ABSTRACT
Structural features of a trisdecamer duplex, [d(GGCAGAGGTGAAA).d(TTTCACCTCTGCC)], in solution are being investigated by proton one-dimensional (1D) and two-dimensional (2D) NMR spectroscopy. This DNA sequence is comprised of the 11-base-pair direct repeat sequence found in the hepatitis B viral genome with an additional base pair from the genome included on each end to minimize end effects on the 11-bp sequence of interest. The direct repeat sequence occurs twice in the viral genome; both are essential for initiation of DNA synthesis. The critical nature of this sequence suggests it may be a target to control replication of the virus. Elucidation of the structure of the direct repeat sequence could prove to be beneficial in targeting efforts. Structural determination via restrained molecular dynamics requires experimentally derived distance restraints. The ability to determine solution structures of biomolecules by NMR spectroscopy is limited by the quality and quantity of distance and torsion angle restraints that can be extracted from the NMR data. Techniques used to establish these restraints are constantly evolving and improving. Modifications in procedure are applied to the trisdecamer duplex to yield improvements in the determination of sugar conformations from COSY data and a substantial increase in the number of distance restraints typically garnered from 2D NOE intensity data. This increase in the number of distance restraints normally obtained from 2D NOE intensities was accomplished by utilizing a new version of the iterative complete relaxation matrix program MARDIGRAS with intensities extracted from a 2D NOE data set acquired in 90% H2O. The exchange rate of the imino and amino protons with the solvent water protons can now be included in the relaxation matrix calculations, thereby providing more accurate distances when utilizing the 2D NOE cross-peaks involving at least one exchangeable proton. In this lab, analysis of two-quantum-filtered correlation (2QF-COSY) spectra to determine the conformational states of the sugar moieties typically employs the program package SPHINX/LINSHA to simulate the scalar coupling effects manifest in the cross-peaks. With enough data, we typically find that a single conformer is inadequate to describe sugar pucker, but a rapid two-state equilibrium is consistent with all the data. In the process, a large number of cross-peaks are simulated with a range of possible sugar conformation ratios, amplitudes, and line widths. A limitation of this procedure is the possibility of producing a nonunique solution. These methods rely on the ability to match the appearance of simulated cross-peaks with real data.(ABSTRACT TRUNCATED AT 400 WORDS)
