Several alloreactive human T cell clones derived from a rejected kidney graft were found to produce in their culture supernatants soluble interleukin 2 receptors (IL-2R) upon specific antigenic challenge (irradiated B cell line from the graft's donor). Among them, the 2B11, a high producer clone, was used to purify a soluble IL-2R preparation which was analyzed, in comparison with the high and low affinity cell-surface IL-2R expressed by 2B11 cells, for its parameters of interaction with a set of anti-IL-2R monoclonal antibodies (mAb) and IL-2. This soluble receptor purified by affinity chromatography (anti-IL-2R mAb column) and sodium dodecyl sulfate gel electrophoresis is composed of a single chain of 35,000 to 45,000 Da. Immunoradiometric assays (IRMA) at equilibrium were set up, using pairs of mAb directed against two separate epitopes on the Tac antigen of the human IL-2R, to measure the respective dissociation constant of these mAb for the soluble IL-2R. They were found to be identical to those found on the cell-surface IL-2R. A 1:1 stoichiometry between the two epitopes were found both on the membrane and soluble species. Competition experiments between membrane and soluble IL-2R for binding the mAb allowed the quantitative analysis of the concentration of soluble IL-2R without the need of amino acid analysis on purified material and set up a quantitative IRMA for the human soluble IL-2R (detection limit 5 pM). The affinity of the soluble IL-2R for IL-2 was determined by various techniques including an IRMA using an anti-IL-2R mAb and radiolabeled IL-2. The results obtained led us to conclude that the soluble IL-2R binds IL-2 with a dissociation constant (KD = 30 nM) identical to that found for the binding of IL-2 to low affinity cell-surface IL-2R (Tac antigen). Whereas 2.5% of cell-surface IL-2R expressed 2 days after antigenic stimulation of 2B11 cells were of high affinity for IL-2 (KD = 25 pM), no (less than 0.07%) high affinity binding sites could be detected on the purified soluble IL-2R. This soluble IL-2R therefore likely corresponds to a truncated, extracellular part of the membrane Tac antigen. The amounts of soluble Tac antigen produced by the 2B11 alloreactive human T cell clone did not exceed 1 nM and, as expected from the binding studies, did not affect IL-2-induced T cell proliferation. The physiologic and pathologic implications of our results are discussed.