We evaluated prime-boost immunization with two recombinant envelope glycoprotein subunit vaccines (HIV-1MN recombinant gp160 vaccine in alum adjuvant [MN rgp160] and HIV-1MN recombinant gp120 vaccine in alum adjuvant [MN rgp120]) for safety and immunogenicity in healthy, HIV-1-uninfected adults. The rationale was to combine the helper T cell memory and binding antibody responses typically induced by rgp160 vaccines with the superior neutralizing antibody responses induced by rgp120 vaccines. In a double-blinded, controlled trial, volunteers were randomly assigned to receive MN rgp160 or adjuvant placebo, and a subset later received MN rgp120. The two vaccines were safe, but reactions to MN rgp160 and its adjuvant placebo exceeded those to MN rgp120. MN rgp160 induced IgG binding antibodies, including all IgG subclasses, to MN rgp160 in all vaccine recipients. HIV-1MN-neutralizing and anti-V3 MN peptide-binding antibodies were observed in a majority of volunteers after the fourth MN rgp160 immunization, but at lower levels compared with immunization with MN rgp120 in historical controls. HIV-1-binding, neutralizing, and fusion inhibition antibodies were boosted to the highest levels among MN rgp160 recipients after MN rgp120 booster injections. MN rgp120 boosting appeared to alter the distribution of MN rgp160 vaccine-induced, anti-MN rgp160 IgG subclass antibodies. MN rgp160 induced helper T cell memory, measured by lymphocyte proliferation, Thl and Th2 cytokine production, and skin testing. Strategies including both subunit vaccines may help maximize antibody and helper T cell memory responses to HIV-1 envelope glycoprotein.