One of the rationales for designing engineered vaccines based on putting together individual defined epitopes is that for HIV, for instance, one can in principle select epitopes that induce neutralizing antibodies, cytotoxic T cells, and helper T cells that might be protective, thereby avoiding epitopes that induce enhancing antibodies, autoimmune responses, or suppressor cells; and, also, that one can combine these defined epitopes in various ways to make them potentially more effective, such as by producing multivalent constructs. I-3 In the course of carrying out work of this sort, we asked the question, can we go one step further? Instead of just using the natural epitopes fiom a virus like HIV, can we improve on these epitopes by tinkering with the internal structure of the antigenic determinant or antigenic site? For both cytotoxic and helper T cells the antigen that is seen is not the whole antigen but a small peptide bound to a major histocompatibility complex (MHC) molecule after cleavage through different pathways in the cell, described by Ron Germain ear lie^^.^ Therefore, the way the peptide binds to the MHC molecule is going to be critical in that recognition.