Computational studies of DNA structure and recognition
This thesis involves the use of large scale molecular dynamics simulations and associated analysis techniques to study DNA structure and recognition. LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator) is a scalable molecular dynamics code including long-range Coulomb interactions that has been specifically designed to function efficiently on parallel platforms. Here we describe the implementation of the AMBER98 forcefield in LAMMPS and its validation for molecular dynamics investigations of DNA structure and flexibility against benchmark AMBER6 code results. Extended MD simulations on the hydrated DNA dodecamer d(CTITIGCAAAAG)₂ and 1:1 and 2:1 drug complexes, which have previously been the subject of extensive dynamical analysis using AMBER6, show that it is possible to obtain excellent agreement in terms of static, dynamic and thermodynamic parameters between AMBER6 and LAMMPS. Also, compared to AMBER6, LAMMPS shows greatly improved scalability in massively parallel environments (Cray T3E). The telomerase enzyme is active in 85-90% of human tumours and is therefore an important target in anti-cancer drug design. Telomerase acts at the telomeric regions of chromosomes adding successive (TIAGGG)n repeats causing immortalisation of the cell. Telomerase can be inhibited by the stabilisation of G-quadruplexes which, in vitro studies show, are formed in these telomeric regions. In order to minimise non-specific toxicity associated with this approach it is important that the drugs preferentially bind to quadruplex over duplex DNA. A series of novel polycyclic acridine salts have been synthesised within our laboratories that show this property. MD studies have been used to study alternative binding relationships of RHPS4 (our lead compound) to quadruplex and duplex DNA and to explore the differences in binding profiles of RHPS4 and its methyl derivative RHPS3. Analysis of extended simulations (≥3ns) has been carried out including evaluation of ΔG from enthalpic and entropic contributions, linear interaction energy, stacking interactions and molecular interaction potentials. "Correct" binding positions for RHPS4 in quadruplex and duplex DNA have been found and simulations and analysis of RHPS3 also carried out. Although the results are not conclusive and do not all agree with the experimental data we can conclude that quadruplex verses duplex selectivity is governed by a subtle balance between many factors, including electrostatic and vdW interactions, DNA flexibility and most probably the models used.