W0276

The Multi-wavelength Anomalous Dispersion Experiment at 0.90 Å Producing Atomic Resolution Experimental Electron Density for 36 kDa Enzyme. A. Joachimiak¹, E. Howard², T. Petrova^2#, B. Chevrier², R. Sanishvili¹, A. Mitschler², and A. Podjarny², ¹Argonne National Laboratory, Structural Biology Center, Argonne, IL 60439, USA, ² UPR de Biologie et Genomique Structurale, IGBMC, 1, rue Laurent Fries, 67404 Illkirch, France, ^# Institute of Mathematical Problems of Biology of the Academy of Sciences of Russia, Pushchino, Moscow Region, 142292 Russia.

MAD has emerged as the most powerful method for de novo protein structure determination. This method has also the potential of providing experimental phases free of model bias and may contribute valuable information about the protein structure, the solvent molecules, unusual chemistry and the multiple conformations. Crystals of a SeMet derivative of human aldose reductase (hAR) have been obtained that are isomorphous to the native monoclinic crystals (P2₁, a=49.324 Å, b=66.956 Å, c=47.378 Å, (=90.000, (=92.262, (=90.000) which diffract x-rays to subatomic resolution (0.62 Å) using synchrotron radiation. Crystals contain one monomer of 36 kDa in asymmetric unit and 6 Se sites. MAD experiments at atomic resolution have been performed on the crystals of SeMet labelled hAR in the ternary complex with coenzyme NADP⁺ and two different inhibitors. Diffraction data were collected at the 19ID beamline of Structural Biology Center at the Advanced Photon Source for three wavelengths (peak, inflexion and high energy remote). The data collection was subject to the constraints of detector size, a fixed wavelength around the Se absorption edge (0.98 Å) and crystal decay; therefore, the resolution was limited to 0.90 Å. Data were shown to be of high quality up to 1.1 Å resolution.

MAD phasing was performed using the SOLVE program. The Se atoms were located and refined, giving rise to experimental phases with figure of merit > 0.9 up to 1.1 Å resolution, and extending to 0.90 Å. The phases and maps obtained from the MAD experiment were compared with those calculated from the model of native hAR refined to 0.66 Å. The experimental electron density MAD maps were found to be highly informative; their analysis showed peaks corresponding to every ordered protein atom, including features such as multiple conformations for side chains, networks of H-bonds and water molecules. There were also finer details, such as H-atom density, particularly in the well-ordered active site region. Some unexpected density peaks near amino acid residues are currently being studied. The SeMet hAR structure was refined using the amplitudes corresponding to the remote wavelength (averaged F+ and F-). The final model includes 312 aldose residues (the last 4 are disordered), the coenzyme NADP+, the inhibitor IDD 594 and 663 H2O molecules.

This work was supported by the Centre National de la Recherche Scientifique (CNRS), by a collaborative project CNRS-NSF, by the Institut National de la Santé et de la Recherche Médicale and the Hôpital Universitaire de Strasbourg (H.U.S), and by the U.S. Dept. of Energy, Office of Health and Environmental Research, under contract W-31-109-Eng-38.