Structural refinement calculation from two- and three-dimensional NMR data by a hybrid full relaxation matrix method
Abstract
The computational study presented is about how to obtain high quality three-dimensional solution structures for a variety of nucleic acids from distance information provided by NMR spectroscopy. The methodology is mainly the hybrid full relaxation matrix in conjunction with molecular dynamics simulation. The whole study has basically two focuses: to apply this methodology to solve structures for new molecules, mainly for nucleic acids; and to further improve and develop this promising approach. In some projects both aspects are involved at the same time. In the first aspect, structures for four molecules of nucleic acids or analogues were successfully refined. A duplex DNA of fourteen base pairs which is used as part of a testing system to gain more understanding about the famous Lac operator has been shown to be largely a B-DNA in the solution while its central part displays certain A-DNA features. The structure of a dithioate-modified DNA hair-pin decamer, an antisense analogue, has finally been refined after some long and hard effort. Two dithioate-modified duplex DNA analogues, one with one phosphorodithioate modification on each strand the other with four modifications on each strand, show consistent bending and unwinding. Detailed analysis suggests that any dithio-DNA duplex will probably contain a sharp bending towards the major groove in the modified region. Two different parameters describing fast internal motion were included in the calculation of spectral density functions in the NOE simulation to improve the refinement quality. The tests with two real molecules suggested a limited effect of internal motion in the current situations. Finally, a new scheme of making use of 3D NOE-NOE data in the hybrid relaxation matrix protocol was tested by simulated refinement. Preliminary results show that the approach is very promising for refining larger molecules that cannot provide good 2D data.
Degree
Ph.D.
Advisors
Gorenstein, Purdue University.
Subject Area
Biochemistry|Biophysics
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.