Human platelet factor 4 (PF4; CXCL4), a heparin-binding CXC chemokine contained in platelet α-granules, is secreted upon activation of platelets. Several reports have identified PF4 as an inhibitor of hematopoietic progenitor and endothelial cell proliferation and angiogenesis. Furthermore, PF4 has been shown to strongly inhibit T cell proliferation as well as IFN-γ and IL-2 release by activated T cells. We have recently reported that human PF4 inhibits the proliferative response of human CD4+CD25− T cells, while inducing expansion of CD4+CD25+ T regulatory (Tr) cells stimulated by anti-CD3 or anti-CD3 and anti CD-28 monoclonal antibodies (mAbs), and that PF4-induced CD4+CD25+ Tr cells lose their potent suppressor function in vitro. The antithetic effects of PF4 on CD4+CD25− T cells and CD4+CD25+ Tr cells in response to anti-CD3 and anti-CD3/28 mAbs appear to be specific since these effects were not mimicked by equivalent quantities of protamine, a positively charged heparin-binding protein used as a control for any putative interaction mediated only by positive charges, or heparin alone. We hypothesize that PF4 acts not only as the target for the antibody, in patients experiencing heparin-induced thrombocytopenia (HIT), but also as a modulator of T regulatory lymphocytes. How PF4 modulates T cell proliferation and CD4+CD25+ Tr cell-mediated suppression of CD4+CD25− T cells remains to be determined. To delineate which domain(s) of PF4 is(are) critical for its antithetical activity on the two T cell subsets, we compared PF4 from rat and cattle, which are about 74% homologous to human PF4, for their ability to affect proliferation of CD4+CD25+ and CD4+CD25− T cells stimulated by anti-CD3 mAb. Both rat and bovine PF4 were recognized by a rabbit polyclonal Ab against human PF4, but neither one, in contrast with human PF4, was capable to induce proliferation of CD4+CD25+ Tr cells. This suggests that the relatively few amino acid (AA) residues at which rat and human differ are critical for the modulatory activity of PF4 on T cells. Therefore, we introduced specific modifications into human PF4 cDNA using site-directed mutagenesis to determine the importance of individual AA residues for PF4 effects on T cell proliferation. Our strategy involved converting selected amino acids in human PF4 (reactive) to the corresponding residues in rat PF4 (non-reactive) and determining the effect of each change on the proliferation of CD4+CD25+ and CD4+CD25− T cells stimulated by anti-CD3 mAb. We created seven mutant forms of PF4. Four of the mutants involved one or more of six residues in the 47 C-terminal AAs, known to comprise the most important heparin binding region. The other three mutants were created at the N-terminus. All the PF4 constructs, expressed in E. coli, assembled properly into tetramers as determined by their HPLC profile and electrophoretic mobility in SDS-PAGE and were comparable to wild-type PF4 in their avidity for heparin. PF4 constructs bearing mutations at the N-terminus (E4S, L11V, and T16S) and three C-terminal mutants (the combined P37A/T38V/A39P, R49S, and L55R) were comparable to wild type recombinant human PF4 for their proliferative effects on the two T cell subpopulations. In contrast, another C-terminal mutant (A57V) completely abrogated PF4 proliferative effects on CD4+CD25+ Tr cells, suggesting the importance of this determinant in the interaction with a specific receptor on T cells.