This study compared immune responses in rhesus macaques immunized with unmodified HIV-1 IIIB Tat, SHIV89.6P Tat, and carboxymethylated IIIB and 89.6P Tat toxoids. Immunization with either IIIB or 89.6P preparation induced high titer and broadly crossreactive serum anti-Tat IgG that recognized HIV-1 subtype-E and SIVmac251 Tat. However, the response was delayed, and titers were lower in 89.6P vaccination groups. Serum anti-Tat IgG recognized peptides corresponding to the amino-terminus, basic domain, and carboxy-terminal region. Cellular proliferative responses to Tat toxoids corresponding to the immunogen were evident in vitro in both IIIB and 89.6P groups. Crossreactive proliferative responses were observed in IIIB groups in response to stimulation with 89.6P or SIVmac251 Tat toxoids, but were much less prevalent in 89.6P groups. The truncated 86 amino acid IIIB Tat appears to be more immunogenic than the 102 amino acid 89.6P Tat with respect to both humoral and cellular immune responses, and may be a better vaccine component. Despite induction of robust humoral and cellular immune responses (including both CD4+ and CD8+ T-cell responses) to Tat, all animals were infected upon intravenous challenge with 30 MID(50) of SHIV89.6P and outcome of vaccine groups was not different from controls. Sequencing both Tat exons from serum viral RNA revealed no evidence of escape mutants. These results suggest that with intravenous SHIV89.6P challenge in rhesus macaques, precipitous CD4+ T-cell decline overwhelms potentially protective immune responses. Alternatively, Tat specific CD8+ T-cell responses may not appropriately recognize infected cells in vivo in this model. In view of evidence demonstrating Tat specific CTLs in the SIV model and in humans infected with HIV-1, results in this pathogenic SHIV model may not apparently predict the efficacy of this approach in human studies. The potency and cross-reactivity of these immune responses confirm Tat toxoid as an excellent candidate vaccine component.