JOINT REFINEMENT OF PROTEIN STRUCTURE AGAINST NMR AND X-RAY DATA

M.L. Raves¹, J.F. Doreleijers¹, H. Vis¹, C.E. Vorgias², Z. Dauter³, K.S. Wilson⁴ and R. Kaptein¹

¹ Bijvoet Center for Biomolecular Research, Utrecht University, The Netherlands, ² EMBL, Hamburg, Germany, ³ Brookhaven National Lab., USA, ⁴ Protein Structure Group, York University, UK.

Joint refinement, i.e. the simultaneous refinement of a structure agains NMR and X-ray data [1], was done in order to investigate the compatibility of the two data sets. Wherever a conflict was found, attempts have been made to determine whether it arose from misinterpretation of the experimental data or from real differences of the protein conformation in solution vs. that in the crystal.
The DNA-binding protein HU was taken as a test-case. Its structure consists of a rigid core and two flexible `arms' and was solved independently by NMR [2] and by X-ray crystallogaphy [3]. The joint refinement procedure started with an unbiased random structure, which was folded using the NMR restraints and then placed in the crystallographic unit cell by molecular replacement. Important differences in main-chain conformations and secondary structure were identified and investigated.

[1] B. Shaanan et al. (1992) Science 257, 961-964
[2] H. Vis et al. (1995) J. Mol. Biol. 254, 692-703
[3] Z. Dauter and K.S. Wilson, manuscript in preparation