Blaine Mooers

University of Oklahoma Health Sciences

Using direct methods starting from random phases, the crystal structure of a 32-base pair RNA (675 nonH RNA atoms in the asymmetric unit) was determined using only the native diffraction data (resolution limit 1.05 Å) and the computer program SIR2014. The RNA’s almost three helical turns in the asymmetric unit introduced two levels of pseudo translational symmetry that modulated the intensities when averaged by the l Miller indices. The helical turns hindered the structure determination when compared to the direct methods structure determination of a single-stranded RNA with no pseudo translational symmetry. More phasing trials were needed to determine the structure of the double-stranded RNA when compared to a hairpin RNA (27-nt, 580 nonH RNA atoms). To investigate this difference, we compared the respective intensity distributions, Patterson maps, translation vectors used to shift misplaced trials structures, and loss of the strongest reflections on the number of trials to a correct structure. The last experiment mimicked the loss of the largest reflections to detector saturation during data collection. Finally, examples are given of using the distribution of the number of trials to a solution to plan phasing experiments with direct methods when starting from random phases.