We have previously shown that human bactericidal/permeability-increasing protein (BPI) is able to inhibit serum-dependent lipopolysaccharide (LPS)-mediated activation of human monocytes and neutrophils in vitro, and to counteract the lethal effects of LPS challenge in vivo. Lipopolysaccharide-binding protein (LBP) is a serum protein which participates in LPS-mediated activation of cells (Tobias, P. S., Mathison, J., Mintz, D., Lee, J. D., Kravchenko, V., Kato, K., Pugin, J., and Ulevitch, R. J. (1992) Am. J. Respir. Cell. Mol. Biol. 7, 239-245). We have proposed that BPI functions in a negative feedback loop which opposes this activation (Marra, M. N., Wilde, C. G., Collins, M. S., Snable, J. L., Thornton, M. B., and Scott, R. W. (1992) J. Immunol. 148, 532-537). We have now cloned and expressed recombinant forms of human BPI and LBP. Here we demonstrate that purified recombinant human LBP can replace the serum requirement for both LPS binding to human monocytes and LPS-mediated secretion of tumor necrosis factor alpha from these cells. These activities of LBP are inhibited by a neutralizing anti-CD14 monoclonal antibody. We further demonstrate that purified recombinant human BPI can inhibit LBP-mediated LPS binding to cells and their subsequent activation. Comparison of the LPS binding properties of BPI and LBP in enzyme-linked immunosorbent type assays and in the Limulus amebocyte lysate assay suggest that BPI has a stronger affinity for LPS than does LBP. Direct competition between BPI and LBP for LPS may explain the inhibition by BPI of the proinflammatory effects of LBP in the presence of LPS.