The ability to induce broadly neutralizing antibodies should be a key component of any forthcoming vaccine against human immunodeficiency virus type 1. One potential vaccine candidate, monomeric gp120, has generally failed to elicit such antibodies. We postulated that gp120 might be a better immunogen if it could be engineered to preferentially bind known broadly neutralizing antibodies. In a first study, we found that four alanine substitutions on the perimeter of the so-called Phe-43 cavity of gp120 could reduce binding of weakly neutralizing CD4-binding site antibodies (R. Pantophlet, E. O. Saphire, P. Poignard, P. W. H. I. Parren, I. A. Wilson, and D. R. Burton, J. Virol. 77:642-658, 2003), while slightly enhancing binding of the potent, broadly neutralizing antibody b12. In the present study, we sought to reduce or abolish the binding of a wider range of nonneutralizing antibodies, by incorporating extra N-glycosylation motifs at select positions into the hypervariable loops and the gp120 core. A hyperglycosylated mutant containing seven extra glycosylation sequons (consensus sequences) and the four alanine substitutions described above did not bind an extensive panel of nonneutralizing and weakly neutralizing antibodies, including a polyclonal immunoglobulin preparation (HIVIG) of low neutralizing potency. Binding of b12, at lowered affinity, and of four antibodies to the C1 and C5 regions was maintained. Removal of N- and C-terminal residues in the C1 and C5 regions, respectively, reduced or abolished binding of the four antibodies, but this also adversely affected b12 binding. The hyperglycosylated mutant and its analogues described here are novel antigens that may provide a new approach to eliciting antibodies with b12-like neutralizing properties.