Deoxyribozymes, or DNA enzymes (DNAzymes), are novel nucleic acids that have the ability to bind to specific sequences of RNA, and to cleave the target site catalytically. DNAzymes are smaller and more efficient enzymatically than ribozymes (RZs), which are catalytic nucleic acids synthesized from ribonucleotides. We have designed three DNAzymes that specifically target the two variants of the p210 bcr-abl gene (splice 1, b3a2; splice 2, b2a2) and the p190 variant (ela2). The cleavage sites for these DNAzymes are located 5 nucleotides (nt) 5' from the fusion site for b3a2, and only 1 nt 5' from the fusion sites for b2a2 and e1a2. We have shown in cell-free in vitro cleavage assays that these DNAzymes efficiently cleave their respective substrates. Mutated DNAzymes, in which only one critical base has been altered, do not cleave these targets. We have used a serum-resistant cytofectin (GS 2888; Gilead) to transfect the DNAzymes into target K562 cells, which express p210bcr-abl. In short-term transfection assays, the DNAzymes specifically inhibited p210bcr-abl protein expression by K562 cells by about 40%, and inhibited cell growth by more than 50% in a 6-day liquid culture assay. We have also transfected freshly isolated CD34+ bone marrow cells from patients with CML with the DNAzymes, which specifically inhibited the growth of bcr-abl-positive CFU-Mix colonies by 53-80%. The potential advantages of anti-bcr-abl DNAzymes over RZs include the following: DNAzymes are much less expensive to synthesize; they are more resistant to serum; and the anti-b2a2 DNAzyme cleaves at a site only 1 nt away from the fusion site, whereas its hammerhead RZ counterpart cleaves this target at a site 8 nt 3' to the fusion site, well within abl exon 2. DNAzymes are novel RNA-cleaving molecules that may significantly improve our ability to inhibit bcr-abl oncogene expression in Ph-positive target cells.