Efficacy and sequence specificity are two major requirements in the use of antisense nucleic acids and ribozymes. For long-chain complementary RNA sequences (> 30 nt), effects in living cells are correlated with the association rate of the complementary RNA in vitro, but not with the stability of the formed double strand. Thus, sequence selectivity of complementary RNA has to be defined as fast versus slow annealing with the appropriate target or non-target sequences, respectively. In this work, we performed a systematic kinetic analysis to evaluate the selectivity of bcl-abl-directed antisense RNA and hammerhead ribozymes with a length of the complementary sequences of between 20 and 80 bases. By kinetic in vitro selection, we identified oligomeric as well as long-chain complementary RNA that annealed at least tenfold faster with the bcr-abl sequence in comparison with either of the wild-type sequences bcr or abl, respectively. In the presence of selected oligodeoxynucleotide sequences and RNase H, the bcr-abl transcript was specifically hydrolysed out of a mixture containing abl and bcl sequences as well. Hammerhead ribozymes were designed such that binding with their target was facilitated either via helix I or helix III-forming antisense arms but not both. Further, cleavage and binding occurred on opposite sides of the bcr-abl fusion point. Target selectivity was found for a ribozyme that annealed fast via abl sequences and cleaved within the bcr portion of bcr-abl RNA. Kinetic probing and calculations of the local folding potential indicate that the bcr-abl fusion point sequences are not easily accessible for complementary nucleic acids. This study supports the need for more detailed structural investigations of the bcr-abl fusion sequence and forms a more rational basis for the therapeutic use of nucleic acid inhibitors of the aberrant, bcr-abl gene expression in Philadelphia chromosome-positive cells.