Recent data have substantially modified our view of natural killer cells. Although maturation of natural killer cells occurs in the absence of a functional thymus, we have shown that clonogenic precursors capable of differentiating into mature CD3-16+56+ natural killer cells exist in CD3-4-8-16- populations isolated from human thymus. Analysis of peripheral blood-derived natural killer clones showed that they can lyse normal cells (e.g., phytohemagglutinin-induced blasts) isolated from some individuals. Importantly, natural killer clones isolated from single individuals displayed different patterns of cytolytic activity against a panel of normal allogeneic cells. These data suggested the existence of a natural killer cell repertoire. A number of observations have revealed that the expression of given HLA class I alleles protects target cells from lysis by different groups of natural killer clones. Evidence has been gained by genetic analysis of the determinants responsible for susceptibility/resistance to lysis by natural killer clones together with analysis, as target cells, of HLA-defective variants or HLA transfectants. Thus, natural killer cells were found to express a clonally distributed ability to recognize HLA class I alleles. The selection of new monoclonal antibodies directed against members of a novel family of natural killer specific p58 molecules allowed the identification of the putative natural killer receptors for different MHC class I alleles. Firstly, a correlation was established between the expression of given p58 molecules (e.g., EB6 and GL183) and the class I alleles recognized. Secondly, anti-p58 monoclonal antibodies restored the natural killer-mediated lysis of class I-protected cells.(ABSTRACT TRUNCATED AT 250 WORDS)