A comparative assessment of the zinc–protein coordination in 2Zn–insulin as determined by X-ray absorption fine structure (EXAFS) and X-ray crystallography

Abstract

Accurate ( ca .0.1Å) knowledge of the metal environment in metalloproteins is essential to understanding their function. Single crystal X-ray analysis has provided detailed descriptions of metal environments in a number of crystallizable proteins but their accuracy has often been limited by their restricted diffraction patterns. The technique of X-ray absorption fine structure (EXAFS) is not limited to crystals and can provide very accurate radial distances between metal ions and their ligands. It has therefore great potential for the study of biochemical metal-containing systems in solution. The method depends on the analysis of the oscillations in the absorption or fluorescence spectrum extending over several hundreds of electron volts above the metal absorption edge. The very intense synchrotron X-ray sources make EXAFS applicable to biological systems where the metal ion concentration is low, typically in the millimolar range. We have determined EXAFS spectra for 2Zn–insulin and a variety of Zn–ligand model compounds of known crystal structures in both the absorption and the fluorescence modes. The X-ray crystallographic refinement of 2Zn-insulin with 1.5 Å data provides estimated standard deviations for well defined atoms ranging from 0.03 to 0.06 Å which thus determine the zinc-ligand distances sufficiently accurately for comparison with those derived from EXAFS. The experimental procedures for obtaining the spectra with the use of the storage ring doris at the Deutsche Elektrononen Synchrotron (DESY) are described. The shapes of the exafs spectra of 2Zn-insulin and the zinc complexes are remarkably similar. These results emphasize one of the major weaknesses of the technique: the difficulty in distinguishing between atoms of similar atomic mass such as oxygen and nitrogen in the present instance. The small but real effect of solvent on the spectrum of one complex has important implications; it reveals the ability of EXAFS to provide evidence for structural changes in the metal coordination and that other structures possibly more relevant to function may exist. For the detailed analysis we used ab initio calculations on the model compounds. The method used is outlined and the programs are cited; the theoretical basis for these calculations is to be found in Lee & Pendry ( Phys . Rev . B 11, 2795-2811 (1975)). Its application to the 2Zn-insulin EXAFS spectrum showed that there is good agreement in both the exafs and the X-ray crystallographic methods for the CE1 and NE2 to Zn radial distance; significant discrepancies, however, exist for the other atoms in the coordinating structure. This failure stems essentially from the smaller contribution of outer atoms to the EXAFS spectrum. To resolve these correctly the amplitudes in the EXAFS spectrum need to be more accurate, and there needs to be a more adequate theory to deal with multiple scattering effects. Until these have been achieved it is probably more profitable to make use of our exact knowledge of the bonding behaviour and geometry of such coordinating groups as imidazole rings.