In order to reduce the nonspecific signal of noncompetitive solid phase immunoassays and to improve their sensitivities, the immune complex transfer enzyme immunoassay has been developed. Antigens to be measured were reacted with 2,4-dinitrophenyl-biotinyl-antibody FabΒ΄ and antibody FabΒ΄-Ξ²-D-galactosidase conjugate, and antibody IgGs to be measured were reacted with 2,4-dinitrophenyl-antigen and antigenΞ²-D-galactosidase conjugate. The immune complexes formed comprising the three components were trapped onto colored polystyrene beads coated with affinity-purified (anti-2,4-dinitrophenyl group) IgG. After washing, the immune complexes were eluted from the colored polystyrene beads with eN-2,4-dinitrophenyl-L-lysine, and the eluates were incubated with white polystyrene beads coated with streptavidin for antigens and coated with affinity-purified (anti-human IgG Ξ³-chain) IgG for antibody IgGs to transfer the immune complexes. By this method, ultrasensitive enzyme immunoassays have been developed for HIV-1 p24 antigen and antibody IgGs to HIV-1 p17 and reverse transcriptase (RT). The nonspecific signals in the absence of the antigen and the antibody IgGs were reduced 300 to 15,000-fold by the immune complex transfer process, but the amounts of the immune complexes decreased only 1.8 to 3.1-fold by the immune complex transfer. As a result, the sensitivities for HIV-1 p24 antigen and antibody IgGs to HIV-1 p17 and RT were improved 100 to 5,600-fold by the immune complex transfer. The detection limit of HIV-1 p24 antigen by 20 hr assay of Ξ²-D-galactosidase activity (10 zmol) was 4,000 to 17,000-fold lower than those obtained with currently available com-mercial kits. The improved sensitivities for antibody IgGs to p17 and RT by 20 hr assay of Ξ²-D-galactosidase activity were 1 Γ 10 5 to 3 Γ 10 5 -fold higher than those of Western blotting for p17 and p66 bands. However, the nonspecific signals in the absence of antigens and antibody IgGs were enhanced to various degrees by two factors. In order to transfer the immune complexes more efficiently within shorter periods of time, the colored polystyrene beads were incubated with the white polystyrene beads in the presence of eN-2,4-dinitrophenyl-L-lysine. Such direct contact between solid phases for trapping and transferring of the immune complexes significantly enhanced the nonspecific signals. In addition, the presence of human serum samples containing neither antigens to be measured nor antibody IgGs to be measured also enhanced the nonspecific signals to various extents. Namely, these two factors limited the effect of the immune complex transfer to improve the sensitivity by 20 hr assay of Ξ²-D-galactosidase activity. By 1 hr assay of Ξ²-D-galactosidase activity, the detection limit of HIV-1 p24 antigen using 10 Β΅l of serum samples (0.24 pg/ml) was 40 to 80-fold lower than those obtained with currently available commercial kits using 100 to 200 Β΅l of serum samples (10 to 20 pg/ml) and the detection limits of antibody IgGs to HIV-1 p17 and RT were 1 Γ 10 4 to 3 Γ 10 4 -fold lower than those by Western blotting for p17 and p66 bands. Finally, the immunoreactions involved in the immune complex transfer enzyme immunoassaysthe formation, trapping, and transferring of the immune complexes-will be performed within 15 to 30 min.