Protein structure determination with paramagnetic compounds

The structure determination of macromolecules by NMR spectroscopy relies currently mainly on the use of short- range distance (NOE) and dihedral angle (scalar coupling) constraints. I plan to establish a new procedure for obtaining distance constraints for NMR protein structure determinations. A paramagnetic compound which is freely mobile in solution and shows no specific interactions with the biomolecule, leads to distance dependent relaxation rate increases. Therefore relaxation rate increases are larger on the surface of the protein and smaller at the core. The hereby obtained distances from the surface can be used as short, medium and long range constraints for structure calculations. In addition to providing distance constraints, the distribution of relaxation rate increases will provide a qualitative parameter for the shape of a macromolecule. Furthermore, paramagnetic complexes specifically modified to give weak interactions with certain functional groups will be used for probing charged or hydrophobic patches on protein surfaces. The suggested techniques will be developed and tested on proteins that are well characterized by NMR to show their validity and then used to obtain the NMR structure of fructose-2,6-bisphosphatase, which is an important regulator in the glycolysis/gluconeogenesis pathway.

A paramagnetic compound in the vicinity of an NMR active nucleus causes the relaxation rates to increase proportional to 1/r6 . If such a paramagnetic center is freely soluble in solution, the relaxation rate increases in a protein depend on the distance of each nucleus to the surface of the biomolecule and therefore contain structurally relevant information. During the course of this project we developed an algorithm which allows the calculation of theoretical relaxation rate increases (Rth ) by paramagnetic compounds through a 1/r6 -averaged integration over a 20 Å thick layer around each nucleus, which includes only the protein-free volume. Subsequently, any descrepencies between these calculated Rth and experimentally obtained relaxation rates can be used to modify the distance from the surface of each NMR spectroscopically detectable nucleus. These algorithms have been incorporated into the programs CNS and X-PLOR, which are routinely used for structure determinations by NMR spectroscopy and X-ray crystallography. A remarkably good correlation between theoretical relaxation rates calculated with this method and experimental ones obtained on structurally well-characterized proteins (lysozyme, ParD) could be found. If the structure of a monomer is known, paramagnetic surface probes can be used to define the dimerization interface. Thereby differences between experimental relaxation rate increases and theoretical ones calculated for the monomer hint towards intermolecular interactions. This procedure was used to define the dimer structure of the bacterial antitode ParD, which was then confirmed through intermolecular NOEs (hydrogen-hydrogen distances). Besides proteins with known structure, we intend to verify the solution structure of fructose-2,6-bisphosphatase with paramagnetic surface probes. Thereby, the crystal structure of a truncated version of this protein can be used as a starting point for structure calculations using only surface distance constraints. The sequential NMR spectroscopic assignment of this protein which is a prerequisite for such studies has been recently finished.