In view of the natural resistance to infection by HIV and occasional delayed clinical manifestation of the disease, as also the fact that the virus is able to enter only cells that express CD4 and a co-receptor, we initiated a search for a soluble co-receptor that might compete with its membrane counterpart. Using a sandwich ELISA system, a soluble human CCR5 receptor (sCCR5) was indeed detected in the circulation. Immunoprecipitation of sCCR5-positive plasma samples from Israelis of Ethiopian and non-Ethiopian origin with mAb 2D7, a conformation-dependent anti-CCR5 antibody, revealed the presence of an approximately 22 kDa protein. A panel of antibodies directed against the membrane receptor was used to characterize the structure of the soluble CCR5: mAb CTC8, recognizing the N-terminal sequence of the protein, 10YDIN13; "multidomain" mAbs FAB181B and FAB183B that are dependent upon the presence of Q93 and D95 in ECL1 and K171 and E172 in ECL2A, and mAb FAB182B, recognizing the stretch 184YSQYQF189, which spans the C-terminal part of the second extracellular loop. The presence of short soluble CCR5 in human plasma has not been previously described. Among HIV-negative non-Ethiopian Israelis, 20.4% were sCCR5-positive, as against only 10.5% in HIV-positives. However, 7.1% of HIV-negative Ethiopian Israelis were sCCR5 positive, as were 5.6% HIV-positives. Plasma concentrations of MIP-1beta, the CCR5 agonist, were twice as high in sCCR5-positives (140.8+/-25.8 pg/ml) as in the sCCR5-negatives (77.6+/-11.0 pg/ml, P=0.0157). A significant positive correlation between plasma levels of sCCR5 and MIP-1beta was found (Fig. 4, r=0.8, P<0.0001).