The absorption edge of protein-bound mercury and a double-edge strategy for HgMAD data acquisition.

The L(III) absorption edge of protein-bound mercury (Hg) has been experimentally determined using X-ray data collection from a crystal. This absorption edge is 12 291 eV, 4 eV higher than the theoretical value of elemental Hg. Considering the possible shift of the Hg absorption edge with the chemical environment in different protein crystals, a double-edge strategy for multiwavelength anomalous diffraction (MAD) data collection has been developed. The approach provides a convenient way to optimize the dispersive signal between a remote wavelength and two edge wavelengths separated from each other by 4 eV. The dispersive signals derived from both edges are used, along with anomalous signals, in MAD phasing and phase refinement. This approach has been used in the crystal structure determination of three proteins containing one Hg atom per 186-196 amino-acid residues at 2.0, 2.6 and 2.7 A resolution. A set of four wavelengths is recommended for HgMAD data acquisition: 1.0087 A (12 291 eV, edge1), 1.0084 A (12 295 eV, edge2), 1.0064 A (12 320 eV, peak) and 0.9918 A (12 500 eV, remote). Although it is no longer necessary to determine the L(III) absorption edge of protein-bound Hg experimentally, an initial fluorescence scan on the crystal for data collection is still necessary to verify the existence of Hg in the crystal.