Anticardiolipin (aCL) and anti-beta 2-glycoprotein I(anti beta 2GPI) antibodies have been shown in animal models as not cross-reacting antibody populations. This observation prompted us to prove if anti-beta 2GPI exist in human sera by using a reliable method and then to investigate if these are independent from aCl antibodies. We have developed a new ELISA for the detection of anti-beta 2GPI antibodies employing the coating of the protein in carbonate buffer to irradiated microtitre plates and the filtration of serum samples, that makes irrelevant the binding to the uncoated wells. IgG F(ab)2 fragments from IgG positive sera were shown bind beta 2GPI, providing that the binding was a specific antibody binding, mediated by the antigen binding site of the antibody molecule: moreover the antibodies were not able to differentiate native and delipidated beta 2GPI coated plates, making a possible role of a phospholipid contaminant unlikely. On the other hand, the phosphorus content of native as well as delipitated beta 2GPI was undetectable. IgG, but not IgM, anti-beta 2GPI antibodies were classically inhibited by the addition of soluble beta 2GPI, while cardiolipin liposomes appear to modify the reaction in a completely different way, possibly by the described interaction between cardiolipin and beta 2GPI.(ABSTRACT TRUNCATED AT 250 WORDS)