The graft-versus-leukemia effect of allogeneic stem-cell transplantation is believed to be mediated by T-cell recognition of minor histocompatibility antigens on recipient cells. For minor histocompatibility antigens HA-1 and HA-2, normal cell expression is restricted to hemopoietic cells, and boosting the immune response to these antigens may potentiate graft-versus-leukemia effect without accompanying graft-versus-host disease. To increase efficacy, expansion of HA-1- or HA-2-specific CTL before transplantation is desirable. However, primary HA-1- or HA-2-specific CTL expanded in vitro are often of low avidity. An alternative approach is to prime specific CTL responses in vivo by vaccination. Clearly, donor vaccination must be safe and specific. We have developed DNA fusion vaccines able to induce high levels of epitope-specific CTL using linked CD4(+) T-cell help. The vaccines incorporate a domain of tetanus toxin (DOM) fused to a sequence encoding a candidate MHC class I binding peptide. This design generates antitumor CD8(+) T-cell responses and protective immunity in preclinical models. For clinical application, we constructed vaccines encoding HLA-A*0201-restricted peptides from human HA-1 and HA-2, which were fused to DOM, and tested their performance in HLA-A*0201-transgenic mice. Priming induced epitope-specific, IFNgamma-producing CD8(+) T cells with cytotoxic function boosted to high levels with electroporation. Strikingly, these mouse T cells efficiently killed human lymphoblastoid cell lines expressing endogenous HA-1 or HA-2. High avidity is indicated by the independence of cytolysis from CD8/MHC class I interaction. These safe epitope-specific vaccines offer a potential strategy to prime HA-1- or HA-2-specific CTL in transplant donors before adoptive transfer.