic structure of even a conformationally stable protein such as cytochrome c. A conventional solid-phase ELISA strongly selects for those antibodies that recognize the unfolded antigen. The results presented warrant serious thoughts about previous reports on anti-peptide antibodies reacting with native whole protein molecules, as tested by those ELISA procedures that have the protein antigen adsorbed to plastic.