The nonrandom chromosomal translocation t(15;17)(q22;q21) in acute promyelocytic leukemia (APL) juxtaposes the genes for retinoic acid receptor alpha (RAR alpha) and the putative zinc finger transcription factor PML. The breakpoint site encodes fusion protein PML-RAR alpha, which is able to form a heterodimer with PML. It was hypothesized that PML-RAR alpha is a dominant negative inhibitor of PML. Inactivation of PML function in APL may play a critical role in APL pathogenesis. Our results demonstrated that PML, but not PML-RAR alpha, is a growth suppressor. This is supported by the following findings: (i) PML suppressed anchorage-independent growth of APL-derived NB4 cells on soft agar and tumorigenicity in nude mice, (ii) PML suppressed the oncogenic transformation of rat embryo fibroblasts by cooperative oncogenes, and (iii) PML suppressed transformation of NIH 3T3 cells by the activated neu oncogene. Cotransfection of PML with PML-RAR alpha resulted in a significant reduction in PML's transformation suppressor function in vivo, indicating that the fusion protein can be a dominant negative inhibitor of PML function in APL cells. This observation was further supported by the finding that cotransfection of PML and PML-RAR alpha resulted in altered normal cellular localization of PML. Our results also demonstrated that PML, but not PML-RAR alpha, is a promoter-specific transcription suppressor. Therefore, we hypothesized that disruption of the PML gene, a growth or transformation suppressor, by the t(15;17) translocation in APL is one of the critical events in leukemogenesis.