Theoretical Studies On The 3D Structures Of The G-quadruplexes And The Interactions With Small Molecules

It is well known that Guanine-rich DNA sequences can fold in the presence of monovalent cations to form a four-stranded structure named G-quadruplex. More attention was paid to quadruplexes since G-rich sequences were found to have the potential to take these structures in several biologically important DNA regions, such as gene promoters and telomeres. In recent years, based on the structures of G-quadruplexes, many small molecular inhibitors of telomerase were exploitated. The investigation on the interaction mechanism between telomeric G-quadruplexes and small molecular inhibitors is very helpful as a platform for rational drug design. In addition, unimolecular G-quadruplex structures of d(GGGTGGGTGGGTGGGT) (G1) and d(GTGGTGGGTGGGTGGGT) (G2) are known as the potent nanomolar HIV-1 integrase inhibitors, thus investigating the 3D structures of the two sequences is significant for structure-based rational anti-HIV drug design. Due to the high flexibility, the number of the G-quadruplexes and their complexes with small molecular ligands obtained by single crystal diffraction and high-resolution solution state NMR is limited, the investigation of the structures of G-quadruplexes and their complexes with theoretical methods is useful for further geometric research of G-quadruplex folding. Molecular dynamics simulations are now used widely, which allow for a description of DNA structure and dynamics at the atomic level. In this thesis, molecular modeling and molecular dynamics simulation methods are used to investigate several structures of G-quadruplexes in detail. Some creative results were obtained from our investigation. The main results are outlined as follows:1. Based on the experimental data of circular dichroism (CD) spectropolarimetry and electrospray ionization mass spectrometry (ESI-MS), the initial models of G1 and G2 were constructed by molecular modeling method. The modeling structures of G1 and G2 are intramolecular parallel-stranded quadruplex conformation with three guanine tetrads. Particularly, the structure of G2 possesses a T loop residue between the first and the second G residues that are the component of two adjacent same-stranded G-tetrad planes. This structure proposed by us has a very novel geometry and is different from all reported G-quadruplexes. The extended (35 ns) molecular dynamic (MD) simulations for the models indicate that the G-quadruplexes maintain their structures very well in aqueous solution whether the existence of K+ or NH4+ in the central channel. Furthermore, we perform 500 ns MD simulations for the models in the gas phase. The results show that all the ion-G-quadruplex complexes are maintained during the whole simulations, despite the large magnitude of phosphate-phosphate repulsions. The gas phase MD simulations provide a good explanation to ESI-MS experiments. The principle component analysis indicates that the conformation spaces of all simulations are sampled well. The results demonstrate that the structures of G1 and G2 built by us are rational and credible.2. The sequences with short loops are able to be aggregated to form stable quadruplex multimers. Using molecular modeling and molecular dynamics simulations, a dimeric G-quadruplex structure formed from a simple sequence of d(GGGTGGGTGGGTGGGT) (G1) and its interactions with a planar ligand of a perylene derivative (Tel03) were investigated. Detailed structure analysis, free energy calculation and principal components analysis show that the structure of the dimer with stacked parallel monomer structures is maintained well during the entire simulation. The Tel03 can bind to the dimer efficiently through the end stacking mode, and the binding mode of ligand stacked with the 3'-terminal thymine base is the most favorable whether for 1:1 or 1:2 complexes. The dominant motions in the free dimer occur on the loop regions and the presence of ligand reduces the flexibility of the loops.3. Cationic meso-tetrakis(4-(N-methylpyridiumyl))porphyrin (TMPyP4) has been of particular interest since it can inhibit the activity of telomerase upon binding to human telomeric DNA quadruplexes. We investigated the binding interactions of TMPyP4 with two antiparallel basket-type human telomeric quadruplexes under K+ ion conditions by molecular modeling and molecular dynamics simulations. Detailed structures analysis, free energy calculation and principal component analysis show that the end stacking mode of A-1, B-1 and A-2 is the best binding mode, and next is the external loop stacking mode of A-9 and B-9 and the intercalation mode of A-4. The remaining external binding complexes are superior to the end stacking complexes of B-2 and B-4 and the intercalation complexes of A-3 and B-4. The sites that can stabilize the three-G-tetrad G-quadruplex structure are more than the two-G-tetrad G-quadruplex in the range of our investigation. In addition, we show the 3D structures of the G-quadruplex-TMPyP4 complexes with almost all possible binding modes.4. An important factor to determine the accuracy of molecular dynamics simulation is force field. Extended explicit molecular dynamics simulations were carried out on four quadruplex molecules with parm99 and parmbsc0 force fields. We provide a full characterization for the flexibility of quadruplexes by analyzing the root-mean-square displacements, average structures, backbone torsion angles, glycosidic torsion angles and phase angle of pseudorotation of the sugar ring. All these analyses reveal backbone torsions of G-DNA located in the noncanonical regions are not retained in both simulations. Most of the glycosidic torsion angles of loop bases deviate largely from the experimental values. Many phase angles of pseudorotation of the sugar rings are transformed to other puckers. Theα/γbackbone torsions have the most important effect to the flexibility of whole structure. The results allow us to find the reasons that G-DNA can not be described accurate enough, especially for the loops, and the rules of transformation of these structural parameters in both force fields are useful for further developing force field of the G-DNA structure.