Soluble forms of the HIV-1 receptor CD4 (sCD4) have been extensively characterized for more than 2 decades as promising inhibitors and components of vaccine immunogens. However, they were mostly based on the first two CD4 domains (D1D2), and numerous attempts to develop functional, high-affinity, stable soluble one-domain sCD4 (D1) have not been successful because of the strong interactions between the two domains. We have hypothesized that combining the power of structure-based design with sequential panning of large D1 mutant libraries against different HIV-1 envelope glycoproteins (Envs) and screening for soluble mutants could not only help solve the fundamental stability problem of isolated D1, but may also allow improvement of D1 affinity while preserving its cross-reactivity. By using this strategy, we identified two stable monomeric D1 mutants, mD1.1 and mD1.2, which were significantly more soluble and bound Env gp120s more strongly (50-fold) than D1D2, neutralized a panel of HIV-1 primary isolates from different clades more potently than D1D2, induced conformational changes in gp120, and sensitized HIV-1 for neutralization by CD4-induced antibodies. mD1.1 and mD1.2 exhibited much lower binding to human blood cell lines than D1D2; moreover, they preserved a β-strand secondary structure and stability against thermally induced unfolding, trypsin digestion, and degradation by human serum. Because of their superior properties, mD1.1 and mD1.2 could be potentially useful as candidate therapeutics, components of vaccine immunogens, and research reagents for exploration of HIV-1 entry and immune responses. Our approach could be applied to other cases where soluble isolated protein domains are needed.