Model building and molecular mechanics studies on G-Tetraplex and B-Z-B DNA motifs

Abstract

For almost 25 years following the proposal of the doublehelix structure by Watson and Crick, this model was believed to be the only structure for DNA. Even though various polymorphic forms such as A, B, C, and D were identified from fibrediffraction studies, all these models were righthanded helices with a uniform structure irrespective of the sequence. The lefthanded structure for DNA, proposed from a theoretical analysis of the conformational flexibility inherent in a nucleotide unit, found little acceptance until the discovery of the lefthanded ZDNA structure in single crystals. However, with the availability of synthetic oligonucleotides and the development of new techniques to probe DNA structure, studies in the last decade have completely changed this view. It is now widely accepted that DNA is highly polymorphic and can adopt a number of different structures depending on the sequence and environment. Some of these structures-such as lefthanded ZDNA, cruciforms, parallelstranded DNA, triplexes, and tetraplexes-are drastically different from the original Watson-Crick double helix and are therefore referred to in the literature as “unusual DNA structures” in contrast to canonical BDNA. The possibility that these unusual structures might have crucial roles in regulating various biological processes has prompted many research groups to focus on their generation, detection, and functional characterization. Although numerous biochemical studies have been carried out on unusual DNA structures, most provide only gross structural features. Only a limited number of NMR and Xray crystallographic studies have yielded detailed conformational information. Since computermodelling techniques provide a powerful tool to study biomolecular conformation, the present work uses modelbuilding and molecularmechanics studies with the AMBER force field to elucidate the conformational features of two unusual DNA structures-Gtetraplex structures of telomeric DNA and B-Z-B DNA motifs. These studies form the major part of the thesis, while results on sequencedependent variability in duplex basepair stacking are presented in the appendix.

Gtetraplex Structures of Telomeric DNA Telomeric DNA contains a doublestranded region with tandem repeats of short guaninerich nucleotides at the 3 end, and the Grich strand terminates with a short singlestranded overhang of this repeat sequence. Although the exact repeat sequence varies between species, the conserved feature is the presence of stretches of guanines interspersed with short stretches of thymines (or occasionally a single adenine). Because of the functional significance of the telomere and the conserved nature of these motifs, several structural studies have been conducted. Initial biochemical studies indicated that telomeric DNA can adopt tetramolecular, bimolecular, or unimolecular structures, all stabilized by Hoogsteenbonded guanine tetrads. When the present work began, no Xray or NMR results on telomeric Gquadruplexes were available. Thus, the modelbuilding and molecularmechanics studies were initiated to address the following questions:

What are the possible arrangements of guanine strands in parallel and antiparallel Gtetraplex structures, and what are the possible guanine conformations within each What is the minimum number of thymines needed to form loops across adjacent or diagonal strands in a Gtetraplex Why do guanine stretches interspersed with thymines form foldedback structures, whereas isolated guanine tracts form parallel tetraplexes What are the possible conformations and orientations of thymines and adenines in foldedback structures of sequences such as dCTGT and dCTAGT

The methods used and detailed results on telomeric DNA are presented in Chapters II-IV, while Chapter I provides an overview of DNA structures reported in the literature.

Summary of Telomeric DNA Studies Chapter II reports modelbuilding and molecularmechanics studies to investigate guanine conformations along a strand and the possible alignments of strands in parallel and antiparallel Gtetraplexes. The studies identify a feasible antiparallel model with all strands adopting identical conformations and an alternating syn-anti arrangement of glycosidic orientations. This model later matched crystallographic and NMR structures. Energetically, parallel tetraplexes with allanti guanines are favoured for isolated Gstretches, whereas alternating syn-anti arrangements are favoured when thymine interruptions exist. Chapter III explores loop conformations. Two thymines are required for loops across adjacent strands; three for loops across diagonal strands. Hairpin dimers form in several orientations, all energetically similar. Stacked thymine tetrads and interloop hydrogen bonds appear in some configurations, consistent with later observed structures. Molecularmechanics results explain why thymineinterrupted sequences fold back, while pure Gtracts form parallel tetraplexes. Chapter IV extends the study to human telomeric sequences (dCTTAG) and Arabidopsis telomeres (dCTAG). Numerous hairpindimer structures are possible. The presence of adenine greatly increases structural variability and enables multiple types of Watson-Crick and Hoogsteen A:T pairing. Structural constraints on thymine syn/anti conformations in loops are identified. Predicted interproton distances suggest specific NOE signatures in NMR experiments.

Conformational Features of B-Z-B DNA Motifs Short alternating pyrimidine-purine sequences capable of forming ZDNA occur naturally, flanked by long BDNA regions, where they may regulate biological processes. Few detailed structural models exist for such B-Z-B motifs. The present work aims to:

Construct a stereochemically satisfactory B-Z-B model with intact base pairing and stacking. Estimate the stability of B-Z-B DNA relative to pure B or ZDNA. Examine how flanking BDNA sequences influence junction conformation and energetics. Estimate the effects of B-Z junctions on DNA bending and twisting.

Chapter V presents modelbuilding and molecularmechanics studies on B-Z-B motifs. A stable, fully stacked, fully hydrogenbonded model is obtained. Interestingly, even without disruptions, a groove reversal occurs at the junction, possibly explaining the hyperreactivity of junction nucleotides. Sequencedependent variations of junction energetics and geometry are analysed using local step parameters consistent with crystallographic data. Effects on endtoend bending and twisting are estimated. The appendix reports stackingenergy calculations for basepair doublets in roll-slide space using a recently developed parameter set suitable for irregular DNA structures. The results correlate well with known crystal structures.