It has been demonstrated in our laboratory that the transition metal complex ((2S,8R)-5-amino-2,8-dibenzyl-5-methyl-3,7-diazanonanedioato)copper(II) (1) is capable of mediating nonrandom double strand cleavage of plasmid DNA. Double strand DNA cleavage is thought to be a biologically significant source of cell lethality and is apparently an important mode of action for both the bleomycin and enediyne classes of antineoplastic agents. We present here the synthesis and characterization of 1 along with mechanistic studies of its DNA scission chemistry. Complex 1, upon reduction to the Cu(I) state in the presence of O2, generates O2 - (HO2) and O2 2- (H2O2) which react to produce hydroxy radical through the Haber−Weiss reaction. Hydroxy radical is known to abstract hydrogen from deoxyribose, leading to strand scission, release of base propenals, and formation of abasic sites. The pseudo first order kinetics of single strand breakage (t 1/2 = 34 min), double strand breakage (t 1/2 = 21 min), and base propenal formation (t 1/2 = 100 min) have been measured. Base propenal release is therefore kinetically distinct from strand scission, ruling out mechanisms which invoke concerted base propenal formation with strand scission.

A copper-based transition metal complex has been designed which performs double-stranded cleavage of DNA in a nonrandom fashion. The complex, ((2S,8R)-5-amino-2,8-dibenzyl-5-methyl-3,7-diazanonanedioate)copper(II), presents an ammonium group on one side of the metal equatorial coordination plane to the DNA backbone phosphate groups, while the aromatic phenylalanine-derived side chains are constrained to the opposite side of the coordination plane toward the DNA groove. This structure was designed to bind at locations where phosphate groups are in proximity to accessible hydrophobic regions of the DNA. We have estimated single-strand break to double-strand break ratios for DNA strand scission by this complex under a variety of activation conditions, and they are substantially lower than that predicted by statistical models for a random DNA linearization process. This means that more double-strand breaks are produced per single strand break than can be accounted for by random coincident single-strand breaks. We have also investigated the formation of abasic sites, and found that at least as many abasic sites can be cleaved to linear DNA as are linearized in the initial cleavage reaction. We interpret this to mean that the complex binds both at the intact DNA surface for strand scission, and binds at nicked sites on the DNA (where the charged end groups of the nick are likely to be proximate to the accessible hydrophobic interior) for reactivation and complementary strand scission. Insofar as double-strand cleavage may be more potent biologically than single-strand cleavage as a source of lethal DNA lesions, the recognition characteristics of this complex may aid in the design of chemotherapeutic agents.

Double-strand breaks in duplex DNA are thought to be significant sources of cell lethality because they appear to be less readily repaired by DNA repair mechanisms. We recently described the design and cleavage chemistry of ((2S,8R)-5-amino-2,8-dibenzyl-5-methyl-3,7-diazanonanedioato)copper(II) (1), which effects nonrandom double-strand cleavage of duplex DNA. After DNA nicking by generation of hydroxyl radicals, the key step in this process appears to occur through recognition by the metal complex of the nicked-abasic site on duplex DNA, followed by delivery of OH(*) to cleave at the opposing strand, forming a double-strand lesion. Through the use of model nucleic acid substrates and comparison to DNA scission chemistry, we have investigated the electrostatic and hydrophobic contributions to DNA binding by complex 1. We have complemented these reactivity studies with studies on the binding of 1 to a model nucleic acid substrate, using (2)H NMR spectroscopy with deuterated 1 and HDO T(1)()()relaxation enhancement methods to study the binding of 1 to nucleotide substrates. With these methods, we have estimated that the association constant for the 1(+).5'-AMP(2)(-) complex is approximately 16 M(-)(1) and that the binding interaction involves both electrostatic and aromatic stacking interactions between the nucleic acid base and the pendant aromatic side chains of 1.

No abstract

No abstract