We have derived an experimental data set containing 234 out of a possible 274 pseudocontact shifts by NMR studies on cobalt and zinc complexes of d(TTGGCCAA)2 with chromomycin. In addition, we measured 812 NOE's on the zinc complex. We have determined the structure of the drug- DNA complex using either or both the experimental shift and NOE constraints, and using a modified X-PLOR containing an additional energy term to describe the shift constraints. We have obtained an ensemble of structures that conform extremely well to the observed shifts, have low energies, and form a very tight family of structures. We are currently investigating the dependence of the final structure on g-tensor parameters which must be known or assumed for the shift calculation, as well as methods for convergence from different starting points, and justifications for differences between the NOE-based and shift-based structures. The facilities and staff of the Computer Graphics Laboratory have been extremely helpful during this project. We have displayed and matched structures, calculated root mean square deviations, and generated a myriad of images representing different facets of the project. These include studies of the metal binding site and the use of the docking features of MidasPlus to identify a fifth and sixth water ligand in the coordination sphere; the use of noeshow in MidasPlus to highlight and examine NOE's which conflict with the shift-based structure; the use of color-by-Bfactor to display shift and precision across the molecular complex. We have used the ability of MidasPlus to overlay different pdb files and to handle very large pdb files to understand characteristics of our shift based structures.
Showing the most recent 10 out of 508 publications
