Despite the widespread use of chemotherapeutic drugs in the treatment of various malignancies, in many cases the mechanism of tumor cell kill remains unknown There is, however, much evidence that suggests that DNA is the major cellular target for many of the agents in current clmical use. A number of physicochemical techniques are available to probe the reversrble and nonreversible interactions of these drugs with DNA, and a wealth of information regarding the sequence specificity of these interactions has been documented using these procedures, as well as molecular-biology-based techniques such as DNA and RNA footprinting (1 ,2 ). These studies have also revealed that binding of these compounds to DNA can interfere with various aspects of DNA replrcation, transcription, and translation. Since these intricate processes involve regulatory proteins and cofactors, another approach to characterizing drug-DNA interactions is to ascertain the ability of DNA binding proteins to recognize then drug-modified DNA consensus sequences (3 –6 ). This approach is particularly relevant because such 5′untranslated regions are unwound when assembled on the nuclear matrix and thus accessrble to drugs during active gene transcription in cells (7 –10 ). The sequence selectivity of the drug will therefore determine which DNA-binding proteins are affected, thus leading to a broad predictive index of which genes are more likely to be affected, and perhaps, more importantly, being able to identify the critical stage of gene expression at which these agents may be most active.