Bisindolylmaleimide Ligands Stabilize c-MYC G-Quadruplex DNA Structure and Downregulate Gene Expression.

G-Quadruplex (G4) structures play a pivotal role in diverse biological functions, including essential processes, such as telomere maintenance and gene regulation. G4 structures formed in functional regions of genomes are actively pursued toward therapeutics and are targeted by small-molecule ligands that alter their structure and/or stability. Herein, we report the synthesis of bisindolylmaleimide-based (BIM) ligands, which preferentially stabilize parallel G4 structures of c-MYC and c-KIT oncogenes over the telomeric h-RAS1 G4 and duplex DNAs. The preferential stabilization of parallel G4s with BIM ligands is further validated by the DNA polymerase stop assay, where stop products were only observed for templates containing the c-MYC G4 sequence. Nuclear magnetic resonance (NMR) titration studies indicate that the lead ligand BIM-Pr1 forms a 2:1 complex with c-MYC G4 DNA with a KD of 38 ± 5 µM. The BIM ligand stacks at the 5' and 3' quartets, with molecular modeling and dynamics studies supporting the proposed binding mode. The ligand is cytotoxic to HeLa cells and downregulates c-MYC gene expression. Collectively, the results present bisindolylmaleimide scaffolds as novel and powerful G4 targeting agents.

Characterization of DNA G-quadruplex Topologies with NMR Chemical Shifts.

G-quadruplexes are nucleic acid motifs formed by stacking of guanosine-tetrad pseudoplanes. They perform varied biological roles, and their distinctive structural features enable diverse applications. High-resolution structural characterization of G-quadruplexes is often time-consuming and expensive, calling for effective methods. Herein, we develop NMR chemical shifts and machine learning-based methodology that allows direct, rapid, and reliable analysis of canonical three-plane DNA G-quadruplexes sans isotopic enrichment. We show, for the first time, that each unique topology enforces a specific distribution of glycosidic torsion angles. Newly acquired carbon chemical shifts are exquisite probes for the dihedral angle distribution and provide immediate and unambiguous backbone topology assignment. The support vector machine learning methodology aids resonance assignment by providing plane indices for tetrad-forming guanosines. We further demonstrate the robustness by successful application of the methodology to a sequence that folds in two dissimilar topologies under different ionic conditions, providing its first atomic-level characterization.