Toward threading polyintercalators with programmed sequence specificity

Synthetic molecules that can bind DNA in a sequence specific manner are an interesting area of molecular recognition research with the potential to develop new therapeutics. A novel class of polyintercalators, composed of 1,4,5,8- naphthalene tetracarboxylic diimide (NDI) intercalating units connected via peptide linkers, have been developed in the Iverson research group. A previous combinatorial library screen identified two different bis-intercalators with altered sequence specificity and subsequent structural studies revealed that these molecules have a unique threading mode of polyintercalation, in which the tethering linker can be in either the major or minor groove, depending on the peptide linker composition. The peptide linkers are believed to play an important role in the recognition of target sequences and several analogues were investigated using DNase I footprinting experiments. These results showed that the position of a lysine residue in the linker plays a crucial role in recognizing internal sequences of 5’-GGNNCC-3’ on DNA. Among those derivatives, the parent compound (1), H2N-Lys-NDI-Gly-Gly-Gly-Lys-NDI-Gly-OH, showed dramatic sequence reading ability on 5’-GGNNCC-3’ sequences with the order of GGTACC > GGCGCC > GGATCC ~ GGGCCC with Kd ~ 125 nM. The previous structural knowledge of the two different bis-intercalators (1, and 1.2) served as binding modules that enabled us to design threading polyintercalators with sequence specific contacts in both grooves. A trimer (4.1), a hybrid molecule of the two bis-intercalator modules was designed to block both grooves and its programmed sequence reading ability was proved in footprinting experiments. As a parallel approach to achieve polyintercalation with programmed sequences, a symmetric tetramer (5.1) that is a conjugate of two identical bis-intercalator units (1.2) with flexible aliphatic chain linker was designed. 1D 1 H-NMR titration experiment revealed that this compound binds to a 14 bp DNA sequence d(GATAAGTACTTATC)2 with 1:1 stoichiometry. 2D 1 H-NMR studies combined with restrained molecular dynamics confirmed a threading binding mode in the three-dimensional structure of the DNA-ligand complex in which β-alanine residues in one type of linker reside in the minor groove as expected, while an adipic acid linker resides in the major groove also as expected. This represents the first structural analysis of any DNA intercalating molecules larger than a bis-intercalator and confirms that the sequence specific binding of longer threading polyintercalators can be predicted from bisintercalator modules. These results will provide the molecular basis for the DNA binding properties of a new generation of threading polyintercalators.