Combinatorial optimization of the DNA cleaving Ni(II) x Xaa-Xaa-His metallotripeptide domain.

A positional-scanning combinatorial protocol was employed to optimize the deoxyribose-based cleavage of B-form DNA by Ni(II) x Xaa-Xaa-His metallopeptides. This procedure employed 18 naturally occurring amino acids (excluding Cys and Trp) to generate two libraries in which the first and second positions of the peptide ligand were varied. Increased direct DNA cleavage relative to Ni(II) x Gly-Gly-His was observed when (1) the amino-terminal peptide position contained Pro, Met, Arg, or Lys (with Pro exhibiting the greatest activity) and (2) the second peptide position contained Lys, Arg, Met, Ser, or Thr (with Lys exhibiting the greatest activity); the optimized metallopeptide, Ni(II) x Pro-Lys-His, was found to cleave DNA an order of magnitude better than Ni(II) x Gly-Gly-His. While metal complexation and the A/T-rich site selectivity of the optimized metallopeptides were not altered, DNA binding affinity was slightly increased relative to Ni(II) x Gly-Gly-His, however, not to an extent necessary to account for the observed increase in reactivity. Examination of molecular models of Ni(II) x Pro-Lys-His bound to the minor groove of DNA via hydrogen bonding of the His N3 imidazole hydrogen to the N3 of adenine or O2 of thymine suggests that the Pro residue can make hydrophobic contacts with the sugars lining the walls of the groove while the Lys residue is able to form a salt bridge with a proximal phosphate; with these interactions, the metal center is poised to abstract the C4'-H of an adjacent nucleotide suggesting that noncovalent interactions result in a positioning which contributes to increased DNA cleavage activity.