Three to five amino acid caspase inhibitors have been extensively used to identify the role of specific caspases in apoptotic pathways triggered by a wide range of cellular insults. Unexpectedly, we have recently demonstrated that the irreversible caspase-9 inhibitor (C9i) Z-LEHD-FMK can promote apoptosis in selected stressed and non-stressed human leukemic cells, and that inhibition of caspase-9 protein expression does not render cells more resistant to stress-induced apoptosis (Shah et al., Blood, in press 11/15/04). The goal of the current study was to analyze the role of caspase-9 in C9i-induced apoptosis, and the role of specific amino acids in the proapoptotic activity of C9i. We first determined whether cells made deficient in caspase-9 were still sensitive to the proapoptotic activity of C9i. Electroporation of the BLIN-4L stromal cell-dependent B-lineage ALL cell line and the RAMOS Burkitt lymphoma cell line with siRNA specific for caspase-9 led to >95% reduction of caspase-9 protein. Yet, both cells still exhibited a dose-dependent apoptotic response to C9i indistinguishable from cells electroporated with control siRNA. Since proapoptotic activity was not a property of inhibitors of caspases-2, 3, 6 and 8 (Shah et al.), we further examined the minimal structural requirements for the proapoptotic activity of C9i. The P1 aspartic acid and P3 glutamic acid are highly conserved in other caspase inhibitors that do not exhibit proapoptotic activity, so we initially focused on the P2 and P4 positions. The parent compound Z-LEHD-FMK was modified to yield Z-LEAD-FMK and Z-AEHD-FMK (synthesized by MP Biomedicals, Livermore, CA). Z-WEHD-FMK (commercially available as caspase-1 inhibitor) and Z-LEED-FMK (commercially available as caspase-13 inhibitor) were also studied. We tested several cell lines previously shown to be sensitive to C9i. As expected (Shah et al.), the BLIN-2, BLIN-3 and BLIN-4L adherent cell-dependent B-lineage ALL, the BLIN-1 pre-B ALL and the RAMOS Burkitt B cell lymphoma were sensitive to C9i as measured by TMRE and Annexin V staining. Strikingly, Z-LEAD-FMK and Z-LEED-FMK induced no apoptosis in the C9i sensitive targets. In contrast, Z-AEHD-FMK was weakly proapoptotic (at the maximum concentration of 100 μM) whereas Z-WEHD-FMK exhibited slightly greater proapoptotic activity than the parent compound Z-LEHD-FMK. Furthermore, Z-WEHD-FMK exhibited an overall pattern of proapoptotic activity against multiple sensitive and insensitive leukemic cell targets that was indistinguishable from Z-LEHD-FMK. Thus, the leucine at position P4 is expendable whereas the histidine at P2 is essential for proapoptotic activity of Z-LEHD-FMK. This pattern of apoptotic sensitivity to Z-LEHD-FMK and Z-WEHD-FMK extended to CD19+ B-lineage cells derived from cord blood HSC plated on the murine MS-5 stromal cell line. As a more stringent test of proapoptotic activity, the peptides were tested by limiting dilution analysis against RAMOS and K-562. Fifty μM Z-LEHD-FMK or 50 μM Z-WEHD-FMK exerted 2.5–3.0 log killing of RAMOS but had no effect on K-562. We conclude that selected peptides previously synthesized and widely used as caspase inhibitors harbor potent proapoptotic activity against human leukemic cells, which is unexpectedly distinct from their capacity to inhibit activated caspases.